# The emergence of American English as a discursive variety

Tracing enregisterment processes in nineteenth-century U.S. newspapers

Ingrid Paulsen

Language Variation 7

# Language Variation

Editors: Martijn Wieling, John Nerbonne (until 2021-07), Alexandra D'Arcy (from 2021-07)

In this series:


# The emergence of American English as a discursive variety

Tracing enregisterment processes in nineteenth-century U.S. newspapers

Ingrid Paulsen

Ingrid Paulsen. 2022. *The emergence of American English as a discursive variety: Tracing enregisterment processes in nineteenth-century U.S. newspapers* (Language Variation 7). Berlin: Language Science Press.

This title can be downloaded at: http://langsci-press.org/catalog/book/341 © 2022, Ingrid Paulsen Published under the Creative Commons Attribution 4.0 Licence (CC BY 4.0): http://creativecommons.org/licenses/by/4.0/ ISBN: 978-3-96110-338-6 (Digital) 978-3-98554-034-1 (Hardcover)

ISSN: 2366-7818 DOI: 10.5281/zenodo.6207627 Source code available from www.github.com/langsci/341 Collaborative reading: paperhive.org/documents/remote?type=langsci&id=341

Cover and concept of design: Ulrike Harbort Typesetting: Ingrid Paulsen, Sebastian Nordhoff Proofreading: Alexandra Fosså, Amy Amoakuh, Cesar Perez Guarda, Fahad Almalki, Geoff Sampson, Ikmi Nur Oktavianti, Janina Rado, Jeroen van de Weijer, Lachlan Mackenzie, Jean Nitzke Fonts: Libertinus, Arimo, DejaVu Sans Mono Typesetting software: XƎLATEX

Language Science Press xHain Grünberger Str. 16 10243 Berlin, Germany langsci-press.org

Storage and cataloguing done by FU Berlin

# **Contents**


# Contents


# Contents


# **Acknowledgments**

I am very thankful for having so many people in my life who have supported me in the process of writing this book and beyond. Most importantly, I am deeply grateful to my doctoral supervisor, Lieselotte Anderwald, for introducing me to the concept of enregisterment and giving me the opportunity to follow my research interests while working as her assistant. Her expertise and guidance were invaluable to me and I would like to thank her for her constant and reliable support. I also thank Elmar Eggert for his willingness to act as my second supervisor and Edgar Schneider for reviewing my thesis as a third supervisor. I would also like to express special thanks to Susanne Heinz who has always been a constant source of motivation and encouragement for me.

I am also very lucky to be surrounded by an amazing group of colleagues and friends at Kiel University who have supported me in many ways. I say thank you to Beke Hansen, Berit Johannsen, Mila Lütjohann, Johanna Gerwin, Robert Daugs and Jana Fischer for their many helpful comments on my work and for sharing my fascination with language with me. I thank Bärbel Schwarz and Tom Schoroth for their administrative, technical and moral support. I would like to thank Victoria Allen very much for the many productive writing sessions that we shared and Benjamin Peter for all the inspiring discussions about language and enregisterment that gave rise to many ideas and insights expressed in this book. I am also very grateful to Stefanie Weymann-Teschke for her good advice, her great humor and her invaluable support in the final stages of writing my thesis.

I would like to express my thanks to the current and former editors of this series, Alexandra D'Arcy, John Nerbonne and Martijn Wieling, as well as Sebastian Nordhoff for making this publication possible. I also thank all the volunteers that have supported the publication process by proofreading one of the chapters. Special thanks are due to Astrid Richardsen for helping me edit the final manuscript. I am also indebted to all the people involved in creating the newspaper databases *America's Historical Newspapers* and *Nineteenth-Century U.S. Newspapers*.

Finally, I would like to thank all of my lifelong friends for their constant encouragement: Mira, Jule, Uwe, Birte and Tine. I thank my partner Christian for his love and patience and my brothers Lars and Pay Sönke as well as my sisters

### Acknowledgments

Irina and Berit for being the best siblings one could hope for. Above all, I thank my parents Rita and Volkert for always believing in me and for their unconditional love and support. I owe you everything.

# **1 Introduction**

In a letter to the editor which appeared in an Arizona newspaper, the*Tucson Daily Citizen*, on July 25, 1882[52] , a reader asks: "Will you kindly inform me what constitutes the American language?" To this question, the editor replies with only one sentence: "It is the English language with the "H's" in their proper places". While the American reader at the end of the nineteenth century thus seemed to be interested in how the "American language" would be defined by the editor, the question is whether the answer given by the editor should be of interest to today's linguists who aim to study the emergence of American English as a new variety of English.

Two influential models of the process of the emergence of new varieties of English give completely different answers: According to Trudgill's (2004) model of new-dialect formation, the question must be answered in the negative, while according to Schneider's Dynamic Model, which he first presented in an article in 2003 and elaborated on in much detail in his monograph *Postcolonial English: Varieties around the world* (2007), it must be answered in the affirmative. This striking difference between the answers can be explained by the fact that the two models are fundamentally different with respect to an important issue: the role of social factors, especially identity. Trudgill's model is essentially deterministic, attributing a major role to frequency distributions of variants, while in the Dynamic Model "identity constructions and realignments, and their symbolic linguistic expression, are [...] at the heart of the process of the emergence of PCEs [Postcolonial Englishes]" (Schneider 2007: 28).

This issue has caused considerable debate among linguists – a major contribution to it was Kretzschmar's (2014) article, in which he integrated the Dynamic Model and his own theory of language as a complex system by arguing that they complement each other: While the *internal* development of a variety is characterized by random interactions between speakers in a complex system, causing frequency distributions to change, the evolution of the *perception* of these frequency distributions can be accounted for by the Dynamic Model. The problem is, however, that it is an essential claim of the Dynamic Model that perceptions of and attitudes towards language influence structural developments because it "predicts that via language attitudes a speaker's social identity alignment will

### 1 Introduction

determine his or her language behavior in detail" (Schneider 2007: 95). In other words, the speaker's recognition of variants and the concept he or she has of a variety influences his or her own linguistic behavior and thus the structural developments. Given the fundamental nature of the disagreement regarding the way that the process of the emergence of new varieties of English can be modeled, it is clear that this debate needs further attention and this study aims to contribute to it in several ways.

The first aim of this study is theoretical: Based on a comparison of Trudgill's, Schneider's and Kretzschmar's models, paying particular attention to how exactly they conceptualize and define the emerging construct, the *variety* or *dialect*, and how they argue for or against the role of social factors, I will present an argument for distinguishing between structural varieties, perceptual varieties and discursive varieties, and for investigating the emergence of these types of varieties in their own right. The focus of this study is on the discursive variety, and I suggest that its development can be studied by using a theoretical framework which originated in linguistic anthropology and redefines and modernizes the notion of *register* as "a cultural model of action [...] which links speech repertoires to stereotypic indexical values", which "is performable through utterances" and which "is recognized by a sociohistorical population" (Agha 2007: 81). Registers are constructed through processes of *enregisterment*, defined accordingly as "processes and practices whereby performable signs become recognized (and regrouped) as belonging to distinct, differentially valorized registers by a population" (Agha 2007: 81). Studying enregisterment thus directs the focus onto the conceptual or discursive level. It is important, however, that this level is not independent from but instead in a dialectic relationship with the structural level of language use – Silverstein (2003) has described this relationship in much detail by using the notion of *indexical order*. Consequently, the framework is not only useful to investigate the emergence of discursive varieties, but it also provides a theoretical account of the interaction between the discursive level and the structural level. Furthermore, theories of enregisterment also discuss the role of identity in the construction of discursive varieties and I will thus explore how they add to the theoretical basis of Schneider's Dynamic Model. The concepts of indexicality and enregisterment have already been fruitfully applied in sociolinguistics (e.g. by Johnstone et al. 2006 and Eckert 2008, to name two prominent examples), and they also play a role in the new field of discourse linguistics that was delineated in detail by the German linguists Spitzmüller & Warnke (2011). The theoretical framework developed for this study thus integrates approaches from linguistic anthropology, sociolinguistics and discourse linguistics to develop a model of the

construction of discursive varieties, which can then inform a general model of the emergence of new varieties.

The second aim is to apply the model in order to contribute to a description of the emergence of American English as a discursive variety in the nineteenth century. The reason for focusing on American English is that it is the case that is discussed in most detail by Schneider (2007). Even though it is the "best researched postcolonial variety of all", Schneider's account is the first one that is "theoretically informed" (Schneider 2007: 250), and using enregisterment as a theoretical framework not only adds to the theoretical basis of such an account, but it also yields insights that shed further light on how American English evolved. The study is restricted to the nineteenth century because it is in this century that Schneider locates (parts of) the phase of nativization and the entire phase of endonormative orientation, claiming that increasingly positive attitudes towards the American variety also fueled its structural differentiation from British English. Other studies on the history of American English also postulate that it was in the nineteenth century that American English became recognized and/or stabilized as a variety of English (e.g. Algeo 2001a, Bailey 2017, Simpson 1986). By applying the model, I aim to show how tracing enregisterment processes leads to insights about the discursive level and how these insights contribute to a theoretically informed description of the history of American English.

As the application of the model requires a methodological approach, the third aim of the study is to develop a method for studying enregisterment processes in a historical context in a systematic and goal-oriented manner. At the heart of enregisterment are reflexive activities – "activities in which communicative signs are used to typify other perceivable signs" (Agha 2007: 16) – because it is through these activities that speech differences become linked to social differences and that linguistic forms acquire indexical meanings. For example, a person uttering the statement *Saying jolly sounds so posh* uses linguistic signs (one type of communicative sign) to create an object, the lexical item *jolly*, that is typified through the assignment of the social value 'posh'. This constitutes an instance of reflexive activity, in this case a very explicit type of reflexive activity, which is metapragmatic because it conveys information about the pragmatic effect of using the word *jolly* (signaling the characteristic of being posh) and also metadiscursive because the object typified is a part of discourse, that is, of "communicative action in the medium of language" (Johnstone 2018: 2). If reflexive activities, which typify linguistic forms, occur repeatedly, these typifications become social regularities and as such lead to registers as large-scale cultural formations. Reflexive activities can take many shapes and forms, but it is essential that they are by definition observable. This makes it possible for the linguist to study them

### 1 Introduction

systematically. In this study I propose to use a model for analysis that was developed by the above-mentioned discourse linguists Spitzmüller & Warnke (2011): the discourse-linguistic multi-layer analysis (DIMLAN). The model is based on an understanding of discourse that goes beyond the sense given above: It is defined as a linguistic practice through which knowledge is negotiated and constituted (Spitzmüller & Warnke 2011: 53). The central element in this process is the *statement* – it is the smallest unit of discourse. Studying discourses empirically means to study statements by taking into account three levels: the intratextual level (based on the observation that statements typically appear in the context of texts), the transtextual level (because statements and texts are essentially related to each other in discourses) and the level of actors (taking into account that transtextuality is a characteristic that is created through linguistic action by individual people, groups of people or non-personal actors like institutions or political parties). Taking the DIMLAN model as a basis, I approach the empirical study of enregisterment processes in nineteenth-century America by focusing on one metadiscursive genre: newspapers. Newspapers contain texts which contain statements about language (constituting reflexive, metadiscursive activities) and they are electronically accessible and searchable because they have been collected in two large databases, *America's Historical Newspapers* (AHN) and *Nineteenth-Century U.S. Newspapers* (NCNP), which comprise close to 78 million articles. By searching these databases for specific pronunciation respellings (representing phonological forms) and for lexical items, I obtained collections of statements and texts which are related on a transtextual level (because the same linguistic forms become typified). The intratextual level was examined by analyzing each newspaper article qualitatively in order to identify the indexical links which are created between the linguistic forms and social values and personae. By comparing the articles to each other and using quantitative analyses to identify larger evaluative patterns as well as the regions and the time periods in which the respective texts were produced and received, I analyzed the transtextual level and the level of actors. I argue that these analyses provide insights not only into which linguistic forms were indexically linked to the value of 'being American', but also into how, when and where the process of delimiting an American register from other registers proceeded.

The structure of this book reflects these aims. In Chapter 2, I will give an overview of the models of the emergence of new varieties of English, followed by a discussion of the definition of the notion of *variety* and the role of social factors in these models. I will describe the theoretical framework of enregisterment and discuss how this framework and one of its key concepts, indexicality, have been applied and developed further in sociolinguistic research, in the field

of perceptual dialectology and in discourse linguistics. Based on this discussion, I will present a model which captures the relationship between two levels that are relevant in the construction of discursive varieties, namely the structural and the discursive level. The final part of Chapter 2 contains a review of previous research on the history of American English, paying particular attention to how the starting point of the variety is determined and the role that is attributed to the discursive and the structural level in this process.

In Chapter 3, I will describe the methodological approach of this study by explaining why I focus on metadiscursive activity in newspapers, by describing the process of data collection and by motivating my choice of linguistic forms: The phonological forms are /h/-dropping and -insertion, yod-dropping, a lengthened and backened bath vowel, non-rhoticity and a realization of pre-vocalic /r/ as a labiodental approximant, and the lexical forms are *baggage* (in contrast to *luggage*) and *pants* (in contrast to *trousers*). I will also give reasons as to why metadiscourses on the phonological forms will be identified by searching the databases for specific pronunciation respellings (*hinglish*, *noospaper/s*, *dawnce*, *deah*, *fellah*, *bettah*, and several spelling variants of trousers). Against the background of the distinction between the structural and the discursive level established in Chapter 2, previous research on these linguistic forms will be reviewed.

In Chapter 4, I will describe the results of the quantitative and qualitative analyses of the metadiscursive activities surrounding the phonological forms (§4.1) and the lexical forms (§4.2) by focusing on the construction of indexical links between the linguistic forms and social values and social personae. These results will be interpreted in Chapter 5 with regard to the question of what they reveal about the enregisterment, that is, the construction of American English as a discursive variety, in the nineteenth century. I will show that three values are of central importance in this process, 'nationality', 'authenticity' and 'non-specificity', and I will explore each of them in turn. Finally, in Chapter 6, I will discuss the theoretical implications of this study for modeling the emergence of new varieties of English, for writing a theoretically informed history of the development of American English and for general theories and models of language change. The study will conclude with a summary of the central findings of the study in Chapter 7, which show that the statement made by the nineteenth-century editor about the characteristics of the "American language" is highly relevant for giving a theoretically informed account of the emergence of American English.

# **2 The emergence of American English: theories, descriptions, and models**

This chapter lays the theoretical groundwork for the present study. Section 2.1 presents an overview and a discussion of theories of the emergence of new varieties of English and the way that the emergence of American English is described in these frameworks. This leads to the development of the essential argument of the study in §2.2, namely that the concept of *enregisterment* provides an important perspective on the emergence of new varieties and should be incorporated in the theoretical modeling of the process as well as used to complement methodological approaches to studying it. A central aspect of the argument is the distinction between three types of varieties, namely structural varieties, perceptual varieties and discursive varieties, and in §2.3, I develop a model which illustrates the difference as well as the relationship between structural and discursive varieties based on the framework of enregisterment. Section 2.4 directs the focus onto the history of American English again by discussing how the development of the variety has been described in works without an underlying theory of the process of emergence and which role is assigned to the structural and the discursive level in these descriptions.

# **2.1 Theories of the emergence of American English as a new variety of English**

There are several works which have been written on the history of American English during the last 100 years. In general, according to Schneider (2007: 250), American English is the "best researched postcolonial variety of all". However, Schneider (2007: 250) also notes a lack of a "theoretically informed history of the language" and he addresses the need for such a history by describing the emergence of American English within the framework of his Dynamic Model. This study continues this line of research, which is why I focus on describing and discussing mainly three theories and models of the emergence of new varieties and the way that they describe this process in the case of American English. In

### 2 The emergence of American English: theories, descriptions, and models

§2.1.1, I compare and contrast Trudgill's (2004) theory of new-dialect formation, Schneider's (2007) Dynamic Model, which draws heavily on Mufwene's (2001a) theory of the "ecology" of language evolution, and Kretzschmar's view of the emergence of new varieties within his theory of language as a complex system presented in Kretzschmar (2014, 2015a,b). Based on this overview, I will discuss two issues in more detail which are crucial to the debate: the definition of the term *variety* (§2.1.2) and the role of social factors in the emergence of new varieties (§2.1.3). In §2.1.4 I finally discuss important consequences of the theoretical debate which are the foundation for my subsequent argument that studying the emergence of new varieties of English needs a careful distinction between structure and discourse in order to be able to investigate the relationship between these two dimensions.

# **2.1.1 Overview of theories and models of the emergence of new varieties of English**

Trudgill's (2004) theory of new-dialect formation and Schneider's (2007) Dynamic Model have in common that they both identify a set of stages or phases that underlie the emergence or formation of new dialects or varieties. Trudgill, who develops his theory mainly based on his study of the history of New Zealand English, argues that "in tabula rasa colonial situations, dialect mixture and newdialect formation are not haphazard processes" (2004: 26) but a predictable development proceeding from stage 1, which he labels "rudimentary levelling and interdialectal development" to stage 2, labeled "variability and apparent levelling in new-dialect formation" and finally to stage 3 "determinism in new-dialect formation". In the initial stage, adult speakers of different dialects come into contact in a new place. The communication between these speakers can lead to the leveling of minority and very localized variants because, as Trudgill suggests, speakers need to make themselves understood and the use of linguistic forms unknown to a majority of speakers can inhibit this aim. Additional reasons for this rudimentary leveling can also be that speakers accommodate to particularly salient forms or that they react to normative attitudes which speakers have brought with them from their home country (2004: 89-93). Trudgill regards this stage as the least important one because in his view, "adults are only capable of limited amounts of accommodation" and it is not them but children who are largely responsible for the formation of a new dialect (2004: 94). It is not accommodation but language acquisition which leads to the apparent leveling in stage 2. Children do not notice low frequency forms (below a threshold of roughly 10%) and do not acquire them, thereby reducing the number of forms which are available to the next generation

### 2.1 Theories of the emergence of American English as a new variety of English

to a considerable extent. This second generation of children plays a crucial role in stage 3, because from an already reduced number of forms they now select the most common ones. The result is a "final, stable, relatively uniform outcome" in the form of "a stable, crystallized variety" (2004: 113). The final stage is completed by a process called *focusing*, first described and labeled by Le Page & Tabouret-Keller (1985) and defined by Trudgill as "the process by which the new variety acquires norms and stability" (2004: 88). However, it is important that focusing only occurs when a stable set of forms has emerged through stages 2 and 3, which is why the process of new-dialect formation is essentially deterministic, the new dialect being "a *statistical composite* of the dialect mixture" (2004: 123).

Because the theory is mainly based on the case of New Zealand English, Trudgill does not describe or analyze the emergence of American English using his theoretical framework in any detail. He notes several times that it is harder to study the formation of American English than that of New Zealand English because the mixture processes underlying the formation of the variety took place such a long time ago (Trudgill 2004: 2). In general, he argues that American English went through the same stages as all other colonial dialects, and he notes only one possible difference, namely that comprehensibility played perhaps a more important role in the rudimentary leveling of the first stage because the traditional dialects spoken by the settlers were more different from one another than those in colonies which were settled later. Although his study does not include American English in the analysis, it is therefore still relevant because his theoretical claims apply to American English as well and it is one of the goals of the present study to discuss their validity.

The second model, Schneider's (2007) Dynamic Model, also assumes that the emergence of new varieties of English is characterized by "a uniform underlying process [which] has been effective in all these [contact] situations and explains a wide range of parallel phenomena from one variety to another" (Schneider 2007: 4). However, the phases which he identifies and the mechanisms operating in the process are in many respects different from Trudgill's. Table 2.1 summarizes the key parameters of the different phases and shows that Schneider does not view dialect formation as a deterministic process, but that linguistic effects constitute only one parameter in his model. They result from sociolinguistic conditions, which are a consequence of speakers' identity constructions, which are in turn caused by the historical and political context.<sup>1</sup> Accordingly, Schneider (2007: 30) speaks of a "monodirectional, causal relationship" operating between the parameters. While his model predicts that all post-colonial varieties go through all of

<sup>1</sup> Schneider (2007) distinguishes between two speech communities in his model: the Settlers speech community (STL) and the Indigenous speech community (IDG).

### 2 The emergence of American English: theories, descriptions, and models

these phases, he does not claim that he is able to predict the precise linguistic forms of the new repertoire. His model is not deterministic but explicitly dynamic. Consequently, Schneider does not use the term *formation* to describe the emergence of new varieties but *evolution*, and he aligns himself explicitly with theories of language evolution, particularly with Mufwene's (2001a) *feature-pool model*. This model postulates that in a contact situation all linguistic features produced by the speakers are in competition (in a "pool") and speakers select from this pool. Their choices are influenced by the "ecology" of the contact situation comprising linguistic as well as non-linguistic factors, such as the demographic and political situation and social factors, particularly identity constructions (which identity speakers want to express) and role alignments (which other speakers they want to align with). In Schneider's Dynamic Model, "identity constructions and realignments, and their symbolic linguistic expression, are also at the heart of the process of the emergence of PCEs [Postcolonial Englishes]" (Schneider 2007: 28). The mechanism that operates in this process is accommodation (Giles 1984), but in contrast to Trudgill, who regards accommodation as an "automatic consequence of interaction" which is "not necessarily driven by social factors such as prestige or identity" (Trudgill 2004: 28), Schneider emphasizes the social nature of the process:

Speakers who wish to signal a social bond between themselves will minimize existing linguistic differences as a direct reflection of social proximity: they will tend to pick up forms used by the communication partner to increase the set of shared features and to avoid forms which they realize are not used by their partner and might thus function as a linguistic separator. (Schneider 2007: 27)

This difference is indicative of the fact that the role of social factors is a matter of considerable debate in theories of the emergence of new varieties and it will therefore be discussed in more detail in §2.1.3.

With regard to the case of American English, Schneider considers it to be "an almost unique opportunity to observe the entire developmental cycle in hindsight" (2007: 251) because it is the oldest and also the best researched variety of all postcolonial varieties. I provide a brief summary of his analysis here by focusing especially on those aspects which illustrate his central thesis that social factors, especially identity constructions, have linguistic effects. In the first phase (roughly from 1587 to 1670), evidence of identity constructions is scarce, but Schneider finds it very likely that the early settlers still perceived themselves as Englishmen (Schneider 2007: 258). Consequently, it is the degree of mixture

Table 2.1: Developmental phases of Schneider's Dynamic Model (from Schneider2007:56)


### 2.1 Theories of the emergence of American English as a new variety of English

### 2 The emergence of American English: theories, descriptions, and models

of speakers coming from different regions and speaking different dialects which has the most effect on linguistic developments. Regions in which the population mixture was highest (as in the case of the Quakers in Pennsylvania) exhibited the highest degree of koinéization, defined by Schneider (2007: 35) as the "emergence of a relatively homogeneous "middle-of-the-road variety"" based on a process in which "speakers […] mutually adjust their pronunciation and lexical usage to facilitate understanding". In regions with culturally and linguistically more homogeneous settler groups (particularly New England, the South and the Appalachian Mountains), less koinéization occurred. This explains the present-day situation, in which the most distinctive dialects are found in the South and in the East, while the mainstream American variety is located in the Midland, the West and the North (Schneider 2007: 261–262). At the same time, Schneider (2007: 262) argues that social similarities between the settler groups in New England and in tidewater Virginia, namely their middle-to-upper-class background and close ties to the home country, were responsible for the fact that southern and eastern dialects share a number of features, for example lack of rhoticity, yod-dropping and lexical forms like *piazza* 'veranda'.

In the second phase (ca. 1670-1773), Schneider (2007: 265) distinguishes two English-speaking groups with different identity constructions. One group was of higher social status and lived on the coast and the other group was of lower social status and lived in more inland regions. While the first group still firmly identified with England, the second one adopted an "English-plus" colonial identity, which was influenced by their more frequent contact with other cultural groups and their "frontier experience" (Schneider 2007: 265). Additionally, Schneider argues that non-English speaking groups had a split identity because even though they wanted to adjust in America and leave problems in their home countries behind, they also wanted to retain their cultural and linguistic heritage (2007: 266). And lastly, African groups were torn between forces to adjust and the desire to resist these forces and maintain their cultural identity. This combination of identity constructions resulted on the one hand in a stable exonormative orientation (as predicted by the Dynamic Model) among the high-status social group and on the other hand in a bilingualism or multilingualism among groups coming from non-English-speaking countries and in a variable sociolinguistic situation for African Americans who sometimes had extensive contacts with white speakers of English and sometimes primarily intra-ethnic contact without an opportunity to acquire English (2007: 266–269). In terms of linguistic effects, Schneider (2007: 269–273) finds a high degree of linguistic homogeneity and lexical borrowings as well as innovations ("Americanisms"). With regard to homogeneity, he notes that this is of course not to be seen in absolute terms, i.e. the complete

### 2.1 Theories of the emergence of American English as a new variety of English

absence of variability. It is rather the case that leveling processes took place as the result of mixing and koinéization, but "cultural and linguistic peculiarities" were retained as well, and English innovations were adopted as a result of the prevailing exonormative orientation. An important example is rhoticity:

The most interesting case in point is postvocalic /r/, a sound which was pronounced even in southern British English well into the eighteenth century and disappeared only then, as in modern RP. In other words, the r-lessness of New England and the South must have developed in America, modeling English linguistic fashion – a strong indication of the exonormative linguistic orientation of colonial America. (2007: 271)

The term *Americanism*, coined by John Witherspoon in 1781 (see Schneider 2007: 272), is another indicator of the exonormative orientation because the lexical items labeled as such were usually evaluated negatively because they did not conform to a British norm. With regard to African American English Schneider finds that there is no evidence of it in the colonial period, although he also notes that "there was room for the development and retention of ethnic speech markers, for the development of linguistic means to signal a non-white, and possibly subliminally counter-European, ethnolinguistic identity" (2007: 268).

The crucial phase in the emergence of American English is the third phase, the nativization phase, which Schneider (2007: 273) dates from ca. 1773 to 1828/1848. According to him, the "birth of American English as a concept and as a variety falls into that period" (2007: 276) – as an effect of the political independence which was accompanied by changing identity constructions. An American nationalism replaced the orientation towards England and Schneider makes it clear that the "close nexus between political events and linguistic developments (via identity rewritings), and the causal role of the former for the latter, are undisputed" (2007: 275). He cites several authors who establish the claim for a language separate from Britain, including Noah Webster's famous words: "as an independent nation, our honor requires us to have a system of our own, in language as well as in government" (Webster 1789: 20, quoted in Schneider 2007: 277). There were also voices resisting the call for linguistic independence by continuing to favor British norms (forming a "complaint tradition") but Schneider argues that "[i]n quantitative terms […] nativization made the balance tip toward the former position" (2007: 277). Schneider postulates a clear relationship between the positive evaluation of linguistic difference and actual structural developments:

[T]he period of structural nativization was the one during which effects inhibiting divergence disappeared and, in contrast, linguistic differences be-

### 2 The emergence of American English: theories, descriptions, and models

came actively promoted or at least positively evaluated. Westward expansion, then, brought an increasingly appreciative attitude toward down-toearth speechways and hence strengthened another powerful factor promoting linguistic nativization. While of course British and American English still have a lot in common and linguistic continuity has also been important, differences between the two major varieties of English kept increasing. (2007: 278)

As evidence for these increasing differences he gives examples of lexical borrowings and innovations, stressing the creativity of word-formation processes characteristic of this phase, which found expression especially in many conversions (2007: 279). On the level of grammar, he finds an "almost endless" number of innovations at the lexico-grammatical interface, e.g. *different than* (vs. *from, to*) (2007: 280). What is also important in this phase is spelling because several differences to British English did not emerge from use (like differences on other levels) but through deliberate language planning and have acquired a symbolic significance in public discourse (e.g. <-or>/<-our>, <-ize>/<-ise>) (2007: 281). The phonological level is the only level on which Schneider finds that hardly any evidence is available and if it is, it is difficult to interpret. Schneider consequently does not give any examples on how structural nativization proceeds phonologically (2007: 279). He does, however, provide more information on how he conceptualizes the process by arguing that

differences between varieties of English, British and American in the present case, not only consist of the ones frequently observed, documented and listed, but they encompass an infinitely larger set of habits and constructions which are hardly ever explicitly noted, most of which are associated with particular lexical items. […] This suggests that structural nativization operates inconspicuously but highly effectively, affecting frequencies and co-occurrence tendencies of individual words and constructions more than anything else. (2007: 282)

This shows that he regards actual structural nativization as proceeding below the level of awareness, but it also raises the question of the exact nature of the relationship between consciously expressed attitudes and evaluations and unconsciously proceeding structural changes, a question that remains open in Schneider's account.

The fourth phase is dated from 1828/1848 to 1898 and, in line with the Dynamic Model, it is characterized by an endonormative orientation based on "a

### 2.1 Theories of the emergence of American English as a new variety of English

new type of national self-dependence and a national pride based on local, American, achievements" (2007: 283). In this phase, the entire continent was settled and controlled by Americans, and this "achievement", at the cost of Native Americans who were killed and forced away from their lands, was not only a source of national pride but also led to "a second heightened phase of koinéization" (2007: 290) resulting in a high degree of uniformity. The uniformity was strengthened by the codification of American English in Webster's *An American Dictionary of the English Language* (1828) and Bartlett's *Dictionary of Americanisms, a Glossary of Words and Phrases usually regarded as peculiar to the United States* (1848). At the same time, regional and social variability obviously continued. Schneider (2007: 289) observes that many literary works which became part of a distinctive American literary canon employed representations of regional and social dialects, but he does not explore the relationship between the literary interest in linguistic variation at a time characterized by the codification of a uniform American variety further. This is something that I will do in the present study, which will show the relevance of this observation to the emergence of American English. Another relevant point that Schneider makes is that the development does not always proceed continually in one direction, but that there can be breaks or even returns to an older phase, as evidenced by "a purist, pro-English movement, which […] gained momentum after the Civil War and in the 1870s and 1880s, after endonormative stabilization" (2007: 288). So exonormative orientations, based on different identity constructions, did not cease to exist, even in a phase of endonormative stabilization, and this study will also shed some light on the interplay between these two types of orientations in the nineteenth century and also on the ways that they interact with regional and social variability as well as with national uniformity.

The starting point of the fifth phase, which continues until the present day, is 1898. Schneider regards the Spanish-American War in 1898 as a turning point because it was the first war fought by the whole unified nation against another power, which strengthened feelings of national unity which were then a prerequisite for cultural fragmentation under the umbrella of the nation. "American society is being transformed into a multicultural mosaic, and this process is mirrored by the emergence of distinct varieties of English, each associated with different identities" (2007: 294). A very important aspect is that Schneider does not postulate that dialect diversity occurred only in the twentieth century, but he finds that it has always existed. In contrast to earlier centuries, however, he finds that the twentieth century is marked by a "*socially indicative* dialect diversity, an ethnic and regional fragmentation of the population along linguistic lines in perception and production" [emphasis mine] (2007: 296). The crucial difference is therefore that diversity comes to index social identities:

### 2 The emergence of American English: theories, descriptions, and models

[T]his diversification happened because the various regional, social, and ethnic groups recognized the importance of carving out and signaling their own distinct identities against other groups and also against an overarching nation which, while it is good to be part of, is too big and too distant to be comforting and to offer the proximity and solidarity which humans require. (2007: 296)

This had an impact on linguistic structure in that dialect differences became more pronounced and, in Schneider's words, more "strictly compartmentalized" (2007: 296). As for the nativization period, he stresses that the process operated subconsciously and that it was not the result of an intentional act. As evidence for the developments in the fifth phase he cites several sociolinguistic studies, e.g. Labov's investigation of Martha's Vineyard (Labov 1972 [1963]) and Wolfram & Schilling-Estes's (1996, 1997) study of the island Ocracoke, which illustrate how people on the islands used traditional regional variants to symbolize and demarcate their island identity against outsiders and to ensure that they are not absorbed in the mainland group identity. Next to these endangered local varieties he also describes the development of further varieties and their connection to identity constructions and realignments: Southern English, a Northern English marked by the Northern Cities Shift, ethnic speech forms in the European-American groups, Native American English(es), African American English, Chicano English and other Hispanic varieties, Cajun English, Hawaiian Creole and Asian Englishes. To give an example, he cites Tillery & Bailey's (2003) research which finds that there were two periods of great social change in the South, the first after the Civil War and the following Reconstruction period (marked among other things by immigration of northern Americans) and the second around World War II (marked especially by urbanization), and in both cases these social developments were accompanied by significant linguistic changes. For example, after World War II "the linguistic expression of a new, modern Southern identity was shaped, [which] affected both "Traditional" and "New" Southern features" (2007: 299). So non-rhoticity and yod-retention were for example features which were associated with traditional, rural, antebellum culture, whereas rhoticity and yod-dropping have come to symbolize the "New South" (2007: 299).

All in all, Schneider's (2007) Dynamic Model stands in stark contrast to Trudgill's (2004) model of new-dialect formation. Trudgill's stages are very much focused on linguistic developments while Schneider's developmental phases emphasize the close connection between the historical situation, social factors and linguistic developments. The type of evidence that both linguists draw on reflects this difference. Trudgill relies largely on data from the *Origins of New Zealand*

### 2.1 Theories of the emergence of American English as a new variety of English

*English* (ONZE) project and on dialectological research on nineteenth-century British English dialects. While Trudgill therefore keeps a strong focus on linguistic data, Schneider cites an abundance of research not only on linguistic developments in seventeen postcolonial varieties of English but also on social developments in order to support his model. With regard to linguistic effects, his case studies provide many examples, but they do not offer as systematic and detailed an overview as Trudgill's analysis of New Zealand English. Instead, Schneider emphasizes the common characteristics of sociolinguistic developments, a dimension that is almost completely neglected in Trudgill's model. These contrary views illustrate the need for more research on the emergence of new varieties of English.

A third theory has been developed by Kretzschmar (2014) as part of his theory of speech as a complex system. In contrast to Trudgill (2004) and Schneider (2007), he does not postulate the existence of separate phases that lead to the emergence of a new variety because this view is not compatible with his theory of language. A complex system is "a system in which large networks of components with no central control and simple rules of operation give rise to complex collective behavior, sophisticated information processing, and adaptation via learning or evolution" (Kretzschmar 2014: 143, citing Mitchell 2009: 13). Applied to speech, these components are linguistic forms, variants realizing a variable, and what is "truly stable and systematic about speech" (Kretzschmar 2014: 151) is that the token frequency distribution of these forms is nonlinear, leading to a typical A-curve when graphed on a chart (Kretzschmar 2014: 147). This nonlinear distribution occurs at different levels of scale, for example at the level of an individual, a community, a larger region or a nation, and variants which occur at the top at one level of scale can occur in the tail of the curve at another level of scale. Consequently, linguists who identify a variety traditionally do so by identifying the top-ranked variants which occur on a specific level of scale, for example at the level of nation, so that American English, for example, is considered to consist of the variants which occur at the top frequency ranks at the national scale. Kretzschmar (2014: 151), however, considers it a mistake to focus only on these top-ranked variants in linguistic analysis because the low-frequency variants are in fact highly relevant in language change and in the emergence of new varieties as well. Change occurs because speakers interact and these interactions can result in changes in the frequency distribution of variants. This in turn leads to new variants ranked at the top of the frequency curve, which can then be described as a new variety at some level of scale. Kretzschmar (2014: 151) supports Schneider's view that the emergence of varieties is *not* a deterministic process:

### 2 The emergence of American English: theories, descriptions, and models

Random interactions between speakers may eventually promote variants at lower frequencies to the top rank, and vice versa, so there is no fixed relationship between input frequencies (say, from settler, indigenous, or adstrate languages) and what will become most common in a new variety.

He argues that not only language is involved in interactions but human perceptions as well and human perception is not restricted to what is most frequent but also encompasses reactions to the use of variants. In this context, Kretzschmar uses the concept of *positive feedback*, which he regards as equivalent to Le Page & Tabouret-Keller's (1985) concept of *focusing*:

The idea of feedback recognizes that the information content in speech is not just the functional message of some utterance or piece of writing, but also an evaluation of who says what when. (Kretzschmar 2014: 151)

Judging by the reactions of other speakers to the use of linguistic forms, speakers evaluate the success of their use and adapt accordingly. Kretzschmar considers this a better view of Schneider's and Mufwene's concept of a *selection process* because, in his view, variants are not selected (and rarely lost). Instead, they simply become more or less frequent through "massive numbers of random interactions between speakers" and "the perceptions of the human agents using language" (2014: 152). Unlike Trudgill, who uses *focusing* to explain the stabilization of the forms of the new variety (basically the last step in the process), Kretzschmar assigns a much more central role to focusing because in his view it explains why the process is not deterministic (i.e. the predictable result of input frequencies). It is noticeable, however, that he does not elaborate on how exactly focusing proceeds and which mechanisms are operative in the process of giving and receiving feedback and evaluating the success of linguistic forms. However, he points out that his model and Schneider's Dynamic Model complement each other precisely because in his view Schneider describes "the evolution of the new society's perceptions" and not "the internal linguistic history of a new variety" (2014: 157). By relegating Schneider's "new varieties" to the level of perception of usage (to be distinguished from actual usage), he questions the traditional definition of *variety*, of the object whose emergence is controversially modeled, which is why in §2.1.2 I will take a closer look at the understanding of the term in linguistics in general and in the models described in this section in particular.

Kretzschmar (2015a) describes the implications of his model for an account of the emergence of American English. For the early phases of settlement, he describes the linguistic situation as a "pool of linguistic features collected from a

### 2.1 Theories of the emergence of American English as a new variety of English

radically mixed settlement population" (2015a: 251), a conceptualization that is in line with Mufwene's feature pool but not with historical accounts on American English such as Fischer's (1989) *Albion Seeds*, which describes the culture and language of the different regions of settlement as having been transplanted from the respective regions in Britain and remaining fairly homogeneous. In this situation, order emerged in the form of the nonlinear distributions characteristic of complex systems. And owing to the scaling property of complex systems, these distributions occurred at different levels of scale, so that Kretzschmar speaks of American English already in this early phase:

Right from the beginning somewhat different sets of variants emerged as top-ranked elements in different localities. Also right from the beginning, a particular set of variants emerged at the highest level of scale, American English. (2015a: 257)

So in contrast to Schneider (2007), who considers American English to nativize in the late eighteenth and early nineteenth century, Kretzschmar suggests that an American variety, conceptualized as a set of top-ranked elements at the highest level of scale, has already been present from the seventeenth century onwards. He views the nativization that Schneider describes as located on the level of perception (people started to notice and describe differences between American and British English) and argues along the same lines that eighteenthcentury comments on uniformity are equally a result of perception and not to be taken as evidence for an actual colonial koiné (Kretzschmar 2015a: 258–259). While Schneider (2007: 270) is also critical of strong versions of the koinéization hypothesis (he cites Dillard 1975 as the strongest one), he nevertheless assumes a "remarkable degree of linguistic homogeneity in the colonies" in phase 2, which is in contrast to the diversification of phase 5. Kretzschmar also disagrees here by stating that diversification also occurred from the seventeenth century onwards, when "noticeable differences, both between American regions and between American and British English" (2015a: 257) emerged. In order to support this view, he gives examples of linguistic forms which were present then and are still associated with specific regions, e.g. non-rhoticity, lexical items like *chunks* and *tote* and grammatical forms like *hadn't ought*. The diversification just became *more* noticeable as time progressed (2015a: 259). Even though Kretzschmar does not postulate the existence of developmental phases in general, he notes a major difference between the eighteenth century and the nineteenth century in the case of American English. While the great population mixture in the early settlements and in the eighteenth century led to the creation of new and independent

### 2 The emergence of American English: theories, descriptions, and models

patterns (2015a: 259), the nineteenth century was rather marked by an extension of these complex systems from the east to the west as a result of westward migration. Groups of settlers who already lived in western parts of the country, especially Scotch-Irish settlers in the Appalachian mountain regions, could contribute to these complex systems, and Kretzschmar cites Montgomery's (1989, 1991, 1997) findings that indeed some Scotch-Irish variants have been retained in these areas, but overall, he finds that Atlas data shows that "major patterns created by historical east-west settlement largely persist" (2015a: 261). Nevertheless, change always occurs in complex systems, but Kretzschmar regards it as located rather at lower levels of scale (neighborhoods and cities) and not at larger regional levels of scale (2015a: 261).

This overview of three models of the emergence of new varieties illustrates that linguists are far from achieving a consensus of opinion on the common, underlying operations and mechanisms behind this process. I argue here that some differences between the models are a result of different conceptualizations and definitions of the term *variety* and that it is important to clarify what the term refers to in order to be able to analyze the emergence of new varieties and discuss and ultimately test the models proposed by Trudgill (2004), Schneider (2007) and Kretzschmar (2014, 2015a). The analysis of the meaning(s) of the term *variety* is therefore going to be the subject of §2.1.2, before I will pay detailed attention to the role of social factors in §2.1.3 and conclude with a final discussion of the models in §2.1.4.

# **2.1.2 Definition of the term** *variety*

A first step in any discussion on how new varieties emerge must be to define what precisely the term *variety* refers to and how a "new" variety can be distinguished from an "old" one. At first, it needs to be noted that the different theories and models foreground different terms. Trudgill (2004) speaks predominantly of *dialects* and particularly of *colonial dialects*, while Schneider (2007) prefers the term *variety* and refers to the varieties that emerge in the different places as *postcolonial varieties* or *Postcolonial Englishes*. 2 It is often indicated that *dialect* and *variety* are used more or less synonymously, with *variety* being a newer and more neutral term. Meyerhoff (2011: 32), for example, defines variety as a "[r]elatively neutral term used to refer to languages and dialects" which

<sup>2</sup> It is important to note that each author uses *both* terms, *dialect* and *variety*, in their books but foreground different terms by choosing them as the main term for the newly formed or emerging entities.

### 2.1 Theories of the emergence of American English as a new variety of English

"[a]voids the problem of drawing a distinction between the two, and avoids negative attitudes often attached to the term *dialect*". Chambers & Trudgill (1998: 5) make a slight distinction between the two terms. They write in their introduction to dialectology that they "shall use 'variety' as a neutral term to apply to any particular kind of language which we wish, *for some purpose*, to consider as a single entity" [emphasis mine], while they regard *dialect* as a more particular term which refers to "varieties which are grammatically (and perhaps lexically) as well as phonologically different from other varieties". It can be inferred from this distinction that Trudgill (2004) foregrounds the term *dialect* in his theory because the main criterion he uses to distinguish a new variety or dialect from an old one is the criterion of structural distinctiveness. He argues that in order to explain the formation of a new dialect, one has to "first decide what the distinctive characteristics of New Zealand English are" (Trudgill 2004: 31). The distinctiveness is established in relation to British dialects, the starting point being a mixture of dialects spoken by the parents of the first New-Zealand-born Anglophones, and the end-point the dialect spoken by the second generation of Anglophone children in New Zealand. Differences which emerged afterwards (since 1890) are not included in the analysis because they are the result of changes occurring only after New Zealand English had been formed (Trudgill 2004: 32). So in order to distinguish and define a "new" variety of New Zealand English against "old" varieties of British English, he looks for evidence of a new set of linguistic features, distinct from other older sets of features, based on two main sources: studies on the history of the English language, especially dialectological research on nineteenth- and twentieth-century English, and the ONZE project, comprising recordings of New Zealanders made by the National Broadcasting operation of New Zealand between 1946 and 1948, which provide insights into the second stage of the formation process because the informants represent the first generation of children born in New Zealand (Trudgill 2004: 33). The linguistic features he analyzes are almost exclusively phonological and located on the segmental level, and the analysis rests heavily on frequencies of use. What distinguishes stage 3, the stage at which he postulates the existence of a distinct New Zealand variety, from stage 2 is that the variety is now characterized by uniformity and stability (Trudgill 2004: 113). Uniformity is achieved because majority variants have survived and minority variants have disappeared, and stability is achieved through focusing. It appears therefore that while the primary criterion in defining the new variety is the structural distinctiveness of a uniform set of features, stability, achieved through focusing, is at least a secondary criterion, which is applicable only after a new set of features has formed. It is noticeable, however, that Trudgill (2004) does not analyze the focusing process in his study on New

### 2 The emergence of American English: theories, descriptions, and models

Zealand English, which again highlights the relative unimportance of the stability criterion.

Schneider (2007) also focuses on the criterion of structural distinctiveness in his model. He speaks of "the birth and growth of structurally distinctive PCEs" (Schneider 2007: 45) and describes a PCE as "a new language variety which is recognizably distinct in certain respects from the language form that was transported originally, and which has stabilized linguistically to a considerable extent" (Schneider 2007: 51). In the phase of structural nativization, the degree of difference to the former input varieties increases the most. He states that "this stage results in the heaviest effects on the restructuring of the English language itself; it is at the heart of the birth of a new formally distinct PCE" (Schneider 2007: 44). S-curves, which typically characterize linguistic changes, have a phase of rapid increase of a variant in the middle of the development over time and this is where Schneider locates the phase of structural nativization. In line with Trudgill (2004), Schneider (2007: 51) regards stability as a second characteristic of a variety, and he argues that it is achieved during a phase which can be seen as corresponding to the later part of the S-curve. However, in stark contrast to Trudgill (2004), speakers' perceptions play a role in his conceptualization of a variety as well. He writes that "regional speech differences emerge, stabilize, and become recognizable in the public mind" (Schneider 2007: 9), which shows that in addition to structural difference and stability, public recognition is also a factor to be dealt with in determining what a variety is and how it emerges. In the case study on American English, he finds that the "birth of American English as a concept and as a variety falls into that period [the period between ca. 1773-1828/1848]" (Schneider 2007: 276). In this statement he explicitly distinguishes between the *concept* of a variety and the variety itself, but he does not elaborate on this distinction any further, which is problematic because he assumes a direct relationship between the development of structural differences and people's perceptions and attitudes:

[T]he period of structural nativization was the one during which effects inhibiting divergence disappeared and, in contrast, linguistic differences became actively promoted or at least positively evaluated. Westward expansion, then, brought an increasingly appreciative attitude toward downto-earth speechways, and hence strengthened another powerful factor promoting linguistic nativization. (Schneider 2007: 278)

So in Schneider's (2007) model, speakers' perceptions and attitudes, which can be seen as formative of the *concept* of a new variety, did not only play a

### 2.1 Theories of the emergence of American English as a new variety of English

role in the phase of endonormative stabilization, but they are also supposed to influence structural nativization, that is the emergence of a *structural* variety. Schneider (2007: 94) argues that people usually focus on a few salient distinctive forms when they perceive and evaluate a variety, but that "those properties of a variety which seem specifically distinctive [are] quantitative tendencies of word co-occurrences, recurrent patterns, speech habits, prefabricated phraseology". Based on this, a conceptual variety can be defined as consisting of a small set of salient linguistic features which are recognized by speakers as distinct and which attract some sort of evaluation. A structural variety can be defined as consisting of a large set of features which make the variety structurally different from another set of features.

Nevertheless, Schneider's idea of what a variety is, and how the concept of a variety and the structure of a variety are related, remains rather vague. Kretzschmar (2014, 2015a,b), on the other hand, discusses the term *variety* in more detail and he is fairly critical of traditional understandings of the term. As pointed out in §2.1.1, he relegates Schneider's (2007) Dynamic Model to the level of perception only and argues that this is the case for most descriptions of varieties:

What actually makes "new varieties" of English or of any other language, in the sense that we usually mean in linguistics, is that linguists from time to time choose to record their perceptions of the usage of some population of speakers. (Kretzschmar 2014: 157)

These perceptions include only the top-ranked variants of the complex system of speech at a specific level of scale which makes the varieties described by linguists "idealized abstractions" (Kretzschmar 2014: 156) which linguists and also lay people are always interested in, but which should not be confused with linguistic reality. According to Kretzschmar (2014), this reality should best be understood as a complex system:

New varieties are not just something to be associated with former colonies; they are emerging all around us every day, as speakers of English form new groups in local neighborhoods, communities of practice, social settings, and new places around the globe in many places besides colonial settings. It is not a process that happens once and is done. The complex system of speech continues to operate, and new order emerges from it all the time. (Kretzschmar 2014: 157)

He does not address Schneider's (2007) distinction between structural variety and conceptual variety, but claims that Schneider views varieties only as

### 2 The emergence of American English: theories, descriptions, and models

"identity-driven discourse constructs" (Schneider 2007: 51, cited in Kretzschmar 2014: 157), a claim which is clearly not justified given that the nativization phase in the Dynamic Model rests crucially on structural differentiation.<sup>3</sup> In an earlier article, Kretzschmar & Meyer (2012) make a distinction similar to the one between structure and concept. They argue that *natural* language varieties need to be distinguished from *ideational* language varieties. They discuss the case of Standard American English and argue that it is an *idea* which does not have an empirical basis in language use. Instead, it is "just another manifestation of a particular culture" (Kretzschmar & Meyer 2012: 156). By calling it an "idea", they relegate Standard American English to a level that corresponds to Schneider's conceptual level, so that, in the end, we can derive a threefold distinction between a) the linguistic reality, which is a complex system, b) the perceptual variety, which is based on the linguists' perceptions of the top-ranked variants on a specific level of scale (e.g. individual, local or national) and c) the conceptual/ideational variety, which is based on people's idea about language. With regard to the ideational variety, it is important to note that even though it does not directly reflect language in use, it is "not a myth" but "a very real […] construct" for its speakers (Kretzschmar & Meyer 2012: 143).

Kretzschmar is not the only one who critically discusses the notion of *variety* in general and in the context of new variety formation in particular. Leimgruber (2013b) also calls for "rethinking the concept of 'geographical varieties' of English". He argues that the starting point for defining a variety, especially in the context of World Englishes, is usually a geographical and political but *not* a linguistic concept:

It may seem impractical to completely do away with such a useful concept as the variety, which has for so long been the basic unit of analysis in many fields of linguistics, including World Englishes. It remains, however, that the concept is often under-defined in works setting out to describe such varieties – terms like 'Singapore English', 'Malaysian English', 'Welsh English', etc., are taken for granted because, after all, they contain a geographical component everyone can relate to. The actual linguistic form of the 'variety' is then described post hoc, with the analytical unit 'variety' conditioning the analysis. (Leimgruber 2013b: 6)

<sup>3</sup> In fact, Schneider (2007: 51) writes that the homogeneity that is often emphasized in descriptions of new varieties is an "identity-driven discourse construct" and not the variety itself. Still, it makes sense to ask to what extent the conceptual variety in Schneider's (2007) Dynamic Model can be defined as a discourse construct and which role identity plays in the process of its emergence. I will address these questions in more detail in §2.2 and §2.3.

### 2.1 Theories of the emergence of American English as a new variety of English

This discussion shows that a clearer definition of what kinds of varieties there are, how they are related and how they can be identified and described is necessary. A recent contribution by Pickl (2016) is helpful in this regard. He draws on the terms *emic* and *etic* to distinguish different types of varieties. Emic dialects are "cognitive concepts of the speakers whose speech is at the same time the object of linguistic investigation" (Pickl 2016: 78) and etic dialects are based on objective linguistic analyses of speech productions. Regarding the relationship between emic and etic dialects he states that

There is no reason why the fundamental linguistic concept of a dialect variety should differ from the folk linguistic concept. In other words, scholarly or etic ideas about geolinguistic entities can and should have the same principal structure as lay persons' implicit ideas about dialects in space while being based on transparent – and, as far as possible, objective – criteria that are not derived from the speakers' ideas, but from scientific reasoning. (Pickl 2016: 78)

The shared "principal structure" is a prototypical one, with the important characteristic of fuzzy category boundaries and linguistic features which have different degrees of typicality. The distinction between emically and etically defined categories is analogous to Schneider's (2007) distinction between a conceptual and a structural variety, but the similarities and differences are elaborated on in more detail. With regard to the emic category, Pickl (2016) refers to perceptual dialectology as a research area, while his own study focuses on the etic category. He defines it more precisely by using a definition by Berruto (2010):

The tendential co-occurrence of variants gives rise to linguistic varieties. Therefore, a linguistic variety is conceivable as a set of co-occurring variants; it is identified simultaneously by both such a co-occurrence of variants, from the linguistic viewpoint, and the co-occurrence of these variants with extralinguistic, social features, from the external, societal viewpoint. (Berruto 2010: 229, cited in Pickl 2016: 79)

In order to identify sets of co-occurring features, Pickl argues for using statistical methods, more specifically factor analysis, which he regards as superior to other statistical methods like cluster analysis, bipartite spectral graph partitioning and multidimensional scaling (2016: 80–83). He supports this view by conducting an analysis on dialect areas in Bavarian Swabia based on data from the dialect atlas *Sprachatlas von Bayerisch-Schwaben* (SBS, König 1996–2009). The

### 2 The emergence of American English: theories, descriptions, and models

variability in the region is reduced to 16 factors which account for 62.21% of the variance in the data. Each factor stands for a recurring pattern of linguistic variants which is particularly strong in a geographical region (and within the region, the locations exhibit different degrees of typicality for the pattern) so that in the end, a combined factor map can be constructed which shows the prototypically structured dialect areas in geographical space. The point is that he finds a way for identifying dialect areas based on a statistical analysis of the data only and without recourse to either lay people's or linguists' subjective judgements. Attempts in this direction have been numerous (they belong to the field of dialectometry), but they did not draw on prototype theory or were skeptical of the existence of dialect areas in general (e.g. Kretzschmar 1996). Pickl's suggestion therefore combines the insight that emically and etically defined varieties share a prototypical structure with a suggestion for describing the structure of etic varieties by means of objective criteria, thereby minimizing the influence of ideas on the outcome of the analysis.

Pickl's (2016) approach counters Kretzschmar's (2014) criticism that the varieties perceived and described by linguists are restricted to top-ranked variants and that all those variants which occur at lower frequencies are ignored and considered irrelevant because he aims at identifying abstract varieties in a way that takes the complexity of speech more strongly into account than other methods. For example, Pickl (2016: 81) states that "it is […] impossible for a cluster analysis to come up with anything more subtle than global, exclusively dominant areas; subordinate, non-dominant areas that are determined by smaller numbers of features cannot be identified by cluster analysis", which is why he proposes factor analysis as a statistical method instead. As his goal is to identify dialect layers overlapping in space, each layer consisting of "congruent distribution areas of co-occurring linguistic forms" (Pickl 2016: 79), he tries to achieve an abstraction that is closer to the linguistic reality than traditional analyses which identified discrete dialect areas based on the presence or absence of distinctive linguistic features.

Pickl's (2016) focus on geographical varieties leaves open the question of how an etic variety can be described which is not only based on regional distributions of data, but also on social ones. It is conceivable, however, that his approach can also identify social varieties if sufficient data are available. Leimgruber (2013b: 6) also points out the need for more data to carry out quantitative studies which complement qualitative analyses conducted by sociolinguists like Blommaert (2010), who do not consider it important anymore to identify and describe varieties at all but are primarily interested in describing and explaining how speakers draw on linguistic resources in their social interaction with others

### 2.1 Theories of the emergence of American English as a new variety of English

(see §2.2.2 for details). In general, Pickl's (2016) approach to defining and identifying structural varieties takes the criterion of structural distinctiveness more seriously than other approaches.

To summarize, it seems that linguists distinguish several types of varieties. Based on the discussion above, I propose a distinction between the linguistic reality, conceptualized as a complex system, and three abstract types of varieties: structural varieties, perceptual varieties and discursive varieties, as illustrated in Figure 2.1.

Figure 2.1: Types of varieties described by linguists

While the linguistic "reality", i.e. the total number of forms used by speakers in particular regional, social and situational contexts, is a complex system as described by Kretzschmar (2014, 2015a,b), it is possible to identify and describe more abstract patterns of co-occurring forms. Structural varieties are tendential co-occurrences of variants in a particular place (and potentially also in a particular social situation or in relation to social factors). They are determined based on production data, collected and analyzed by the linguist using statistical measures and Pickl's (2016) prototype approach is a very convincing one because it allows for fuzzy category boundaries and does not just identify one dominant pattern but also its overlaps with other non-dominant patterns. Perceptual varieties are tendential co-occurrences of forms in lay people's perceptions of language variation. They are determined based on perception data, again collected and analyzed by the linguist. Here, perception in a cognitive sense (what is perceivable by the senses) and in a conceptual sense (what is perceivable because people have a cultural concept of the pattern in mind) are both included and it is convincing to assume a prototypical structure of these perceptual categories as well. Discursive varieties take the ideational character of varieties into account, which is for example discussed in Kretzschmar & Meyer (2012). They are tendential co-occurrences of forms in speakers' discursive constructions of patterns of variation and it is

### 2 The emergence of American English: theories, descriptions, and models

the aim of §2.2 and §2.3 to show how the concept of *enregisterment* is useful in defining and investigating discursive varieties.

It is important to note that the different types of varieties overlap. In determining structural varieties linguists might be influenced by their perceptions already in the process of collecting data, so it is questionable whether a purely fact-driven abstraction of co-occurrence patterns is even possible.<sup>4</sup> At the same time, perception is of course related to structure in that highly frequent variants could be more easily perceivable than low-frequency ones. It is, however, more than doubtful that frequency of use is the only factor influencing perception. If variants are often the focus of *discourses* on language, for example, they are more likely to be perceived by speakers even if they are *used* infrequently. And lastly, perception influences the construction of discursive varieties because speakers are more likely to engage discursively with variants that they perceive. However, it is equally possible that linguistic forms remain part of discursive varieties even though they are neither produced nor cognitively perceived anymore. Against the background of this distinction I compare and discuss the role of social factors, particularly of identity, in the emergence of new varieties of English in the next section.

# **2.1.3 The role of social factors in the emergence of new varieties**

The roles attributed to social factors in the emergence of new varieties by Trudgill (2004) and Schneider (2007) could not be more different. The following citations illustrate this well. Schneider (2007: 95) writes that

The Dynamic Model, supported by accommodation and identity theory, predicts that via language attitudes a speaker's social identity alignment will determine his or her language behavior in detail. Note that there is no implication made here that these developments have anything to do with consciousness: accommodation works irrespective of whether the feature selected and strengthened to signal one's alignment is a salient marker of which a speaker is explicitly aware or an indicator which operates indirectly and subconsciously.

Trudgill (2004), on the other hand, is skeptical of social factors influencing the formation of the new variety. Especially identity, which assumes a central role in the Dynamic Model, is ruled out as a relevant factor:

<sup>4</sup>Leimgruber (2013b: 6) for example criticizes that the collection of corpora for studying World Englishes take a "conceptual linguistic system tied to a particular locale" as a starting point.

### 2.1 Theories of the emergence of American English as a new variety of English

And it is clear that identity factors cannot lead to the development of new linguistic features. It would be ludicrous to suggest that New Zealand English speakers deliberately developed, say, closer front vowels in order to symbolise some kind of local or national New Zealand identity. This is, of course, not necessarily the same thing as saying that, once new linguistic features have developed, they cannot become emblematic, although it is as well to be sceptical about the extent to which this sort of phenomenon does actually occur also. For example, we can say that the twentieth-century innovation in New Zealand English, whereby the kit vowel became centralised might perhaps now constitute a symbol of New Zealand identity and that the vowel might for that reason in future even become more centralised. But I have to say that I would, personally, find even this unconvincing. Why do New Zealanders need to symbolise their identity as New Zealanders when most of them spend most of their time, as is entirely normal, talking to other New Zealanders? But in any case, we most certainly cannot argue that New Zealanders deliberately centralized this vowel *in order to* develop an identity marker. (Trudgill 2004: 157)

A further argument which he puts forward against the role of social factors relates to the key role played by children in the process of new-dialect formation. In his view, children are not influenced by social factors like prestige or stigma, but "[t]hey simply selected, in most cases, the variants which were most common" (Trudgill 2004: 115). Adults influenced the new variety only marginally through accommodation in the early stages of contact, but even this accommodation "is not necessarily driven by social factors such as prestige or identity, but is most often an automatic consequence of interaction" (Trudgill 2004: 28).

These two extreme positions have attracted the interest of a number of linguists and consequently, a discussion section in *Language in Society* 37 (2008) has been devoted to the issue. Schneider (2008) and Trudgill (2008b,c) defend their respective positions and support it with more arguments and evidence. Mufwene (2008) and Tuten (2008) partly agree with Trudgill's position; Bauer (2008), Coupland (2008), Holmes & Kerswill (2008) disagree with Trudgill and argue in favor of Schneider's (2007) view. One of the key issues of the discussion is the role of identity and it is noticeable that it is not understood by all participants in the same way. Trudgill's (2004, 2008a) argument focuses on "national identity" and all authors concede that if identity is reduced to this national perspective, Trudgill rightly doubts its relevance. However, apart from Mufwene (2008), all authors emphasize that the question of identity should *not* be reduced to some sort of abstract national identity. Bauer (2008: 273), for example, argues

### 2 The emergence of American English: theories, descriptions, and models

that "complex kinds of identity are being expressed in the choice of a particular phonetic variant. It will not be as simple as feeling that one is 'British' or 'New Zealand'; it will be much more local and much more specific". Tuten (2008: 259) also suggests that it is more likely that local or regional identities develop and influence the variety formation. Coupland (2008: 269) goes beyond the local/national distinction by pointing out that identity in general is "often less coherent, less rationalised, more elusive, more negotiated, and more emergent […]. Identities are known to be often multiple and contingent". This view is echoed by Holmes & Kerswill (2008: 274) who state that "[…] to imply, as Trudgill seems to be doing, that "national identity" can stand for all types of identity deflects our attention from the real sociolinguistic issues". Schneider (2008: 265) also points out that identity plays a role in *all* phases of the model, also before the stage of nation-building. Nevertheless, he still emphasizes the importance of national identity (next to all other identity constructions):

Of course, there are linguistic forms considered diagnostic of individual postcolonial Englishes, and it is difficult to see how precisely these forms rather than any others should have been selected on a purely deterministic basis, excluding national identity as a factor. The strongest argument for the impact of identity in these processes is the observation that the origin and/or recognition and spread and/or scholarly documentation of these forms typically fall into periods of heightened national or social awareness. (Schneider 2008: 266)

Related to the question of what kind of identity is supposed to play a role (national, local, individual) is the question of intentionality. In Trudgill's (2004) view, arguing for a role of identity means arguing that people change their linguistic behavior *intentionally* and Mufwene picks up on that point by stating that

Trudgill is certainly correct in refuting the position that colonial identity drove the structural divergence of "new dialects" both from their metropolitan kin and from each other. This would be tantamount to claiming that evolution is goal-oriented and the colonists had really planned to be different linguistically. (Mufwene 2008: 257)

While it is convincing that settlers in a colony did not come together and developed the goal to intentionally change their speech in order to mark themselves as different from their country of origin, it is nevertheless unclear why Trudgill (2004, 2008b) and Mufwene (2008) reduce the question of identity to the

### 2.1 Theories of the emergence of American English as a new variety of English

abstract national level and interpret effects of identity construction as the result of intentional moves because this is not at all what Schneider (2007) or other sociolinguists claim.

On the contrary, Schneider (2008: 264) clarifies his position again by emphasizing that he sees identity constructions and linguistic accommodation as closely related. He regards identity as an "individual stance with respect to the social structures of one's environment" and accommodation as "a process [in which] individuals approach each other's speech behavior by adopting select forms heard in their environment, thus increasing the set of shared features" and concludes that accommodation is therefore "one of the mechanisms of expressing one's identity choices" (Schneider 2008: 264). As a counterargument to Trudgill's (2004, 2008a) view that accommodation is automatic (and not social) because it is biologically given, he states that it is actually the social nature of the process that is biologically given because human beings are by nature social beings who strive to create group cohesiveness to ensure their survival in a hostile world (Schneider 2008: 264). Holmes & Kerswill (2008: 275) argue similarly for the importance of identity not only in influencing the direction of accommodation but also in determining the frequency of interactions between people:

[N]on-demographic social factors bear directly on frequencies of interaction. Because people bring to each encounter their personal and social identities, as well as knowledge and beliefs about intergroup relations and about the social marking of linguistic variants, interactions with certain social groups will be sought out or avoided. Thus, social factors influence *both* the frequency of interactions *and* the direction of accommodation.

This is reminiscent of Kretzschmar's (2014, 2015a) argument that complex systems evolve and change through massive interactions between speakers, which are naturally constrained by the place that they live in and their social background, which also has an influence on how much they travel and which communication channels they use to get in contact with people who live in other places. While Kretzschmar (2015a: 262) is very skeptical of simple correlations between linguistic features and regional and social features, he nevertheless argues that "positive feedback, or focussing, creates the A-curve for every feature" (Kretzschmar 2014: 152) and cites from Le Page & Tabouret-Keller's (1985) *Acts of Identity*:

We see speech acts as acts of projection: the speaker is projecting his inner universe, implicitly with the invitation to others to share it. […] The feedback that he receives from those with whom he talks may reinforce him, or

### 2 The emergence of American English: theories, descriptions, and models

may cause him to modify his projections, both in their form and in their content. To the extent that he is reinforced, his behavior in that particular context may become more regular, more focused. (Le Page & Tabouret-Keller 1985: 181–192, cited in Kretzschmar 2014: 152)

This shows that Kretzschmar (2014, 2015a) argues for a role of social factors and identity, albeit not in an abstract national sense, but in an individual sense and that he regards people's reactions and social consequences of speech in actual communicative situations as crucial for the creation of A-curves in complex systems of speech.

Even Mufwene (2008), who strongly argues against a role of identity, regards accommodation as a social process. He states that "[t]he speakers' mutual accommodations are certainly the social aspect of the mechanisms by which selection from among the competing variants (and language varieties) proceeds" (Mufwene 2008: 257). Why he sees accommodation as a social process influencing the evolution of varieties while at the same time ruling out identity as a factor is less clear; perhaps it is the restriction of identity to an abstract colonial one that makes him skeptical. He does, however, speculate that "[i]f identity has a role to play, it must be in resisting influence from outside one's community" (Mufwene 2008: 258). As he does not elaborate on this any further, his view on the role of social factors remains very general and it is also not clear how they interact with other linguistic factors like frequency and "simplicity, perceptual salience, semantic transparency, regularity, and more familiarity to particular speakers" (Mufwene 2008: 257) and the role of children as "affective filters" (Mufwene 2008: 258).

Trudgill (2008b) defends his position against the criticism by emphasizing again that children play the crucial role in new-dialect formation and he draws on Pickering & Garrod's (2004) interactive alignment model to support his argument that young children *automatically* accommodate to each other, without any influence of social factors, and that therefore the majority variant survives (Trudgill 2008b: 279). However, Tuten's (2008) discussion of the interactive alignment model emphasizes that its applicability seems rather restricted to young children while "older children and adults could adopt a strategy of non-alignment when appropriate" (Tuten 2008: 261). He argues that when children grow older, they change their social orientation away from their parents towards the peer-group and that they become aware of similarities and differences between them and others. Furthermore, in the process of growing up, people "are heavily socialized to perform in certain ways" (Tuten 2008: 260) so that "it may be that accommodation (or automatic interactive alignment) and identity formation (among

### 2.1 Theories of the emergence of American English as a new variety of English

older children and adolescents) are closely linked" and that "community identity formation and koine formation are simultaneous and mutually dependent processes" (Tuten 2008: 261).

It can be concluded that the overwhelming majority of the arguments are in favor of a position that includes social factors, particularly identity constructions, in the process of new variety formation. This is not to say that identity is unequivocally seen as the main driving force. Schneider (2008: 265) for example concedes that other factors may be as strong as identity and Coupland (2008: 267–268) warns that "there are dangers in running too freely to causal explanations around identity". Trudgill's (2008b: 279) strongest point remains that there is no "feature-by-feature social-reasons account" which shows convincingly how social factors and identity influence the shape of a new variety. Schneider (2008: 266) equally concludes that "[D]esigning a study that will test a straightforward connection between socio-psychological attitudes (including national identity) and the use of specific linguistic forms in these contexts will certainly be a worthwhile task".

# **2.1.4 Conclusion**

This section has shown that theories of the emergence of new varieties of English still differ to a considerable extent with regard to the underlying mechanisms and phases involved in the process. Even though Kretzschmar (2014) claims that his account is complementary to Schneider's (2007) Dynamic Model, the analysis has shown that this is only true to a limited degree. The only feature-by-feature account is provided by Trudgill (2004) on the emergence of New Zealand English, but as he regards his model to be applicable to all new dialects in tabula rasa situations, his claims can be tested in the American context as well. I have raised two important issues in §2.1.2 and §2.1.3: The first issue was that the different theories and models conceptualize the emergent new variety in different ways, which has a considerable impact on their claims, and the second issue was that one of the key differences between the theories is the role attributed to social factors in the process. It is not hard to see that these issues overlap. Trudgill (2004), who regards a new variety as a new linguistic system that is structurally different from other linguistic systems (particularly from those that the settlers brought to the new country), views social factors as irrelevant in the process. They could only become important *after* the formation process in the way that people single out new linguistic features as characteristic of the variety and make them emblematic of it. I interpret this as claiming that social factors could contribute to the emergence of a perceptual and/or discursive variety, but only as

### 2 The emergence of American English: theories, descriptions, and models

a *consequence* of the formation of a new structural variety. Schneider (2007) on the other hand claims that not only a structural variety emerges but a perceptual and/or discursive variety as well, and that the emergence of the latter acts as a driving force in the emergence of the former. The perception, recognition, public documentation and discussion of a variety is therefore in his view *not* just a consequence of already present structural differences, but it is one of the *causes* of structural differentiation. The linguistic features which make the new variety distinct from other varieties are of course much more numerous than those perceived and discussed as distinct and Schneider (2007) claims that they are found on different linguistic levels as well: While the majority of forms of the perceptual and/or discursive variety are located on the phonological and lexical level (with a focus on individual phonological segments and particular words or different spellings of the same word), the majority of structural differences can rather be found on the lexico-grammatical interface. The link between the two levels is supposedly found in the social realm – in the phase of structural nativization, a positive evaluation of forms recognized as distinct leads people to increasingly use distinct forms and patterns because they can express a new (national) identity by aligning themselves with a new model of speech, a discursive variety which subsequently stabilizes in the endonormative phase. However, the description of this process, that is the "micro-level of the relationship between attitudes and the evolution (i.e. selection or avoidance) of individual linguistic forms" (Schneider 2007: 95) remains rather vague and is only supported by some examples which do not represent a "feature-by-feature social-reasons account" that Trudgill (2008b) demands. The same is true for Kretzschmar's (2014, 2015a) claims that positive feedback in people's interactions influences the non-linear frequency distributions which form the basis for the idealized abstract structural varieties described by linguists. Concrete evidence or detailed descriptions as to which forms receive feedback, how this feedback is expressed and perceived and how people adapt their linguistic behavior based on the feedback is not provided.

In my view, the question of how new varieties emerge needs to be approached by first of all distinguishing more systematically between the different kinds of varieties because the investigation of each kind of variety requires a different methodology. Structural varieties should be identified based on bottom-up analyses of data, but it needs to be acknowledged as well that a completely datadriven, bottom-up analysis might be difficult to conduct in practice and this type of analysis is especially difficult in historical contexts because of the scarcity of data. Perceptual varieties should be described based on data obtained by methods used in perceptual dialectology, and discursive varieties should be investigated by means of discourse-linguistic methods (see §2.2.4). The systematic investiga-

### 2.2 Enregisterment

tion of these different types of varieties should then form the basis for exploring the connection between them. The concept of *enregisterment* is helpful in this regard because it involves a theory of how people construct *registers* as models of linguistic (and social) behavior based on language use in everyday social situations. To conceptualize discursive varieties as registers which are based on structural varieties because they are formed as a result of social interaction (both immediate and mediated) helps to clarify the link between the different types of varieties. It is the aim of the present study to gain insights into the enregisterment of American English by showing how the construction of a discursive variety can be traced systematically in nineteenth-century America. In §2.2, I will therefore describe the concept of *enregisterment* in detail and discuss its relation to several research areas of linguistics.

# **2.2 Enregisterment**

*Enregisterment* is a concept that has been developed by the anthropological linguists Michael Silverstein (2003,1979,1993) and Asif Agha (2003, 2007). In §2.2.1, I compare and contrast their definitions of register and enregisterment and sketch the advantages that their models hold for theorizing the emergence of new varieties. In §2.2.2, I outline and discuss the integration of enregisterment and central concepts to which it is tied (e.g. indexicality and orders of indexicality) in sociolinguistic research in general and show what potential it holds for analyzing social factors in the emergence of new varieties in particular. Taking up Kretzschmar's (2014) observation that Schneider's Dynamic Model (2007) is primarily concerned with the perception of new varieties (and not with actual linguistic usage), I sketch the theoretical assumptions and important findings of the field which is primarily interested in the perception of varieties by non-linguists, namely perceptual dialectology, and discuss its relation to enregisterment in §2.2.3. Finally, I suggest in §2.2.4 that the newly emerging field of discourse linguistics, which is concerned with the social negotiation of knowledge through linguistic practice (Spitzmüller & Warnke 2011: 53), can contribute to a study of enregisterment, both from a theoretical and a methodological point of view.

# **2.2.1 The origins of the concept** *enregisterment* **in linguistic anthropology**

The term *enregisterment* was originally introduced by the anthropologist and linguist Michael Silverstein in the mid-1980s (see Silverstein 2016 for details on the

### 2 The emergence of American English: theories, descriptions, and models

earlier uses of the term), but elaborated on in most detail in his 2003 article on indexical order.<sup>5</sup> He claims that an analysis of any sociolinguistic phenomenon requires an analysis of the indexicality of the linguistic forms used. Indexicality means that the linguistic forms possess the quality to point to aspects of the micro- as well as the macro-context in which the forms are used. For example, in a specific micro-context of wine tasting evaluative phrases like *beautifully complex, very pronounced yellow* and *assertive backbone* index connoisseurship (even though they are not part of specialists' vocabulary, which comprises words like *bouquet* and phrases like *slightly pasty/acidic texture*). At the same time, this connoisseurship is "macro-sociologically locatable" (Silverstein 2003: 226) in that the use of these specific forms also indexes a social distinction between those who know how to describe wine and people who do not. Therefore, the use of the evaluative terms indexes more than connoisseurship: It also indexes social traits of the speaker, such as being well-bred or being at least upwardly mobile and having an interesting character. These two types of indexicality are located by Silverstein on different orders: The connoisseurship indexed by the phrases is on the *n*-th order and the social traits on the *n*+1st order of indexicality. This example illustrates that these orders are in a dialectic relationship. The existence of *n*-th order indexicality makes it available for what Silverstein terms "ethno-metapragmatic evaluations" (Silverstein 2003: 214) which are embedded in a larger cultural schema shaped by social and linguistic ideologies: Being an expert of wine is associated with being part (or trying to be part of) an elite social group and this in turn is connected to expectations about the character and the social and linguistic behavior of that group. Describing a wine as *beautifully complex* therefore comes to index social traits of the speaker such as a high social standing, educatedness and cultivation. This is what Silverstein calls *essentialization*: The social traits are ideologically constructed as essences of persons and thus become "predictable-as-true". When people believe that the phrase *beautifully complex* is uttered by members of an educated, cultivated and upper-class elite, the utterance of the phrase points to these qualities. As such, it can be used by speakers to signal this macro-social identity. The essentialization (and sometimes even naturalization) of the values indexed by the form makes it possible that *n*+1st order indexicality blends with *n*-th order indexicality or even replaces it: Speakers have an idea about "wine talk" and about people engaging in it even though they may never have been part of a wine tasting situation themselves. This in turn opens up the possibility for new *n*+1st order indexical values:

<sup>5</sup>This article is both based on and an elaboration of ideas presented in Silverstein's prior publications (1979, 1993).

### 2.2 Enregisterment

The use of the forms can for example be evaluated negatively within a cultural schema that is skeptical or even highly critical of elites and elitist behavior.

The idea of indexical order is a prerequisite for Silverstein's definition of enregisterment and register. Enregisterment is the process by which "*n*-th- and *n*+1st-order indexicalities are dialectically mediated" through "culturally construing and interpreting contextual formal variation as "different ways of saying 'the same' thing"" (Silverstein 2003: 216). Enregisterment is therefore a process of cultural construction. A wine can be described with the phrase *beautifully complex with an assertive backbone* or with the phrase *very good and tasty*; even though both phrases do not denote the exact same thing, they are culturally constructed as expressing the same meaning. Silverstein (2003: 212) calls this phenomenon "metapragmatically imputed denotational equivalence". Through the cultural schema and metapragmatic evaluations described above, the first phrase which indexes connoisseurship (*n*-th order) is enregistered (*n*+1st order) and becomes part of a lexical register which Silverstein (2003) labels as *oinoglossia*. Registers are defined by Silverstein as "alternate ways of "saying 'the same' thing" considered "appropriate to" particular contexts of usage" (2003: 212). This means that even though both phrases are constructed as having the same basic meaning, they are different in that only the phrase *beautifully complex with an assertive backbone* is deemed appropriate in a wine tasting situation – not only because it indexes knowledge about wine but also because it indexes knowledge about appropriate linguistic behavior in social circles where wine tastings are common. Silverstein writes further that "the register's forms being extractable from the sum total of all possible texts in such a context, a register will consist of particular register shibboleths, at whatever analytic plane of language structure (phonologico-phonetic, morpholexical, morphosyntactic, grammaticosemantic, etc.)" (2003: 212). Applied to the example of the oinoglossia register, this means that this register can be recognized by shibboleths which are found on the lexical level (I assume that this is why Silverstein calls it a lexical register). A phrase like *beautifully complex* indexes the use of the oinoglossia register, but in order to produce a coherent text it must co-occur with other linguistic forms which are also part of the register. To illustrate this, I suggest the following example sentences:


These sentences show that the address terms and the adjective phrases are in a paradigmatic relationship and that their syntagmatic co-occurrence is deter-

### 2 The emergence of American English: theories, descriptions, and models

mined by the register to which they belong. In (1) and (3), the address term and the adjective phrase belong to the same register or registers which are compatible with each other, whereas in (2) they belong to different registers which are incompatible with each other. The oinoglossia register is created by a higher-order indexicality linking *beautifully complex* to a high social status and the prestige activity of wine tasting - these indexical values are not compatible with the values of American masculinity and non-conformity indexed by the address term *dude* (Kiesling 2004). Sentence (2) is unlikely to be produced by speakers, unless they want to create irony and achieve a humorous effect. It is therefore possible that linguistic forms which belong to different registers co-occur syntagmatically, but these combinations are usually marked and open to interpretations involving an even higher-order indexicality ((*n*+1)+1). Silverstein therefore concludes his definition by saying that "[w]hile such shibboleths are strongly salient as indexes that the register is in use, the overall register itself consists of these plus whatever further formal machinery of language permits speakers to make text, such as invariant aspects of the grammar of their language. (A *language* is thus the union of its *registers*.)" (2003: 212). This definition implies that registers are recognized by means of salient linguistic forms (register shibboleths) co-occurring syntagmatically, but that they consist of other forms as well which are invariant or at least not register shibboleths of other incompatible registers.

The aspect of *recognition* is also at the heart of Agha's (2007) concept of enregisterment. He defines it as "processes and practices whereby performable signs become recognized (and regrouped) as belonging to distinct, differentially valorized registers by a population" (2007: 81). He therefore regards registers as entities that come into existence through recognition by a group of people in a particular cultural and social context and during a particular time period. It is also noteworthy that registers are not restricted to linguistic forms but comprise all kinds of signs which are performable and therefore visible to others. *Semiotic registers* in a wider sense are consequently distinguished from more specific *registers of discourse*. In contrast to Silverstein, whose focus is on the development of a semiotic model of enregisterment (and ultimately a model of language as a union of registers), Agha is very much concerned with modeling the cultural and social processes by which registers are created and changed continually in a socio-historical context. The definitions of semiotic register and register of discourse reflect this (Agha 2007: 81):

	- (a) which links speech repertoires to stereotypic indexical values

2.2 Enregisterment


A *semiotic register*: a register where language is not the only type of signbehavior modeled, and utterances not the only modality of action. A register of discourse is a special case.

While the indexicality of (linguistic) forms is as central for Agha as for Silverstein in the emergence of registers, it is crucial that Agha defines registers not as repertoires of signs per se, but as cultural models of action. A model of action is only an abstraction of the actual, observable action, but it is necessarily based on that action and serves as a point of reference for all participants in the action and therefore influences the action that it models. The emphasis on action means that only that which is performable and therefore observable can become part of the model. In all these observable instances of action which necessarily happen in a social context, indexical links are created between signs involved in the action and aspects of the context. Agha's main contribution is to provide a theoretical model for how these links between (linguistic) forms and indexical values, created in every instance of observable action ("the micro-time of interaction", 2007: 103), become cultural models of actions ("macro-social regularities of culture", 2007: 103), which then influence actual action again. I outline this process of enregisterment in the following paragraphs.

The key activities in enregisterment are speakers' *reflexive activities*, "namely activities in which communicative signs are used to typify other perceivable signs" (Agha 2007: 16). This means that several signs are grouped together and assigned a metalinguistic predicate which relates to types of persons, types of interpersonal relationships or types of behaviors. Using the wine example again, the speaker's use of *beautifully complex with an assertive backbone*, the elegant and expensive jacket that he or she wears and the way he or she greets people with a smile and a handshake are evaluated and typified by participants in the same situation as 'wine-connoisseur', 'superior social standing' and 'polite'. In this process, disparate cross-modal signs become icons of categories of personhood, behavior and relationships and the use of these icons comes to index characteristics of its users. The icons are therefore classified by Agha as *indexical icons*, which are *emblematic signs*. It is important to note that while these typifications are perceivable by definition, this does not mean that they are always expressed explicitly and linguistically. On the contrary, they are often implicit and only mediated by overt signs, e.g. by the deferential behavior of others (2007:

### 2 The emergence of American English: theories, descriptions, and models

103). Individual face-to-face encounters between people (as in this example) are part of larger "communicative chain processes" and "communicative networks" (2007: 69) through which typifications are transmitted within a population. Agha (2007: 151) provides a list of common typifications of language use (Table 2.2) which reflects the difference between reflexive activities in direct interpersonal interactions (1.) and in larger more indirect cultural forms of communication (3.). Typifications also occur as the result of interventions of experts (2.).

> Table 2.2: Typifications of language use in Agha's framework of enregisterment (from Agha 2007: 151)


It is essential in enregisterment that one instance of metapragmatic activity cannot constitute a register, but that typifications have to be recurrent in the behavior of many speakers. It is only through the transmission and recurrence of evaluations of forms that they become recognized and distinguishable as a register. The frequently recurring typifications become "stereotypes of indexicality" or "metapragmatic stereotypes" (2007: 151-153) which means that they become social regularities. It is evident that especially the reflexive behaviors in (i)-(q) of

### 2.2 Enregisterment

Table 2.2 contribute to this development because they involve mass communication of some sort where linguistic forms and their indexical values are brought to the attention of a large audience. It is possible that these stereotypes become so widely known and accepted that they become "a routinely background reality for very large groups of people" and therefore "socially routinized metapragmatic constructs (such as beliefs, habits, norms, ideologies)" (2007: 29). In the transmission process, however, typifications are also negotiated and transformed. Here, Agha draws on Silverstein's orders of indexicality: If an evaluation of a set of forms becomes so common that it becomes presupposable by many people, it can become subject to reanalysis which could affect both the forms (regrouping) and the indexical values (revalorization). For example, if the indexical link between *beautifully complex with an assertive backbone* (together with other signs) and the image of an 'expert of wine' and 'polite' behavior and social 'superiority' becomes transmitted to a large audience (possibly through advertisements), it may become so strong that it becomes subject to evaluation itself, for example it could be revalorized as an index of snobbery. This does not imply that the former valorization disappears; it is rather the case that there may be competing valorizations which co-exist and depend on the evaluator. What is 'polite' and 'sophisticated' for a person who is part of higher social circles may be 'snobbish' for a person who is not part of these circles. This is why Agha defines registers as dependent on the people recognizing its forms and evaluating them as different from other forms. It also explains why Agha finds questions of boundaries of registers "fruitless and misplaced" (2007: 168) because any boundaries associated with registers are continually negotiated and reset in the processes of enregisterment. Table 2.3 provides an overview of the dimensions of register organization and change that Agha (2007: 169) postulates:

Registers can be characterized with regard to the three dimensions A–C, but "any such account is merely a sociohistorical snapshot of a phase of enregisterment for particular users" (Agha 2007: 170). Even though Agha places particular emphasis on this processual perspective, he nevertheless also stresses that there are mechanisms which work towards a relative stability and persistence of registers. Institutions of various kinds play an important role here; it is possible, for example, that the forms of a register are codified in dictionaries and grammars and used in educational institutions. This is obviously the case for standard registers whose size is usually not only much larger than that of other registers, but also more resistant to change. Agha uses the example of Received Pronunciation to illustrate how phonological forms come to be enregistered as a national standard of pronunciation in England.<sup>6</sup> Furthermore, processes of *essentialization* and

<sup>6</sup>This article was first published in 2003 and has been re-published in 2007 as Chapter 4 in *Language and social relations*.

### 2 The emergence of American English: theories, descriptions, and models

Table 2.3: Some dimensions of register organization and change (from Agha 2007: 169

# A. **Repertoire characteristics**

*Repertoire size*: number of forms *Grammatical range*: number of form-classes in which forms occur *Semiotic range*: variety of linguistic and non-linguistic signs associated with use

# B. **Social range** of enactable (pragmatic) values *Indexical focus*: Stereotypes of speaker-actor, relation to interlocutor,

occasion of use, etc.

*Images* (or icons) stereotypically attached to indexical sign-forms: for speaker-focused indexicals, persona types (male/female, upper/lower class, etc.); for interlocutor-focal indexicals, types of relationship (deference, intimacy, etc.)

Positive or negative values associated with the registers

C. **Social domain(s)**: Categories of persons acquainted with the register formation

*Domain of recognition*: persons who recognize the register's forms *Domain of fluency*: persons fully competent in the register's use

*naturalization*, which also play an important role in Silverstein's model of enregisterment, establish a natural motivation of the link between the register's forms and values (Agha 2007: 74). Metapragmatic activity which repeatedly presents the values indexed by the forms as natural qualities of its users causes people to disregard the fact that the indexical link has in fact been socially constructed. This is for example the case when the use of standard forms is so commonly linked to the attribute 'intelligent' in metapragmatic activity that language users start considering intelligence as a *natural* characteristic of the people using the forms. Using standard forms even comes to be seen as an essential quality of the group of intelligent people. It is easy to see how normative criteria are based on these essential qualities: In such a scenario, people must use standard forms in order to be seen as intelligent by others. Stabilizing mechanisms are reinforced when they are formulated and backed by authority, for example by institutions of expertise, or by people who have been assigned an expert role. Despite these mechanisms of stabilization, however, registers are always subject to reanalysis and transformation; the difference lies in the speed and the extent to which that happens.

### 2.2 Enregisterment

What, then, is the effect of registers? The circulation of metapragmatic stereotypes and images of personhood associated with them provide the ground for role alignments of speakers in interaction. This means that they can either signal their sameness and their co-membership in a social category (symmetric alignment) or their difference (asymmetric alignment) (Agha 2007: 133). This happens in every instance of interaction, but the existence of culturally shared stereotypes also leads to social regularities in role alignment. Alignments are particularly motivated by so-called *characterological figures* linked to registers. Such a figure is "any image of personhood that is performable through a semiotic display or enactment (such as an utterance). Once performed, the figure is potentially detachable from its current animator in subsequent moments of construal and re-circulation" (Agha 2007: 177). This means that links between forms and their indexical values become less abstract, but embodied, and as such more readily inhabited by speakers to signal their social identity.

Comparing Silverstein's (2003) and Agha's (2007) definition of register, it becomes clear that they are very similar in many respects. They see the metapragmatic engagement with perceivable signs (linguistic and others) as the key process in the emergence of registers as it is through this engagement that the signs come to index social values (on the *n*+1st indexical order). They both stress the interrelation between micro-contexts and macro-contexts (Silverstein) and individual face-to-face encounters and large-scale cultural processes (Agha) because registers cannot come to exist on one of these levels only. However, there is also an important difference, which can be explained by the fact that they pursue slightly different aims with their theories. Silverstein's aim is more (socio-) linguistic in that he wants to show how sociolinguistic analysis needs to be completed by studying not only the *n*th-order of indexicality but also the *n*+1st-order. Agha's theoretical orientation is more sociological as he aims to describe and explain the role that language has in social life and its impact on social relations. For Silverstein, accordingly, a register comprises all linguistic forms needed to make a text which are judged to be appropriate in context, but while some forms are salient and point to the existence of registers, some forms are not salient or simply invariant. He elaborates on this view in a recent article where he states that

Language users evaluate discourse with intuitive metrics of coherence of enregistered features of form co-occurring in text-in-context across segmentable stretches of discourse such as an individual's contribution to discursive interaction, generally focusing on highly salient 'register shibboleths' that reveal a basic register setting around which cluster the untrou-

### 2 The emergence of American English: theories, descriptions, and models

bled compatibility or indexically marked lack of compatibility of other aspects of usage. (Silverstein 2016: 59–60)

His statement that a language is a union of its registers is basically a theoretical claim about the nature of language. Agha's definition of registers as models of conduct is not so much language-theoretical but rather social. By emphasizing that models are reference points for social (including linguistic) behavior, he distinguishes them at the same time from actual conduct and language use. Therefore, a register does not comprise all forms needed to produce text in an actual context, but only those forms which are recognized by people as part of the register. It can therefore vary in size – some registers comprise only few forms, while others, especially standard registers, comprise a very large number. It seems as if Silverstein locates registers more on the level of actual language use, while Agha locates them on a discursive level: It is language through which such models are formulated (the terms *discourse* and *discursive* will be discussed extensively in §2.2.4). Nevertheless, as both stress the interrelation between the level of language use and the discursive level, this difference is more a reflection of their different research aims than a difference between their theories: Silverstein's registers, located on the level of language use, only exist because of the metapragmatic engagement of language users with some of its forms; Agha's registers, located on the discursive level, have a bearing on actual language use and in these instances of actual use the register's forms of course co-occur with other forms. These forms can be congruent with the register or they may "by degrees, cancel the stereotypic values" indexed by the register's forms, as in the example sentence (2) above.

This discussion of Silverstein's and Agha's conception of registers and enregisterment shows why enregisterment is a useful theoretical framework for studying the emergence of new varieties: First of all, it provides a way of linking the two levels identified in Schneider's Dynamic Model, the level of the structure and the level of the concept of a variety, in one theoretical framework, the structural one corresponding to actual language use and the conceptual one corresponding to the discursive level of metapragmatic activity and engagement with linguistic forms (and other signs). At the same time, the theoretical framework of enregisterment also provides a basis for systematically distinguishing between these levels in a study of the emergence of a new variety. This in turn is the basis for providing empirical support for the important role of social factors and people's perceptions, beliefs and attitudes in the emergence of new varieties if a changing statistical correlation between linguistic forms and social and geographic categories can be shown to correlate with changing metapragmatic ac-

### 2.2 Enregisterment

tivities surrounding these forms. Secondly, Agha and Silverstein elaborate central ideas of the Dynamic Model in much more detail: the interrelation between identity construction and the construction of linguistic difference as well as the emergent *recognition* of forms as differential by a population, which is important because it is part of Schneider's definition of the term variety itself (see §2.1.2). The third advantage of the theoretical framework of enregisterment is that it provides a methodological basis for studying the discursive level. By emphasizing that metapragmatic activity is by definition observable it becomes clear that it can be studied empirically. Agha's and Silverstein's case studies and examples are mostly qualitative and intended to underline their theoretical arguments; in this study, however, I develop a methodology for extensive case studies which combine both quantitative and qualitative methods to analyze enregisterment processes in nineteenth-century America and contribute to a theoretically informed account of the emergence of American English.

# **2.2.2 Indexing varieties: enregisterment in (historical) sociolinguistics**

Given its emphasis on theorizing the relation between language and social behavior, it is not surprising that the concept of enregisterment has caught the interest of sociolinguists. Silverstein himself links his order of indexicality explicitly to sociolinguistic research by aligning it with Labov's (1972) order of linguistic variables as indicators, markers, and stereotypes as well as with the difference between dialectal and superposed variability posited by Gumperz (1968: 383–384). But it was essentially a change of orientation in sociolinguistics, which Eckert (2012) describes using the metaphor of three succeeding waves, that has sparked the interest first of all in indexicality and increasingly also in enregisterment. While the first wave of sociolinguistics was concerned with discovering the systematic relations between linguistic variation and macro-social categories like social class, age, gender as well as race and ethnicity by using mainly quantitative methods, second wave studies rather looked at how language use correlated with social categories which were relevant to a specific group of speakers on a local level by adding qualitative and ethnographic methods to the research design. Third wave studies shifted the focus from correlations to agency: Instead of viewing linguistic variation as a reflection of social categories they concentrate on the way that speakers make use of variation to construct and express social categories and identities. Consequently, the social meaning created through stylistic practice is at the center of interest of third wave studies and the concept of indexical order is very helpful in this respect because it shows how linguistic forms come to index social attributes and how these indexical meanings co-exist but

### 2 The emergence of American English: theories, descriptions, and models

also change over time. Eckert (2008) develops the concept of the *indexical field* which she defines as "a field of potential meanings" of variables, or "constellation of ideologically related meanings, any one of which can be activated in the situated use of the variable" (2008: 454). It is evident here how the emphasis on activation reflects the importance of agency: By activating one or more indexical meanings of a variant, the language user contributes to the maintenance but also to the change of the indexical field, which is in constant flux. Eckert illustrates the concept by using the released /t/ as an example. Drawing on several studies conducted in American contexts, Eckert constructs an indexical field (see Figure 2.2) consisting of social personae indexed by released /t/, qualities, which are seen as permanent, and stances, which are rather momentary and tied to specific situations but can become constructed as part of people's identity if they are habitually expressed.

Figure 2.2: Indexical field of released /t/: social personae (in circles), permanent qualities (in grey boxes) and stances (in italics), my own illustration based on Eckert (2008: 469)

The indexical field created here emphasizes again the importance of social personae (similar to Agha's characterological figures) which, according to Eckert (2008: 470), anchor the process of interpretation because they are less fluid than permanent qualities and situated stances. Eckert's (2008: 454) proposal to study "variation as an indexical system, taking meaning as a point of departure rather than the sound changes or structural issues that have generally governed what variables we study and how we study them" is therefore a programmatic

### 2.2 Enregisterment

statement for third wave sociolinguistic studies. Using the field metaphor, she makes the point that Silverstein's indexical order is not to be understood as linear (as it is the case in the first sociolinguistic studies on enregisterment which I will discuss below), but as a continual reconstrual of indexical value. At the same time, she also draws on Agha by claiming that "variables combine to constitute styles" (2008: 472) and that styles are "the product of enregisterment" (2008: 456). However, she does not employ the term *register* because she regards the common definition of register as "a static collocation of features associated with a specific setting or fixed social category" (2008: 456) as too established in sociolinguistics. However, as her own definition of *style* is also very different from the established definitions of the term in traditional variationist studies, her argument against the use of the term *register* is not very convincing – it rather underlines the necessity to develop and redefine established terms and concepts.<sup>7</sup> This is precisely what Agha does for the term *register* by criticizing earlier views (2007: 167–170). Nevertheless, it becomes clear that Eckert places herself firmly in the field of sociolinguistics and aims at developing her own theoretical framework of style (instead of using a framework from linguistic anthropology) and testing its usefulness for the study of sociolinguistic variation. At the same time, she also stays within the variationist sociolinguistic tradition by focusing on the indexical potential of single variants and by studying actual usage and behavior, not metadiscursive activities. It is interesting that she does identify a "need to examine a far greater range of variables than is commonly done in the field" (2008: 472) and a need to address questions of the structure of styles and to model the process of *bricolage*, the process whereby individual (linguistic) resources are "interpreted and combined with other resources to construct a more complex meaningful entity" (2008: 456–457), since it is precisely these issues that Agha and Silverstein address in their theory of enregisterment. In a recent article, she elaborates more on the relation between her conception of style and Agha's register: She describes Agha's register as "a style that is enduringly associated with some widely recognized character type such as Posh Brit or Surfer Dude" and therefore as "an outcome of stylistic practice" (Eckert 2016: 76) in which people do not *use* registers, but refer to them and draw on them in their actual language use when they make smaller or larger interactive moves. By describing register as "a sign at a particularly high level of consensuality and metadiscursivity" (Eckert 2016: 76) she locates it on a conceptual level and distinguishes it from actual linguistic and stylistic practice. Although Eckert captures a very important point

<sup>7</sup> For an extensive overview of the development and definitions of the concept of *style* see e.g. Coupland (2007).

### 2 The emergence of American English: theories, descriptions, and models

in Agha's theory here, namely that registers are *models* of action and not action per se, it is an account of Agha's theory that is too reductionist. In Agha's model, a differential metapragmatic treatment of sets of forms is already sufficient to indicate the existence of a register because it points to a differential evaluation of these forms: Users associate one set of forms with different indexical values than another set of forms. In Agha's view, explicit metadiscourses on speech forms and the values that they index *can* exist and they often do exist, especially in the form of characterological figures embodying these links, but it is not a prerequisite for the existence of registers. Registers may acquire a very large social domain (especially with respect to the domain of recognition), but their domain can theoretically also be much smaller. Instead of viewing registers as specific types of style, Agha distinguishes them in a different way and adds the concept of *enregistered style*. Styles are "patterns of co-occurrence among semiotic devices" (Agha 2007: 186) and these devices include linguistic and non-linguistic tokens. Every utterance can therefore be described as a co-occurrence style because in an actual interaction tokens cannot occur in isolation, but they always co-occur and therefore create an observable formal pattern. Only when a formal co-occurrence pattern is differentially evaluated, Agha speaks of enregistered style*s* because they have acquired a cultural intelligibility and significance. It is these enregistered styles that are "reflexively endogenized to a register model" (Agha 2007: 186). Figure 2.3 illustrates this integrative view of styles and registers.

Figure 2.3: Style, enregistered style and register based on definitions by Agha (2007)

### 2.2 Enregisterment

Third wave sociolinguistic studies basically study the relation between style and enregistered styles because their aim is to identify the indexical meaning of (combinations of) linguistic variants, thereby paying attention to how they are linked to other signs in the social landscape. This indexical meaning is used by speakers to position themselves socially by making stylistic moves, thereby changing styles and creating the possibility for new indexical meanings.<sup>8</sup> Agha adds another dimension to these studies by explaining how this process works: Indexical meaning is created through reflexive activities, resulting in a register which is a model of action and therefore not to be equated with a pattern occurring in actual use. Nevertheless, it has an influence on language use as the model is performable in actual stylistic practice. In addition to an empirical study of language use, Agha sees an empirical study of reflexive activity as crucial in identifying how people create and transform social identity and social relations through language.<sup>9</sup>

Another theoretical framework in which the notion of indexicality plays an important role and which is closely related to Eckert's theory of style and the indexical field is Bucholtz & Hall's (2005) framework for the analysis of identity. They regard identity not as an internal psychological phenomenon, but as a social and cultural phenomenon produced through linguistic and other semiotic practices. Identity, defined as "the social positioning of self and other" (2005: 586), is linguistically indexed in various ways: through labels, implicatures, stances, styles, or linguistic structures and systems. This incorporation of linguistic structures and systems in their *indexicality principle* is particularly interesting because it is reminiscent of Agha's (2007: 18) idea that "acts of value ascription to language can […] acquire much more generic discursive objects (e.g., entire speech varieties), and become habitual for large groups of evaluators". Bucholtz & Hall (2005: 597) do not elaborate on that point in more detail, nor do they employ the concept of enregisterment, but they do mention the works on language, nationalism and ideology that inform this view, especially that of the linguistic anthropologists Gal & Irvine (1995). These two researchers elaborate on the semiotic processes by which differences between languages and dialects are constructed through

<sup>8</sup> In my view, Silverstein's notion of register can therefore be equated with Agha's notion of enregistered style.

<sup>9</sup> In her well-known classification of sociolinguistic studies in three waves, Eckert also states that "every case of variation [discussed in her article] involves enregisterment" (2012: 96), but she does not clearly distinguish or relate enregisterment and style in this article. This makes it difficult to follow her argumentation that enregisterment loses its analytic force at some point because nuances of sound, such as fortition or lenition, cannot fruitfully be regarded as components of registers because it is *only* "in continual stylistic practice that nuances of sound take on sufficient meaning to participate in processes of enregisterment" (2012: 97).

### 2 The emergence of American English: theories, descriptions, and models

linguistic ideologies: "[I]deologies interpret linguistic structure, sometimes exaggerating or even creating linguistic differentiation" (1995: 993). They argue that this construction process is not only shaped by the speakers ("the immediate participants in a sociolinguistic field", 1995: 977), but also by scholars ("professional observers", 1995: 993) and criticize that

Although it is now a commonplace that social categories—including nations, ethnic groups, races, genders, classes—are in part constructed and reproduced through symbolic devices and everyday practices that create boundaries between them, this analysis is only rarely extended to language. Despite a generation of sociolinguistic work that has persistently provided evidence to the contrary, linguistic differentiation—the formation of languages and dialects—is still often regarded as an a social [sic] process. (Gal & Irvine 1995: 969)

As a potential reason for this, Gal & Irvine identify nineteenth-century ideologies which equated one language with one culture – an equation which was then used as a basis to claim nationhood and territory. They argue that precisely because of the idea that language is independent of social activities it could be used for the identification of nations. As the unity of the nation was identified based on the unity of language, linguistic homogeneity became an important characteristic of nations, and ideologically-driven processes of erasure led to the eradication of internal variation to such an extent that people viewed language as free of variation and as a fixed system that should not change. Borders between languages and dialects are therefore socially constructed – a line of argumentation that is continued by Agha's proposal that these construction processes must be paid attention to and complement (if not replace) traditional variationist analyses:

The terms dialect and sociolect describe forms of variation in the denotational system of a language community. […] Dialects may exist and be describable by linguists but groups speaking these dialects may be separated in various ways so that cross-dialect contact among persons does not occur and the existence of dialect differences is not even suspected by most (in principle, by any) speakers of the language. Dialect differences are relevant to social life only insofar as they are experienced through communicative events. How such relevance is construed society internally is an empirical question that will have different answers in different sociohistorical locales. (Agha 2007: 132)

### 2.2 Enregisterment

What people view as languages and dialects are registers (in Agha's sense of the term) which are constructed through an interpretation of linguistic variation through reflexive models – Agha (2007: 135) calls this "the reflexive construal of such '-lects' *as* registers". Bucholtz & Hall's statement that identity can be indexed through linguistic structures and systems can therefore be interpreted by using the framework of enregisterment: Linguistic forms have the potential for being grouped with others and evaluated as a distinct speech variety (a dialect, a sociolect, a language) by speakers in a process which links these forms to aspects of the social identity of speakers using these forms. And it is in fact this potential of forms to become indexes of speech varieties that has been the focus of sociolinguistic studies of enregisterment to date. In the following paragraphs I outline and discuss the most important studies on enregisterment in order to demonstrate that although these studies provide numerous important insights, some theoretical and methodological clarifications and adjustments are necessary to make enregisterment a fruitful theoretical framework for the emergence of new varieties of English.

The first sociolinguistic study using Silverstein's orders of indexicality and Agha's concept of enregisterment as a theoretical framework is Johnstone et al.'s (2006) study on the enregisterment of *Pittsburghese*. They relate their study to Agha's (2003) case study on the historical enregisterment of Received Pronunciation by emphasizing that enregisterment can also be used to explain how varieties which do not carry overt prestige can nevertheless also become standardized through the development of vernacular norms (2006: 80). Furthermore, they take up Silverstein's point that Labov's (1972) trichotomy of indicators, markers and stereotypes can be captured by orders of indexicality, the latter being more abstract and providing the advantage of a more nuanced understanding of how relationships between linguistic forms are formed and stabilized (2006: 81). Their study is of great value because it links three fields of research: dialectology, sociolinguistics and linguistic anthropology. They use dialectological evidence to describe linguistic variants which are used in Pittsburgh or in the Pittsburgh metropolitan area, but which are at the same time not limited to this geographical area (the only variant with a small area of occurrence is monophthongal [aː] in the lexical set mouth), and can therefore not be used to identify a set of linguistic forms distinctive of this area. They use sociolinguistic interviews to determine an index score for the use of the monophthongal variant of mouth by five speakers to measure the extent to which they use this local variant. At the same time, the interviews are used to study their perception and their evaluation of this variant. They find that there are two groups of speakers with respect to usage: The two oldest speakers use the local monophthongal variant most of the

### 2 The emergence of American English: theories, descriptions, and models

time, while the middle-aged speakers and the youngest speaker use it hardly at all. In terms of attitudes and perception, they find that one of the older speakers is not aware of using the variant and does not recognize it as local, while the other speaker has learnt to recognize this variant as local and, while knowing about negative evaluations of the variant, regards it as rather neutral (he states that sounding like a working-class Pittsburgher is not "a big deal" (2006: 89). In the second group, the two middle-aged speakers regard the variant as local, incorrect and as signaling working-class membership. One of them, however, also indicates a potential of signaling solidarity by using the variant. The youngest speaker in the second group regards the monophthongal variant as an indicator of local speech and local identity, but in contrast to the other speakers, he does not associate any negative values with it. These findings, which are summarized in Table 2.4, are interpreted by drawing on the anthropological concepts of orders of indexicality and enregisterment.

John K.'s almost invariable use of [aː] and his complete lack of recognition of the form in his own speech is interpreted as the variant being a first-order-index in his speech: He does not use it to express social identity but because he is from the region. According to Johnstone et al., this situation is typical of the time until the 1960s, when the localness of the variant [aː] was observable to outsiders, but not to speakers themselves. Dottie X. confirms this experience but recognizes the variant now as local and working-class. Since she evaluates it neutrally (it is not a problem for her to sound working class), she also does not use the variant to do social work and it is therefore also seen as a first-order index. Arlene C. on the other hand, evaluates the monophthongal variant negatively as incorrect and working-class and uses the diphthongal variant to the extent that even a pattern of hypercorrection can be observed (2006: 91).

In Johnstone et al.'s view, this marks the second-order-indexicality of the variant, as it is avoided by Arlene C. to avoid stigmatization. Barb E.'s use of [aː] is very low, but her evaluation of [aː] is interpreted as an indicator of stylistic variability in her own speech because she regards the use of the variant negatively in some contexts (with its potential to index incorrect and working-class speech), but also positively in other contexts (with its potential to index solidarity among fellow Pittsburghers). This potential to index localness in a positive way is foregrounded in the evaluation of [aː] by Jessica H. As older, negative evaluations of the forms are now in the process of being replaced by a new, positive one, the form is now interpreted as being a third-order index by Johnstone et al. (2006). A further indication of its third-order status is that it is not stylistically variable in Jessica H.'s own speech and not used in her everyday interactions, but *only* in explicit performances of identity, which she describes in the interview. This

### 2.2 Enregisterment


Table 2.4: Speakers' use and perception and evaluation of [aː] in Johnstone et al.'s (2006) study on the enregisterment of Pittsburghese and interpretation in terms of orders of indexicality

### 2 The emergence of American English: theories, descriptions, and models

performance of local Pittsburgh identity by drawing on a specific set of linguistic forms, including but not limited to [aː], is supported or even made possible by a growing number of metadiscursive activities. Johnstone et al. (2006) analyze twenty newspaper articles about local speech and observe that they have appeared more regularly since the 1950s and 1960s and that they evaluate the supposedly regional forms they describe in a disparaging way (2006: 95). In the 1960s and early 1970s, the attitude shifted mostly due to experts like the University of Pittsburgh dialectologist Robert Parslow, who legitimized these forms and assigned them to a local dialect in several interviews. Other materials analyzed by Johnstone et al. (2006) also constitute evidence of an increasingly favorably evaluation of (supposedly) local forms which are listed in a folk dictionary titled *Sam McCool's New Pittsburghese: How to Speak Like a Pittsburgher* (1982, cited by Johnstone et al. 2006: 96) and which became commoditized by being written on T-shirts, mugs, postcards, shot glasses and similar products. It is in this context that younger speakers like Jessica H. now recognize forms like [aː] in mouth and associate them with Pittsburgh speech and use them in performances of local identity even though they do not use them in their everyday speech.

The reason for summarizing this study in such detail is that it has been very influential – it is cited in every sociolinguistic work drawing on enregisterment – and has therefore shaped the understanding of theoretical issues, such as the delimitation of the different orders of indexicality and the relationship between these orders and enregisterment, as well as methodological approaches to the study of enregisterment within a broader sociolinguistic framework. There are three theoretical points made by Johnstone et al. (2006) that need closer attention and which I will discuss in detail below: They relate to the role of awareness in delimiting different orders of indexicality, the relationship between orders of indexicality and enregisterment and the relationship between production and recognition of language forms.

Awareness (or consciousness) is an important factor for Johnstone et al. (2006) in delimiting the different orders of indexicality. While a correlation between a linguistic form and an extra-linguistic characteristic is not noticeable to speakers on the level of first-order-indexicality, it is crucial for second-order-indexicality that speakers start to notice the correlation and use it to do social work. When a linguistic form is explicitly discussed in metadiscursive activities and used consciously in performances of identity, it is a third-order index of a particular aspect of identity (here of *local* identity). The difference here seems to be not only one of awareness but also of intentionality: Only when a linguistic form is intentionally and reflexively used to signal a social characteristic is it a third-order index. Their view is in line with Labov's classification of indicators, markers and stereotypes,

### 2.2 Enregisterment

which also rests on differing levels of awareness, from not being aware of social correlations (indicators) to not necessarily being aware of them, but definitely reacting to them (markers), to being very aware and explicitly discussing them (stereotypes). It is not in line with Silverstein (2003), however, who explicitly argues against Labov's views by stating that

Where the Labovian sociolinguistic marker differs from the mere indicator is the inherent interaction of whatever SEC-indexing rates of production of standard with what we might term *register demand* (a species of tasks demands in the normal psychological sense, and having nothing inherently to do with "consciousness," contra Labov's speculation). (Silverstein 2003: 218)

This shows that he also regards style shifts (i.e. intra-speaker variation in different situations) as evidence for the respective forms being second-order indexes because speakers react to the "demands" of an existing register by changing those forms which are linked to it. But the process of indicators becoming markers (or first-order indexes becoming second-order indexes) is independent of whether speakers do this consciously or not. This view on the role of awareness in style shifting can also be found in sociolinguistic theories belonging to the third wave. In a recent theoretical article, Eckert (2016) argues against the traditional sociolinguistic emphasis on consciousness and awareness, which is also visible in distinctions like *change from below* and *change from above* as well as *overt prestige* and *covert prestige*. In her view, "consciousness and awareness are not simple matters, and agency does not equal or require awareness" (2016: 78). Rather than being external to cognition, "the social is embedded in the unconscious to the same extent, in the same way, and along the same timeline, as the linguistic" (2016: 78). To support her argument she cites experimental studies which show that speech perception is influenced by social information about the speakers (D'Onofrio 2015) and nonlinguistic information like the presence of stuffed toy kangaroos and koalas or toy kiwis invoking associations to Australia and New Zealand respectively (Hay & Drager 2010). Using eye-tracking, D'Onofrio (2015) shows that persona-based information affects early and automatic speech processing (which is not under conscious control by speakers) because it leads speakers to expect a particular vowel which they associate with the persona. Even though these studies focus on perception, they suggest that speakers' sensitivity to social information is not necessarily conscious but can be automatic as well. Given these findings, it is doubtful whether consciousness or awareness are helpful constructs in delimiting different orders of indexicality. What is important instead is the metapragmatic engagement with language

### 2 The emergence of American English: theories, descriptions, and models

which is empirically observable evidence for *n*+1st-order indexicality and therefore of enregisterment. If we can observe speakers' reflexive activities in which they typify perceivable signs, we have evidence of the construction of a register which then influences speakers' conscious or unconscious perceptions and productions. What makes *n*+1st-order indexicality different from the (*n*+1)+1st-order of indexicality is in Silverstein's view not primarily a matter of consciousness but of presupposition:

Labovian sociolinguistic 'stereotypes', of course, are markers that have tilted in the direction of ideological transparency, the stuff of conscious, valueladen, imitational inhabitance – consciously speaking "like" some social type or personified image […]. The values of stereotypes are presupposed in the social-structure-as-indexed according to an ideological model, pure and simple; *n*+1st-order indexicality has become presupposing, in other words, in effect replacing an older n-th-order indexical presupposition. (Silverstein 2003: 220)

When the values indexed on the *n*+1st-order become presupposing, they become open for reanalysis and reinterpretation. The existence of metadiscursive activity involving explicit comments on language use and the use of the features in stylized performances cannot be regarded as evidence for (*n*+1)+1st-order of indexicality per se, but if it occurs repeatedly, it shows that the form-value links are transmitted and become more and more stable and therefore potentially presupposing. Using the example from §2.2.1 again, it is possible that the adjective phrase *beautifully complex with an assertive backbone* is explicitly commented on in a magazine article giving readers suggestions as to how they can describe wine at wine tastings. This contributes to the transmission and stabilization of the oinoglossia register but it is not an indicator of a (*n*+1)+1st-order of indexicality as the values associated with the form are not subject to reinterpretation. However, if an article makes fun of the use of such phrases and suggests that time is better spent drinking wine instead of describing it, the wide recognition of the phrase as belonging to the oinoglossia register is the basis for its negative revalorization. This is an instance of (*n*+1)+1st-order of indexicality. To sum up this discussion, it is not convincing that consciousness or awareness would be helpful concepts in separating the different orders of indexicality. Metapragmatic engagement with language, including (but not restricted to) all types of reflexive activities listed in Table 2.2, is best seen as evidence of second-order indexicality. Metapragmatic activities which build on second-order indexicality to reanalyze and reinterpret these indexical values are evidence of (*n*+1)+1st-order of indexicality.

### 2.2 Enregisterment

The second issue is the relation between different orders of indexicality and enregisterment. Johnstone et al.'s (2006) study seems to equate the enregisterment process with succeeding orders of indexicality, such that a form is not enregistered when it is a first-order index, somewhat enregistered when it is a second-order index, and fully enregistered when it is a third-order index. This conception is implicitly or explicitly found in most other works on enregisterment. Beal (2012: 137–138) for example describes the process of enregisterment as follows:

Enregisterment comes about through the 'indexing' of linguistic features as associated with social characteristics of speakers. Applying a languageideological approach first developed by Silverstein (1976,1998), Milroy (2000, 2004) asserts that there are three orders of indexicality whereby linguistic forms are associated with social categories. The orders relate to ascending levels of awareness within and beyond the speech community:


This characterization illustrates again the problematic distinction between the different orders of indexicality based on different levels of awareness and the close parallel to Labov's indicators, markers and stereotypes. It is especially unclear why speakers need a lower level of awareness for rationalizing and justifying the link between the linguistic form and a particular social category than for overtly commenting on it. Milroy (2004: 167) in fact states that "*second-order indexicality* is a metapragmatic concept, describing the noticing, discussion, and rationalization of first-order indexicality". Her definition therefore includes the *discussion* of linguistic forms, which is by definition overt and explicit, on the level of *second*-order indexicality, but also other forms of metapragmatic engagement without recourse to notions of awareness, while third-order indexicality is not defined or discussed in this article at all. In line with Silverstein (2003), this definition of second-order indexicality rather emphasizes the important role of

### 2 The emergence of American English: theories, descriptions, and models

language ideologies, which are defined by Silverstein (1979: 193) as "sets of beliefs about language articulated by users as a rationalization or justification of perceived language structure or use" (cited by Milroy 2004: 166), in the emergence of second-order indexicality. Even though she does not use the concept of enregisterment in her article either, it is telling that her main focus is on first- and second-order indexicality because, as shown in §2.2.1, this is where Silverstein locates enregisterment. Following his view, forms become already enregistered by becoming second-order indexes. In fact, the data presented and discussed by Johnstone et al. (2006) and Beal (2009b, 2012) can be interpreted in a different way. In the Pittsburgh case, the analysis shows that the linguistic forms have first been enregistered as non-standard, based on a linguistic ideology associating local forms as deviating from a national prestige variety. A form like [aː] therefore became a second-order index of primarily social characteristics and the resulting register is rather a more general register of *non-standard speech*. The explicit metapragmatic judgments elicited through the interviews and the metadiscursive activities identified in newspaper articles are evidence for this enregisterment process. The results of the interviews (see Table 2.4) also show that the social domain of the register has expanded over time. While Dottie X. did not recognize the variant [aː] and did not associate it with local and incorrect speech when she was young, she recognized the variants and the values indexed by it at the time of the interview. John K. still does not recognize the variant as local. So the social domain of the register has expanded so that it includes Dottie X. but not John K. yet. The expansion of the social domain is a potential factor influencing the transformation of the register and the revalorization of the register's forms. While localness has played a role before in that it marked the forms as deviating from a national standard, the social repercussions of this deviation were in the foreground and not the indexicality of place. But against the background of ideologies attributing a positive value to regional identity, the index of localness was foregrounded and positive social evaluations were added. They existed parallel to the negative ones or even disappeared for some people, especially young people like Jessica H., who regards it as positive to have a linguistic form which signals her regional identity. The register label *Pittsburghese*, which first appeared in a newspaper article published in 1967 (Johnstone et al. 2006: 95), shows the transformation of the social range of values with its increasing focus on local identity and its positive evaluation. Pittsburghese is therefore a register which builds on a prior register whose form-value links ([aː] indexing local, nonstandard, incorrect, uneducated speech) have become a social regularity and as such open for reanalysis in the given sociohistorical context marked by social and geographical mobility ([aː] indexing local identity). This change in social

### 2.2 Enregisterment

range and social domain is typical for enregisterment and emphasizes the processual character of the concept and its fluidity. In fact, in Johnstone's (2009) article on the commodification and enregisterment of Pittsburghese, she interprets the results in this way by stating that the local forms are already enregistered by becoming second-order indexes. In her book on Pittsburghese published in 2013, Johnstone even uses the term *re-enregisterment* to label the transition from second-order to third-order indexicality. Even though, on principle, the prefix *re*is not necessary here, as the term enregisterment already encompasses fluidity and change (as does the dialectic relationship between *n*th-order and *n*+1st-order of indexicality described by Silverstein 2003), it is useful because it adds emphasis to the transition process of one order of indexicality to the next.

The second case is Beal's analysis of enregisterment in a northern English dialect, which she presents in several articles (e.g. 2009b, 2012, 2017). It has a strong historical focus and shows how linguistic forms come to index localness and are evaluated positively. One example is the creation of "symbolic working heroes" in dialect literature "with the characters of the weaver in Lancashire and Yorkshire, the pitman and keelman in the northeast, embodying the symbolic virtues of the "gradely" or the "canny lad"" (2012: 136). While these are certainly instances of metadiscursive activity and therefore evidence of the enregisterment of northern dialects, I suggest considering them as third-order indexes not simply because they link form and values explicitly, but because they had already been enregistered as northern and non-standard before. Beal (2012: 137) herself notes parallels between the historical development in Pittsburgh and in northern England: The increasing mobility led to contact between speakers of different dialects and the increasing exposure to different variants increased the potential for higher-order indexicality and enregisterment of forms as belonging to specific regional dialects as well. She describes the rising number of metadiscursive activities in northern England particularly in the second half of the nineteenth century, but in contrast to Johnstone et al. (2006), she does not elaborate on prior negative evaluations of these local forms as non-standard, incorrect and uneducated. Her argument is that the enregisterment of northern dialects is a reaction against leveling processes, defined by Trudgill (1986: 98) as the "reduction or attrition of *marked* variants". While Trudgill considers variants as marked which are of relatively low frequency, Beal also considers variants as marked which are socially stigmatized. As an example she cites Watt's observation that some phonological forms of Tyneside speech are "stereotyped as parochial, unsophisticated, old-fashioned (etc.)" (2002: 55, cited in Beal 2012: 127). Interestingly, this implies the presence of prior negative evaluations of variants which are indicative of second-order indexicality and therefore of enregisterment before the

### 2 The emergence of American English: theories, descriptions, and models

second-half of the nineteenth-century. In fact, in a more recent article Beal (2017) argues that second-order indexical links between linguistic forms and the North of England can already be identified in the sixteenth century and that a *new* order of indexicality started to emerge in the eighteenth century because "forms which have been indexed at the *n*+1-order become associated with another ideological schema" (2017: 28). To sum up, this discussion shows that the traditional emphasis on the Labovian distinction between indicators, markers, and stereotypes and the accompanying different levels of awareness is not helpful to study enregisterment processes. For this reason, scholars most prominently engaged in promoting the use of enregisterment in sociolinguistics like Johnstone and Beal have changed their definitions and interpretations in the last decade at least to some extent to follow Silverstein and Agha more closely. In my study, I will do so even more by disregarding the concept of awareness completely and by viewing enregisterment as the discursive construction of a cultural model of action which rests on an emerging second-order indexicality, empirically observable through instances of reflexive activity, with an inherent potential for third-order indexicality. Instead of theorizing enregisterment as proceeding on a cline from first- to second- to third-order indexicality (perhaps even implying some sort of completeness at the last order), I emphasize the dialectic relationship between the orders of indexicality, which is the basis for the processual character of enregisterment.

The third issue is the relationship between perception and production. It is related to the question of the role of consciousness and awareness and also to the relationship between orders of indexicality and enregisterment discussed above. Johnstone et al. seem to imply that it is the difference between using forms in "unself-conscious speech" and using forms in self-conscious, stylized performances of identity which marks the difference between forms which are enregistered and forms which are not (2006: 97–99). They distinguish "the variable, second-order use of regional variants in everyday interaction" from "third-order performances of a person's knowledge of the sociolinguistic stereotypes that constitute "Pittsburghese"", and they seem to suggest that only the latter case is tied to enregisterment. Their example case to illustrate this is the youngest speaker, Jessica H., who evaluates the variant positively as a marker of Pittsburgh speech, but her frequency of use is as low as for those speakers who evaluate the variant negatively (see Table 2.4). She has a middle-class background and little contact with people using local forms but an "explicit awareness of "Pittsburghese"" (2006: 97), which makes it possible for her to use the forms of the register Pittsburghese to index her identity when it is called for (she describes a situation where a group of college students from different places compare and perform their accents). While

### 2.2 Enregisterment

this case undoubtedly provides evidence of enregisterment because Jessica H.'s account of her own linguistic behavior is an account of reflexive activity, it is only one potential way of how speakers can position themselves socially through language use. In fact, considering Agha's definition of enregisterment, it is not the production of forms which is crucial for defining registers but the perception of forms. A register is a register when its forms are recognized by a sociohistorical population – this is independent of whether speakers of the same population actually produce the forms. Nevertheless, Agha suggests analyzing the extent to which speakers are competent in the use of the register as well and therefore distinguishes two social domains: the domain of recognition and the domain of fluency (see Table 2.3). His analysis of the enregisterment of RP illustrates this necessity as the asymmetry between receptive and productive competence is very high. The point is that Agha's suggestion to view registers as cultural models of action implies that speakers *can* align with them when they use language, but that they do not have to. Spitzmüller (2013) has developed a model of social positioning which extends a sociolinguistic model of stancetaking (Du Bois 2007) to include what he calls metapragmatic stancetaking as well.

Figure 2.4: A model of social positioning through language (my own illustration based on the German and English versions in Spitzmüller 2013, 2015, 2016)

This model integrates the two dimensions which are visible in Figure 2.3: The triangle on the left captures the dimension of actual language use in a communicative situation in which an actor uses and evaluates linguistic forms in a specific way. The triangle on the right, by contrast, captures the dimension of the register as a cultural model of action which links linguistic forms to abstract

### 2 The emergence of American English: theories, descriptions, and models

types of persons (social personae) and practices. So by using a linguistic form, a speaker not only aligns himself or herself with other actors, but also to types of persons and behavior which are indexed by the linguistic forms that he or she uses. The model emphasizes the relation between the concrete linguistic interaction and the abstract register as a social regularity – a relation which is not fixed but highly dynamic. At the same time, it also allows for a separation of the two dimensions to study their interaction in specific cases. The implications for the role of production and perception are that while language use shapes registers, actors can position themselves with respect to registers through evaluation as well as through production.

Spitzmüller (2013) draws on Bucholtz & Hall's (2006, 2005) proposition regarding the different ways in which actors construct identity in positioning their selves in relation to other actors and to abstract register models. The first way is a process of *adequation* or *distinction*. Adequation means that an attempt is made to create the impression of similarity to the other actors by downplaying differences and foregrounding similarities. Distinction is complementary to adequation. In this process, differences are foregrounded and similarities are downplayed to create the effect of differentiation to other actors. The second possibility is a process of *authentication* or *denaturalization*. The key concept here is authenticity, but Bucholtz & Hall emphasize that authenticity as an inherent essence needs to be distinguished from the social process of authentication through which the realness and genuineness of an identity is verified discursively (2005: 601). Denaturalization is the opposite process because it involves the deconstruction and subversion of claims to authenticity. Assumptions of naturalized and essentialized links between identity and social and other characteristics (e.g. biological characteristics like skin color) are called into question in this process. Actors can therefore establish a relation to other actors by claiming authenticity of their own language use or by denaturalizing it e.g. through parody. The third way is a process of *authorization* or *illegitimation* where identities are affirmed and possibly even imposed or, by contrast, dismissed or ignored. In this process, social institutions play a great role as they give power to ideologies which underlie both processes. All these processes show the relationality of identity, which requires the social positioning of actors in relation to other actors through linguistic and other means. It is for example possible that they evaluate linguistic forms that other speakers use as authentic for the group of people they recognize as indexed by the forms, but that they do not feel the need to align with this group and therefore do not use the features themselves. The decision as to whether to align with a group or not can also be a matter of situational context as in the case of Jessica H. She might feel the need to align with Pittsburgh speakers when es-

### 2.2 Enregisterment

tablishing her identity in relation to other college students, but obviously not in the interview situation. Coming back to the issue of consciousness and awareness, it needs to be stressed again that Bucholtz & Hall (2005) argue in the same way as Eckert (2016) that the construction of identity and social positioning is not by definition a conscious process: "identity may be in part intentional, in part habitual and less than fully conscious, in part an outcome of interactional negotiation, in part a construct of others' perceptions and representations, and in part an outcome of larger ideological processes and structures" (2005: 585). To sum up, this discussion shows that it is fruitful to integrate enregisterment with current sociolinguistic theories and models which belong to the third wave rather than with first-wave ideas and concepts like Labov's indicators, markers and stereotypes.

Since the classic studies discussed above, several sociolinguistic studies have described enregisterment processes in varying contexts.<sup>10</sup> In studies on enregisterment in English-speaking contexts, most attention has been given to the role of region and regional dialects in this process. Analogous to Johnstone et al.'s (2006) study on the enregisterment of Pittsburghese, the enregisterment of several other regional dialects was investigated, with an almost exclusive focus on England and the United States: in England, Sheffieldish and Geordie (Beal 2009b), the Black Country dialect in the West Midlands (Clark 2013), the Yorkshire dialect (Cooper 2013) and, more specifically, the Barnsley dialect (Cooper 2019) and other "distinct sub-'Yorkshire' repertoires" (Cooper 2020: 128) as well as the Lancashire dialect (Ruano-García 2020); in the United States, Copper Country English or Yooperese in Michigan and Wisconsin (Remlinger 2009, Remlinger et al. 2009, Remlinger 2017), the Cleveland accent or the northern accent in Ohio (Campbell-Kibler 2012, Campbell-Kibler & Bauer 2015) as well as Southern speech (Cramer 2013). While social factors necessarily play a role in all the enregisterment processes described in these studies, the labels given to the registers clearly underline the prime importance of positive values associated with the regional culture and language. An interesting case is the "northern accent" in the United States because it demonstrates that a perceived *lack* of social meaning is actually a social

<sup>10</sup>The following overview is restricted to studies of enregisterment processes in Englishspeaking contexts and communities, but it should be noted that there are studies of enregisterment in other contexts as well: Cole (2010) and Goebel (2007, 2008, 2012) study enregisterment processes in Indonesia, Babel (2011) and Romero (2012) in South America, Dong (2010) in China, Eggert (2017) in France, Elmentaler & Niebuhr (2017) in Germany, Frekko (2009) and Peter (2020) in Spain, Madsen (2013) in Denmark, Managan (2011) in Guadeloupe, Newell (2009) in Côte d'Ivoire, Park (2016) in Korea, Slotta (2012) in Papua New Guinea, Wilce (2008) in Bangladesh.

### 2 The emergence of American English: theories, descriptions, and models

meaning in itself and therefore relevant for enregisterment processes. Campbell-Kibler & Bauer (2015) explicitly argue against Johnstone et al.'s (2006) view that forms like *yinz* and monophthongal [aː] lacked social meaning before they were noticed by speakers by suggesting that "they likely existed within a larger register, with meanings such as normative, unremarkable, or (in the right contexts) American, non-Southern, and native English" (Campbell-Kibler & Bauer 2015: 98). In their view, the prevailing language ideology in the United States leads to a positive evaluation of unmarked speech. Consequently,

the alternative to a model in which *yinz* indexes Pittsburgh or raised trap/ bath indexes Cleveland is not one in which they index nothing. Rather, it is one in which they both, together with *you*, lower trap/ bath, high front tokens of fleece, the lexical item *cat*, and so on, index normative, unremarkable American English. (Campbell-Kibler & Bauer 2015: 98)

This observation is particularly relevant in that it shows enregisterment processes of regional varieties to be embedded in and linked to other enregisterment processes. In a country where a standard language ideology prevails and interacts with the ideology of nationalism, the enregisterment of that standard, unmarked speech proceeds in relation to the enregisterment of non-standard speech, which is marked as regionally or socially restricted and not adequate to be associated with the speech of the nation as a whole. It is essentially this assumption that the present study builds on: Investigating historical enregisterment processes of American English requires not only an analysis of which forms are associated with American speech, but also, and perhaps even more importantly, which forms are excluded from an American register. I will argue that the construction of what is American proceeds against an existing British English register which leads to forms marked as American and not British. At the same time, however, it also proceeds through an internal differentiation process constructing some variants as regionally or socially marked and, consequently, the alternative variants as unmarked forms which are fit to represent the speech of the nation.

The attention paid to region in studies on enregisterment is actually a bit surprising given that that Agha's (2003) seminal study on the enregisterment of *Received Pronunciation* (RP) in England focuses mainly on the establishment, transmission and transformation of *social* values associated with linguistic forms, which led to RP becoming "a status emblem in British society" (Agha 2003: 231) and not an emblem of regional belonging. For English, there are only two studies investigating the emergence of a register that is constructed primarily through

### 2.2 Enregisterment

the creation of indexical links between linguistic variants and *social* characteristics of the speakers: Schintu Martínez (2016) and But (2017) describe the enregisterment of *cant* in eighteenth-century England, highlighting the association of linguistic forms and low speaker status, involvement in criminal activities and negative character traits like maliciousness. A third study, Pratt & D'Onofrio's (2017) analysis of parodic perfomances of Californian characters, covers a middle ground. The author's main interest is in the social nature of enregisterment processes as they describe the creation of social stereotypes (in this case the Californian Valley Girl and the Surfer Dude) and how they are linked to a repertoire of phonetic forms. They argue, however, that this creation process is a part of a broader enregisterment process of Californian English, thus showing that social aspects and region are intertwined. The social complexity of the process is also underlined by Eberhardt (2012), whose study of the enregisterment of Pittsburghese in the African American community demonstrates the importance of Agha's (2007) dimension of social domain: In contrast to the white community in Pittsburgh, local African Americans recognize not only a limited set of linguistic forms as indexical of Pittsburgh speech, but they also recognize different social values: whiteness and, connected to that, negative associations with oppression and racism (Eberhardt 2012: 367).

It should be noted that most of the studies above also include a historical dimension of some kind, doing justice to the processual character of enregisterment. Most notable with respect to this dimension is Cooper's (2013) study because of the methodological framework he developed for studying enregisterment in historical contexts. His use of quantitative and qualitative analyses of the representation of dialect forms in historical texts and their discussion in metapragmatic discourse serves as an inspiration for the methodology of the present study that has a historical focus as well. Furthermore, he describes actual changes in the register's repertoire of forms over time, using the term *deregisterment* (a term that was first introduced by Williams 2012).

In general, most studies on enregisterment that have been conducted to date investigate these processes in England and the United States. In that regard, the present study is not different, as it also aims at tracing enregisterment processes in the United States in the nineteenth-century. However, the focus is not on a particular region in the United States, but on its emergence as a new variety of English, thereby relating theories modeling this process to enregisterment. This has not been done before, even in the few studies that look at enregisterment processes in the context of World Englishes: Moll's (2017, 2014) analysis of the enregisterment of what she calls "Cyber-Jamaican", a very specific "digital ethnolinguistic repertoire" (Moll 2014: 216), Henry's (2010) investigation of

### 2 The emergence of American English: theories, descriptions, and models

the enregisterment of Chinglish in China and Hodson's (2017a) analysis of the enregisterment of American English in British novels published between 1800- 1836. For the present study, Hodson's analysis is nevertheless highly relevant. Although it is limited in scope to a rather short time period and only six novels, it nevertheless provides important insights into the enregisterment of American English as a new variety of English in a British context, so *outside* of the United States, in a social domain different from this study, but nevertheless connected to it in several ways. Especially the key value of inferiority/superiority, embodied prominently through the figure of the "vulgar American", plays a crucial role in enregisterment processes in an American context as well, as the results of the present study will show. Despite the very limited number of studies on enregisterment of new varieties of English, the importance of one of its essential concepts is recognized: In the recently published *Cambridge Handbook of World Englishes*, Schleef (2020) makes a strong case for using indexicality to study identity constructions and the role of these constructions in the development of New Englishes. An insightful application of indexicality is provided by Leimgruber (2012, 2013a) who uses it to explain the particular mix of linguistic features employed by Singapore English speakers (instead of one of the varieties which are assumed to exist in Singapore, i.e. Singlish, Hokkien or Standard English).

The growing interest in enregisterment is accompanied by even more sociolinguistic studies investigating aspects that are highly relevant for studying enregisterment processes. Three of those deserve particular attention: commodification, authenticity and communication technology. First of all, commodification is concerned with economic motivations for representations of language, thereby highlighting the possibility that language becomes commodified, i.e. "organized and conceptualized in terms of commodity production, distribution, and consumption" (Fairclough 1992: 207, cited in Johnstone 2009: 161). This is the case when linguistic forms are printed on T-shirts or mugs (for concrete examples see Beal 2009b, Cooper 2013, Johnstone 2009, Remlinger 2009) or when dialect representations are used to increase the popularity of novels (see Picone's 2014 analysis of literary dialect in nineteenth-century local color novels depicting the American South). This is in part a consequence of enregisterment, as the producers of these goods will only use forms which they expect to be recognized and whose indexical values they judge as attractive, but commodification is also part of the enregisterment process as it increases the domain of recognition, stabilizes the repertoire and can also contribute to changes in evaluation. When linguistic forms are on display as part of economic goods, people who might not have recognized them before because they have a different socio-economic or regional background can learn about them and are therefore enabled to position themselves socially in

### 2.2 Enregisterment

relation to the forms and the values linked to them. This shows the importance of analyzing economic interests and their effects on enregisterment processes, which will also be a relevant factor in the present study with its focus on newspapers.

The second aspect, authenticity, is central to enregisterment processes as well. The traditional variationist view on authentic speech is that it is the kind of speech which is the least self-conscious and the least influenced by standard norms in the speech community (see Eckert 2014 for an overview). In this framework, speech in performance, defined as "verbal art" by Bauman & Briggs (1990), cannot be authentic because by consciously adapting speech to foreground the poetic function of language a high amount of attention is given to speech. Third wave sociolinguistics have challenged this view of authenticity, however, by arguing that authenticity is not an inherent quality but a claim that speakers make (Eckert 2014). These claims involve attributes that speakers need to possess and based on which speakers can construct themselves or others as authentic or inauthentic. In this framework, speech in performance can be constructed as authentic even though the performing speakers do not use the same speech forms in other natural contexts. Beal's (2009a) analysis of the British Indie Band Arctic Monkeys is a good example of this. The speech of the lead singer during performances of their songs is claimed to be authentic even though he is known to speak differently in contexts which are not part of the performance (interviews, other parts of radio broadcast). Linking her study to Spitzmüller's model, it becomes clear how the Arctic Monkeys position themselves not just in relation to one register but to several registers in developing their unique singing style. It is crucial that, in this view, authenticity is not equated with non-reflexive language use, but it is the reflexive nature of language use which is essential in constructing authenticity – by engaging in a particular set of cultural practices and using a particular set of linguistic forms a speaker can make claims about being authentic. This view of authenticity therefore highlights the role of performance as "a highly reflexive mode of communication" which "puts the act of speaking on display – objectifies it, lifts it to a degree from its interactional setting and opens it to scrutiny by an audience" (Bauman & Briggs 1990: 73). As reflexive activities constitute evidence for enregisterment processes, performances are therefore a highly relevant source of data (see Johnstone 2011a for a detailed analysis of such a performance). Performances also allow for the creation and transmission of embodied stereotypes or what Agha calls *characterological figures*. Such a figure is defined as "any image of personhood that is performable through semiotic display or enactment (such as an utterance). Once performed, the figure is potentially detachable from its current animator in subsequent mo-

### 2 The emergence of American English: theories, descriptions, and models

ments of construal and re-circulation" (Agha 2007: 177). Characterological figures render social personae visible in performance and therefore have an important consequence for the interaction between register models and actual language use – they "motivate patterns of role alignment in interaction" (Agha 2007: 177).<sup>11</sup> It is convincing that language users are more likely to align their speech with register models linked to characterological figures than to those models consisting of indexical links between speech and only abstract social values like 'correct' and 'proper'. Several studies have already identified such figures: the Lanky (Lancashire), the Tyke (Yorkshire) and the Geordie (northeast England) (see Beal 2017), the Yooper in Michigan (Remlinger 2009), the Yinzer in Pittsburgh (Johnstone 2017) and the Valley Girl and the Surfer Dude in California (Pratt & D'Onofrio 2017).<sup>12</sup> However, it needs to be noted that enregisterment can also proceed without recourse to a characterological figure embodying imagined speakers of that variety. Campbell-Kibler & Bauer (2015) find that when northern Ohioans were asked to describe the northern accent, they offered only diffuse descriptions and did not mention any characterological figures. This can be explained by their argument that "lack of accent [is] a valuable social meaning in and of itself" (2015: 115) – even though enregisterment proceeds through the indexical link between linguistic forms and the values of 'accent-free', 'unmarked' and 'normal' speech, these values do not invite an association with a particular figure because such a restriction to a particular speaker would stand in contrast to the generality and unmarkedness of the variants in question. However, characterological figures are still important because they are linked to those varieties which *are* evaluated as accented. Campbell-Kibler (2012: 301) cites Preston (1997) here, who found that widely circulating figures of the hillbilly and the gangbanger are associated with

<sup>11</sup>It is important to point out here that the appearance of characterological figures is of course not restricted to performances – they can also be found in static images like cartoons, which is obviously important for a historical study for which recordings of performances are not available (see e.g. Clark's 2020 detailed study of the role of cartoons in enregisterment processes in the West Midlands).

<sup>12</sup>The Yooper is an excellent example of such a characterological figure. It is described by Remlinger as "stereotypically male: a backwoods, independent do-it-yourselfer who hunts and fishes, rides a snowmobile, drinks beer, spends time at deer camp, and is suspicious of outsiders" (2009: 119). It is noticeable that this description encompasses mostly activities and a general stance towards life; the appearance of this figure seems to be less important. This is supported by a bumper sticker described by Remlinger which reads "Yooper it's not just a word, it's a lifestyle" (2009: 119). Remlinger argues for the importance of such a figure in enregisterment by stating that the Yooper and the way of speaking attributed to this figure "reinforces the notion that a distinct and unified dialect exists in the Copper Country, despite the variability of English throughout the area and despite the widespread use of many of these features throughout the Upper Midwest and places as distant as Alaska" (2009: 119).

### 2.2 Enregisterment

a "Southern accent" and "Ebonics", two varieties which she regards as "highly enregistered in the U.S. context" even though a systematic investigation of the enregisterment of these varieties has not been conducted so far. All in all, it needs to be noted that performances and characterological figures are crucial for the development of registers as cultural models of action and their links to language use by inviting speakers' role alignment. Every one of these reflexive activities is then part of a larger process of enregisterment and because claims to authenticity are processes (referred to as *authentication* by Bucholtz & Hall 2005) which are not global but usually selective (not all practices and all forms are used to claim authenticity), they possess the potential for change. Speakers not only select those attributes and qualities which are important to them, but they also bring in new ones which have the potential of being interpreted by others as authentic as well and become enregistered in the process. This leads Eckert (2014: 44) to conclude that "[t]o the extent that a linguistic variable is deployed in an authenticity claim, the process of adequation will contribute to its ever-changing indexical field". Finding an answer to the question which qualities and linguistic forms marked an American as authentic in the nineteenth century will consequently be an important part of the present study.

The last aspect to be discussed in this section is the role of communication technology in enregisterment. Johnstone (2011b) provides an important study, again in the context of Pittsburghese, which shows that it is important to take into account the medium in which instances of reflexive activity appear. Even though it does not necessarily *determine* who gets access and what kind of content is selected, it nevertheless has an influence on it. At the same time, it has to be considered to what extent the media are controlled by individuals, families, companies or institutions. Last but not least, the study indirectly demonstrates that it is not primarily the medium that shapes people's perception and evaluation of the content, but that other factors play a more important role. Johnstone emphasizes this point in her discussion of the website *Pittsburghese* (http://www.pittsburghese.com) by pointing out that even though the website was intended to be entertaining rather than informative, it turned out that at least some people were willing to take the information on the site to be technical expertise (2011b: 10). Johnstone cites one case of a person who searched for information about Pittsburgh speech and preferred this website over a website provided by linguists (Johnstone herself and Scott F. Kiesling) because he considered the latter to be hard to understand at times and not entertaining enough (2011b: 6). In the context of my study, this has important implications: Even though I analyze only one medium, printed newspaper articles, it is to be expected that other factors, especially the text type (whether the article is for example a news story,

### 2 The emergence of American English: theories, descriptions, and models

a humorous anecdote or an advertisement), have an influence on what people expect from the content, how they interpret it and how their beliefs and views are shaped by it.

Overall, three important conclusions can be drawn from the discussion presented in this section: First of all, sociolinguistics has benefited from the concepts of indexicality and enregisterment, but at the same time it has also contributed greatly to their development. Indexicality proves to be a useful concept in third wave sociolinguistic studies because it helps sociolinguists to theorize and analyze how social meaning is created through stylistic practice in specific instances of language use, and enregisterment can be used to explain how many instances of stylistic practice lead to enregistered styles and registers. At the same time, the development of the concept of the indexical field to identify and describe indexical meanings has advanced the study of indexicality to a great extent, and several studies have underlined the importance of indexicality in the construction and negotiation of social meaning and identity. All the studies which have used the framework of enregisterment illustrate the usefulness of the concept and provide empirical support for and a theoretical elaboration of important factors in enregisterment processes: commodification, performance, characterological figures, authenticity and communication technology. Taken together, the studies show that various research objectives, research methodologies and kinds of data can be integrated when studying enregisterment.

Secondly, it has become clear that especially the integration of enregisterment into sociolinguistic theory can still be advanced further. The distinction between a register as a model of action located on the discursive level and an enregistered style located on the level of actual language use is helpful for differentiating the concepts of *style* and *register* (which are still often used synonymously) and for theorizing and analyzing the process of social positioning modeled by Spitzmüller (2013). According to this model, speakers' stylistic choices in communicative situations are not only influenced by the other actors present in the situation but by registers as well. At the same time, these stylistic choices and therefore instances of language use in particular situations also shape registers. Furthermore, newer sociolinguistic findings suggest that while agency is important in this process, the heavy reliance on levels of awareness, which goes back to Labov and which is an attribute of first wave studies, is not fruitful in analyzing orders of indexicality and stages in the enregisterment process, not least because it is difficult to operationalize. It is, however, not only necessary to integrate enregisterment further into sociolinguistics, but enregisterment also provides an excellent chance for a further integration of sociolinguistics, perceptual dialectology and the field of research labeled discourse linguistics (e.g. by Warnke &

### 2.2 Enregisterment

Spitzmüller 2008 and Spitzmüller & Warnke 2011). Particularly the model for a discourse-linguistic multi-layer analysis (DIMLAN)<sup>13</sup> developed by the just mentioned authors and outlined in §2.2.4 is an excellent way of systematizing different ways of accessing and analyzing enregisterment and of providing guidance and orientation. Consequently, I will develop and justify my research questions and methodology with reference to this model (see Chapter 3).

The third conclusion concerns the relevance of enregisterment for the study of the emergence of new varieties of English. In §2.1.3, I have outlined the debate on the role of social factors, particularly of identity, in the emergence of new varieties, and the results of the sociolinguistic studies discussed in this section strongly suggest that these factors play an important role not only in everyday interaction but also in large-scale linguistic change and differentiation processes. In her recent theoretical article, Eckert (2016: 68) argues that "social meaning in variation is an integral part of language and that macrosocial patterns of variation are at once the product of, and a constraint on, a complex system of meaning", leading her to conclude that while the role of identity in language change still requires further research, it is unlikely that it does not play a role at all. Consequently, she points out that it is crucial to investigate the construction of social meaning(s) at the micro-level and hypothesizes that "[a]ccommodation in colonial situations may have more to do with emerging local social types or stances in the colonial situation than with some abstract colonial identification" (Eckert 2016: 82). I would add that rather than foregrounding one or the other it seems fruitful to take a closer look at how exactly local social types and stances might actually be connected to a more abstract colonial identification.

The following section will shed light on the intersection between enregisterment and a branch of linguistics which has already surfaced in this section because it also has a considerable overlap with sociolinguistics: perceptual dialectology.

# **2.2.3 Perceiving varieties: enregisterment and perceptual dialectology**

One of the leading scholars in the field of perceptual dialectology, Dennis Preston, makes an important point in a recent handbook article, namely that the term *perception* refers to two different things: On the one hand, it refers to ideas that people have about linguistic facts around them and on the other hand, it

<sup>13</sup>In the German original, the model is abbreviated as DIMEAN (*diskurslinguistische Mehr-Ebenen-Analyse*). It was first described in Warnke & Spitzmüller (2008) and presented in more detail and with minor modifications in Spitzmüller & Warnke (2011). An English version was presented by Spitzmüller in Gothenburg in 2014.

### 2 The emergence of American English: theories, descriptions, and models

refers to perceptual abilities, that is the abilities to detect even subtle differences in speech and to identify and differentiate varieties based on these perceptions. Perceptual dialectology is interested in both kinds of perception (Preston 2018: 199). He notes that while earlier studies focused more on the ideas in people's mind, later studies were increasingly interested in finding out how the acoustic reality is perceived by speakers and how these perceptions relate to production and to people's ideas. This development was accompanied by a shift from considering more global properties of speech to analyzing the perception of fine linguistic details, by testing for example whether people could perceive slight differences in the degree of diphthongization of a vowel and match these differences to different places. A general finding is that the relationship between ideas, psychological perception and production is complex.

A second important point that Preston makes relates to the relevance of social factors in perceptual dialectology. As suggested by the term *dialectology*, the field is concerned with the identification of dialects which have traditionally been understood mainly in regional terms. In contrast to traditional dialectology, which bases the identification of dialects on speakers' linguistic productions, the aim of the early studies was to add speakers' perceptions to the picture and consequently, to identify perceptual dialect areas, either based on respondents' ratings of degrees of difference of their own dialect to other dialects or based on maps drawn by the respondents themselves (see Preston 2018 for details). Despite this early focus on regional dialects, social considerations have also played an important role in perceptual dialectology. In the beginning, perceptual dialectologists mainly studied attitudes towards the perceived dialects and evaluative judgements were elicited through a variety of methods, for example by asking respondents to label and provide comments on the maps they have drawn or by asking them to rate perceived dialect regions or samples of actual speech either based on descriptors provided by the linguist (often involving the two important dimensions of pleasantness and correctness) or based on descriptors provided by the speakers themselves (see Preston 2018 and Cramer 2016 for details). More recently, elaborate experimental methods have been developed to test the interaction between attitudes and perception, and studies have revealed that speakers' ideas about and attitudes towards speech can lead to a mismatch between the actual acoustic reality and the perception of this reality. I have already cited some of these studies in §2.2.2, but I add two older studies here summarized by Preston (2018: 197–199). The first one was conducted by Niedzielski (1999), who found that people from Michigan listening to the recorded speech sample of a fellow Michigan speaker did not recognize that this speaker used a higher realization of the vowel /æ/, which is characteristic of the so-called Northern Cities

### 2.2 Enregisterment

Shift, but identified a lower and more central variant as the form actually produced by the speaker. In her view, the idea that Michigan speakers use *standard* American English and not regional forms influenced the informants' perception of actual data. Other studies showed that speech perception can also be very accurate and sensitive to fine details. The second study, by Plichta & Preston (2005), shows that American respondents could distinguish different degrees of diphthongization of /ɑɪ/ and also place the slightly varying vowel realizations on a north-south accent continuum. This demonstrates that they not only recognized the monophthongal [ɑː] as a stereotypical southern variant, but that they could also accurately perceive the variants. Nevertheless, the authors could also show an influence of social factors on speakers' perceptions because when they separated the results by sex of speaker, it became clear that even though the degree of monopthongization was the same (through resynthesis), women's vowel realizations were rated as more northern and men's realizations as more southern. As there was no acoustic basis for these differences in rating, the authors concluded that ideas about women's speech being more correct than men's speech and northern speech being more standard than southern speech influenced the informants' perceptions.

Although this overview of the field of perceptual dialectology is not very detailed, it nevertheless illustrates an important point, namely that studies conducted in this field share many theoretical assumptions and aims with studies using enregisterment as a theoretical framework. First and foremost, they assume that nonlinguists' ideas about and perceptions of language are worth studying and by focusing on lay people's point of view, both frameworks can be considered part of *folk linguistics*. Secondly, studies on perceptual dialect areas as well as on enregisterment focus on the perception or recognition of linguistic forms, either as part of a perceived regional dialect or register. And thirdly, they share an interest in the evaluation of (sets of) linguistic forms, which is an indispensable part of enregisterment studies because the recognition of difference is assumed to emerge through a differential evaluation of linguistic forms, but which is also central to most studies in perceptual dialectology because of the insight that attitudinal factors can trigger and influence speakers' perceptions.

One difference between perceptual dialectology and enregisterment is their slightly different focus, the first paying more attention to regional dialects and the second to social models of action, which is also a result of their evolution from different fields of study (dialectology on the one hand and linguistic anthropology on the other hand). Perceptual dialectologists rather emphasize the relevance of their results for explaining language variation and change, while scholars using enregisterment as a theoretical framework stress that their studies are im-

### 2 The emergence of American English: theories, descriptions, and models

portant for explaining how people are connected in social relationships through language. Nevertheless, as shown in §2.2.2, many studies on enregisterment also focus on the evaluation of specific linguistic forms as indexical of region. Another major difference concerns the methodological approach. For the most part, perceptual dialectologists analyze data which were obtained through their own intervention, while enregisterment studies are more often than not based on data which are not the result of intervention by the researcher. This means that with regard to Agha's overview of different kinds of typifications of language use (see Table 2.2), perceptual dialectology can provide the best insights into the typifications in category 2, while enregisterment studies contribute most by analyzing typifications in category 1 and 3. This is accompanied by a much heavier reliance on quantitative methods in perceptual dialectology and qualitative methods in studies on enregisterment. However, as shown in §2.2.2, some enregisterment studies combine quantitative and qualitative methods and Preston (2018: 194) explicitly discusses discourse as a source of evidence for perceptual dialectology, and he illustrates qualitative discourse-analytic methods by extracting presuppositions underlying a short exchange between two friends.

In general, it follows from this discussion that perceptual dialectology and enregisterment are two frameworks which are highly compatible and have a lot of potential to complement each other. In the conclusion of his article, Preston (2018: 200) notes that

Respondents delineate areas as distinct or different on the basis of their likes and dislikes with respect to speakers and the stereotypes that respondents hold of them, giving concrete expression to Silverstein's notion of higherorder *indexicality*, in which the attributes of people (slow, smart, fun-loving, etc.) are assigned to their language variety and, in fact, become intrinsic parts of that variety's description.

Indexicality and enregisterment can therefore be used by perceptual dialectologists to explain the processes that lead to the recognition of perceptual dialect areas and the specific evaluations of speech forms connected to these regions, while perceptual dialectology can add a cognitive perspective to enregisterment studies and, together with sociolinguistics, provide the link between abstract register models and actual linguistic production, as modeled by Spitzmüller (2016, 2013).

It is therefore not surprising that some studies on enregisterment in an American context have been conducted by researchers with a background in perceptual dialectology (see the articles by Campbell-Kibler 2012, Campbell-Kibler & Bauer

### 2.2 Enregisterment

2015, Remlinger 2009, Remlinger et al. 2009 discussed in §2.2.2), but there is definitely more potential for a fruitful integration of these frameworks.

To conclude, perceptual dialectology contributes in important ways to the definition of a variety that relies not on patterns of production, but of perception, including cognitive as well as ideational aspects, and can be integrated with enregisterment, which offers a primarily social and cultural perspective. Before describing a model which brings together different perspectives on the term variety in §2.3, I address the relationship between enregisterment and discourse linguistics in §2.2.4. As the time frame that this study focuses on is the nineteenth century, the majority of methods used by perceptual dialectologists cannot be used because there are no informants available from whom data could be elicited. Discourse linguistics, however, provides a theoretical framework for integrating a vast amount of methods to study typifications of language in historical texts and other material and is therefore an indispensable resource to give systematicity and coherence to a study of (particularly historical) enregisterment.

# **2.2.4 Constructing varieties: enregisterment and discourse linguistics**

The concept *discourse* is used in many different fields of research and it is consequently defined, interpreted and analyzed in a variety of ways. Angermüller (2014) provides a comprehensive overview of research on discourse and divides the approaches into two inseparable fields: discourse theory and discourse analysis (visualized in Figure 2.5).

The former comprises approaches which are primarily concerned with the connection between language and knowledge, power and subjectivity and which contribute to the development of theories in the social and political sciences as well as cultural studies. The latter focuses on methodology and comprises works which are concerned with empirical approaches to studying discourse as linguistic practice in social contexts. This separation between discourse theory and analysis is reflected in different (yet crucially connected) conceptions of discourse which are described by Johnstone (2018). The first one is discourse as "actual instances of communicative action in the medium of language" (2018: 2) and it is these instances (or to be more precise, *records* of these instances) which are studied by discourse analysts. The second conception regards discourses as enumerable entities, as "conventional ways of talking that both create and are created by conventional ways of thinking. These linked ways of talking and thinking constitute ideologies (sets of interrelated ideas) and serve to circulate power in society" (2018: 2–3). An example of an approach which combines the two perspectives is Blommaert (2005), whose central goal is to explain what he calls

2 The emergence of American English: theories, descriptions, and models

Figure 2.5: Discourse research as discourse analysis and discourse theory (my own illustration based on Angermüller 2014: 26)

"language-in-society" (2005: 16), which manifests itself in the shape of discourse which he defines as

all forms of meaningful semiotic human activity seen in connection with social, cultural, and historical patterns and developments of use. Discourse is one of the possible names we can give to it, and I follow Michel Foucault in doing so. What is traditionally understood by language is but one manifestation of it; all kinds of semiotic 'flagging' performed by means of objects, attributes, or activities can and should also be included for they usually constitute the 'action' part of language-in-action. (2005: 3)<sup>14</sup>

This analysis of language-in-society should be "critical" in that it "needs to focus on power effects, and in particular on how inequality is produced in, through,

<sup>14</sup>This definition of discourse even goes beyond a purely linguistic conception of discourse by including other kinds of semiotic activity – even though this view is not generally shared, the importance of including other ways of meaning-making in the analysis is also pointed out by others. Johnstone (2018: 2) for example notes that "discourse analysts often need to think about the connections between language and other such modes of semiosis, or meaning-making".

### 2.2 Enregisterment

and around discourse" (2005: 233). In contrast to earlier approaches belonging to the school of Critical Discourse Analysis (CDA), he argues for putting more emphasis on the study of "context", which should not only include "linguistic and textual explicit forms" but also "modes of production and circulation of discourse" (2005: 233). Indexicality plays a crucial role in his approach to discourse analysis because indexical meanings connect linguistic signs and contexts and ultimately link language to larger social and cultural patterns. Blommaert (2005) draws on Silverstein's orders of indexicality to emphasize that indexical meanings are not randomly created but "systematically reproduced, stratified meanings often called 'norms' or 'rules' of language, and always typically associated with a particular shape of language" (2005: 73). Although Blommaert does not draw on the concepts of *register* or *enregisterment*, it is not difficult to see how it can be integrated in his approach: The "particular shape of language" associated with indexical meanings is what Agha calls a register. Furthermore, two of the fundamental theoretical principles of discourse analysis developed by Blommaert can be extended by using Agha's concepts of register and enregistered style.

[Principle 3] Our unit of analysis is not an abstract 'language' but *the actual and densely contextualised forms in which language occurs in society*. We need to focus on varieties in language, for such variation is at the core of what makes language and meaning social. Whenever the term 'language' is used in this book, it will be used in this sociolinguistic sense. One uneasy by-effect of this sociolinguistic use is that we shall often be at pains to find a name for the particular forms of co-occurrence of language. The comfort offered by words such as 'English', 'Zulu', or 'Japanese' is something we shall have to miss. We shall have to address rather complex, equivocal, messy forms of language.

[Principle 4] Language users have *repertoires*, containing different sets of varieties, and these repertoires are the material with which they can engage in communication; they will determine what people can do with language. […] [W]hat people actually produce as discourse will be conditioned by their sociolinguistic background". (2005: 15)

The notions *repertoire* and *variety* could be described and analyzed in a better way by using the terms *style*, *enregistered style* and *register* as discussed in §2.2.1 and §2.2.2 because they reflect precisely Blommaert's first principle, namely that "[i]n analysing language-in-society, the focus should be on what language means to its users" (2005: 14). This means that by identifying which repertoires of forms

### 2 The emergence of American English: theories, descriptions, and models

people recognize based on their connection to a specific set of indexical meanings, discourse analysts do not have to give up using words like *English*, but they have to clarify what this word means to language users, i.e. describe the respective register. Registers as recognizable repertoires do not only constitute constraints on discourse, as Blommaert points out in his fourth principle, but they are also constructed through discourse, and this is in my view the essential link between enregisterment and discourse theory and analysis. Because of this theoretical compatibility, discourse analysis can offer a methodological framework for studies on enregisterment, and the results of these studies can in turn be used by discourse analysts in their study of discourses on issues other than language. In what follows, I outline Spitzmüller & Warnke's (2011: 185) theory and methodology of *discourse linguistics*, which is an excellent way of systematizing the variety of approaches to studying enregisterment discussed in §2.2.2.

In contrast to Johnstone (2018), who explicitly considers discourse analysis a research method and not a discipline or subdiscipline of linguistics, Spitzmüller & Warnke (2011: 2) claim that it is justified to establish a discipline which they call *Diskurslinguistik* (discourse linguistics) because it can contribute to studying the phenomenon of discourse in the same way as other disciplines dealing with discourse (e.g. philosophy, sociology, history, literary studies) and because it can also contribute to gaining insights into language which cannot be generated by other subdisciplines of linguistics (e.g. lexicology, semantics, text linguistics, sociolinguistics). While emphasizing the close relation between discourse linguistics and the Anglo-American tradition of discourse analysis, they distinguish them by claiming that the focus of discourse linguistics is much more on the function of language to constitute society and knowledge (*gesellschafts- und wissenskonstituierende Funktion*). The center of discourse linguistics is therefore the conception of discourse which rests on Foucault, consequently putting more emphasis on the goal of discovering how language shapes the world (as Johnstone 2018 puts it), whereas the focus of discourse analysis is more on the conception of discourse as instances of communicative action and an inquiry of how language is shaped by the world. Consequently, Spitzmüller & Warnke's (2011) theoretical framework is very much concerned with a definition of the term *knowledge*, which they describe as "the result of a continuous negotiation, recognition and rejection of insights in discursive practice" (2011: 42, my translation). Knowledge is seen as discursively formed and this is only possible through language with the *statement* as the smallest unit (Foucault: *énoncé*, Spitzmüller & Warnke: *Äußerung*). Consequently, knowledge is constituted in discourse through statements and it is for this reason that Spitzmüller & Warnke (2011)suggest an adaptation of Jakobson's model of factors of verbal communication:

2.2 Enregisterment

# PROJECTION

SPEAKER WRITER HEARER READER STATEMENT MEDIUM

# KNOWLEDGE

Figure 2.6: Factors in the discursive constitution of knowledge (from Spitzmüller & Warnke 2011: 54, my translation)

In the center of the model is Foucault's *statement*, a terminological decision which is meant to reflect that in this model the message is conceived of as instances of communicative action (and not abstract sentences). The speaker or writer is the actor in this process, but it is important to note that actors do not have to be persons, but can also be institutions, specific groups or individuals representing specific social functions (e.g. politicians). The statement links the actor and the hearer or reader. Another important change concerns Jakobson's *context*. Spitzmüller & Warnke (2011) speak of *projections* instead, to emphasize that statements to do not refer to any ontological reality, but that they refer to and at the same time evoke projections of this reality. Instead of *contact*, Spitzmüller & Warnke (2011) use the term *medium* to emphasize that every message is transmitted by means of a medium and the precise nature of the medium has a bearing on how the statement is perceived by the hearers or readers. Spitzmüller & Warnke (2011) urge discourse linguists to identify precisely which media are used and to include visual elements and interactions between the visual and the linguistic in their analyses. Lastly, perhaps the most important adaptation of Jakobson's model concerns the *code*. They replace this term by *knowledge* because in order to understand and interpret statements not only a shared code is needed but also shared knowledge. They summarize the factors in their model by stating that

the discursive constitution of knowledge is a result of statements which are produced by actors in a medial form, which are perceivable by other actors and which refer to mental content on the basis of shared knowledge which is relevant for the understanding of the statement. (Spitzmüller & Warnke 2011: 57, my translation)

Based on these six factors they deduct six functions of the discursive constitution of knowledge and, again with explicit reference to Foucault, six *regulatives*

### 2 The emergence of American English: theories, descriptions, and models

of this process, which illustrates that statements do not exist separately from power structures in society but are embedded in them. The factors, functions and regulatives are summarized in what Spitzmüller & Warnke call a *field model* in Table 2.5.


Table 2.5: Factors, functions and regulatives of the discursive constitution of knowledge (from Spitzmüller & Warnke 2011: 63, my translation)

This model shows how language, knowledge and power are intrinsically connected. As statements are linked to speakers or writers, they have an argumentative function: These actors argue for or question knowledge and in doing so create knowledge or change it. The distributive function is equally important because in a model which assumes that language has the function to evoke projections (and not to refer to an ontological reality) statements need to be distributed and shared. Both of these factors are regulated by power structures because in order to participate in the discursive constitution of knowledge, speakers have to be able to make themselves heard and hearers have to gain access to the statements made by speakers. In this context, norms of expression regulating the statement and its rhetorical function also come into play because the form of the statement also crucially influences whether speakers are not only able but also willing to hear the statement. It is here that Spitzmüller & Warnke explicitly mention enregisterment, but they do not explore the connection between enregisterment and norms of expression in any detail (2011: 61). However, they stress that norms of expressions are crucial because the forms of language and the values attributed to the forms in a community have a decisive influence on access to and participation in discourse. Following my argumentation above, it is therefore important to take norms of expression into consideration when studying the discursive constitution of knowledge, but it is helpful to use enregisterment and the concept of register in addition to norms of expression to investigate this

### 2.2 Enregisterment

regulative.<sup>15</sup> Furthermore, it is important to note again that registers do not only regulate discourse, but that they are themselves discursively constructed. The parallels between Agha's model of enregisterment and Spitzmüller & Warnke's model of the discursive constitution of knowledge are evident: Language itself can become the topic of discourse through reflexive activities and a reflexive model of speech is constructed by actors typifying linguistic forms and making these typifications heard by other actors. The typifications have a specific form, are bound to a specific medium and evoke a projection which is bound to shared knowledge. As pointed out in §2.2.1, the circulation and transmission of typifications is crucial for the formation of a register. One typification does not form a register and this statement is also true for discourse: One statement does not constitute a discourse, but, as Spitzmüller & Warnke (2011: 187) point out, "[t]he discourse is […] only discourse where intratextual phenomena, actors and transtextual structures interact" [my translation]. The methodological framework that they develop for analyzing discourses is structured based on these three dimensions and presented in Figure 2.7.

As registers are discursively constructed, this methodological framework is also useful for studying enregisterment. At the center of discourse analysis as conducted by discourse linguists are statements in concrete temporal-spatial contexts (Spitzmüller & Warnke 2011: 123), which means that when studying enregisterment the focus has to be on statements which typify other linguistic signs. Note here that Agha defines reflexive activities more widely, as the use of "communicative signs" to "typify other perceivable signs" (2007: 16), but this integration of linguistic and other signs is also important in discourse analysis where other signs occurring in combination with statements are also seen as part of the process of constituting knowledge. Statements are the smallest units in discourse analysis and they typically occur in texts, a unit which is defined by Spitzmüller & Warnke as "a multiplicity of statements with syntactic-semantic relations and one/several thematic center/centers in a formal or situational frame" (2011: 137). The intratextual layer therefore comprises an analysis of a single text, where the focus can be on words, propositions or the structure of the text. If

<sup>15</sup>Register is a wider concept than that of a *norm*. Agha (2007: 126) considers a *norm* to be an "externally observable pattern of behavior", a "statistical norm or frequency distribution in some order of behavior". If this pattern is reflexively recognized as normal by a population, it becomes a "normalized model of behavior" which constitutes a norm for the group of people who recognizes it. The normalized model can then become a normative model, which is "linked to standards whose breach results in sanctions"; it becomes "a norm codified as a standard". In this line of argumentation, 'normal' or 'standard' are therefore social values that can be linked to linguistic and other forms of social behavior in enregisterment processes.

### 2 The emergence of American English: theories, descriptions, and models

Figure 2.7: Layout of the discourse-linguistic multi-layer analysis (DIM-LAN) (my own illustration based on Spitzmüller & Warnke 2011: 201 and Spitzmüller 2014)

### 2.2 Enregisterment

the focus is on textual structures, the visual structure of the text should also be taken into account, comprising non-linguistic elements like images (and their relation to the rest of the text), typography and the material to which the text is bound. Discourse analysis as understood by Spitzmüller & Warnke does not stop at the intratextual layer, but regards the transtextual layer as equally important. On this layer, a multiplicity of texts produced by different actors and occurring in different media is analyzed; the relation between these texts is established through communicative/discursive action which is why a third layer is included in the framework, namely the agent layer. They see this layer as mediating between the intra- and the transtextual layer because at this layer statements are "filtered" (2011: 173–174). This means on the one hand that discursive actions determine which statements occur as part of the discourse and which position and importance they have in the discourse (this filter is referred to as "discourse rules"). On the other hand, every text produced by an actor is shaped by one or several discourses (this filter is referred to as "discourse shape"). Their intention behind the DIMLAN model is not to provide a set of instructions that needs to be followed consecutively and comprehensively, but they consider it a framework which helps the discourse linguist to decide and argue for a specific method or set of methods. They are convinced that "fixed procedures cannot do justice to the multimodality and linguistic-systematic heterogeneity of discourses" (Spitzmüller & Warnke 2011: 135, my translation). Instead, they argue for the principle of triangulation and a mixture of multiple methods and approaches to studying discourses. In addition to the DIMLAN model summarizing the different *layers* of discourse-linguistic analysis, they also distinguish several objects of study as well as methods and procedures for studying it. These are summarized in Table 2.6 and intended to support discourse linguists planning a study. In this section, I will not elaborate on all these objects, methods and procedures in detail, but I will come back to them in Chapter 3 when I develop the aims and the methodology of the present study.

In conclusion, this section shows that discourse linguistics (as developed by Spitzmüller & Warnke 2011 by drawing essentially on Foucault's work but also many other theoretical and methodological approaches to discourse analysis, e.g. by Blommaert 2005) is especially beneficial for studies on enregisterment because an understanding of a register as something that is discursively constructed allows researchers to use the methodological framework developed by Spitzmüller & Warnke (2011) to describe precisely how they approach the analysis of this construction process. This framework has a solid theoretical foundation and while it keeps the necessary openness to a variety of approaches and methods it provides helpful systematicity at the same time. A discourse-linguistic

### 2 The emergence of American English: theories, descriptions, and models

Table 2.6: Objects, methods and procedures of discourse-linguistic studies (based on Spitzmüller & Warnke 2011: 124–135, my translation).


approach to enregisterment is particularly important for studying historical enregisterment processes because methods which rely on the intervention by the researcher to obtain reflexive activities on language for example by interviewing speakers (as used in perceptual dialectology) are obviously not an option in historical contexts.

A further conclusion which can be drawn from §2.2 as a whole is that enregisterment can be located at the intersection between several overlapping areas of linguistic research. It is this overlap (visualized in Figure 2.8) which is responsible for the wide variety of methods and approaches to enregisterment and also for different nuances of conceptualizing enregisterment.

In this study, the conception of enregisterment remains very close to Agha (2003, 2007), but important additions and refinements of concepts and methods provided by sociolinguistic studies (e.g. characterological figures, social indexicality, identity construction and social positioning through language, the role of performances and authenticity) will be taken into account in the development of the methodology, the analysis and the interpretation of the data. As the focus of this study is on the nineteenth century, methods used by perceptual dialectologists play no role in this study, but I will heavily rely on the discourse-linguistic framework developed by Spitzmüller & Warnke (2011). In combination with the theory of enregisterment and indexicality, this framework is the basis for devel2.3 The construction of discursive varieties through enregisterment: a model

Figure 2.8: Areas of linguistic research and enregisterment

oping the concept of the *discursive variety*, in contrast to the *structural variety*, and for modeling the relationship between them.

# **2.3 The construction of discursive varieties through enregisterment: a model**

Structural and discursive varieties are both fuzzy and not discrete entities. For purposes of description and investigation, however, they are idealized and abstracted as separate entities. On a structural level this is done by linguists who base their descriptions on frequency distributions of variants they observe in actual linguistic behavior. On a discursive level, the speakers themselves construct a variety through reflexive activities – their metapragmatic and metadiscursive engagement with language. Frequency plays a role here as well. The more frequently typifications of linguistic forms occur, the more they turn into metapragmatic stereotypes and the more stable and visible a register appears.

Figure 2.9: A model of the construction of discursive varieties through enregisterment (from Paulsen (forthcoming), reprinted with kind permission from John Benjamins Publishing Company, Amsterdam/Philadelphia)

### 2.3 The construction of discursive varieties through enregisterment: a model

Figure 2.9 illustrates the process of how a register, or discursive variety, is constructed and it therefore synthesizes the discussion in §2.2. <sup>16</sup> First of all, the model proceeds from the assumption that there is a level of structured variation underlying all abstractions. This level consists of all the forms which are realized in social and communicative contexts and linguists can abstract structural varieties by using statistical methods. If people engage reflexively with linguistic forms, they elevate the forms to a discursive level by making them subject to metapragmatic and metadiscursive activities. In these activities, forms are indexically linked to social and pragmatic values which are linked to larger ideologies, linguistic and non-linguistic, within a cultural and sociohistorical context. In terms of orders of indexicality, the forms therefore move from *n*th-orderindexicality to *n*+1st-order-indexicality in this process. It is usually the case that metapragmatic and metadiscursive activities focus on a few salient forms and mark one set of salient forms as different from another set of forms, thus creating different registers consisting of all forms recognized as co-occurring with the set of salient forms, the so-called *register shibboleths*. As pointed out above, if forms are frequently subject of reflexive activities, this contributes to their salience, but other factors can increase the salience as well, e.g. if the form is foregrounded against other forms or if a form is connected to a conspicuous characterological figure. The key idea behind this model is that there is a potential relationship between the structural and the discursive level not only because forms found on the structural level can become part of a register but also because the register potentially influences speakers' choice of forms in interaction when they position themselves socially (see Spitzmüller's (2013, 2015) model of social positioning, described in §2.2.2). Overall, the model captures Agha's (2003: 234) idea that "phonetic varieties have now become objects – or, object discourses – in relation to a metadiscourse linking speech to social classifications".<sup>17</sup> By linking sets of co-occurring forms to social values, people discursively construct knowledge about the forms, about social characteristics of speakers using the forms as well as about situations and contexts in which the use of the forms is appropriate. It is important to emphasize again that even though the model suggests that there are boundaries, these boundaries are only constructed, either by the

<sup>16</sup>I have developed this model in collaboration with my colleagues Benjamin Peter and Johanna Gerwin. It is also presented in Paulsen (forthcoming) and in Peter's study of the discursive construction of Andalusian (2020: 116). Peter also created a second model that brings particularly the processes of revalorization of variants and thus of re-enregisterment into focus (2020: 157).

<sup>17</sup>In this case, Agha (2003) refers to the case of Received Pronunciation in England, which is why he restricts the statement to *phonetic* varieties. In principle, this statement can of course be extended to varieties comprising all structural levels.

### 2 The emergence of American English: theories, descriptions, and models

linguist, based on objective criteria, or by the speakers themselves, based on the need to build social relations to other speakers. While discrete boundaries seem to be useful or even necessary, they are in fact fuzzy and subject to constant change. As pointed out above, the stability of both structural and discursive registers depends on several factors, e.g. the amount of interactions between people using different sets of linguistic forms, the frequency of typifications of speech, the salience of characterological figures, and the degree of institutionalization and codification.

The model forms the basis for my case studies which focus on the discursive level by investigating when and how American English was enregistered in the United States in the course of the nineteenth century. Before describing my methodology in detail in §3, I will end this chapter by coming back to prior research on the historical development of American English in §2.4. The aim of the section is to analyze how historical linguists have determined the beginning(s) of the variety and which importance they attribute to the structural and the discursive level in describing the major developments of the variety.

# **2.4 Structure and discourse in descriptions of the history of American English**

Historians describing the development of American English face the problem of determining when American English emerged as a new variety of English, that is, as a variety that is distinct from an old variety. This problem is illustrated in a recent article by Richard W. Bailey, which was published as part of a collection of articles on the history of varieties of English edited by Bergs & Brinton (2017). On the one hand, he finds that "[h]istorically considered, American English begins to emerge in the 16th century, even before any English speakers reached the shores of the North American continent" (2017: 9). On the other hand, he notes that "America was slow to develop a distinctive linguistic identity" (2017: 10) and that it was only "by the beginning of the 19th century [that] American English had become a recognized variety of the language" (2017: 12). Finally, he argues that "[w]hat is less recognized [by scholars] is the emergence of the meaning of "American English" as a distinct language with certain distinctive properties" (2017: 13). These statements show that Bailey's concept of American English is rather vague: The reason for postulating such an early beginning, which seems paradoxical at first, is probably based on lexical items which had been used by Native Americans before the arrival of English settlers (he cites for example *canoe* and *maize*) and which were borrowed by the English settlers later

### 2.4 Structure and discourse in descriptions of the history of American English

and constituted a difference between their speech and the language forms used in England. In contrast to that, the American English with a "distinctive linguistic identity" seems to be based on the *recognition* of linguistic differences, which according to Bailey occurred on lower levels of the social scale: While "[h]igh status Americans spoke just like high-status Britons" (2017: 12), slaves, servants as well as Scots and Irish immigrants spoke differently. American English as a distinct *language* is defined based on the "certain distinctive properties" accorded to American English by contemporary commentators. One of the most important commentators was Noah Webster, whose views that he proposed and argued for in his dictionaries influenced the perceived properties of American English to a great extent. Bailey (2017: 16) summarizes these properties as follows:

Early in the 19th century, the reputation of American English had been settled – at least for some. The language was free of regional variation, at least in comparison to Great Britain. And it was remarkable for its purity which had been achieved through the preservation of the good old ways of Shakespeare and Addison and through the efforts to regularize it by analogy (so *deaf* was like *leaf* ), preservation (continuing to employ *air* and *heir* as homophones), or transparency (in a preference for *meeting house* rather than *auditorium*).

The example of Bailey's overview article thus illustrates how vaguely American English is defined – structural distinctiveness is not the primary criterion for postulating the existence of a new variety, but only the basis for a discursive construction of the *variety* or the *language* that relies on recognition and valorization. It is not clear, however, how exactly the recognition of different forms of speech in America (used by slaves, servants, Scots and Irish) leads to *one* American variety and it is equally unclear which set of forms is evaluated as constituting a 'pure' and 'homogeneous' American language.

The issue of defining American English is also addressed by Algeo (2001b) in his preface to the sixth volume of the *Cambridge History of the English Language*, which is concerned with the history of English in North America. He states that

[a]ll languages have internal variation ranging in scope from idiolects (the particular ways different persons use the language) to national varieties (standardized forms of the language used in a particular independent political unit). [...] Between the idiolect and the national variety are dialects, regional and social, on various dimensions. (2001b: xviii)

### 2 The emergence of American English: theories, descriptions, and models

In this vein, he states with regard to the "beginning" of American English that while the "process of differentiation between the English of Britain and that of America began with the first settlement in America" (2001b: xvi), "with the American Revolution, the variations that had developed in the colonies became a new national variety, contrasting with what from this point can be called the British national variety" (2001b: xviii). The existence of independent political units thus seems to be taken as a basis for postulating the existence of two separate national varieties. On the other hand, he problematizes the terms *dialect*, *language variety* and *language* as being abstractions that are usually used metaphorically and thus not to be taken literally: "To talk about language in such metaphors is useful and not to be avoided. But it is wise to remember that such talk is metaphorical, not literal". That he finds it difficult to define a variety based on the criterion of structural distinctiveness becomes clear when he says that "[b]ecause of the complexities of linguistic systems, it is impossible to speak with confidence about how much alike or how different two speechways are or to compare two dialects with respect to their overall rate or degree of change" (2001b: xix). Consequently, much like Bailey (2017), Algeo (2001b) also bases the emergence of a distinct American variety mainly on discursive factors and identifies the period from 1776 to 1898 as crucial in that regard because in this "National Period" "the sense of a distinct variety arose, which was standardized especially in dictionaries and spelling books and spread over the continent during the westward expansion" (2001b: xx).

A focus on recognition and thus on discursive factors can also be found in Davis' (2003) introduction to his collection of texts relating to American English between 1781 and 1921. He attributes a "pivotal position" to American English because "it is the first form of English to be recognized as a distinct new variety" (2003: xi). He describes this recognition process further as "the process by which a distinct form of English was picked out of the mass of variation existing on opposite sides of the North Atlantic and on the Atlantic itself, and the various social and political values assigned to that form once its identity had been instituted" (2003: xi). In some ways, this description of the process is reminiscent of the definition of enregisterment because it is essentially about a set of forms being recognized as distinct; however, Davis argues here that social values are assigned to the set of forms *after* they have been recognized, whereas Silverstein and Agha emphasize that the differentiation process is essentially based on a differential evaluation of forms, so that the emerging registers are a *consequence* of valorization processes: Forms are recognized *because* they are assigned social values in the process of enregisterment.

### 2.4 Structure and discourse in descriptions of the history of American English

Like Algeo (2001b), Davis dates the starting point of this process of recognition to the eighteenth century, following not only the American independence from Great Britain but also the standardization of English in England. The texts edited and published in this series largely constitute expert discourses on the language spoken in America, starting with selected papers in the first volume, which span the publication of John Witherspoon's "The Druid" in 1781 (in which he coined the term *Americanism*) to Charles Whibley's article "The American language", published in the Blackwood's Magazine in 1908. The second volume contains glossaries of Americanisms by Elwyn (1859), Fallows (1883) and Norton (1890), the third and the fourth volume exhibit the works of the verbal critic Richard Grant White (1870, 1880), followed by Schele de Vere's (1872) re-interpretation of Bartlett's (1848) *Dictionary of Americanisms*. The sixth volume is concerned with works by language planners, including Molee (1888) and Williams (1890). The seventh and eighth volumes focus on articles and books by academics: Brander Matthews (1892, 1901, 1909), George P. Krapp (1919) and Gilbert M. Tucker (1921). Based on these texts, Davis identifies five "loosely overlapping periods, often corresponding to Americans' changing political awareness" (2003: xii). The first period is marked by an emphasis on linguistic innovation and opposition to British English, which was connected to the new political independence. During the second period, however, many of these presumably innovative forms were identified as older British forms, reflecting a change of focus from innovation to conservatism. Davis links this to "a cultural backlash against the perceived coarseness of frontier-dominated political life in the 1820's and 1830's" as well as to "the concerns of the new philology" (2003: xii). For the third period, Davis notes an increase in the awareness of American English as a distinct variety, which is reflected in authors trying to define standard American English and establish their authority over other competing positions (a notable example being Richard Grant White). Davis characterizes this period as being marked by great changes, the most important ones being "industrialization, urbanization, intensive immigration, and the expansion of the middle class and its educational institutions", and these changes were accompanied by great "cultural insecurity" (2003: xii). This not only increased the pressure to define a standard but also the difficulty of the task. The fourth period was different from the previous one because the emphasis shifted to the status of American English as a standard and a model in the world, a development which "corresponded to the awakening and growth of the United States as an imperial power from 1890–1914, and its involvement in Latin America and the Pacific" (2003: xiii). Davis dates the last period as starting after the first world war because of a "renewed focus" on the distinctiveness of American English, which was especially noticeable in Mencken's (1937

### 2 The emergence of American English: theories, descriptions, and models

[1919]) work on the *American language* (Davis 2003: xii). Regarding expert discourses on language, his interpretation thus suggests that the crucial period for the emergence of an American standard variety is the third one, which loosely spans from the middle of the nineteenth century to World War I.

Another attempt to define the beginning of a distinct American English variety can be found in Simpson's (1986) *The politics of American English, 1776-1850*. He claims that

American English as we recognize it today had been essentially established by 1850. That is, its major deviations from British English had by that time both been proposed in theory (mostly by Noah Webster) and adopted into relatively common (though not uncontested) practice. (Simpson 1986: 11)

This statement shows that on the one hand, Simpson distinguishes between theoretical ideas about language (a discursive dimension) and practice (a structural dimension). On the other hand, he implies a connection between the two – not only temporally, by saying that American English was established by 1850 on both levels, discourse and practice, but also by using the term *adopted*, which suggests that the theoretical ideas have shaped actual practice. The importance he attributes to discourse is highlighted further in his statement that "[t]hanks to the efforts of two generations of linguistic pioneers, Noah Webster foremost among them, and to the spectacular rise in national self-confidence, America had, by about 1850, a version of English that was recognizably its own" (Simpson 1986: 3). Furthermore, he also argues for the date 1850 primarily on the discursive level: First of all, he notes a change in discourse from a focus on social and political differences and tensions to a focus on unity and homogeneity, which is related to democracy having become "the dominant ideology or self-image, so that, in the continuing development of a self-declared pluralistic culture, a struggle of languages has been the harder to perceive where it does exist" (Simpson 1986: 7). Secondly, and this is his central argument, he finds a difference between two traditions of American literature and their representations of language. The first tradition, before 1850, has James Fenimore Cooper as a particularly representative writer, while the second tradition is marked by Transcendentalist authors:

In Cooper's world, language and society are presented as mechanical; the parts remain parts, without combining into any grand whole. Language is always made up of different languages in conflict, and they do not resolve themselves into any democratically representative common language. These are precisely the implications that the Transcendentalists avoid or cover over. (Simpson 1986: 252)

### 2.4 Structure and discourse in descriptions of the history of American English

Simpson's study is thus essentially a study to be located on the level of discourse, with a focus on literary works published between 1776 and 1850. The importance he attributes to the late eighteenth and first half of the nineteenth century in the creation of a distinctly recognized American English is acknowledged by Jones (1999) in his study *Strange talk: The politics of dialect literature in Gilded Age America*; however, as the title suggests, Jones argues that the latenineteenth century is even more important for the formation of American English: "I agree with Simpson that the real political clash of dialects that we find in Cooper disappears – by about 1850 – with the idealism of the transcendentalists [...]; it reappears, however, in an even more various and extreme way in post-Civil War literature" (Jones 1999: 215). Regarding the cultural background in which such dialect literature was published, he notes that after the Civil War, two changes could be observed with regard to the development of ideas about language: First, "the idea of an American English came into its own" and second, "this recognition of a new national language was accompanied by a new acceptance of dialect" (Jones 1999: 16–17). This acceptance and (scientific) interest in dialect was motivated by ideas that regarded language as crucial "to understand the mind and culture of its speakers", but, on the other hand, there was also a fear of dialect diversity connected to a fear of "contamination and fragmentation" (Jones 1999: 17). This new literary movement attracted the interest of a wide range of readers, reflecting a "cult of the vernacular" (Jones 1999: 39). Representations of dialect were not only popular in mainstream literature but they were also part of a large number of dialect sketches included in "highbrow literary magazines", which had a refined readership, and they were used by minority writers (such as Abraham Cahan and Paul Laurence Dunbar) in their works, which were read not only by white but also by Black Americans (Jones 1999: 7). His main conclusion is that "[d]ialect literature rose to prominence in the Gilded Age because it was integral to developing a cultural debate over the state of American English" (Jones 1999: 210). This shows that, like Simpson (1986), his study focuses on the discursive level by investigating the ideas surrounding American English, which led to the creation and transformation of linguistic ideologies, and by taking a closer look at the role that literature played in this process. Unlike Simpson, however, he regards the late-nineteenth century as the crucial period for the formation of American English.

David Simpson and Gavin Jones are both literary scholars and even though Simpson (1986: 11) makes claims about the influence of ideas about language on actual linguistic practice, their interest is mainly on the discourses on language and not on language use. The importance of the discursive level for investigating the historical development and especially the emergence of an American stan-

### 2 The emergence of American English: theories, descriptions, and models

dard has also been explored by several linguists in more detail. Cooley (1992) provides an important contribution by asking the question: "[U]nder what circumstances are some variant systems recognized as dialects while others are not? In other words, when and how is variation accorded perceptual status and named a language variety itself, a dialect?" (Cooley 1992: 167). She thus distinguishes between the presence of variation and its perception by speakers and uses this distinction to shed new light on the debate about whether early American English is marked by uniformity or diversity (on a structural level) – a debate that also plays an important role in accounts on the emergence of American English that are based on different theories of the emergence of new varieties (see §2.1.1). Cooley (1992) argues that there was "a co-existence of both diversity and uniformity in early American English" and that "[t]his co-existence may be reconciled by psycholinguistic or sociolinguistic principles rather than denied by arguments for a single state of uniformity or diversity or for sequences of one followed by the other" (1992: 168). An important point to consider here is the type of evidence – she finds that for the early American period most sources are descriptions and comments by grammarians, orthoepists and journalists as well as literary dialect, all of which constitute secondary sources which are not independent of the beliefs and perceptions by those who produced them. A change in these beliefs therefore necessarily had an effect on their descriptions, comments and representations of language and such a change was caused by the Revolutionary War and the War of 1812, which led to a "change of social and political allegiance, through which the colonists began to consider themselves Americans rather than Englishmen" (Cooley 1992: 180). In her view, a consequence of this development was a "psycholinguistic, perceptual, change of the standard variety" during the latter part of the eighteenth century and the early part of the nineteenth century, which is similar to Schneider's postulated change from an exonormative to an endonormative orientation. Her main argument is that the recognition and acknowledgement of language diversity requires the existence of a standard variety, marked by uniformity, against which the diverse sub-varieties can be delimited (Cooley 1992: 180). She regards the appearance of literary dialect in the late 1780s as another indicator of an increasing recognition of an American standard because only the existence of a standard would guarantee the recognition of representations of forms differing from the standard. Overall, her main argument is thus that historical evidence pointing to the uniformity as well as to the diversity of language in the United States needs to be interpreted by taking into account the beliefs of the commentators and writers, which are in turn shaped by social and political changes. It is important to note, however, that while Cooley (1992: 183) suspects that the growing recognition of a standard

### 2.4 Structure and discourse in descriptions of the history of American English

also influenced language use (even if unconsciously), she does not explore these connections in more detail.

The question of the relation between literary dialect and the emergence of an American standard is also addressed by Minnick (2010), who summarizes her main claim as follows:

An American standard for English, then, emerges through the contrast to it that the literary representations of vernacular speech provided, since they are replete with information about what is not standard by their very markedness. This contrast contributes to an image of invisibility for 'standard American', a prestige variety defined by no identifying characteristics of its own but rather only by what it is not: regionally or socially or racially marked. The 'otherness' of the vernacular-speaking characters of the local color tradition, then, is part of an increasingly vivid background against which an image of a 'standard' American English that otherwise has no appreciable identity of its own is rendered visible. (Minnick 2010: 181)

Minnick bases her argument on the one hand on the works by Anne Newport Royall, which were published early in the nineteenth century and in which she represented the speech of provincial characters from rural New England and the south, and on the other hand on writers in the Old Southwestern tradition, who created the figure of the rugged uneducated frontiersman or backwoodsman. Interestingly, Minnick (2010: 178–179) finds a strong ambivalence here because this figure and his speech are not only constructed as the "other" against which the narrator's speech is presented as standard, but the independence and the individualism of the figure is also celebrated in these texts, so that it becomes "a popular if stereotypical symbol for American national identity and particularly for American masculinity". This leads her to conclude that a positive evaluation of a standard did not automatically coincide with a depreciation of vernacular speech because these "varieties" could "index culturally popular values like masculinity, independence, bravery, and physical strength" (Minnick 2010: 180-181) as well. The situation is different for the third group of speakers whose speech is constructed as the "other" in literary works: African Americans. Representations of their speech functioned as implicit evidence of Black inferiority and there were no positive social values connected to it, at least not in works written by white authors. In general, Minnick's analysis thus locates the emergence of an American standard in the nineteenth century and the emphasis on its construction through contrasting it with "other" speech varieties in literary works clearly locates her work on a discursive level. Minnick's study also provides an important

### 2 The emergence of American English: theories, descriptions, and models

contribution to the analysis of the discursive construction of American English from a methodological point of view because she advocates the use of quantitative corpus methods for analyzing literary dialect to avoid interpretations and conclusions which are limited because they are only based on an impressionistic analysis of the data.

A very detailed study on the standardization of American English which analyzes discourses on language in America from a sociolinguistic point of view is Bonfiglio's (2002) *Race and the rise of standard American*. His focus is on pronunciation, and thus on the emergence of a standard American accent rather than a standard variety comprising all linguistic levels, and he aims to show "why and how [the] mid(western) accent rose to be perceived as the standard" (Bonfiglio 2002: 1). The linguistic form that is central to his analysis and discussion is the post-vocalic /r/: The two cultural centers of the country, New York and Boston, were marked by the absence of post-vocalic /r/ and yet its presence became a defining feature of the national standard, for which he uses the label *American network standard* because it came to be used by broadcasting media. To uncover reasons for this rather unusual development (in comparison to England and other European countries, where the speech of the cultural centers became the standard), Bonfiglio analyzes on the one hand a multitude of texts to identify the linguistic ideologies of "influential figures in the United States" and how they relate to ideologies of race and ethnicity. On the other hand, he also investigates how the views expressed in these texts were perceived and received by speakers that were heard by many people: popular actors, announcers and politicians. One important result of his study relates to the time period during which the standard became recognized: He regards the early twentieth century as the crucial period and thus explicitly argues against Labov's (2006: 225) speculation that the shift in the prestige of /r/ occurred in the 1940s in connection with changes resulting from World War II. He rules out an earlier starting point for the emerging pronunciation standard because he regards the radio as crucial for the transmission of knowledge of pronunciation: "In 1920, radio had not yet begun its programmed broadcasting. Thus knowledge of pronunciation remained largely local, and impressions of the speech of other regions was not gained directly but spread largely by word of mouth" (Bonfiglio 2002: 47). This affected also the discourse on /r/:

[…] both the description and prescription of pronunciation of /r/ remained largely local and tended to generalize based upon the regional custom until the advent of regular radio broadcasts. Non-rhotic /r/ is observed and

### 2.4 Structure and discourse in descriptions of the history of American English

recommended on the east coast, while rhotic /r/ is preferred in the central states. (Bonfiglio 2002: 54–55)

This view must be challenged based on the studies noting the enormous interest in representations of dialect in the nineteenth century. Even though it is clear that the impression of pronunciation gained through these representations is not as direct as that which can be gained by listening to actual speech, it nevertheless provides the readers with an idea of what people in other regions sounded like. In the following analysis, I will also show that representations of non-rhoticity (and, implicitly, also rhoticity) circulated in nationwide newspaper discourses especially in the last two decades of the nineteenth century, so that discourses on /r/ do not seem to be as local as Bonfiglio suggests.

In addition to the advent of radio broadcasting, Bonfiglio identifies two other reasons for (mid)western speech patterns, particularly rhoticity, to become the standard accent in the early twentieth century. The first are "xenophobic and anti-Semitic movements" (2002: 4) at that time. They built on an increasing consciousness of race and ethnicity that Bonfiglio sees emerging in the postbellum period and ultimately leading to a shift in prestige of eastern cities, especially of New York City, from being positively viewed as cultural centers to being negatively viewed as "contaminated" by poor immigrants. This shift in prestige also affected the prestige of the eastern speech patterns, including non-rhoticity, which became seen as racially different and as impure. This emphasis on race, which is underlined by the title of his study, is explicitly directed against the view that the emergence of an American pronunciation standard, or more specifically the shift in prestige of post-vocalic /r/, is a result of differentiation from British English speech – a view that he labels a "myth" (Bonfiglio 2002: 2).

Connected to this is the second reason, namely the construction of the (mid) western region and their speech patterns as an ideal:

The (mid)western accent was constructed and desired by forces external to the area itself that projected a preferred ethnicity upon that region and defined it within a power dynamic of difference, i.e. it was precisely not the speech of the ethnically contaminated areas of the northeast metropolis and the south. (Bonfiglio 2002: 8)

His argument is thus that it was not the (mid)west per se that was attractive, but it was the negative image of the south and the northeast that led to its positive valorization – he speaks of "antigravitational forces" (2002: 72) here. Similarly, this valorization was not pushed by the (mid)westerners themselves, but was

### 2 The emergence of American English: theories, descriptions, and models

the result of the cultural power exerted by the northeastern population (2002: 72). This is for example visible in "the decision of Harvard and other Ivy League Universities to seek the sustenance of their proper ethnicity and culture in rural western regions", which indicates "the onset of a shift in the prestige discourse of the educated man" (Bonfiglio 2002: 230). An important figure in this valorization process is the "western hero as an instantiation of the proper American male" whose "speech patterns came to function as metonymies of the condition of nostalgia, sentimentality, and tradition" (Bonfiglio 2002: 231). There are interesting parallels to Minnick's (2010) finding that there are positive evaluations of a similar type of masculinity, characterized by physical strength and toughness needed for a life at the frontier, in the dialect representations in the literature of the Old Southwestern tradition. She takes this as evidence that while the speech of these figures was marked as non-standard, it was nevertheless not condemned but indexed positive characteristics as well. This suggests that this positively evaluated masculinity noted by Bonfiglio (2002) has its roots in the early nineteenth century.

It is obvious that Bonfiglio's analysis must also be located primarily on the discursive level. His basic argument is "that folk linguistic beliefs determined the national standard", which is why he aims to "illuminate the larger cultural factors that informed the folkish linguistic beliefs in question" (Bonfiglio 2002: 73). However, this argument shows that he also makes claims about changes in language use – the standard is not only an idea, but a model that influences people's choice of linguistic forms. This is visible, for example, in the statement that "Americans gravitated toward the pronunciation associated with a "purer" region of the country, and they did so in a largely non-conscious manner" (Bonfiglio 2002: 4). This implies that speakers aligned with a model of speech that they evaluated positively, and that this alignment process was not the result of a conscious effort. Similarly, the following statement illustrates his view that a regional diffusion of a linguistic form resulting in a linguistic change in a region is caused by non-linguistic factors: "The migration of the American continuant postvocalic /r/ from the western states eastward, its supplanting of the dropped postvocalic /r/ of the east coast, and its rise to standardization began in the twenties and was precipitated by the axis of radio, anti-immigration, and westward nostalgia" (Bonfiglio 2002: 60). Bonfiglio's claims and arguments are thus reminiscent of the claims underlying the theoretical framework of enregisterment outlined in §2.2.1 and the model of social positioning developed by Spitzmüller (2015, 2016) (see Figure 2.4 in §2.2.2). It thus supports my position that an analysis of the discursive construction of American English is an indispensable part of studying the emergence of American English. His focus on a linguistic form, /r/,

### 2.4 Structure and discourse in descriptions of the history of American English

and its "culturally constructed value" (Bonfiglio 2002: 7) is also interesting from the point of view of enregisterment. However, the strong focus on one form, even if it has a "high cultural visibility" (Bonfiglio 2002: 6), is also problematic as it does not reveal anything about its relation to other forms – if American English is investigated as a process of enregisterment, it is necessary to study how *sets of* perceivable signs, linguistic forms and other non-linguistic signs, become linked to social and cultural values and thus differentiable registers.

Bonfiglio's major claims, namely that race played an important role in the standardization of American English and that language ideology, that is beliefs about language, are central to the standardization process, are also shared by Milroy & Milroy (2012). In their third edition of *Authority in language*, an often-cited work on processes of language standardization and the role played by prescriptivism in these processes, they conclude that

In the US, bitter divisions created by slavery and the Civil War shaped a language ideology focused on racial discrimination rather than on the class distinctions characteristic of an older monarchical society like Britain which continue to shape language attitudes. Also salient in the US was perceived pressure from large numbers of non-English speakers, from both long-established communities (such as Spanish speakers in the South-West) and successive waves of immigrants. This gave rise in the nineteenth and twentieth centuries to policies and attitudes which promoted Anglo-conformity. (Milroy & Milroy 2012: 160)

Even though they do not cite Bonfiglio's (2002) study, their review of the literature thus confirms his emphasis on race in contrast to the emphasis on class in Great Britain. However, concerning the periodization of the process of standardization in the United States, they regard a different period as particularly important: "Heavy immigration to northern cities between 1880 and 1920 gave rise to conflicts of dominance between immigrant groups and older élites, and to labour conflicts which had the effect of crowding out democratic ideals of equal rights in both north and south" (Milroy & Milroy 2012: 160). Especially the importance of radio broadcasting is thus not regarded as an important factor in their analysis, which constitutes a major difference to Bonfiglio's account. However, it needs to be noted that Milroy & Milroy do not focus on pronunciation alone, which makes radio broadcasting perhaps less important as a factor. On the other hand, they do not analyze the standardization process by focusing on the linguistic forms themselves – it is not clear which forms become standard and which

### 2 The emergence of American English: theories, descriptions, and models

do not, so that the reasoning about the causes for standardization remains on a rather general and abstract level.

With regard to the importance of language ideologies, Milroy & Milroy (2012: 162) not only agree with Bonfiglio (2002), but they also link their findings on standardization processes to the understanding of language ideologies proposed by the linguistic anthropologists Woolard, Silverstein and Irvine and Gal. They particularly point out that these beliefs have central social significance by recognizing their essential function of helping language users "to make sense of the socially structured language variability which they observe every day" (2012: 162) and also, beyond the individual level, their role in "delimiting and defining salient social groups and indeed whole nations" (2012: 163). It is precisely this delimitation process that is at the heart of the theory of enregisterment and this study seeks to identify not only which linguistic forms were constructed as 'American' but also which underlying ideologies made this construction possible.

The overview thus far shows that there is much interest in the history of American English which focuses on the discursive level, but the views with regard to the crucial periods for the recognition of American English as a distinct or standard variety vary and depend, at least to a great extent, on the material studied: publications by "experts" on language or other influential figures on the one hand, and literary works, especially dialect literature, on the other hand. This underlines the need for further studies investigating different materials – a need which is met by this study, as it focuses on newspaper articles.

With regard to the structural level, i.e. the structural differentiation between American English and British English as well as between distinct American varieties, it needs to be noted that systematic and detailed studies on the historical development of linguistic forms in America are actually rather scarce. Montgomery (2001: 152) notes for example that "[t]he field has many simplistic statements and assumptions about what must have occurred in new-dialect formation in the American colonies, rather than documentation of input varieties and the extent to which these were maintained". His article is mainly concerned with summarizing what is known about British and Irish antecedents, but in doing so he also offers theoretical and methodological insights. He criticizes for example that some hypotheses which are not supported by enough linguistic evidence have "achieved a life of their own" as "part of constructed American memory" (Montgomery 2001: 111). An example is the postulation of linguistic ties between Massachusetts, Virginia, or Appalachia with southeastern England, southern or southwestern England, and Ulster respectively (Montgomery 2001: 109). In his view, linguistic studies are required in which several types of evidence are carefully analyzed and interpreted: This evidence should not only comprise

### 2.4 Structure and discourse in descriptions of the history of American English

secondary sources like popular observations by outsiders, commentary of grammarians and lexicographers and literary attestations but also primary sources like poetic rhymes and original texts and manuscripts because the latter have a higher value in reconstructing structural changes and differences than the former (Montgomery 2001: 97). An example for such a collection of original texts providing insights into the variation present during the earliest phase of settlement is the Early American Corpus containing texts from New England from the 1620s to the 1720s, among others records of Salem witchcraft trials from 1692 (Kytö 2004). In a pilot study, Kytö (2004) finds that there is hardly any evidence of grammatical and phonological forms that can be classified as *dialect input*, that is as forms found in local vernacular British dialects. The few dialect forms that do occur are found in speech-related records. The other forms reflect instead an "early prestige language" that also has its origin in Britain and that was taken by the settlers to America. The fact that these prestige forms constitute the majority of forms in the corpus is explained by Kytö (2004: 151) by the educatedness of most of the authors of the documents included in the corpus. In general, Kytö does not expect "major differences from the language of the mother country" and her analysis and interpretation rather supports the continuity between the (socio-)linguistic variation present in the areas that the settlers originated from and the variation present in the New England settlements.

The difficulty of determining the beginning of structural differentiation based on primary sources can for example be illustrated by the discussion revolving around the beginning of White Southern American English. Bailey's (1997) article addresses the question "When did Southern American English begin" by analyzing several types of primary data, among them phonetic records of southern speakers born in the nineteenth century from the Linguistic Atlas of the Middle and South Atlantic States (LAMSAS) and the Linguistic Atlas of the Gulf States (LAGS), as well as orthographic evidence from the Tennessee Civil War Veterans Questionnaires (TCWVQ), a collection of documents written by white male Tennessee speakers, most of whom were educated (Bailey 1997: 260). He focuses on twelve phonological features and seven grammatical features, which he regards as distinctively southern (he calls them "long-established stereotypes of SAE", 1997: 258) and finds that only six phonological and four grammatical features were clearly part of Southern American English by the middle of the nineteenth century. The other six phonological and two grammatical features appeared or were increasingly used in the nineteenth century, particularly in the last quarter (in the period after 1875). He thus concludes that Southern American English begins in the last quarter of the nineteenth-century and hypothesizes that drastic social changes in the south during that time created a situation which was

### 2 The emergence of American English: theories, descriptions, and models

particularly conducive to the diffusion of linguistic changes (Bailey 1997: 271). The most important changes were the increasing number of villages and towns which became connected through railroad tracks and the increasing geographical mobility.

Bailey's hypothesis was tested seventeen years later by Montgomery et al. (2014), who published their results in an article entitled "When did Southern American English really begin?" Between the two studies, new evidence had been found based on which the Corpus of American Civil War letters was created, which provided new primary evidence to shed light on the beginning of Southern American English. Based on an analysis of the grammatical features investigated by Bailey (1997), the authors find that "the crucial period for the developing distinctiveness of Southern American English must be pushed back at least one generation". They evaluate Bailey's argument that the social changes in the last quarter of the nineteenth century correlate with the beginning of Southern American English as not very convincing because based on research on the history of the American South they assume that "other periods probably witnessed substantial innovation and diffusion, too" (Montgomery et al. 2014: 334). In support of their own argument, they stress instead that historians "have argued that by 1830 the South had become a self-conscious region increasingly at odds with the nation at large", which is why they ask the question of whether "regional consciousness [could] have played a role in the formation of regional standards of speech" (Montgomery et al. 2014: 345).

These two articles discussing the "beginnings" of White Southern American English thus show that especially the scarcity or even absence of primary evidence makes it difficult to reconstruct linguistic changes, especially on the phonological level. Furthermore, differentiation may have proceeded differently on different levels of structure (phonology, grammar, lexicon), which makes the question of when a variety becomes distinct difficult to answer. Montgomery et al. (2014) test Bailey's (1997) hypothesis only on the grammatical level, so that it is not clear whether the earlier beginning that they postulate would also apply to the phonological level. Quantitative statistical measurements, as proposed by Pickl (2016), are difficult or impossible to apply given the amount of data that is available for such analyses, but this only strengthens Montgomery's (2001) call for finding and analyzing more primary evidence. Lastly, their discussion of extra-linguistic factors influencing structural differentiation shows how a case can be made for both the late nineteenth century and the first half of the nineteenth century, which underlines the need for investigating this relationship in more detail.

### 2.4 Structure and discourse in descriptions of the history of American English

This need has been acknowledged and addressed by Montgomery (2015). He provides further support for locating the beginning of "distinct (type)s of English" (2015: 99) in the South in the period between 1750 and 1850 by not only investigating more primary evidence (letters and a testimony written by semiliterate commoners), but also secondary evidence: 51 primary-level confederate schoolbooks. So despite his emphasis on investigating primary sources, Montgomery suggests here (albeit tentatively) that "the development of distinctive Southern English may have involved ideology leading to *perception* of the South as a distinct region perhaps as much as the reality of one" (2015: 99). His argument is that those forms that were subject to comment in the textbooks must have been salient in some way, and they must have existed since at least the 1840s. He finds that most comments pertained to forms of pronunciation, and he values this evidence not only as providing "glimpses of many features, some of which (such as the drawl) come from a period earlier than previously documented" but also as evidence of southerners' awareness "of linguistic contact and competition [...] and [of] the new-dialect formation that was the result" (Montgomery 2015: 114). This study is valuable because it draws attention to the discourses surrounding particular linguistic forms in the south, especially their evaluation as correct or incorrect. It can thus be determined which linguistic forms played a role in the discursive construction of southern American English. However, the material can only function as a starting point for further investigations and his article shows that there are still many open questions, for example relating to the prestige (and distribution) of non-rhoticity (see §3.3.5 for a detailed discussion of research on this form) and the relation between this early awareness of southern forms and the "new, modern Southern identity" identified by Schneider (2007: 299) which led to the recognition and increasing use of innovative features which mark the present-day American south as a distinct dialect region.<sup>18</sup>

The difficulties of investigating the emergence of new varieties on a purely structural level also become visible in the case of African American English. In a recent overview and discussion of research on "the origins and history of African American Language", Lanehart (2017) not only reviews several research positions but also criticizes that research focuses too much on a set of salient linguistic features and that several perspectives on the development of African American Language "use these salient features for various purposes and sometimes in con-

<sup>18</sup>Schneider (2007) distinguishes between "traditional" and "new" southern features – the former are associated with the rural pre-Civil War culture and the latter with the modern, urban culture resulting from the social changes following the defeat in the Civil War. For southern (linguistic) identity, the middle of the nineteenth century therefore marks a turning point.

### 2 The emergence of American English: theories, descriptions, and models

tradictory ways to support their argument" (2017: 86).<sup>19</sup> She finally argues that with regard to structural developments, "we simply do not have the artifacts and hard evidence (recordings of nascent AAL) to make a definitive assessment about the origins and history of AAL" (2017: 91). In her discussion, she also deals implicitly with the discursive construction of African American Language by focusing on the contribution of linguistic research to this construction. She describes the Deficit position in the nineteenth century, according to which Blacks are biologically inferior to whites and thus not able to acquire English in the way that whites do – this position continued to be supported throughout the twentieth century. Lanehart argues that not only this position but also the following ones (the two most prevalent ones being the Anglicist position, which claims that African American English is based on British English varieties, and the Creolist position, which purports that the language spoken by African Americans developed from an earlier creole) are influenced by the "ideological and epistemological perspectives of their originators and supporters", which shows that research is influenced by social, political and cultural circumstances. Research discourse also has an influence on the recognition and evaluation of African American English, but this influence also has its limits, as Lanehart points out: "[W]hen I tell people outside of linguistics about AAL, they seem dumbfounded that anyone would believe that AAL is not historically rooted to Africa since the people who speak it are, hence the African Diaspora" (2017: 92). Her overview thus strengthens the view developed in §2.1.2 that investigations of structure are very often influenced by discourse, and while it might not be possible to completely disentangle structure and discourse in the investigation of language in general and of the historical emergence of new varieties in particular, it is important to critically reflect the ways in which one's own investigation of structural developments is shaped by the discursive construction of the variety in question.

Lanehart also addresses the issue of defining African American Language – an issue that is also discussed by Mufwene (2001b) in much detail. Mufwene draws the following conclusion:

So far, we have done poor jobs either in not reconciling some of our definitions of AAE [African American English] with our analyses, in overemphasizing extreme differences and disregarding similarities with other English vernaculars, or in proposing definitions that ignore the sentiments of native speakers. We might even be better off not even trying to define AAE and

<sup>19</sup>Lanehart prefers the label *African American Language* over *African American English* because *language* is "less limiting" than *English* (2017: 86).

### 2.4 Structure and discourse in descriptions of the history of American English

just speaking of peculiarities observable among African Americans. There is probably no way of defining AAE – if a language variety can be defined at all – that does not reflect a particular bias, and this problem is true of any language variety in the world. (2001b: 37)

He suggests instead a vague characterization of African American English as an ethnolect – as "English as it is spoken by or among African Americans" (2001b: 37). This puts emphasis on the speakers as a basis for defining the object of investigation – a perspective that is also adopted by Lanehart (2017). This reflects their research position, namely the importance of viewing language not only as a set of linguistic forms, but as more than that: "[L]anguage is more than the sum of its parts or the handy grammar that we all like to turn to [...]. If language could be learned from reading a grammar book, we could all be multilingual" (Lanehart 2017: 91). In addition to linguistic forms, aspects of perception and recognition are thus crucial, as pointed out by Mufwene (2001b: 37):

I doubt that African Americans utilize just one rigid battery of structural features to identify a person as speaking English in a manner that corresponds to their own. For the purposes of group identity, I think that being able to recognize speech as African American on the family resemblance model, based on a disjunction of kinds of peculiarities, is more realistic than doing so on the basis of whether its speaker has more or fewer specific nonstandard features.

To conclude this overview of the role of structure and discourse in prior research on the emergence of American English and, related to it, other varieties in North America, it can be stated that even though it was not possible to give a complete overview here because of the large amount of literature on their historical development, the question of when American English became a distinct variety has been answered in different ways. As studies on the distribution and change of linguistic features, especially those based on primary evidence like original texts and manuscripts, are rather scarce, the "beginning" of American English has usually been determined primarily on the discursive level, by identifying the point in time that American English came to be *recognized* as distinct from British English. While the Revolutionary War and the following independence from Britain are usually seen as important events because they mark the starting point of the American nation and are thus a prerequisite for the recognition of a *national* variety, some have stressed that this recognition process was completed by 1850, while others found the late nineteenth century or the

### 2 The emergence of American English: theories, descriptions, and models

early twentieth century to be crucial periods for the definition of an American (standard) variety. Despite those differences, there is general agreement that the recognition process not only proceeds through delimitation from British English, but also through the recognition of American sub-varieties against which a uniform, national American variety is constructed. The linguistic differences related to race and ethnicity have been identified as particularly important and also as constituting a striking difference to standardization processes occurring in England. Another difference between England and America that has motivated investigations is that in America it was not the speech of the political, cultural, and economic centers in the northeast that became the standard, but the speech of the rural (mid)west, while in England the London speech patterns came to be recognized as the standard. Despite a focus on discourse, claims have been made about its influence on language use – on the other hand, studies focusing on language use also make claims about the correlation between the changes they observe and social and cultural developments happening at the same time. Bonfiglio's (2002) study is special in this regard because he links discourses surrounding a specific linguistic form, post-vocalic /r/, to what is known about changes in the use of that form by American speakers. As post-vocalic /r/ is also one of the forms investigated in this study, I will summarize the research findings on this particular form in more detail in §3.3.5. Finally, not only contemporary observers are influenced by ideologies, resulting for example in different assessments of the uniformity and diversity of American English, as shown by Cooley (1992), but also linguists contributing to the current research debate, particularly (but definitely not restricted to) the investigation of African American English. In this study, I will use enregisterment as a theoretical framework to contribute to the question of how American English was constructed as a discursive variety – but in contrast to prior studies, my analysis focuses on particular linguistic forms and how the social and cultural meanings they acquired led to the recognition of a set of forms as distinctly American. Rather than following the majority of studies investigating discourses on language shaped by language experts, influential figures or authors of literary works, I will focus on newspaper articles because they had a wide and varied readership and because newspapers contained several different text types, for example editorials, news reports and advertisements as well as anecdotes and humorous paragraphs intended to entertain the readers (see §3.1 for more details).

2.5 Conclusion

# **2.5 Conclusion**

The theoretical framework developed in this chapter is intended to provide a basis for gaining deeper insights into the role of social factors and identity in the emergence on new varieties of English. I have demonstrated in §2.1 that this point is one of the most contested issues in discussions about models that have been proposed so far: Trudgill's model of new-dialect formation, Schneider's Dynamic Model and Kretzschmar's model of the emergence of varieties in speech as a complex system. I have shown that the arguments they present interact in crucial ways with their definition of the emerging construct, the *variety*, and that an investigation that aims to shed light on the role of social factors needs to distinguish carefully between different types of varieties: structural varieties, perceptual varieties and discursive varieties. This postulation of different types of varieties does not imply that they are to be understood as existing independently of each other – on the contrary, they are crucially connected. However, it is the main argument of this study that in order to explore these connections, they have to be investigated in their own right and this requires a sound theoretical and methodological framework that does justice to the different types of varieties. As this study is primarily interested in the role of social factors, I proposed to focus on the emergence of discursive varieties, and I demonstrated in §2.2 that the theory of enregisterment provides a useful framework for this task because it describes how speakers construct registers through engaging reflexively with linguistic forms (and other perceivable signs) and evaluating these linguistic forms in different ways. Linguistic forms thus become indexically linked to different social values and social personae, and the more frequently reflexive activities occur, the more salient these indexical links become and the more speakers are likely to recognize these links and contribute to their persistence or to their change through their own reflexive activities. It is through this process that registers are constructed – discursive varieties which can be understood as cultural models of action consisting of a set of linguistic and non-linguistic forms which are recognized as distinct from other sets of forms by a group of speakers.

As shown in §2.2.2, the theory of enregisterment and indexicality has already been fruitfully applied in sociolinguistic research, but especially the distinction between the concepts of register and style as well as the role of awareness in the creation and recognition of indexical links and registers deserves closer attention. I have argued here that it makes sense to distinguish style, enregistered style and register: Whereas the first two concepts are located on the level of language use, the third concept operates on a more abstract level because it is essentially a *model* of speech (and action in general) that cannot be identified

### 2 The emergence of American English: theories, descriptions, and models

by investigating speakers' use of linguistic variants but only by investigating their reflexive activities, in other words their typifications of linguistic forms in metadiscursive practice. In line with Agha's theory of enregisterment, sociolinguists and linguistic anthropologists have suggested to understand identity as being produced through social practice – practices in which speakers position themselves and others socially. How registers affect this process of social positioning is explained and modeled by Spitzmüller (2013, 2015, 2016).

With regard to the perceptual variety and its relation to enregisterment, I demonstrated in §2.2.3 that studies in perceptual dialectology have also drawn on the concept of enregisterment. I have suggested that theories and methods of perceptual dialectology can mainly be used to add a cognitive perspective to the question of how linguistic forms become enregistered. They cannot be used, however, to study enregisterment in a historical context, which is why they do not play a role in the present study. In contrast to that, the new field of discourse linguistics can contribute to the study of enregisterment and thus to the study of the construction of discursive varieties in crucial ways. This can be done first of all by understanding this construction process as the result of discursive action, that is, of linguistic action that constitutes knowledge. Secondly, the methodological framework DIMLAN that takes into account the intratextual layer, the transtextual layer and the agent layer, which functions as a filter between the intra- and the transtextual layers, provides a reference point for developing a methodological approach for the systematic study of discursive activities that bring about socially shared knowledge about language in the form of cultural models of action and thus of registers.

The model presented in §2.3 visualizes the enregisterment process and thus the relation between the structural level and the discursive level that have been distinguished in the Dynamic Model. As it is the central aim of the next chapters to apply this model to the investigation of the enregisterment of American English, I have presented an overview of previous research that has described (aspects of) the development of American English and addressed the question of the "beginning" of this variety. I demonstrated that this question has been answered in very different ways and, more often than not, based on data that belongs to the discursive level. The *recognition* of American English as a new variety was foregrounded, whereas the structural differentiation was much less studied. But even in studies that can be located on the discursive level, not much attention has been given to the specific linguistic forms that the recognition of the new variety was based on. Furthermore, it has not been sufficiently investigated which speakers actually recognized linguistic forms as distinct and when and where these processes of recognition could be observed. This underlines the need for

2.5 Conclusion

a study that aims to investigate the emergence of a discursive *American* variety systematically and empirically. However, I have also outlined important suggestions by Cooley (1992) and Minnick (2010) concerning the manner in which the recognition of a standard American variety proceeded. That these processes can also be identified in the following study of enregisterment in nineteenth-century America will be shown in the remaining chapters of this book.

# **3 Tracing enregisterment processes of American English: aims and methodology**

This chapter describes the methodology used for the present study of enregisterment processes of American English in the nineteenth century. It begins by justifying the focus of the study on one type of material, newspaper articles, and continues by describing the databases and method of data collection and analysis. Section 3.3 provides an overview of prior research on the phonological variables which are the focus of this study, followed by an overview of what is known about the historical development of the lexical forms investigated in this study in §3.4. Based on this, Section 3.5 develops the specific research questions of the present study.

# **3.1 Newspapers as a source for enregisterment processes**

Prior studies on enregisterment have focused on a variety of material (see the overview in §2.2.2). Many studies have not just analyzed one type of material, but several; by doing so, they not only do justice to Agha's observation that different genres have a different circulation and therefore contribute in different ways to the creation of speech chain linkages (2003: 259) but also to the discourse-linguistic call for taking into account a multiplicity of texts, produced by different actors and occurring in different media (Spitzmüller & Warnke 2011: 187–188). Using only newspaper articles for studying enregisterment therefore seems like a restriction which limits the amount of insight that can be gained. While this is of course true to some extent, there is, however, also an advantage to such a restriction: The procedure adopted here allows for a comprehensive and extensive analysis of one genre by including larger quantities of texts. So far, the only studies which included newspaper articles in their analysis analyzed a small number (20 articles by Johnstone et al. 2006 and a similar number of articles by Remlinger 2009), whereas the present study is based on close to 1,200

### 3 Tracing enregisterment processes of American English

articles. The availability of large electronic databases comprising close to 78 million newspaper articles published in the nineteenth century allows for systematic and focused searches and combinations of quantitative and qualitative analyses of the data. For example, the large number of articles makes it possible to analyze the development of speech chain linkages both temporally and regionally and these insights provide important contextual information for analyzing the evaluation of speech forms based on a detailed qualitative analysis of a smaller number of articles.

The large number of newspaper articles, collected in archives and made electronically accessible in databases (see §3.2), has further advantages for studying enregisterment processes. Agha (2003) describes the genre of penny weeklies, popular periodicals which reached a wide readership because of their low price, and argues that they were largely responsible for the expansion of metadiscourses on accent in England due to their large circulation. They did not form an independent speech chain structure but were closely linked to other texts because writers publishing texts in penny weeklies drew on a variety of sources and genres with a lower circulation (prescriptivist works, popular handbooks as well as novels and literary work). I argue that newspaper articles in the United States of the nineteenth century came to fulfill the same important function: They enlarged the social domains of registers by acquainting a large number of people with their linguistic forms and indexical values. Evidence for this claim is provided by historical research on the development of the number of newspapers and their circulation. By 1800, there were already 234 newspapers, most of them weeklies, which together had a circulation of 145,000 copies (Copeland 2002: 149). Huntzicker (1998: 453) estimates that at the beginning of the nineteenth century the readership of each newspaper ranged from at least a few hundred to one thousand people. In the first quarter of the nineteenth century, the number of newspapers increased from about four hundred in 1810 to more than eight hundred in 1825, which "made the United States by far the greatest newspaper country in the world" (Nord 2001: 88) at that time. But it was essentially the 1830s in which major changes in the newspaper business started to occur. Not only did the number of newspapers continue to increase to more than 2,300 in 1900 (Huntzicker 1998: 453), but they were increasingly published daily, printed in larger quantities and, in addition to subscriptions, sold as single copies on the streets for a low price (sometimes for just one cent, leading to the term *penny press*). While some newspapers still reached only a few hundred people in 1900, other newspapers (especially those published in big cities) reached more than 100,000 readers (Huntzicker 1998: 453). This development would not have been possible without technological advances. Particularly important were the steam press, which

### 3.1 Newspapers as a source for enregisterment processes

was first used by *The Times* in London in 1814 and which accelerated the speed of printing, so that the output, the number of copies, could be increased greatly, and the steam railways, which made it possible for papers to be distributed across much wider geographical regions (Barker & Burrows 2002: 6). But it was not only the case that eastern papers could increase their distribution to new western territories. Settlers moving westwards usually took newspapers with them because, as Cloud (1998: 232) points out, "having a local newspaper represented stability and legitimacy". The so-called *frontier press* had of course more functions than this rather symbolic one: Next to several political and economic functions, it also provided information and reading material in regions characterized by a high literacy rate but also by a lack of books and libraries (Cloud 1998: 232). In addition, western newspapers often assumed the role of "town boosters" (Cloud 1998: 232): They promoted their communities to outsiders with the aim of getting their articles published in eastern newspapers and to attract the interest of potential new settlers. In the far west of the country, the number of newspapers therefore increased greatly especially in the second half of the nineteenth century. Cloud (1998: 233) notes that there were 11 newspapers in the 11 western states and territories in 1850, and more than 1,000 by 1890. Not just in the West, but in the whole country it was especially in the last two decades of the nineteenth century that the number and circulation of newspapers rose extraordinarily. Nord (2001: 228) regards this period as "the genesis of the modern mass-circulation newspapers in America" because the number of newspapers increased by 78 percent in the 1880s and "the circulation of all dailies jumped 135 percent, from 3.6 to 8.4 million per day". In addition, he points out that the beginning of the "modern, mass-circulation national magazine" can be found in that period as well (Nord 2001: 228), which is important because magazine articles were often reprinted in newspapers.

While the large number of newspapers, their wide circulation and their presence in even remote regions of the country are certainly indicative of a large social domain, it could be argued that the domain of newspaper readers is nevertheless restricted to particular social groups. Unfortunately, not much is known about the readership of newspapers. Nord (2001: 225) points out that "[n]early all of the research in the history of newspapers and magazines and much of the research in the history of the popular book has centered on the production, not the consumption, of reading materials". However, one possibility to find out more about newspaper consumption is used in Nord's (2001) study: He analyzes a statistical survey of the family cost-of-living budget, which includes the amount of money spent on books and newspapers.<sup>1</sup> The survey was conducted by the Bu-

<sup>1</sup>Expenditures on books and newspapers are not distinguished in the survey, but the assumption is that if people spent money on books, they would also spend money on newspapers.

### 3 Tracing enregisterment processes of American English

reau of Labor Statistics headed by Carroll D. Wright from 1890-1891, and Nord (2001) extracted a random sample of one hundred working-class families in the cotton textile industry from the 1891 report to find out to what extent they spent money on newspapers and how this amount was affected by income, region, nationality and type of community.<sup>2</sup> His findings show that working-class families did indeed spend money on newspapers and books, but not all to the same degree. One factor influencing the reading expenditures was the degree to which the family's income depended on children. The more children contributed to the family's income, the less the family spent on reading materials. Region was an influential factor as well: Southern families spent much less on newspapers and books than northern families, even if the factor of the children's contribution to the family income was controlled for. Nationality, that is, the difference between native-born workers and immigrants, played a role in the North, where nativeborn Americans spent more on reading materials, even though they earned less money on average than immigrant working-class families.<sup>3</sup> With regard to the last factor, community, Nord (2001: 239) distinguishes between "traditional interpersonal communities" (*Gemeinschaft*, in Ferdinand Tönnies's terms) and "modern contractual society" (*Gesellschaft*) and finds that reading as an activity is rather associated with the latter (Nord 2001: 240). This finding leads him to conclude that "[t]hough reading was a common activity for all groups in my sample of working-class families, the more avid readers seem to have been more at home with the institutions of the modern industrial society" (Nord 2001: 240) and the less avid readers, especially southerners and a particular immigrant group in the North, French Canadians, were more committed to the institutions of the *Gemeinschaft*, such as family and church. Even though Nord's (2001) study is restricted to working-class families in the late nineteenth century, it provides a very detailed insight into the factors that influenced whether people read newspapers or not and it can be assumed that they also played some role in the rest of the nineteenth-century.

With regard to the early nineteenth century, historians also assume that the readership of newspapers was already quite large and varied. Copeland (2002: 141) investigates the role of the press in the creation of an American public sphere in the period from 1750 to 1820 and finds that

<sup>2</sup>The sample focuses on one industry, the cotton textile industry, because "it was the *only* industry well represented across regions and ethnic groups" (Nord 2001: 229). The working-class families come from six states in three regions where this industry was thriving: New England (Massachusetts and New Hampshire), the mid-Atlantic region (New York and Pennsylvania) and the South (Georgia and South Carolina).

<sup>3</sup> In the South, all families of the sample were native-born Americans so that an influence of nationality could not be observed here.

### 3.1 Newspapers as a source for enregisterment processes

[t]he prominence of newspapers and other forms of print reflected both the breadth of popular involvement in public debate and the widespread use of the press to facilitate and promote this process. But it was not just the backcountry farmer, middle-class merchant or elite, educated planters, lawyers and politicians who had access to, and used the public forum afforded by, the press. High literacy rates in America, which exceeded 90 per cent in some regions by 1800, meant that even the poor were more often literate than not, and ensured that access to the public sphere was restricted neither by gender nor by race.

However, as already indicated by Nord's (2001) study, it is clear that even though access to newspapers was not restricted per se, there were nevertheless huge differences between social groups and between regions. In addition to the factors analyzed by Nord (2001), the difference between whites and African Americans needs to be considered as well. In an overview of the historical development of literacy in the United States, Pawley (2010: 49) states that "[w]hile for most white Americans the nineteenth century saw an expansion of reading opportunities, for African Americans the picture was more somber". This was especially true in the South. Even though some slaves were taught how to read or taught themselves, literate slaves were also feared and seen as rebellious and dangerous. Following South Carolina 1740 and Georgia 1755, several slave states passed anti-literacy laws in the first half of the nineteenth century to make literacy learning among African Americans illegal. It is therefore not surprising that the first African American newspaper, *Freedom's Journal*, was founded in the North, in New York City, in 1827, and that it was aimed at "a small community of free African Americans who lived mostly in northern cities" (Amana 1998: 26). However, after the Civil War, the situation changed. Amana (1998: 26) states that roughly 4 million African Americans were freed and formed new communities that were in need of newspapers. After the end of the Reconstruction period, many African Americans left the South and migrated to the North and the West. The increasing demand for newspapers by an increasingly literate African American readership led to the rise of African American newspapers from 40 papers in the pre-Civil War period to almost 200 papers at the end of the nineteenth century (Amana 1998: 26). This shows that the social domain of newspaper readers increasingly included African Americans as well.

Databases consisting of a large collection of newspapers are therefore a good source for studying enregisterment processes with a potentially large social domain. But for the analysis it is not only important to ask who the readers were

### 3 Tracing enregisterment processes of American English

but also who was involved in the production and publishing of articles. Dicken-Garcia (1998) states that modern journalistic practices developed in the nineteenth century. While newspapers in the eighteenth century were often produced by a printer and an assistant only, the nineteenth-century production process involved editors and reporters. The development of the mass press in the 1830s led to an increasing demand for nonpolitical news and changed the way that news were gathered. Before the 1830s, it was common to take articles from other papers, print word-of-mouth reports and notes of congressional sessions. While this practice continued, it then became the task of reporters to identify news stories that could be of interest to readers, to travel to the places where the stories were taking or had already taken place, to gather information about them and to construct an account for the public (Dicken-Garcia 1998: 585). Eyewitness reporting and correspondent reporting also became important. Essays which expressed the printer's (political) opinion and which had dominated newspapers before the nineteenth century became separated from news sections and were labeled *editorials*. So in general, reporters wrote most of the news articles, but ultimately the editors and owners of the papers were the ones who decided which articles to print. While most owners, editors and reporters were white and male, there were also approximately 300 female reporters by the 1880s and the growing number of African American papers also suggests that African Americans increasingly became actors in newspaper discourses as well. In addition to news articles, there were other text types which appeared in newspapers and were produced by other actors. Literary texts, such as poems or longer novels, were often printed in newspapers (novels were printed serially), company owners could publish advertisements, and readers themselves had the chance to write letters to the editor, which could potentially be printed as well. During the Civil War era the *column* developed as "an article of moderate length that appears on a regular schedule under the byline of its author" (Riley 1998: 438). The goal of a column was to be interesting and, according to Riley (1998: 439), the earliest columnists were women because owners hoped to attract a wider female readership. There were also many male columnists; literary authors and humorists started to contribute articles on a regular basis (Riley 1998: 439). Beginning in the 1850s, techniques to print illustrations were taken over from England and improved, so that artists became another group of actors that contributed to newspapers (Everett 1998: 267). Not only were there specialized illustrated newspapers like *Frank Leslie's Illustrated Newspaper* and *Harper's Weekly*, founded in 1855 and 1857 in New York City, but illustrations appeared more and more frequently in other newspapers as well, often as reprints of those appearing in illustrated newspapers and magazines. Illustrations, columns and humorous short

### 3.2 Databases, data collection and method of analysis

texts, which were often published in rubrics called for example "Multiple News Items", reflect the development that newspapers increasingly "turned into cheap consumer products to be sold for profit" since the 1830s and that this new penny press' idea of reporting the news was "to tell interesting stories of occurrences" (Nord 2001: 104). It is therefore important to consider that one important aim of the owners was to sell newspapers and to publish articles that promised to be attractive to a large audience.

As shown above, the variety of actors producing and publishing texts went hand in hand with the variety of text types that appeared in newspapers, which makes them a multi-faceted resource for studying enregisterment. Newspaper articles did not only potentially increase the social domain of a register, but they also contributed to the creation and development of the social range of values through the diversity of text types. Newspaper articles were, to use Agha's words, "linked to earlier genres by a speech chain structure" (2003: 259). Many oral and written texts which were produced outside of newspapers, like prescriptivist texts, fictional and non-fictional books, political speeches, plays and other performances, conversations on the street and many others, were discussed in newspaper articles, which made these texts not only known to a large number of people, but also provided them with an interpretation and evaluation of these texts. Consequently, studying newspaper articles can potentially also provide insights into how other genres and texts contributed to enregisterment processes. In §3.2, I will give an overview of the databases used for the study and the methods of data collection and analysis.

# **3.2 Databases, data collection and method of analysis**

As pointed out in Chapter 2, an important goal in any study of enregisterment is the identification of metapragmatic stereotypes, that is "social regularities of metapragmatic typifications [which] can be observed and documented as data" (Agha 2007: 154). Agha observes that

[t]he fact that metapragmatic stereotypes are expressible in publicly perceivable signs is not just a matter of convenience to the analyst interested in identifying and studying registers. It is *a necessary condition on the social existence of registers*. (2007: 154)

The analyst must therefore develop a way to obtain data points which provide evidence for metapragmatic stereotypes and for reasons pointed out in §2.2.4, I adopt a discourse-linguistic approach to achieve this. With reference to the

### 3 Tracing enregisterment processes of American English

possibilities summarized in Table 2.6, I outline my approach to data collection and analysis in this section.

First of all, I define the objects of the study as discourses on language in nineteenth-century America which are observable through statements in newspaper articles published in the United States. These articles are products of action and constitute a series of discursive events which then constitute larger discourses on language. As illustrated in Figure 3.1, the objects of study are therefore located at the intersection of discourses appearing in newspaper articles and discourses on language in general. The initial access to discourses on language is consequently deductive, but to avoid the circularity that often comes with such an approach (studying discourses on language implies that such discourses exist in the first place), I complement this approach by an inductive one by focusing on linguistic variables which I identified through an analysis of statements and not based on prior knowledge about differences between British and American English (see §3.3 and §3.4 for details). The focus on linguistic forms is very useful in a study on enregisterment because it is essentially the recurrent evaluation of linguistic forms which is at the heart of the construction of a discursive variety and because the forms provide transtextual links between a set of statements and texts.

Figure 3.1: Objects of study

Secondly, the concrete methods of data collection follow from this definition of the objects of study. As a basis for the study, I use two large databases which contain close to 78 million newspaper articles published in nineteenth-century America: *America's Historical Newspapers* (AHN, ~60 million articles) and *Nineteenth-*

### 3.2 Databases, data collection and method of analysis

*Century U.S. Newspapers* (NCNP, ~18 million articles).<sup>4</sup> The exact number of newspapers, issues and articles in the two databases is listed in Table 3.1. 5


Table 3.1: Number of newspapers, issues and articles in each database.

While these collections constitute a corpus of newspaper articles, they cannot be regarded as representative because providing a representative sample of newspaper articles was not the goal of the compilers of the database. The group of experts responsible for choosing the newspapers (five journalism historians for the NCNP and eleven scholars and experts in various fields, mainly historians and librarians, for the AHN) rather aimed at including as many aspects as possible. The company Gale advertises its NCNP database as providing "an asit-happened window on events, culture and daily life in 19th-century America" and states that it features

publications of all kinds, from the political party newspapers at the beginning of the nineteenth century to the mammoth dailies that shaped the nation at the century's end. Every aspect of society and every region of the nation is found in the archive – rural and urban, large cities and small towns, coast to coast, etc. Includes major newspapers as well as those published by African Americans, Native Americans, women's rights groups, labor groups, the Confederacy, and other groups and interests. (Gale 2018)

Similarly, the flyers advertising the AHN database emphasize the "unparalleled breadth and depth" of the collection (Readex, Early American Newspapers, series

<sup>4</sup>The database *Nineteenth-Century U.S. Newspapers* is a product of Gale, a Cengage Company, and is only available for purchase by institutions. The database *America's Historical Newspapers* was created by Readex, a division of NewsBank, and can also only be accessed through a subscription (usually via libraries or institutions).

<sup>5</sup>The AHN database includes articles from 1690 to 1922, but the present study only includes articles from 1800-1899. There is a small number of newspapers which are included in both databases. It was not possible to determine the exact number of articles which appeared in both databases, but compared to the overall size of the databases, I estimate that the number is so small that it will not skew the results.

### 3 Tracing enregisterment processes of American English

1) and the focus on the "record of daily life in hundreds of diverse American communities", so generally including "local and national perspectives" (Readex, Early American Newspapers). Figure 3.2 and Figure 3.3 show the temporal and the regional distribution of articles.<sup>6</sup> Especially the temporal distribution illustrates that the sample is not representative of the development sketched in §3.1. While the NCNP database shows an expected increase of articles over time, the AHN database includes an unexpectedly large number of articles from the first two decades. With regard to the regional distribution, the NCNP database seems more balanced (if also not representative). The AHN database clearly has a stronger focus on the northeast. Nevertheless, it is important to note that if both databases are combined, newspapers from every region of the United States are included.<sup>7</sup>

Figure 3.2: Number of newspaper articles per decade in the databases AHN and NCNP and in both databases

<sup>6</sup>The maps showing regional distributions of articles were created using the R packages maps (Becker et al. 2016) and ggplot2 (Wickham 2009) as well as an R script developed by DeSante & Sparks (2012).

<sup>7</sup> For technical reasons, Alaska and Hawaii are not depicted on the maps. Alaska is the territory with the lowest number of articles in the database (2,539 articles, compared to 13,111,797 articles from Massachusetts, the state with the highest number of articles). 62,574 articles were published in Hawaii, which makes the state number 46 on the frequency list (see Appendix). In addition to the 48 states depicted on the map and Alaska and Hawaii, the District of Columbia is listed separately in the newspaper databases. It is much higher in the frequency list (number 12), with 1,940,043 articles published in the two databases. As Hawaii, Alaska and the District of Columbia are not found on the maps, I will list the number of articles underneath the maps which show the results of my searches.

### 3.2 Databases, data collection and method of analysis

Figure 3.3: Number of newspaper articles per state in AHN and NCNP and both databases

The advantage of having such a large corpus is that it not only increases the likelihood of finding articles containing statements relating to language, but that it also minimizes the possibility that only a very specific point of view is taken into account. Furthermore, it is not only possible to say when and where statements occurred, but also when and where they did not occur. So in order to obtain statements I collected articles containing the phrase *American language* to identify linguistic forms which were associated with this label. This thematic approach served as a basis for a systematic collection of data, which involved a strong focus on these linguistic forms. I used one or a combination of search terms which represent the linguistic form in question by using a different spelling (for example *hinglish* 'English' representing /h/-dropping and -insertion) and searched all articles in the corpus which contained the search terms (see §3.3.1 for more details on pronunciation respellings and their methodological value).<sup>8</sup>

<sup>8</sup>The articles have been digitized from microfilm using Optical Character Recognition (OCR) technology. According to the information provided by the compilers of the AHN database, newspapers are particularly difficult to digitize because of the wide variety of fonts, formats and other elements and because of the fact that the paper and/or the ink have often deteriorated over time. Consequently, sometimes articles containing words with a similar shape appear in

### 3 Tracing enregisterment processes of American English

As the linguistic forms and the search terms guiding the data collection were not pre-defined but the result of an initial data analysis, this approach is not only corpus-based but also corpus-driven. The case studies which involve the analysis of the collected articles containing the search term(s) form the backbone of the study on which the identification and description of discourses on language and therefore the enregisterment of an *American* variety rests. It is therefore obvious that a diachronic approach is used because the question of when enregisterment processes can be observed and how they developed in the nineteenth century is a key question of the present study. The key components of data collection are illustrated in Figure 3.4.

Thirdly, when it comes to actual procedure, the emphasis is clearly on a focused analysis of articles containing the search term(s). Focusing the discourse analysis on the datasets (each dataset consisting of all articles containing a specific (combination of) search term(s)) makes it possible to systematically analyze discursive structures, including regional distributions and temporal developments. While this approach puts an emphasis on the transtextual layer, the intratextual layer is not neglected. Qualitative analyses of individual articles follow a heuristic procedure in order to find out how the linguistic forms are evaluated, which strategies are employed in this evaluation process and how these evaluations contribute to the emergence of an American register. This heuristic analysis forms the basis for further focused and quantitative procedures because it is then possible to find out how many articles employ the same evaluation and/or the same strategy. Due to the fact that this project is conducted by an individual and not the result of a collaborative research project, the restriction to case studies involving seven linguistic forms was necessary. Nevertheless, I included some additional analyses to assess the reliability of the results, e.g. by comparing the results in the two databases to each other in order to ensure that they are not an artefact of the database used, and by adding datasets based on different search terms representing the same variable to ensure that it is indeed the phonological form and not the lexical item which is subject to evaluation in the article.

Before stating my precise research questions in §3.5, I will describe the linguistic variables which are at the center of the case studies and provide an overview of prior research on these variables in §3.3 and §3.4. In these sections, I also explain which search terms I used for collecting each dataset and justify that choice.

the list of results and have to be excluded manually. More importantly, it is possible that the computer did not find *all* of the articles containing the search term(s) in the database. It should also be noted that I included results in the dataset in which the search term appeared as part of a compound, e.g. in *Hinglishman*.

### 3.3 Phonological forms and their written representations

Figure 3.4: Key components of data collection

# **3.3 Phonological forms and their written representations**

# **3.3.1 Pronunciation respellings and their methodological value**

Finding instances of phonological forms which are part of enregisterment processes in the nineteenth century is of course difficult because no primary spoken data is available which could provide a record of the exact phonological forms. Historical linguists often rely on spelling to study phonological variation and change (see for example the contributions in Hickey 2010), but in nineteenthcentury newspaper articles the spelling is already standardized to a high degree. However, there is an exception: spellings which explicitly draw attention to phonological variants and signal a differential pronunciation. Picone (2016) labels them "pronunciation respellings" and emphasizes the importance of distinguishing them from *eye dialect*, a term which also denotes a non-standard spelling, but without a corresponding difference in pronunciation. For example, if *was* is spelled <wuz>, both spellings, <was> and <wuz>, represent the *same* phonological form [wəz] (the weak form of [wɑːz]), which makes <wuz> a case of eye dialect. But if *that* is spelled <dat>, it signals a phonological form ([dæt]) which is *different* to the one indicated by <that> ([ðæt]), which makes <dat> a pronunciation respelling signaling TH-stopping.

These pronunciation respellings are a valuable source for enregisterment processes because they provide a clue as to which phonological form is evaluated and therefore part of the register's repertoire. This view is shared by Honeybone & Watson (2013: 335), who investigated pronunciation respellings of Liverpool accent features in humorous dialect literature to assess their sociolinguistic

### 3 Tracing enregisterment processes of American English

salience and who suggest that especially those phonological features which they have identified as being salient are likely to be enregistered for Liverpool English. In general, pronunciation respellings could be subsumed under the category *literary representations* in Agha's (2007) classification of typifications of language use (see Table 2.2 in §2.2.1) even though they do not only appear in literary texts, but also in other text genres, as I will show below. It is typical of this category that even though the phonological forms are explicitly indicated, the evaluations can be very implicit. If they are used in a literary text, for example in the representation of the speech of a character, the form is linked to the social attributes of the character, his or her social relation to other characters and his or her social behavior. These attributes and social relations and their links to language are rarely explicitly described but have to be inferred by the reader. However, explicit evaluations of linguistic forms are also a possibility, for example in editorials containing discussions of language and using such respellings as examples which are supposed to illustrate the pronunciation of the variant. The advantage of using pronunciation respellings is that a search can be conducted which includes all instances of a particular spelling in a corpus. For example, searching the databases for *dawnce* yields all newspaper articles containing this particular form, which makes it possible to identify exhaustively all social values which become indexically linked to it and also to get an overview of when and where they occurred. Due to the fact that pronunciation respellings are by far not as frequent as the same words spelled the regular way, the datasets were also of a manageable size – large enough for quantitative analyses, but small enough for reading every article for qualitative analyses. An analysis of such a set of articles is then comparable to the dataset obtained by a different search term, which makes it possible to find out whether forms are evaluated similarly or differently and therefore belong to the same or a different register.

However, there is also a possible disadvantage to using pronunciation respellings to study the enregisterment of phonological forms. The problem is that respellings rely on already established norms of spelling so that, as Agha (2007: 197) points out, "the reader can only construe defective spelling as an implicit comment on defects of pronunciation". Similarly Jaffe & Walton (2000: 582) write that

Our research suggests that it is almost impossible to avoid stigma in the nonstandard orthographic representation of others' low-status speech varieties. *Light* orthography cues voice, but it does so by using stereotyped forms whose meanings are inescapably linked to their use in texts whose aim is to denigrate the speakers being represented. *Heavy* orthographies require too

### 3.3 Phonological forms and their written representations

much investment and decoding to allow voice to come through; few readers are able or prepared to sustain the work of attending closely to spelling as a vehicle for voice.

This implies that the means of representing the phonological forms already bring about a negative social evaluation, signaling a deviation from a norm, defectiveness and incorrectness. In my view, however, even though a negative interpretation of pronunciation respellings is certainly likely, the possibility of a positive interpretation is not ruled out. For example, if the form is innovative or at least has not long been an object of metadiscursive activity, the writer could use the respelling primarily to illustrate the form. Since spelling is a way of rendering abstract descriptions of phonology more vivid and illustrative, and as it is also the best way of representing phonological changes if the genre is written and does not allow for long descriptions (e.g. dramatic texts consisting mostly of dialogue), there is often no alternative for indicating a particular phonological form, even in cases where the evaluation of the form is positive. Moreover, if a new standard register is in the process of being constructed, a deviation, which is normally seen as defective, has at least the potential to be reinterpreted as positive and as an index of a newly emerging norm. Lastly, even if the evaluation of the form is indeed negative, it still constitutes evidence for positive indexical values linked to the alternative phonological form and contributes to the strengthening of the standard register as well as to the clarification of the repertoire and the social range of that register. In any case, it is important to look closely at the context in which the respelling occurs. As the citation by Jaffe & Walton (2000) shows, it makes a difference whether there is only one or few respellings or whether there are so many respellings (possibly also in combination with eye dialect) that the text becomes difficult to read and that the voice that is being represented is actually almost silenced in the process. In the analysis, it is therefore important to determine how many other respellings accompany the one that serves as the basis for the collection of the dataset, which ones are particularly salient (through their sheer frequency or due to their foregrounding by other means), which phonological forms these salient respellings represent and whether they are constructed as belonging to the same or a different register.

In the following sections, I will describe the phonological forms and the respective search terms used for the present study and give reasons for the choice of these forms. I will also provide a brief summary of prior research on the historical development of these forms, both on a structural and a discursive level. While it can be argued that the structural development is not relevant in the context of the present study, as it is concerned with the discursive level only, I still see the

### 3 Tracing enregisterment processes of American English

two levels as linked to each other. Linguistic forms that become enregistered do not come out of nowhere, and a change in the use of these forms (possibly as a consequence of enregisterment processes) might in turn lead to the creation of new indexical values. This is why I will not only discuss what is known about attitudes and evaluations of these forms but also about their use by speakers of English around the world.

# **3.3.2** *hinglish***: /h/-dropping and -insertion**

The search term *hinglish* is a differential spelling of the word *English*. It draws attention to a process called /h/-*insertion*: the insertion of unetymological /h/ in syllables withouth an onset (Lass 2006: 65). This process is closely linked to the process of /h/-*dropping*, also labeled /h/-*deletion* or /h/-*loss*, which describes the non-existing (or "silent") pronunciation of the consonant. In an overview on the historical development of /h/, Minkova (2014: 104–108) dates the emergence of /h/ as a distinctive phoneme to the end of the tenth century. Throughout history, it was marked by great instability. First, it became increasingly deleted, then it became used again. The exact dating and the extent of these processes is subject to debate (see Milroy 1983, Lass 2006, Crisma 2007, Schlüter 2009, 2012), but it is clear that the variability between [h] and Ø was influenced by etymological factors (Romance vs. Germanic origin of the word), linguistic factors (stressed vs. unstressed syllable), regional factors (Northern vs. Midland or Southern varieties) and social factors (Minkova 2014: 105–108). Research on the Present-Day English situation in England shows that /h/-dropping only occurs in unstressed syllables (mainly function words like *he*, *her*, *his*, *her*, *have*, *has* etc.) and is otherwise restricted to very few lexical items (e.g. *heir*, *honest*, *hour*) in Standard Southern British English. Nevertheless, /h/-dropping in stressed syllables has remained present in England until today and is subject to regional and social variation (see Ramisch 2010). The existence of /h/-insertion is usually interpreted as a sign of linguistic insecurity of the writers and therefore as evidence for /h/ dropping because the writers seem to have been unsure about where /h/ occurs and therefore inserted it in words which never have been pronounced with an /h/ historically (Milroy 1983). For this reason, I argue that the spelling *hinglish* represents both insertion and deletion, even though it only marks the insertion explicitly.<sup>9</sup>

<sup>9</sup>This is also supported by Milroy's (1983: 42) assumption that they have "their origin in a single phonological change: loss of underlying /h/ in speech". Furthermore, the qualitative analysis of newspaper articles shows that they are regularly treated as one phenomenon in metapragmatic reflexive activities.

### 3.3 Phonological forms and their written representations

The development of /h/-dropping and -insertion in North America has hardly received any attention. Descriptions on the historical development of American English usually only note the absence of /h/-dropping (if it is mentioned at all). Krapp (1925, 2: 206) for example writes that

[T]he [town] records do not indicate that at any time or in any region was the loss of *h* [h] in words with this sound in the initial position, or the addition of *h* at the beginning of words with initial vowels, familiar to all in Cockney speech, current in American use.

In the light of these findings, it is not surprising that there are no studies on the development of attitudes towards and evaluations of the form in an American context. Krapp's (1925, 2: 206) statement which I cited above indicates, however, that the form was well-known in America and that it was associated with the speech of the Cockney, a stereotypical English figure. The mention of this figure suggests that Americans were familiar with discourses on language in England (see also the examples cited by Bailey 1996: 131). These discourses have been studied in most detail by Mugglestone (2003) and her analysis shows that /h/ dropping and -insertion occupies a very prominent position in them. She finds that the first comments on the form revealed a rather tolerant attitude towards it and to variation in general. However, this period was followed by

a transition, both in terms of comment on /h/ and its attendant connotations [which] seems to come decisively with the second half of the eighteenth century as the nuances of accent were incorporated into the prescriptive consciousness, and the renewed interest in elocution forced a new awareness of the ideals of speech. Fashion, together with a heightened responsiveness to the role of external markers in assigning social status (real or intended), was also to play its part. (Mugglestone 2003: 99)

The evaluation of /h/-dropping since the mid to late eighteenth century has been very negative. Mugglestone (2003: 99) cites a prominent example of such a comment: Thomas Sheridan's statement that "There is one defect which more generally prevails in the counties than any other, and indeed is gaining ground among the politer part of the world, I mean the omission of the aspirate in many words by some, and in most by others" (Sheridan 1762: 34, cited by Mugglestone 2003: 99). This citation illustrates the association between /h/-dropping and defectiveness and incorrectness and it shows the reasoning behind it. Dropping the /h/ is marked as rustic and regional (a feature of speech in the counties as opposed

### 3 Tracing enregisterment processes of American English

to the fashionable speech in the metropolis, in London) and seen as an index of a lower social standing (people who drop the /h/ were generally not considered to be a member of the "politer part of the world"). According to Mugglestone (2003: 115), the negative social meanings came to prevail and the absence of /h/ became an important marker of a lack of education (with the underlying argumentation that it signaled the illiteracy of the speaker) and a lack of refinement. As mentioned above, /h/-dropping became linked to the figure of the Cockney, a lower-class Londoner, and it was regularly used by novelists, such as Charles Dickens, to mark figures as socially inferior (Mugglestone 2003: 122). The negative social meanings of /h/-dropping and -insertion consolidated until the 1850s and 1860s and the form was firmly established by then as a "prime symbol" of the social divide between upper and lower classes (Mugglestone 2003: 113). With regard to /h/-insertion, Mugglestone (2003: 108) observes that it was used particularly often to mark "the social (and linguistic) stereotype of the 'new rich', and the 'self-made-man'", a figure who seeks to become part of the elite social circle, but even though he has acquired wealth, he has not acquired the socially acceptable manners and language, which is made visible by his failed attempt to produce the /h/ word-initially, resulting in frequent /h/-insertions. All in all, Mugglestone (2003)shows convincingly that the omission and insertion of /h/ became stigmatized to a considerable extent and that the form became a linguistic shibboleth which features prominently in discourses on language. It is therefore not surprising that it is also one of the two linguistic forms which are part of Agha's (2003) analysis of the enregisterment of Received Pronunciation.

This overview shows that while /h/-dropping and -insertion is characterized by variability on the structural level and relatively stable negative evaluations on the discursive level in England, it seems to be of no relevance in an American context. This raises the obvious question of why the form should be part of a study of enregisterment processes of American English at all. The answer is related to the methodological approach used in the present study: It was the initial exploratory survey of newspaper articles containing the search term *American language*, which led to two interesting articles pointing to the importance of /h/-dropping and -insertion. The first one is a letter to the editor which appeared in the *Tucson Daily Citizen* on July 25, 1882, and in which a reader asks: "Will you kindly inform me what constitutes the American language?" To this question, the editor replies with only one sentence: "It is the English language with the 'H's' in their proper places". This answer suggests that the *absence* of /h/-dropping and -insertion in America is used as a defining feature of the newly emerging variety and that it deserves close attention in the present study. In an earlier article, attention to this form is given more indirectly: It is a letter "from an

### 3.3 Phonological forms and their written representations

American gentleman at Paris" which was published in five newspapers in 1830, among them in the *Daily National Intelligencer*, a Washington D.C. newspaper. The American describes the "ignorance that prevails in Europe, with regard to our country" and especially with regard to the language spoken in America. He gives anecdotal evidence to support this: "A Yorkshire man, who was my fellow traveller on the top of a coach, upon learning I was an American, complimented me by saying, "Yees talks ez gued Hinglish az hi duz"". This not only indicates an expectation by Englishmen that Americans do not speak "good English", but also a view held by Americans that they speak English better than Englishmen. That /h/-dropping and -insertion plays a prominent role here is supported by the spellings <Hinglish> ('English') and <hi> ('I') in the representation of the speech of the Yorkshire man and also by the following imitation of this speech by the American, who replies ironically: "Yees, sur, hand hize cum to Yorkshire, said I, tu perfect my hedication". The choice of *hinglish* as a search term is based on this article, in which *hinglish* appears a second time as part of the compound Englishman, spelled <Hinglishman>.

The spelling of the word *education* as <hedication> in the letter points to another difference between the speech of the American and the Englishman. The representation of the sounds /juː/ in the second syllable by the letter <i> could indicate the loss of the palatal glide /j/ before the vowel (in addition to the central realization of the vowel). This so-called *yod-dropping* is the second phonological form investigated in the present study, albeit with a different search term, for reasons which I will provide in the next section.

# **3.3.3** *noospaper/s***: yod-dropping**

*Yod-dropping* is the term that is often used for a process that involves the elision of the consonant /j/ in onset clusters of a class of words that historically had either /iu/ or /ɛu/.<sup>10</sup> According to Lass (2006: 88), this elision process has started in the eighteenth century in words where /j/ followed /r/ (e.g. in *rue*, *true*) or /l/ (e.g. in *blue*, *glue*). It occurred less commonly after /s/ and the deletion of /t/, /d/, /n/ "has never caught on in the British standards" (Lass 2006: 88). Concerning the present-day variability in the British Isles, Hughes et al. (2012: 68–69) find that yod-dropping after /r/, /l/ and /s/ is almost complete in Received Pronunciation and in many other accents and that in several other regional accents, /j/ is also dropped after other consonants – this process is most advanced in eastern

<sup>10</sup>This class of words is different from the class of words deriving historically from Middle English /oː/, e.g. *boot* (Lass 2006: 87).

### 3 Tracing enregisterment processes of American English

England where yod-dropping occurs before all consonants. Regarding metadiscourses on yod-dropping in England, Lass (2006: 88) notes that it is stigmatized after /t/, /d/ and /n/ and cites Walker (1791) as an early source: He evaluated *noo* 'new' and *doo* 'dew' as 'corrupt' Londonisms. In the nineteenth century, there were different voices: Sweet regarded the pronunciation of *news* without /j/ as 'vulgar' (Sweet to Storm, January 21, 1889), whereas Ellis (1869: 601) listed both pronunciations, with and without /j/, without favoring one of them (both sources are cited in MacMahon 1998: 473). The situation in the twentieth century is interesting because in contrast to the retention of /j/ after /t/, /d/ and /n/, which is regarded as a form which is part of Received Pronunciation and which is therefore evaluated as standard and correct, the retention of /j/ after /s/ is "considered amusingly old-fashioned" and affected (Hughes et al. 2012: 69).

In America, yod-dropping has been described as occurring after all coronal consonants, so that it only remained after velars and labials – this development has been labeled "Later Yod Dropping" by Wells (1982a: 247). While he notes that it is a feature of General American today, he also points out that there is variability: "Some easterners and southerners, however, have either /ju/ or the diphthong /ɪu/, and GenAm usage is not entirely uniform" (1982a: 247). An indicator for such variability being present from at least the early nineteenth century is a comment by Grandgent (1899: 224–226, cited by Krapp 1919: 196), which indicates, in Krapp's words, "a confusion in the use of [u] and [ju] in New England speech which was at its height about 1820 and which affected both polite and dialect speech" – a sign of this "confusion" was the insertion of [j] in the onset clusters of words like *bruised* and *smoothed*. This comment suggests that [j] after [r] was already so uncommon that the presence of [j] is interpreted as being as odd as its insertion in a word where historically a [j] had never been present. That this variability continues to at least the early twentieth century is indicated by Krapp's (1919: 98) observation that yod-dropping correlates with region: "It seems to be less general in the South, than it is in New England, whence it has spread to all sections of the country". This is in line with Schneider's (2007: 299) classification of yod-*retention* as a traditional southern form, which has started to disappear in the new south. Phillips (1994) notes that Kurath & McDavid (1961: 174) found that yod-retention after [t], [d] and [n] was a form found mostly in the south and the south midland, but she adds that "it is also preferred by some cultivated Northerners". This indicates that yod-retention was and is not an exclusively southern form and with regard to the change that can be observed regarding yod-dropping she also reports on a study by Pitts (1986) who found that while southerners increasingly drop the yod, there are some northerners, especially newscasters, who produce it (Phillips 1994: 115). A recent study by Feagin (2015)

### 3.3 Phonological forms and their written representations

focuses on the development in the south, more specifically in Anniston, Alabama, and finds that while yod-retention was a form used by all social classes in the south, "it has developed into an emerging feature of working-class speech in the younger generation" (Feagin 2015: 361–362)). This suggests that there are three developments with regard to the evaluation of yod-dropping: One is the association of yod-dropping with a national norm, which, according to Schneider (2004: 272), motivates the general increase of yod-dropping in the south. Secondly, there seems to be a retention of an old link between yod-retention and cultivation which seems to have motivated some northerners (particularly newscasters) to use forms with the glide. In contrast to that, the third development is a southern one: Here, yod-retention marks the speech of young working-class speakers, which suggests a possible association between the presence of yod and workingclass values that seem attractive to young speakers. However, Feagin finds that, in contrast to other variables (e.g. non-rhoticity), yod-retention or -dropping is not subject to overt comment and therefore part of changes occurring below the level of awareness, which are "interesting as unwitting reflections of community norms" (Feagin 2015: 363). In her view, the overt prestige of forms like rhoticity leads to its adoption by speakers of all social classes, while the covert prestige of forms like yod-retention leads to their maintenance by working-class speakers:

These larger economic pressures mean that the local upper class tends to accommodate culturally and linguistically to national pressures, including to the national prestige variety, while the working class does not. In stark contrast, working-class speakers align themselves with the very Southern white rural culture of country music, stock car racing, and hunting, and with speech which could be called the local covert prestige variety. This may explain why the two varieties are diverging as much as they are. (Feagin 2015: 364)

However, Phillips' (1981: 72) report on her own experience as a student at a southern university in the 1960s shows that yod-dropping is not unnoticed by southern speakers:

Before I entered Duke University from my south Georgia high school in 1967, my older brother instructed me that the correct way to pronounce the university's name was [djuk]. When I arrived on campus, however, I found that all my new acquaintances called it [duk], and that soon became my pronunciation as well. The social significance of the change was not driven home to me, however, until I saw hostile slogans from the nearby University of North Carolina at Chapel Hill reading "Beat Dook". The antagonism

### 3 Tracing enregisterment processes of American English

between the Southern state university and the "Northern school set in the South" was crystalized in the pronunciation of that one word.

Several comments showing that metadiscourses on yod-dropping in America have existed since at least the eighteenth century can be found in Krapp (1919, 1925, 1, 1925, 2). With regard to prescriptivist comments, he observes that

[t]he dictionaries generally authorize only the first of these pronunciations [with [j]] after [d], [t], [θ], [n], [s], and academic authority is very likely to condemn the pronunciation of [uː] as uncultivated, in spite of the fact that it occurs widely in the speech of educated and informed people. It has long been current in America, as is evident from Noah Webster's defense of [uː] in **duty**, etc., as the best pronunciation. (Krapp 1919: 95–96)

His statement is thus indicative of an early controversy regarding the evaluation of the presence or absence of yod: While Webster, a very influential figure, advocated yod-less forms, the majority of prescriptivists were opposed to them. With regard to the representation of dialects in literature, Krapp (1925, 1: 239) cites C. Alphonso Smith, a "Southern scholar and historian", who regards the "vanishing *y*-sound" as being "almost a shibboleth of the Southerner to the manner born and helps to differentiate him from the Westerner and the Northerner". Opposed to this view, Krapp (1925, 1: 240) comments that yod-dropping "is so general as to have little value as a mark of local dialect, especially literary dialect, for dialect writers have rarely made any effort to distinguish by spelling or otherwise between the pronunciation of *tune* with the vowel in *boot* and with the vowel of *mute*".

Given this controversy, yod-dropping is thus an interesting form for an investigation of discourses on language and the discursive construction of American English – that it played at least some role is suggested by the use of the form *doocid* 'deuced' in some articles that contained other search terms. I have chosen a different lexical item, however, because the lexical item *deuced* is linked to a particular characterological figure, the American dude, and I wanted to rule out lexical influence here. I chose *noospaper*, in its singular or its plural form (hence the precise search term *noospaper* OR *noospapers*, represented here as *noospaper/s*) for three reasons: First, the lexical item *newspaper* was not likely to be connected to a specific group of people or to a specific context. Secondly, I expected it to be fairly frequent (especially in newspaper articles) but not too frequent to make a qualitative analysis impossible. The third reason was that it

### 3.3 Phonological forms and their written representations

is a fairly long word and therefore more likely to be identified correctly by the OCR software.<sup>11</sup>

# **3.3.4** *dawnce***: realization of the bath-vowel**

What I refer to here as the bath-vowel is the vowel which appears in words of the standard lexical set bath, which is defined by Wells (1982a: 133) as "comprising those words whose citation form contains the stressed vowel /æ/ in GenAm, but /ɑː/ in RP". This difference between these two current standard accents is the result of a historical development which Wells (1982a: 134) terms the trap–bath split. Lass (1999: 103–108, 2006: 89–90) dates the beginning of this split back to the seventeenth century, based on Cooper (1687), who notes a lengthening of /æ/ in words like *passed*, *cast*, *gasp*, *barge*, and *dart*, while the vowel remained short in words like *pass*, *bar*, and *car*. It seems like consonant clusters (/-rC/ and /-sC/) following the vowel have favored the initial lengthening. In the following years, lengthening expanded to more phonetic contexts and there is evidence for a shift in quality as well. Lass (1999: 106) cites Flint (1740) who describes a lowering of the lengthened vowel to [aː] before /r/ (in *art*, *dart* and *part*) and variably in *bath*, *castle*, *calf* and *half* (see Kökeritz 1944). Another historical source cited by Lass (1999: 106) is Nares (1784), who has [aː] in *after*, *ask*, *ass*, *bask*, *mask*, *glass*, *pass*, *plant*, *grant*, *advance*, *alms*, *calm*, and *palm*. Jones (2006) and MacMahon (1998) provide detailed analyses of eighteenth- and nineteenth-century comments on this vowel change and they agree with Lass (1999, 2006) that no clear pattern of the distribution of the lengthened and lowered vowel can be discerned based on the sources. Although the phonological environment clearly had an influence, some contexts seem to have had a stronger effect than others. While [aː] occurred fairly generally before tautosyllabic /r/, it was much more variable before tautosyllabic voiceless fricatives and nasals. The orthoepists often disagree on which words contain the lengthened and lowered vowel, which makes it even more difficult to identify phonological constraints on the distribution of the vowel. Jones (2006: 198) therefore concludes that the "BATH/TRAP split was manifested primarily through lexical distribution". With regard to the temporal development of the split, MacMahon (1998: 436-437) describes it as "a gradual shift in favor of /aː/" and he finds that "the last decade of the eighteenth century and the first two of the nineteenth must be regarded as the critical period of change". Next to

<sup>11</sup>An explorative search for the term *noos* 'news' yielded more than 5,000 articles in the NCNP alone, many of which contained *news* or other similar lexical forms. A manual check of all newspaper articles to identify those that actually contained *noos* would have been too timeconsuming.

### 3 Tracing enregisterment processes of American English

the phonemic split, which resulted in minimal pairs like *Sam* and *psalm*, the realization of the new phoneme /aː/ has also been subject to variation and change. Innovative retracted variants of /aː/ (e.g. [ɑː]) have been noted by orthoepists since the late eighteenth century, and in many accounts quite detailed descriptions of different vowel realizations are given (MacMahon 1998: 455–456; Jones 2006: 193–198). Jones (2006: 189-190), however, warns of any certainty in assessing the exact quality of a vowel based on historical sources:

Vowel quality identification in this area of the phonology is notoriously difficult, the low vowel space being relative cluttered with segments distinguished by relatively small shifts in the Hz value of F<sup>2</sup> . [æ], [a], [ɐ], [ɑ] among unrounded variants, with rounded forms in [ɒ] and [œ]. Even for the trained modern observer, it can be difficult without instrumental analysis to be certain about the precise phonetic values of segments like this, especially in connected allegro speech contexts, so that assessing their actual shape from written historical evidence is well-nigh impossible with any degree of certainty.

Regarding language-external factors influencing the bath-trap split and the different vowel realizations, sources point to region and social class. What can be deduced from the orthoepists' comments is a restriction of the split to the South of England and a more extreme quality shift of the vowel in London and surrounding areas (Lass 1999: 106–107). A variety of insights can also be gained based on a study of current variability of /æ/ and /aː/ in Australian English. Bradley (1991) investigates sixty morphemes whose pronunciations are mainly invariant in southeastern British English today, but vary in Australia between the trap vowel /æ/ and the palm vowel /aː/ (e.g. *graph*, *dance*). Based on synchronic data he reconstructs how the change proceeded in southeastern England. He summarizes the interpretation of his findings in Bradley (2008):

[P]laces settled by the early nineteenth century, and primarily by people of lower-socio-economic status, use more palm as in Sydney, Hobart, and Brisbane. Melbourne, settled in the mid-nineteenth century, with a more mixed population, shows a higher proportion of trap. Adelaide, settled later in the nineteenth century primarily by people of middle or higher socio-economic status, uses the highest proportion of palm, and shows a more advanced stage of the shift before nasal + obstruent than elsewhere in Australia, though not as far advanced as New Zealand or modern southeastern British English. This implies that the change in southeastern England was underway during the settlement of Australia, and that palm was

### 3.3 Phonological forms and their written representations

a lower-status form in the late eighteenth and early nineteenth century, but had reversed its social value and become a high status form by the late nineteenth century. Furthermore, the change in the nasal + obstruent environment must have followed the prefricative environment by quite some time.

Furthermore, he suggests that the backing of /aː/ happened fairly late in London and southeastern England because the phonetic realizations of /aː/ in Australia at the time of his study all range between central to front qualities (Bradley 1991: 227).

The bath-trap split and the precise realization of the vowels used in the words belonging to these sets have been commented on and evaluated extensively in works on language in England since the late eighteenth century; Mugglestone (2003: 78) even considers it to have been "one of the most prominent topics for discussion". Lass (2006: 90) states that while Cooper (1687) and Flint (1740) gave a neutral, descriptive account of the vowels, Walker's (1791: 10-11) comments are the first to contain a clear social evaluation. While he neutrally noted the presence of the "long sound of the middle or Italian *a*" before /r/ (*car*) and before <l> + labial (*calm*, *calf* ), he considered it vulgar to use the same vowel in other environments (e.g. in *glass*, *fast*, *after*, *plant*). In the following century, the emerging picture was quite complex. Bailey (1996: 117) describes the commentary as "muddled" because it is often difficult to distinguish between description and evaluation and to determine which vowel quality the commentators refer to. Nevertheless, there are two general and overlapping developments that can be observed. The first is that the bath-trap split continued to be subject of much criticism throughout the nineteenth century. Bailey (1996: 118-119) cites three authors, the anonymous author of *Vulgarities of Speech Corrected* (1826: 256–267, Savage (1833: xxvi–xxvii) and Leigh (1840: 25), who consider any vowel but the traditional [æ] in the bath-set to be vulgar and improper and associated with Cockney speakers. Mugglestone (2003: 81) finds that Walker has shaped these prevailing negative attitudes considerably, which is for example visible in Longmuir (1864). However, a second general tendency that can be observed is a change of social evaluation. Jones writes in his *General Pronouncing and Explanatory Dictionary* (1798: ii-iii) that "giving to those and similar words [= the bath set] the flat dead sound of *a* in *lack*, *latch*, *pan* &c. is encouraging a mincing modern affectation" (cited in Jones 2006: 194). This comment indicates a positive evaluation of the innovative lengthened, lowered and retracted variants as 'normal', whereas conservative [æ] is linked to a restricted set of affected speakers who aim at

### 3 Tracing enregisterment processes of American English

sounding elegant and mannered. The attention then started to focus on the precise realization of the vowel in the bath-set and on the question of which words actually belonged to the set. As a sort of compromise solution, many commentators developed the concept of an "intermediate" vowel, which was long and had a quality somewhere between [æ] and [ɑ]. This vowel was evaluated positively because it meant that the negative associations linked to the other vowels could be avoided (Bailey 1996: 122–123). A characteristic example of such an evaluation is Smith's (1858: 34) comment, which is cited by Mugglestone (2003: 83):

Avoid a too broad or too slender pronunciation of the vowel *a*, in such words as *command*, *glass*, *pass*, &c. Some persons vulgarly pronounce the *a* in such words, as of written *ar*, and others mince it as to rhyme with *stand*. […] Equally avoid the extremes of vulgarity and affectation.

A comment by Jespersen ([1909] 1948: 428) shows that the low back realization remained stigmatized at least until the early twentieth century: "A vulgar retracted or rounded [aˑ] is sometimes represented by novelists as *aw*". Similarly, Ripman (1906: 55, cited by Mugglestone 2003: 88) observes that "precise speakers" often prefer a "delicate" middle sound and use [aː] instead of [ɑː] due to "an excessive desire to avoid any Cockney leanings in their speech".

Mugglestone (2003: 83) calls it a paradox that "it was, in the end, not to be the iterated 'correctness' of this 'middle' variant, or even Walker's favoured [æ], but instead the theoretically 'vulgar' [ɑː] which stabilized as one of the dominant markers of RP and the non-regional accent". In her view, this is a sign that "prevalent attitudes to language will not necessarily affect the ultimate outcome of a linguistic change" and that "the real direction of linguistic change can and does instead regularly run counter to such precepts [found in prescriptive writings]". How [ɑː] became an index of high social status, education and refinement is not answered in her study. Similarly, Agha (2003) lists [ɑː] as a marker of Received Pronunciation in the twentieth century, but he does not offer any suggestion as to when and how the vowel became enregistered as standard and prestigious.

The historical development of the trap-vowel in America has not been studied in much detail. Eighteenth- and nineteenth-century descriptions and modern evidence from dialect atlases suggest that a split of the class into trap and bath through a lengthening, lowering and backing of the vowel to [aː], [ɑː], or even [ɒː] in the bath set occurred only on the American East Coast. Krapp (1925, 2) provides a detailed analysis of early accounts regarding the split and describes the pronunciation as he perceives it in the early twentieth century. He cites Webster's works as early evidence for a low back vowel in "words of the type of *artist*,

### 3.3 Phonological forms and their written representations

*arm*, *clasp*, *after*, *balm*, *grant*, *advance*, etc." occurring in New England speech (Krapp 1925, 2: 63) and concludes that

one can infer from Webster's remarks […] that in New England, as Webster heard the speech of New England, which was probably with a considerable degree of accuracy, the sound of [ɑː] had established itself in several large groups of words, whereas elsewhere in America the sound [æ] or [æː] prevailed in these words. (Krapp 1925, 2: 68)

Based on a variety of other sources exhibiting descriptions and comments on language (mostly dictionaries), Krapp (1925, 2: 69) finds that the bath-trap split and the use of [ɑː] in bath remained regionally restricted, mainly to New England: "the pronunciation of words like *laugh*, *dance*, *branch*, etc. with [ɑː] has established itself nowhere as a popular custom outside of New England". Regarding the frequency of use he cites Grandgent (1899), who estimates that [ɑː] was most widely used in New England between 1830 and 1850 and that its use declined towards the end of the century. However, in the first half of the twentieth century, the split is still observed by fieldworkers for several dialect atlas projects on the East Coast of the United States. Kurath & McDavid (1961) summarize the results for the vowel in *aunt* by stating that it is predominantly a low-front vowel in New England and a low-back to low-central vowel in Tidewater Virginia, whereas outside of these areas, the front vowel [æ] is "nearly universal" (Kurath & McDavid 1961: 135). The exception are "cultured speakers" in larger cities like New York City, Philadelphia and Charlottesville (Kurath & McDavid 1961: 135). The results for words like *calf*, *glass* and *dance* (the vowel followed by a voiceless fricative or by /n/ plus a dental consonant) show a similar pattern. In Eastern New England, the low-front vowel [aː] is often found, but usage varies considerably and speakers "freely shift" between this vowel and [æ], which is common elsewhere, except again for "some cultured urban speakers" who use a low-central vowel occasionally (Kurath & McDavid 1961: 136). In addition, the low-front [aː] is also used by speakers along the coast of South Carolina and Charleston, and Kurath & McDavid (1961: 136) point out that these speakers are not "cultured" but "common folk". Low front [aː] is also used in the word *can't* in large parts of Eastern New England by speakers of all social classes, but it is especially common in important cities such as Boston, Providence, and Newport. The vowel is again a "prestige pronunciation" in some cities outside Eastern New England (Hartford, Manhattan, Rochester, and Philadelphia) and it is used by a few cultured urban speakers in Virginia and occasionally by common speakers in coastal South Carolina and Georgia. That a retracted vowel in bath persisted in Boston until the

### 3 Tracing enregisterment processes of American English

second half of the twentieth century is reported by Nagy & Roberts (2008: 56) based on Laferriere's (1977) finding that low-central [aː] occurs in words before /f/ and /θ/ and in some words before /n/ (e.g. *aunt*) and much more variably in words before /s/ and in other words before /n/ (e.g. *dance*). In the southern areas, the use of a low-central or back vowel has declined, and it is only used in a few words like *aunt* or *rather* (Thomas 2008: 97).

Krapp's (1925, 2) account also offers a variety of insights into the evaluation of the bath-trap split and of the different vowels used in bath. He attributes high importance to the fact that Webster recommended a lengthened and retracted vowel in bath in his *American Spelling Book* and in his further dictionaries. Webster explicitly evaluates [æ] as a feature of "mincing, affected pronunciation" in his *Compendious Dictionary* (1806, cited in Krapp 1925, 2: 68), which reflectsJones' (1798) comment cited above. Krapp's (1925, 2: 68–69) interpretation of Webster's comments is that he "regarded [ɑː] as the good-home spun pronunciation", which was to be seen as positive even though it "was at least open to the charge of being a somewhat rustic one". In the course of the nineteenth century, Webster's books and New England culture in general became increasingly influential, which, according to Krapp (1925, 2: 69), led to [ɑː] being evaluated as the refined and elegant variant and losing associations with rustic, common speech. Krapp (1925, 2: 75) attributes an important role to instruction in schools in the establishment and dissemination of positive attitudes towards [ɑː] because the "New England school book and also the New England school and school teacher exerted an enormous influence over American education during the second and third quarters of the nineteenth century". As in England, Krapp (1925, 2: 74–75) notes the introduction of an "intermediate sound of *a*" in the debate, which he transcribes as [aː]. A well-known commentator who advocates the use of this variant is Richard Grant White, a New Yorker, who writes in his *Words and Their Uses* (1870: 62) that "the broad *ah* sound of *a*" indicated "social culture which began at the cradle" whereas the use of "the thick throaty sound of *aw*" or "oftenest, […] the thin, flat sound which it has in "an," "at," and "anatomy"" is a sign of a lack of culture and refinement (cited in Krapp 1925, 2: 76). Regarding the development of evaluations, Krapp (1925, 2: 76–77) speculates that the "period of highest esteem of [ɑː], or [aː], outside New England, perhaps falls in the two decades from 1870 to 1890", so forty years after its greatest popularity *within* New England. Despite the similarity to the development of attitudes in England, he decidedly argues against any British influence on American evaluations: "There is no evidence to indicate that any single pronunciation which has become general in America has become so through imitative influence of British pronunciation" (Krapp 1925, 2: 80). Instead, he argues for a parallel development which "was caused by the ac-

### 3.3 Phonological forms and their written representations

cident that the sound [ɑː] became established as a popular speech sound in that very section of America which came to be recognized as our first seat of culture, and that in England, by a curious similarity, the situation was exactly the same" (Krapp 1925, 2: 80). However, he also notes that towards the end of the century not only the use of [ɑː] declines, but that the sound is also increasingly evaluated negatively as affected. The comment by Lounsbury (1909: 101, cited by Krapp 1925, 2: 76–77) that the sound [ɑː], "when heard, is already looked upon by many as an affectation" therefore indicates an evaluation which is in complete opposition to Webster's evaluation of [æ] as affected a century earlier. Next to [æ], [aː] and [ɑː], Krapp (1925, 2: 36) briefly mentions a fourth variant, a back and rounded vowel, which he finds to be "one of the marks of the grotesque pronunciation of the haw-haw type of comic Britisher on the stage" which indicates that there was indeed a variant of the bath-vowel associated with British speech. It is at least theoretically possible that the variant targeted in these humorous performances was [ɑː], but that it was exaggerated to an extent that it was in fact [ɔ] or [ɔː].

The view that British influence did not play a role in the evaluation of the vowel is not shared by Montgomery (2001: 141), who proposes an alternative account. He argues that

[æ] in *fast*, *bath*, *aunt* is a colonial lag reflecting earlier emigration, and that alternative pronunciations gained currency in the eighteenth century as New Englanders and Virginians attempted to keep pace with English speech, facilitated by the contact of coastal American cities with England, which was closer than that of the hinterland.

In this view, the prestige of [ɑː] or [aː] is attributed to the association of these variants with English speech norms, or, as Montgomery (2001: 140) puts it, with "British fashion".

Kurath & McDavid (1961: 136) also confirm a perception of [aː] as "refined" and the fact that it was used primarily by "cultured", urban speakers in those areas where it was not commonly used can also be regarded as evidence of an association of the variant with education and a higher social status. The "intimate commercial and cultural relations with London and with cultured Englishmen throughout the Colonial period and after" are also proposed by Kurath & Mc-David (1961: 136) as a likely reason for the high prestige of the variant. However, they also note that there are speakers in Eastern New England who regard the variant as rustic, which indicates the presence of conflicting evaluations in the first part of the twentieth century. That these conflicting evaluations continue at

### 3 Tracing enregisterment processes of American English

least in Boston until the 1970s is shown by Laferriere (1977: 105–106), who identifies "nonlinguistic labels" for what she calls the "backing of [æ]".<sup>12</sup> She finds that it has a "complex social profile": It is first of all associated with older people, which matches the production data. Secondly, it is linked to Boston identity by middle-aged and older people, who associate themselves with the city. Thirdly, it is stigmatized by speakers "whose ties to the city are more precarious because of higher education or prolonged residence outside the New England area" (Laferriere 1977: 105), even though they tend to use the variant themselves. And lastly, Laferriere (1977: 106) finds that the backed variant is used "as a hypercorrect indication of erudition" and gives examples of older speakers using the backed variant in words in which it normally does not occur (e.g. *Master's*) to impress others.

This overview shows that the vowel in bath has quite a complex history, not only regarding changes in use but also changes in evaluation. The supposed associations of the innovative variant with New England speech, with culture and refinement *as well as* rusticity, and also with fashionable British speech make it an interesting variant to explore in a study on the enregisterment of American English: Which of these associations can be found in newspaper discourse? Was the situation in nineteenth-century America the same as today, when lengthened and retracted [ɑː] is not considered to be a "general" American variant? The first analyses of articles containing *hinglish* revealed pronunciation respellings representing a lengthened variant of the vowel. In a few cases <h> was used to mark the lengthening (*cahn't*, *pahdon*) in several cases <r> can be found (*carn't*, *blarsted*, *'arf* 'half'). As a lengthened variant has been well-established before <r> (e.g. in *car*, *arm*, *part*), it seems to have been a good indicator of vowel quality. However, there is also the possibility that <r> was used to represent hyper-rhoticity (see §3.3.5 for details on rhoticity and hyper-rhoticity), so <r> is not as reliable an indicator of vowel quality as <h> or <w>, which is the third way of representing the realization of the bath vowel (*rawther*). This last way of representing the vowel additionally indicates a back quality because the combination <aw> usually represents back and rounded [ɔː] (e.g. *awe*, *claw*, *paw*). It is this last spelling variant which I choose to focus on in the present study because it indicates a representation of a more extreme quality within the range of possible variants of the bath vowel and it is therefore more likely to be used to draw attention to a back quality of bath than other variants. As a search term, I decided to use *dawnce* because first of all, the vowel is more variable before /-nC/ clusters than in other environments and secondly because the lexical item *dance*

<sup>12</sup>Unfortunately, Laferriere (1977) does not give any information on how exactly she elicited the labels from the informants.

### 3.3 Phonological forms and their written representations

is fairly frequent and has more semantic content than for example *can't*, *rather*, or *last*. An initial search in the NCNP database revealed that *dawnce* occurs 44 times, while *darnce* occurs three times and *dahnce* only twice, which confirms the hypothesis that the variant <aw> is more often used than other variants for representing the quality of the bath vowel.

The analysis of articles containing *dawnce* revealed articles in which *dawnce* co-occurs with representations of another linguistic form: non-rhoticity. Interestingly, spelling representations of non-rhoticity overlap with representations of the bath vowel. The article "National What-Is-Its" (October 25, 1887[83] ) contains for example the forms *dawnce*, *cawn't*, *pawth* and *chawnce*, in which <aw> represents a back vowel, and the forms *rathaw* and *fathaw*, in which <aw> represents non-rhoticity. Before discussing the consequences of this overlap for the choice of search terms, I will summarize prior research on non-rhoticity in the following section.

# **3.3.5** *deah, fellah* **and** *bettah***: non-rhoticity**

One of the most studied phonological forms in English (historical) linguistics is non-rhoticity, which is defined in slightly different ways but generally constitutes the absence of /r/ in a post-vocalic and non-prevocalic position (e.g. *car* /kɑː/,*cart* /kɑːt/). Research on the structural development of non-rhoticity, which has often been termed */r/-loss*, has identified different phases based on the linguistic context in which the loss of /r/ occurred. Lass (2006: 91–92) distinguishes an earlier phase which is characterized by the loss of /r/ without an accompanying lengthening of the preceding vowel (e.g. *arse* /æs/) from a later phase which is characterized by a lengthening of the preceding vowel (e.g. *arm* /ɑːm/). In his view, it is only in the second phase that /r/-loss occurs systematically; before that it is sporadic and lexically restricted. Minkova (2014: 124), however, regards vowel lengthening only as "one possible outcome rather than an essential stage in the process of /r/-loss". In her analysis, it is rather the position of /r/ in the syllable which is crucial for distinguishing two developments which not only differ in their temporal occurrence, but also in their underlying mechanism. The first development is the loss of /r/ in the syllable coda preceding one or more tautosyllabic consonants (as in *cart* /kɑːt/), which she terms "pre-consonantal /r/-loss". The earliest evidence for it stems from the eleventh century: words spelled without an <r> that had contained an <r> before (e.g. OE *gorst* - ME *gost* 'gorse').<sup>13</sup> In addition, poetic rhymes of words spelled with and without <r> point to a loss of

<sup>13</sup>Minkova (2014) argues against several others scholars here, who do not regard these early instances as part of a general development of non-rhoticity.

### 3 Tracing enregisterment processes of American English

pre-consonantal /r/. The place of articulation of the following consonant seems to affect the change: /r/ was lost first before coronal consonants ([t], [d], [n], [l], [θ], [ð], [s], [ʃ], and [dʒ]), while it remained before the other non-coronal consonants until the seventeenth century (Minkova 2014: 122, citing Hill 1940). Nevertheless, Minkova (2014: 124) concludes that "the earlier and later cases of /-rC/ simplification represent a single historical process stretching over more than six centuries" because she assumes that assimilation is the mechanism underlying all cases of pre-consonantal /r/-loss. She also adds that /r/-loss affected "different dialects and different lexical items unevenly". The second development is the loss of /r/ in the syllable coda without a following consonant (as in *car* /kɑː/), which she terms "post-vocalic /r/-loss". The reason for separating it from the first development is a different mechanism of change which involves analogy rather than assimilation. The starting point of this type of /r/-loss is a matter of debate, which largely ranks around the interpretation of unetymological <r> in seventeenth-century spellings. Minkova (2014: 126) cites <winder(e)s/wynders> 'windows' (1601, 1613), <feller> 'fellow' (1639), and <pillars> 'pillows' (1673). Insertions of <r> have been (and still are) often interpreted as evidence for loss of post-vocalic /r/. They are either regarded as signs of hyper-correction and linguistic insecurity (using the same argumentation as in the case of /h/-insertions) or, when they appear wordfinally preceding a vowel in the following word, as an intrusive /r/ which is typical for non-rhotic, but not for rhotic accents. However, a very detailed study by Britton (2007) shows that it is much more convincing to interpret these spellings as indicating hyper-rhoticity, which he defines as "the appearance, in rhotic accents, of epenthetic, unetymological rhyme-/r/" (2007: 525).<sup>14</sup> His analysis of the *Linguistic Atlas of England* (LAE, Orton et al. 1978) reveals that there is modern dialectological evidence for hyper-rhoticity in the comma set in rhotic accents of the South and the south-west Midlands (see the maps for meadow, yellow, and window).<sup>15</sup> This suggests that hyper-rhotic forms also occurred in the history of English, beginning in the seventeenth century with the examples cited above. This is a convincing argument against the loss of post-vocalic /r/ in the

<sup>14</sup>It is important to note that Britton (2007: 527) distinguishes hyper-rhotic /r/ from intrusive /r/ because the former is "not attributable to a sandhi development, being probably […] first established in the lexicon, rather than having phonological origins as a linking, hiatus-breaking device".

<sup>15</sup>He also mentions the occurrence of hyper-rhotic forms in some items of the bath and palm sets (exemplified in maps of last, calf and half) and of the thought and cloth set (found in maps of sawdust, slaughterhouse, straw, walk, daughter, broth, cross, off, fox). The regional distribution of these forms is more restricted than those in the comma set, however, and largely confined to west Shropshire (Britton 2007: 528).

### 3.3 Phonological forms and their written representations

seventeenth-century, and it also has repercussions for accounts which suggest that /r/-loss became generally common in the eighteenth century.

With regard to the eighteenth century, Britton (2007: 530) argues that the interpretation of comments by orthoepists (Walker 1791: 37, Elphinston 1786: 264, Douglas 1779, see Jones 1991) on forms like *winder* and *feller* as constituting evidence for intrusive /r/ (and, connected to that, non-rhoticity) is not necessarily the only possible one, but that it is even more reasonable to assume that these forms are instances of hyper-rhoticity. The mechanism which Britton assumes to underlie this unetymological insertion of /r/ is analogy. An analysis of an early and mid-eighteenth century corpus by Sóskuthy (2013) reveals that words with an etymological /r/ following a final /-ə/ (e.g. *better*) were much more common than those ending in /-ə/ without being followed by /r/ (e.g. *idea*).<sup>16</sup> As a consequence, speakers extended the more common /-ər/-pattern in the former class of words to the relatively few words in the latter class. Sóskuthy uses the results mainly to explain the occurrence of intrusive /r/ in non-rhotic accents, but he also agrees with Britton (2007) that they can be used to account for hyper-rhoticity in rhotic accents (Sóskuthy 2013: 77). With regard to the early nineteenth century, Britton (2007: 529) cites Savage's (1833) reports of forms like *pillar* 'pillow' and *eye-dear* 'idea' as features of popular London speech and polite usage. As Savage only provides very few examples of non-rhotic forms in his account, Britton (2007: 529) finds it reasonable to assume that again hyper-rhoticity plays a greater role here than it has commonly been assumed. Britton's (2007) analysis therefore seriously challenges the view that unetymological insertions of <r> in the spellings are always evidence for non-rhoticity. His interpretation suggests that the spread of non-rhoticity in London and the south-east of England is largely a phenomenon of the middle to late nineteenth century.

This view is backed by other research, as for example Jones' (2006) analysis of orthoepists' comments on /r/ which leads him to conclude that "post-vocalic [r] loss was very much a minor characteristic of the phonology of late eighteenthcentury English" (2006: 261). Of the few contemporary comments on the development of non-rhoticity in England, Walker's observation that "[i]n England, and particularly in London, the *r* in *lard*, *bard*, *card*, *regard*, &c. is pronounced so much in the throat as to be little more than the middle or Italian *a*, lengthened into *laad*, *baad*, *caad*, *regaad*" and that it is "sometimes entirely sunk" in London (1791: 50) is probably most well-known. It is often cited as clear evidence for an ongoing change towards non-rhoticity. It should be noted, however, that in all

<sup>16</sup>The same difference holds for /ɔː/ and /ɑː/ where the token frequency of words with /ɔːr/ and /ɑːr/ is far higher than that of those with /ɔː#/ and /ɑː#/ (Sóskuthy 2013: 60).

### 3 Tracing enregisterment processes of American English

examples but one (*baa* 'bar') the /r/ is weakened or lost in pre-consonantal position before /d/. It is therefore more indicative of pre-consonantal /r/-loss than of post-vocalic /r/-loss. In the latter position, /r/ was probably still present which is also supported by Savage's (1833) extensive discussion of /r/-insertions. Jones' (2006: 342) conclusions therefore need to be restricted to word-final /r/-loss:

Savage's data on *r* seem to suggest that in the early part of the nineteenthcentury [r]-loss was not yet an active, ongoing change (even among the 'vulgar' speakers, and certainly not in the *Orthoepy*). On the other hand, syllable-final [r]-adding was ongoing and active and gaining ground everywhere, not least in socially acceptable circles.

Further evidence for the view that loss of /r/ was by no means general in the middle of the nineteenth-century comes from research on the historical development of (non-)rhoticity in New Zealand English. Many first generation New Zealand-born Europeans whose speech has been documented and analyzed as part of the Origins of New Zealand English project were rhotic to some degrees, which leads Hay & Sudbury (2005: 804) to conclude that the "lack of rhoticity in New Zealand English [...] is not an inherited feature from British English, but rather the result of a change that occurred (or at least went to completion) in New Zealand".

The picture which emerges is evidently not complete yet, but it seems as if post-vocalic /r/-loss began in the eleventh century when /r/ was followed by a coronal consonant in the same syllable. The loss extended to non-coronal contexts in the seventeenth century. It is likely that early unetymological insertions of <r> in unstressed syllables point to hyper-rhoticity and not to a beginning loss of /r/ in postvocalic and non-preconsonantal positions in the seventeenth century. At what time the loss of /r/ in the latter position became more widespread is not clear – orthoepists' comments, especially Ellis (1869), suggest that it was common in London, even among educated speakers, by the middle of the nineteenth century. However, based on a careful interpretation of several types of evidence, MacMahon (1998: 475) concludes that "[t]he evidence for considerable variation between rhoticity and non-rhoticity, with intermediate semi-rhoticity, especially during the later nineteenth century in the educated South of England, is, clearly, very strong". The continuing presence of rhoticity is also supported by first-generation New Zealand speakers exhibiting some degree of rhoticity. The change to non-rhoticity did not affect all regions. Rhoticity is maintained in Ireland and Scotland to this day (even though there are also studies showing an increasing non-rhoticity at least in some areas and groups of speakers). The

### 3.3 Phonological forms and their written representations

LAE maps show that rhoticity has also persisted in some regions in England for a long time, especially in the south-west and the south-west Midlands, but again, non-rhoticity is on the rise here as well.

The discussion on when and how non-rhoticity began to spread in England is also relevant for the reconstruction of the development of non-rhoticity in North America. Two points are important here: first, the assumed rhoticity or non-rhoticity of the settlers based on their origin in England and, second, the two different types of /r/-loss. As shown above, /r/-loss in preconsonantal position was already attested by the time that the first settlers emigrated to America, so it has been argued by Kurath (1971), Fisher (2001) and Montgomery (2001) (among others) that these early settlers must have brought this type of non-rhoticity to America as well. Evidence for early non-rhoticity is provided by Krapp (1925, 2: 228–230), who lists 35 words in which <r> has been omitted in the spelling which he found in American town records and in one collection of diaries in the seventeenth and eighteenth century. He also lists 24 instances of unetymological <r>-insertions which he interprets as evidence for non-rhoticity, but in the light of the detailed discussion about these insertions above, it is more likely that these cases are instances of hyper-rhoticity as well.<sup>17</sup> Furthermore, he cites numerous examples of this early type of /r/-loss in poetic rhymes. This contradicts the assumption by Newman (2014: 136) that non-rhoticity emerged only at the end of the eighteenth century in England and that therefore all early settlers must have been rhotic. Given the evidence described above, this view is only possible if non-rhoticity is defined as necessarily encompassing both types of /r/-loss: pre-consonantal *and* post-vocalic. The question when and why /r/-loss extended to more environments and spread in the United States is more difficult to answer. Kurath & McDavid (1961: 171) state with regard to the regions where *door* is pronounced without /r/ by the informants interviewed for the Linguistic Atlas of the Eastern United States that

four geographically separated subareas have a mid-central semivowel /ə/:̯ Eastern New England as far west as the Connecticut Valley (to the crest

<sup>17</sup>This is supported by a comparison of omissions of <r> in the letter set and insertions of <r> in the comma set: Omission of <r> in the letter set occurs only three times (of all 35 examples) and in two of these examples, <r> is omitted before <s> (*Whiticurs* > *Whittacus*, *Rogers* > *Roges*), that is, in a linguistic context (preconsonantal, before /s/) which is typical of early /r/-loss. So while /r/-loss in letter is hardly attested at all, 16 out of the 25 insertions of <r> (64 %) occur in the comma set (e.g. *feler* 'fellow', *famerly* 'family'). To argue that <r> insertions reflect the writer's non-rhoticity because he is unsure about where to use an <r> in the spelling is not convincing if there are no instances where the <r> is also omitted in the context where it is most frequently inserted.

### 3 Tracing enregisterment processes of American English

of the Green Mountains in Vermont), Metropolitan New York, the Upper South (the southern peninsula of Maryland, Virginia to the Blue Ridge, and north-central North Carolina), and the Lower South (South Carolina and Georgia).

With regard to the historical development of the non-rhotic pronunciation of words like *door* in America, they speculate that

[i]n English folk speech of today, *door*, *care*, *ear* end in unsyllabic /ə/ in ̯ the eastern counties north of the Thames, in /r/, articulated as a constricted [ɚ], in the south and the west. In all probability both types came to this country with the first colonists and could be heard in all of the colonies. With the acceptance of /doə, keə ̯ , iə ̯ / in Standard British English during ̯ the eighteenth century, it would seem that this type acquired prestige in the chief American seaports on the Atlantic coast—Boston, New York, Richmond, Charleston—and spread from there to the hinterland. In the inland, on the other hand, the post-vocalic /r/, common from the beginning, came to be generally established, as also in the Quaker-dominated port of Philadelphia and vicinity. (Kurath & McDavid 1961: 171)

So on the one hand, they stress that non-rhotic forms have been a result of inheritance, but on the other hand they also stress a parallel development in England and in some English regions. Based on the discussion above, it is plausible to specify that what is inherited is the pre-consonantal /r/-loss, whereas the later parallel change concerns the post-vocalic /r/-loss. However, there is still considerable disagreement about the timing and the causes of the later change. One influential hypothesis is that the increasing prestige and use of non-rhotic forms (in all environments) in England also influenced developments in America. Bailey (1996: 105) also argues along these lines that

[i]n North America, only New England, southern New York, and the coastal South (from Baltimore to New Orleans) took part in the general weakening of noninitial *r*, though even in these districts the loss of *r* was not universal. There is nothing surprising in this parallel development for, despite political differences, the cities were part of a single speech community, increasingly so in the first half of the [nineteenth] century, as business and cultural contacts developed between British ports and the sea-coast cities of North America.

### 3.3 Phonological forms and their written representations

Not only the increasing non-rhoticity in the east, but also the retention of rhoticity in western parts of the United States is explained by drawing on the prestige that non-rhoticity supposedly used to have as a British form:

Not everyone rushed to mimic London fashion. Little in the first two centuries of the transatlantic migration of northerners induced a desire to follow London fashions in English. Most North American settlers, had little direct experience of London speech and less interest in imitating it. (Bailey 1996: 107)

Given the finding that by the end of the nineteenth century there was still considerable variability in the use of non-rhotic forms in southeastern England, dating the increasing use of non-rhoticity in the eastern United States to the late eighteenth and early nineteenth century based on developments in England is not convincing – in this line of argumentation it is more likely to have occurred later in the century.

However, there are also scholars who argue against any influence of British speech. For example, Krapp states that even though "the weakening and disappearance of [r] as a final sound and before consonants was particularly noted at about the same time, that is, at the end of the eighteenth century" in England and in America, this is only "a coincidence in the *critical record of* historical development" and not "an actual coincidence in historical development" (Krapp 1925, 2: 227, emphasis mine). Similarly, Dillard (1992: 49–50) is very doubtful about British influence on American developments:

Most Americans on the East Coast – and even more so those to the west – were busy with concerns of their own rather than with connections to England. Knights (1969) stresses the 'incredible' mobility of the Boston population between 1830 and 1860, roughly the period in which the change should have been taking hold. […] Whether such mobility would contribute to a stable, invariant shift of /Vr/ to /V0/ appears to be something less than a matter of socio-linguistic certainty. […] The picture of a stable community, with the top group accepting a British innovation and the lower classes placidly following suit, does not seem to follow.

Bonfiglio (2002) also looks at the factor of mobility with regard to the case of New York City and argues that with the opening of the Eerie Barge Canal in 1825, immigration to New York from western parts, especially the Great Lakes region, increased significantly and, as a consequence, New Yorkers came into

### 3 Tracing enregisterment processes of American English

contact with a large number of rhotic speakers. Against this background, he argues that "one must conclude that the dropped /r/ of New York City was not the result of British influence in the nineteenth century, but instead the result of the resistance of indigenous pronunciation to external influences" (2002: 51). However, this view requires that non-rhoticity has already been firmly established in New York City by the early nineteenth century – so firmly that the immigrants coming to New York adapted to the speech pattern there.

Linguists not only disagree about when and why non-rhoticity became a feature of the speech in eastern regions of the United States, but also about when and why this process reversed. Very prominently, Labov argued in his study on New York City English in the 1960s that "[t]he introduction of /r/ in the New York City system as a dominant element in the prestige dialect may certainly be traced in the 1930s, but apparently made a great step forward in the years coinciding with World War II" (Labov 2006: 376). In his view, the increasing use of rhotic forms and the shift in norms towards rhoticity being the prestigious form "reflected the abandonment of the earlier prestige form of Anglophile English" (Labov 2006: 296). As already discussed in §2.4, Bonfiglio (2002) disagrees with this view because he not only finds that the crucial shift in prestige occurred already in the first half of the twentieth century, but also because he links it to a sinking prestige of New York City, which becomes viewed as ethnically "contaminated" because of large numbers of African Americans and immigrants living there. Fisher (2001) makes yet a different case: He claims that non-rhoticity lost its prestige after the Civil War because economic and political power shifted away from regions which had traditionally been most closely linked to England (non-rhotic Virginia and Boston) towards New York, Pennsylvania, and the trans-Appalachian Middle West. The last two regions have traditionally been rhotic, but he also finds that more and more rhotic speakers migrated to New York, where the influence of the old non-rhotic Colonial elite, with its ties to England, weakened in the last decades of the nineteenth century, "so that nonrhotic pronunciation lost its prestige" (Fisher 2001: 77). However, Bonfiglio's study is the only one that is supported by a systematic and detailed analysis of sources, whereas Fisher's and Labov's views with regard to causes of the shift of prestige of rhoticity remain rather speculative. Recent studies show that the change towards rhoticity has continued in New York City (Becker 2014) and that is has occurred in Boston and New Hampshire as well (Nagy & Irwin 2010).

In White Southern American English, non-rhoticity is one of the forms that Schneider considers to be traditionally southern, while rhoticity marks the new South (Schneider 2007: 299). Thomas (2008: 107) summarizes the research find-

### 3.3 Phonological forms and their written representations

ings on the southern development, which have identified World War II as a turning point:

Before World War II, non-rhoticity was prestigious, appearing most frequently among higher social levels and spreading (except, perhaps, in nurse words). Afterward, rhoticity became prestigious and non-rhoticity became most common among lower social levels. Females have forged ahead of males in this change.

Feagin's (2015) study on the development of several linguistic variables in Anniston, Alabama, also includes non-rhoticity. She observes that non-rhoticity is almost categorical in the speech of the two older upper class speakers (born 1882 and 1890), but in the speech of the four older rural and urban working class speakers (born 1895, 1899, 1907 and 1911) rhotic forms are already more frequent than non-rhotic forms (63% rhoticity for the speaker born in 1907 and 84-87% rhoticity for the rest of the speakers). The two working class speakers born after World War II (1956 and 1957) also have a high degree of rhoticity (86% and 92%), but a drastic change can be observed for the upper class: The speakers born in 1953 and 1955 use rhotic forms at a rate of 65% and 91%. Even though no major conclusions can be drawn based on this low number of speakers in one location, it only partially confirms the development described by Thomas. While the middle of the twentieth century does seem to constitute a turning point, this only affects the upper class speakers – and even though the rate of rhotic forms in the speech of the upper class speakers is drastically higher than that of the upper class speakers born in the nineteenth century, it does not surpass the rate of the working class speakers born before World War II.

Finally, it has to be noted that the development of non-rhoticity was different for African American English. Even though evidence is scarce, non-rhoticity has shown to be a fairly stable feature. Thomas (2008: 107) states that

[i]t should be noted that the dramatic increase in rhoticity applies only to white Southerners; African Americans remain largely non-rhotic, except in the nurse class, and [...] social polarization of the two ethnicities magnified during the civil rights movement may be related to the divergence in rhoticity.

The discussion about the reasons for the development of the presence or absence of post-vocalic /r/ in the United States shows that the factor of prestige is often drawn on to explain changes. In this regard, the prestige of non-rhoticity

### 3 Tracing enregisterment processes of American English

in England also matters because it is assumed by some to affect the prestige of the form in America. However, most studies usually take the increase in nonrhoticity in (southeastern) England as a sign of the prestige of the variant – but they do not take studies on metadiscursive activity on /r/ into account which provide a more detailed picture of the social meanings of non-rhoticity or rhoticity respectively. I will thus summarize findings on metadiscourses on /r/ in England before turning to the rather scarce investigations of metadiscursive activities surrounding (non-)rhoticity in America.

Mugglestone's (2003: 86–94) overview of attitudes towards the emerging nonrhoticity in England emphasizes the great discrepancy between reality and prescriptive comments. She notes that throughout the nineteenth century the loss of /r/ was commented on negatively. Its use was considered vulgar and attributed to illiterate speakers from lower classes. The stereotype often invoked was that of the Cockney speaker. Even in the second half of the nineteenth century, when postvocalic [r] was probably dropped by large numbers of speakers in London and southern England more generally, its absence was still considered incorrect. An important reason for this attitude was probably the high valuation of education and literacy. The occurrence of <h> in the spelling is an important argument for demanding the pronunciation of [h]; it is only logical that this applies to <r> as well. Mugglestone (2003: 88) cites a comment by the poet Gerard Manley Hopkins ([1880] 1935) on Keats's rhymes of, for example, *higher* and *Thalia*, which illustrates a way to reconcile the logical incorrectness of non-rhoticity with its obvious widespread presence: He evaluates the rhymes as "most offensive, not indeed to the ear but to the mind". So even though they *sound* right, they are still to be regarded as incorrect by the educated and literate speaker. Mugglestone (2003: 90) concludes that

[t]he use of post-vocalic and final [r] […] was evidently able to achieve the paradoxical status of being present as a prescriptive norm in the 'best' English—even while descriptive discussion of non-localized pronunciation was forced to acknowledge its absence in precisely the same positions.

Jones' (2006: 259–260) analysis suggests that the negative evaluation of the realization of /r/ as an alveolar or uvular trill played an important role in the metadiscourse on /r/ as well. On the one hand, this realization was described as foreign and provincial and often attributed to Scottish and Irish speech. On the other hand, it was described as rough, harsh, and canine (it was compared to a dog's snarl). The commentators preferred the 'softness' of the alveolar approximant which is regarded as typically English, but also warned about its complete omission. Walker for example comments that "this letter is too forcibly

### 3.3 Phonological forms and their written representations

pronounced in Ireland" but at the same time "it is often too feebly sounded in England, and particularly in London, where it is sometimes entirely sunk" (1791: 50). This shows the link between the negative evaluation of the trilled /r/ and non-rhoticity: They represented opposing poles on a continuum between a highly sonorant realization of /r/ and no constriction at all. This is reflected in the nineteenth-century comment by Smart (1836: 33, cited by Jones 2006: 339) who again states that both extremes have to be avoided. However, Ellis' (1869: 603) remarks can be interpreted as indicative of a change in attitude, if not in prescriptivists' writings then in society more generally: He states that in London *farther*, *lord*, *stork*, and *drawers* are "reduced" and pronounced like *father*, *laud*, *stalk*, and *draws* "even in the mouths of educated speakers". He acknowledges the prevailing prescriptivist tradition by writing "I have usually written (ɹ) final in deference to opinion" but at the same time he emphasizes that the prescribed difference between the word pairs is only upheld by careful speakers "when they are thinking particularly of what they are saying", which suggests that /r/-less pronunciations had become more and more accepted in everyday life. Those speakers who tried to follow prescriptivist opinion and not omit the /r/ were also criticized, as for example in the anonymous usage guide *Hard Words Made Easy*: "Some of our public speakers, who push the accuracy of utterance beyond a wholesome limit, get the habit of trilling the *r* so much that one would think that they wished to be thought unlettered Scotch or Irish peasants" (Anonymous 1855: 4, cited by Mugglestone 2003: 90). In my opinion, it is reasonable to argue that the avoidance of the Scottish or Irish 'rough' /r/ was considered to be most important, even if it led to the omission of /r/ in post-vocalic contexts. A text on teaching reading and pronunciation, entitled *The First Part of the Progressive Parsing Lessons* and printed in 1833 (cited by Mugglestone 2003: 248) exemplifies this: "*Ar* must be pronounced with the tip of the tongue pressed against the gums of the under teeth, to prevent the *r* having its rough or consonant sound". It is highly unlikely that any constriction can be achieved if these instructions are followed and it is definitely impossible that an alveolar approximant can be produced if the tip of the tongue is not allowed to come near the alveolar ridge. Mugglestone does not really explain why the loss of post-vocalic /r/, which is condemned throughout the nineteenth century, then becomes part of "a set of regionally neutral 'standard pronunciation features'" (Mugglestone 2003: 4) in the late nineteenth century which are linked with values like correctness and educatedness in the metadiscourse. I suggest that it is plausible that the stigmatization of the trilled /r/, which Ellis calls "decidedly un-English" because it "has a Scotch or Irish twang with it" (Ellis 1869: 603), played an important role here. This can be seen in Jespersen's (1922: 244) account of the change involving /r/:

### 3 Tracing enregisterment processes of American English

There is one change characteristic of many languages in which it seems as if women have played an important part even if they are not solely responsible for it: I refer to the weakening of the old fully trilled tongue-point *r*. I have elsewhere […] tried to show that this weakening, which results in various sounds and sometimes in a complete omission of the sound in some positions, is in the main a consequence of, or at any rate favoured by, a change in social life: the old trilled point sound is natural and justified when life is chiefly carried on out-of-doors, but indoor life prefers, on the whole, less noisy speech habits, and the more refined this domestic life is, the more all kinds of noises and even speech sounds will be toned down. One of the results is that this original *r* sound, the rubadub in the orchestra of language, is no longer allowed to bombard the ears, but is softened down in various ways, as we see chiefly in the great cities and among the educated classes, while the rustic population in many countries keeps up the old sound with much greater conservatism.

The omission of /r/ is now regarded as a natural consequence of the weakening of /r/ which is associated with refinement and culture. Consequently, there are no traces of an evaluation of /r/-loss as vulgar. Furthermore, it becomes visible that non-rhoticity is linked to femininity. An earlier example which shows that /r/-loss carries connotations of refined speech and not vulgarity is the 1866 publication entitled *Mind your H's and Take Care of your R's: Exercises for Acquiring the Use & Correcting the Abuse of the Letter H with Observations and Additional Exercises on the Letter R* written by Charles William Smith, professor of elocution. He writes that

[t]he English language, independently of its copiousness, &c., is in this respect nearer to perfection than any other modern tongue, and next after the Greek; I mean in sound, for our language is not the same in sound, when well spoken, as it appears upon paper. The English language is most expressive if properly spoken. It abounds in words which seem to paint the thing for which they stand. Of its many elements of strength the aspirate and the *r* are the most important, but unfortunately they are most villainously abused by the vulgar, the educated, and the refined. The two former either leave out the aspirate where it should be given, or prefix it where it should not be given—many committing both faults, while the last drop the *r*, or convert it into *w*. One robs his "'am" to enrich his "hegg," while the other "*w*enders himself *w*emarkably *w*idiculous" by his senseless affectation". I know not which is the worse, the Wellerism of the East or the Dundrearyism of the West. (Smith 1866: 5)

### 3.3 Phonological forms and their written representations

Even though /r/-dropping is still evaluated negatively as incorrect, it is interesting that its use is explicitly not attributed to 'vulgar' speakers ("Wellerism of the East" invoking the Cockney stereotype) but to 'refined' ones. With regard to the latter speakers Smith alludes to the figure Lord Dundreary which was invented by the English playwright Tom Taylor for the very successful play *Our American Cousin* (1858). Lord Dundreary, played by the English actor Edward Askew Southern, is a rather dumb and eccentric English nobleman who apparently lisps and realizes the /r/ as a labiodental approximant [ʋ].<sup>18</sup> This particular realization of /r/ is actually foregrounded which can be seen in the following elaboration on the letter R in a later chapter:

After H, the letter R is the worst used letter in the English Alphabet. By most persons, it is very imperfectly articulated, being given softly at the beginning of words, and altogether omitted when occurring at the end; such words as court, form, lord, being pronounced ca*w*t, fa*w*m, la*w*d. Foppish affectation changes its fine, manly sound into that of *W*, even at the beginning of words. This Dundrearyism is "t*w*uly *w*idiculous," and more worthy a monkey than a man, yet it has become fashionable, not only among a certain number of brainless "swells," but has been assumed, together with many other effeminate habits and mannerisms, by brave-hearted gentlemen; whose minced speech and foppish manner completely belie the inner man. (Smith 1866: 29)

Here, weakening and loss of /r/ is attributed to "most persons", which implies that he does not associate the omission with any particular social class. The focus of his criticism is rather on the realization of /r/ as a labiodental approximant which he evaluates as foppish, affected, ridiculous and unmanly. Interestingly, the association of the changes with softness and femininity found in Jespersen's comment are also present here. In contrast to Jespersen, however, this association is one of Smith's main arguments against these changes.

To conclude, metadiscourses on /r/ in England are rather complex and seem to shift over time. While non-rhoticity is condemned as incorrect throughout the century and associated with vulgarity, illiteracy and uneducatedness, there is also some evidence for new perceptions and attitudes in the middle of the nineteenth century which associate non-rhoticity and 'soft' realizations of /r/ with refined but effeminate and unmanly speech and behavior, especially when

<sup>18</sup>These forms are indicated by pronunciation respellings in the play: e.g. *yeth* 'yes' and *wath* 'was' for the lisp and *welations* 'relations' for the realization of /r/ as [ʋ]. The latter form will be discussed in more detail in §3.3.6.

### 3 Tracing enregisterment processes of American English

it is combined with a snobbish and affected labiodental /r/. I suggest that the ultimate change of attitude which led to the positive evaluation of non-rhoticity as one of the desirable neutral standard forms (very much visible in Jespersen's (1922) comment) was spurred by the negative attitude towards the realization of /r/ as a trill which was regarded as unpleasant, unrefined and un-English. This is a tentative suggestion, however, which could be investigated more systematically with the methodological framework developed for the present study.

Some associations and evaluations found in metadiscourses on /r/ in England can also be found in metadiscourses on /r/ in the United States. The most comprehensive study on metadiscourses on phonological forms in the United States is provided by Bonfiglio (2002). His focus is on /r/, and he notes, for example, that Webster, like Walker in England, described the presence of non-rhotic forms but did not recommend their use:

Some of the southern people, particularly in Virginia, almost omit the sound of *r* as in *ware*, *there*. […] But there seems to be no good reason for omitting the sound altogether; nor can the omission be defended on the ground, either of good practice or of rules. It seems to be a habit contracted by carelessness. (Webster 1789: 110, cited in Bonfiglio 2002: 38)

Despite the negative view of non-rhoticity expressed in this influential dictionary, Bonfiglio notes that the form has acquired prestige in the northeast and the south by being associated with cultivation, refinement and higher social circles. Bonfiglio (2002: 41) makes the interesting observation that even Krapp (1925, 2: 230–231) still calls the absence of post-vocalic /r/ a feature of cultivated speech in the early twentieth century. With regard to the beginning of the high prestige of non-rhoticity, I have pointed out above that it is often dated to the late eighteenth century. However, Montgomery (2015) finds that this view must be challenged, at least for the south, based on his analysis of confederate textbooks published in the south in the 1860s. In these textbooks, readers are instructed to "[s]ound the R's–Poo*r*, not poo-ah; matte*r*, not mattuh; mothe*r*, not mothuh; wa*r*m, not wāäm, &c" (Chaudron 1863a: 13, cited by Montgomery 2015: 108). At the same time, however, southern children are supposed to practice "to enunciate [r] without harshness" (Chaudron 1863b: 1, cited by Montgomery 2015: 108). This suggests a similar ambivalence found in prescriptive texts in England: The /r/ should not be omitted, but the 'harsh' /r/ (usually a uvular trill) is to be avoided as well.

Another value linked to non-rhoticity in England and America is femininity. In America, however, this link becomes part of an evaluative pattern which links

### 3.3 Phonological forms and their written representations

non-rhoticity to weakness – physically, as in the case of women, but also culturally, as in the case of southerners, yet more importantly, in the case of African Americans. Bonfiglio regards this "equation of the weakening of /r/ and cultural feminization and degeneration" as "one of the main thematic motifs in the rise of network standard and the discussion of standard American pronunciation in the broadcast media" (2002: 46). In contrast to England, where the negative evaluation of the trilled /r/ did not make this variant seem desirable as an alternative to a non-rhotic pronunciation, the retroflex /r/ came to be associated with very positive values in America. According to Bonfiglio, it was linked to virility, strength, vitality and, very importantly, to a purity that was understood in ethnic terms: While urban eastern areas became 'contaminated' by immigration and racial heterogeneity, the western rural parts were constructed as 'uncontaminated' areas and thus representing ideal American values. Especially in nonrhotic areas, these conflicting evaluative patterns led to an ambivalence on the part of contemporary observers which continued to exist in the twentieth century. Bonfiglio (2002: 47–48) analyzes, for example, a publication by the Harvard educated Charles H. Grandgent (1920), entitled "The dog's letter" and notes that while Grandgent still regards non-rhoticity more highly than rhoticity, he also associates it with softness, weakness and decay.

The extensive research on the historical development of non-rhoticity summarized above shows that it was and continues to be a highly salient form. At the same time, there are still several open questions, not only with regard to structural but also to discursive developments. First and foremost, the question as to how non-rhoticity became constructed as a prestigious 'standard' form in England despite being largely condemned in prescriptivist texts has not been answered convincingly based on an analysis of metadiscursive activities. Similarly, the situation in the United States has shown to be quite complex, which is why the present study aims to complement Bonfiglio's (2002) analysis by focusing on metadiscursive activity in nineteenth-century newspapers.

With regard to the choice of search terms, non-rhoticity is the only phonological form for which I have chosen two different search terms and carried out two separate analyses. The first search term is a combination of two forms, *deah* and *fellah*, and the second search term is *bettah*. There are two reasons for searching for articles in which both *deah* and *fellah* occur: The first is that when I searched for *deah* alone, the recognition software also found articles containing *death*. The number of articles resulting from this search was thus very high, and it would have been too time-consuming to manually exclude articles containing *death*. Adding *fellah* to the search made it possible to obtain a collection of articles in which *deah* was accurately identified because *death* was unlikely to

### 3 Tracing enregisterment processes of American English

occur with *fellah* in the same article but likely to co-occur with *fellah* since I had found this combination in previous analyses. At first sight, the term *fellah* seems like an unusual choice for investigating non-rhoticity as the alternative spelling <fellow> does not indicate a rhotic pronunciation. However, another common spelling variant was <feller>, and this spelling could indicate a hyperrhotic form. Schneider (2004: 279–280) draws attention to evidence of a rhotic pronunciation of words like *wash* and *Washington* as well as of words ending in final -*ow* (among others) in early southern America English. He finds that rhoticity in these words is "derived from English sources" (which is in line with Britton's (2007) study of hyper-rhoticity discussed above) and "expanded and redistributed in the American South" (Schneider 2004: 280). Even though Schneider also allows for the possibility that writers did not represent hyper-rhoticity but simply misspelled the words because they were non-rhotic speakers, a representation of hyper-rhoticity is at least a possibility. While <fellah> therefore clearly highlights the *absence* of rhoticity, the alternative <feller> creates an opportunity for writers to draw attention to a (hyper-)rhotic pronunciation. In the following analysis and discussion, I will represent the search term as *deah* AND *fellah* to indicate the co-occurrence of the terms in the same article. However, the first results suggested that not only non-rhoticity was indexically linked to specific social values and personae but also the lexical item *fellow* (often occurring as part of the phrase *my dear fellow*, spelled <me deah fellah>). This is why I conducted a second search using the search term *bettah*. In this case, <bettah> clearly indicates a non-rhotic variant in contrast to the rhotic <better>. Like *deah* and *fellah*, *bettah* also represents a case of non-rhoticity occurring in post-vocalic and nonpreconsonantal position, so that all of these search terms allow me to focus on the later phase of the development of non-rhoticity.

# **3.3.6 twousers: non-rhoticity and phonetic realization of /r/**

In contrast to non-rhoticity, the realization of /r/ as a labiodental approximant [ʋ] has not attracted much interest in linguistics. Foulkes & Docherty's (2000) study on the development of this realizational variant is the most comprehensive one to date. They describe how the variant is described and often dismissed as a speech defect, an infantilism and an upper-class affectation by linguists and other language observers from the nineteenth century onwards. Jespersen ([1909] 1948: 354–355) cites Christmas' (1844) edition of Pegge's *Anecdotes of the English language*, in which the editor adds a footnote stating that some people invariably substitute a *w* for the letter *r* because of their "inability to pronounce the letter" and illustrates this disability with the phrase "Awound the wagged wocks the

### 3.3 Phonological forms and their written representations

wagged wascals wun their wure-wall wace" (1844: 66). Next to this evaluation of [ʋ] as a speech defect, Jespersen ([1909] 1948: 355) also notes that "[t]his *w* is found in some novelists as a constant feature of the speech of noble swells" and he gives the examples of *gwandfather*, *thwee*, *scweeching*, *wight*, *cwied* and *Fwank* from Thackeray's *Pendennis*, which was published in monthly parts between 1848 and 1850. While research on phonetics and language acquisition confirms the higher complexity of the alveolar approximant [ɹ] in comparison to the labiodental [ʋ], which makes it likely that the variant actually occurred and still occurs in children's speech (Foulkes & Docherty 2000: 55), it is not so clear to what extent the stereotypical upper-class usage of [ʋ] was a reality in the nineteenth century. Wells' (1982b: 282) intuition is that at least in the twentieth century perception and reality diverge: "Although this [variant] is often regarded as an upper-class affectation, I am not convinced that it is nowadays found more frequently among upper-class speakers than among those of other social classes". In stark contrast to the stereotype of the [ʋ]-pronouncing swell, Foulkes & Docherty (2000) offer the tentative hypothesis that [ʋ] emerged as a feature of non-standard London accents in the late nineteenth and early twentieth century. They argue that the large number of Jews who migrated to London between 1880 and 1890 were responsible for an increasing frequency of variants similar to [ʋ] in London speech. These variants were probably the result of Yiddish speakers attempting to produce [ɹ] and they might have led to the persistence of the "developmental [ʋ]" used by children in the acquisition process. The assumption underlying this argument is that the adult norms become weaker when variability increases, so that the pressure on children to produce [ɹ] instead of [ʋ] decreases in the light of the higher presence of the latter variant in the community. Therefore, the result is a higher frequency of [ʋ] in non-standard south-eastern accents and a more recent spread of the variant to other urban accents of England. Altendorf & Watt (2008: 212) summarize this development by stating that

there is plentiful evidence of a dramatic rise in frequency of the labiodental approximant [ʋ] in southern England, and indeed in parts of the North. This feature, formerly regarded as an affectation, a speech defect, or an infantilism, is now heard frequently in the accents of a wide range of English cities, and appears generally to be more favoured by young working-class speakers than by middle-class ones.

However, research on the development of attitudes towards this form is scarcer. Foulkes & Docherty (2000) note that with the increase of the variant in London it became associated with the Cockney accent and that the "[ʋ] pronunciation is

### 3 Tracing enregisterment processes of American English

now commonly used to stereotype a Cockney accent" (2000: 37). This represents an enormous shift: A form that had been associated with affected upper-class swells is now perceived as part of a well-known London working-class accent which is increasingly evaluated positively. Foulkes & Docherty (2000: 37) somewhat vaguely describe a "present atmosphere where young people are on the whole positively oriented towards the non-standard south-eastern variety" and therefore towards [ʋ]. They support the impression of a positive attitude towards this variant by listing some public figures (musicians, sports stars, actors as well as television and radio presenters) that can be heard using it. At the same time, they emphasize that the variant is still used to achieve a comic effect, for example, in a television advert for Pizza Hut (Foulkes & Docherty 2000: 32). This advert evokes and plays with the older negative association of the form with defective speech because when Jonathan Ross, a British television presenter, greets the American supermodel Caprice in a Pizza Hut store by pronouncing her name [kæpˈʋiːs] she remarks "You can't pronounce your *r*'s?", marking his pronunciation as flawed. However, she immediately adds "I love a man who can't pronounce his *r*'s". This phrase creates the humor of the advert because the positive evaluation of the variant as attractive and appealing to a beautiful woman runs counter to the negative values of incorrectness and defectiveness invoked by her first remark. This advert is a particularly interesting example because it reveals the ambiguity in the evaluation of the variant in England at the end of the twentieth century. The older negative values were still well-known, but the use of the variant by a popular public figure and his attractiveness to an American model indicate a change of attitude, even though the positive evaluation of the form is still clearly marked as humorous. Whether intentional or not, there is also an additional American dimension included in the advert: Even though the nationality of the main characters (English vs. American) is not foregrounded, it could play a role in reconciling the two conflicting evaluations. Being American, the model Caprice is an outsider to the British English speech community, which might be used as an explanation for why she is presented as finding the form attractive.

In linguistic research on American English, however, the labiodental approximant [ʋ] has hardly played a role, neither on the structural nor on the discursive level. In my first case studies on the linguistic forms presented above, however, I came across many instances of representations of this variant which are indicated in the spelling by the replacement of the grapheme <r> by the grapheme <w>. The prominence of this variant can be illustrated by the following example: In the humorous newspaper article entitled "A Martyr of the Modern Type, but None the Less a Real Hero", published on April 20, 1895[151] in the *Idaho Statesman*,

3.4 Lexical forms

a young man called Percy Paddleford gives an example of a modern martyr in the following way:

"Young Hawold Montmowenci, who was a very deah fweiend of mine—the deah boy!—Hawold Montmowenci was one of the gweatest mawtahs this world has ever pwoduced; 'pon my honah, he was! […] He worked for eight dollahs a week, doncher know. But in spite of the extweme pwices chawged for clothes, he chwished the pwaiseworthy ambition of keeping himself dwessed in the vewy latest fashion. [He continues to describe how his friend spent all his money on new clothes to keep up with the changing fashion.] Yaas; but, doncher know, the noble cweature perwished in the owah of his twiumph. He weally perwished of starvation, but he was the best dwessed wemains that my eyes evah wested upon. Don't twy to tell me about the old mawtahs! Hawold Montmorwenci was the gweatest mawtah in the whole history of wecorded time".

It therefore seemed worthwhile to investigate the role of this variant in the enregisterment of American English in more detail. As a search term I chose the lexeme trousers, and I included several spelling variants in my search. Next to those variants indicating the realization of /r/ as [ʋ] by replacing the first <r> by a <w>, I also included those which additionally indicated non-rhoticity by replacing <er> by either <aw> or <ah>. In addition, I included variants using the older spelling <ow> instead of <ou>. As I also wanted to find out how many variants occurred which indicated non-rhoticity without the realization of /r/ as [ʋ], I included spellings which only altered <er> but not the first <r>.<sup>19</sup> I therefore used ten search terms: *twousers*, *twowsers*, *twousaws*, *trousaws*, *twowsaws*, *trowsaws*, *twousahs*, *trousahs*, *twowsahs*, and *trowsahs* which are all represented in the following analysis by twousers. An additional advantage of using twousers as a search term is that it is also of interest on a lexical dimension as I will show in §3.4.2.

# **3.4 Lexical forms**

# **3.4.1** *baggage* **vs.** *luggage*

According to the *Oxford English Dictionary* (2021), *luggage* is a noun derived from the verb *lug*, which is probably of Scandinavian origin. An early meaning of *luggage*, which is now obsolete, is "what has to be lugged about; inconveniently heavy baggage" and also "the baggage of an army". The current use of the word is indicated as being restricted to Great Britain (where it is "the ordinary word") and its first sense is given by the *OED* as "The baggage belonging to a traveller

<sup>19</sup>This makes it possible to assess the prominence of [ʋ] in comparison to non-rhoticity with regard to the one specific lexical item. If many spelling variants occurred which just signal non-rhoticity, the prominence of [ʋ] would be rather low; however, if only few of these spelling variants occur, the prominence of [ʋ] will be rather high.

### 3 Tracing enregisterment processes of American English

or passenger, esp. by a public conveyance". This shows that the word *baggage* is used to define *luggage*, which indicates that the two terms are basically synonymous. Not surprisingly, *baggage* is also defined by using the word *luggage*: "The collection of property in packages that one takes along with him on a journey; portable property; luggage". Regarding its use, the *OED* states that *baggage* is "[n]ow rarely used in Great Britain for ordinary 'luggage' carried in the hand or taken with one by public conveyance; but the regular term in the U.S." (2021). *Baggage* originates from the Old French word *bagage*, meaning "property packed up for carriage" and it is first attested in 1430, whereas the first occurrence of *luggage* is dated to 1596, so more than a hundred years later. It thus seems that two different forms with different origins but a largely overlapping sense have become used in England at different points in time; and while *luggage* has become the majority form in England, *baggage* has come to dominate in the United States.

That *baggage* and *luggage* are two forms which constitute a linguistic difference between Great Britain and the United States today makes them promising forms to investigate in the present study. However, the primary reason for studying them is again not based on the linguistic situation found today but on newspaper articles which suggest that the forms played a role in nineteenth-century discourses on language. An example for such an article, which I found in the collection of articles containing *hinglish*, is the following anecdote taken from the magazine *Harper's Round Table* and published in the *Daily Inter Ocean* on March 19, 1896[158] (and with minor modifications in the *Denver Evening Post* on March 24, 1896[159] ):

# **A Boy's Observation in Europe.**

We had a great time when we landed. All our trunks had to be opened by the custom house inspectors to see if we had any cologne or cigars in 'em. I don't see why they call them custom house officers though. Their costumes weren't anything wonderful. It took Pop a half an hour to get his trunks all through because he said the inspector didn't know the language. Pop says he asked him what nation he belonged to and the man said he was Hinglish and Pop told him he'd never heard of any such people, where did they live. In Hingland, the man said. Where's that asked Pop, and the man nearly fainted and then Pop gave him a half crown and the man said he guessed he needn't open any more trunks, because a man as ignorant as he was wouldn't have sense enough to try to smuggle anything in anywhere.

After the trunks were all passed Pop asked a man **where the baggage car was** and **that man couldn't speak English either**. He asked Pop what, and Pop says again where's the baggage car, and **just then an American that had been over before says to the man he means the luggage van**, and the man says oh, wy didn't ee si so. Pop says he thinks that's Welsh, which is a language he never liked anyhow. [...] [emphasis mine]

This anecdote exaggerates the difficulty of communication between Englishmen and Americans in a humorous way, and it shows that the difference between *luggage* and *baggage* (here as part of the compounds *baggage car* and *luggage*

3.4 Lexical forms

*van*) are salient forms which are used to illustrate the difference leading to a breakdown in communication.

# **3.4.2** *pants* **vs.** *trousers*

*Pants* is described by the *OED* (2021) as a shortened form of *pantaloons* – *pantaloons* being a word borrowed from French and Italian and first attested in the late sixteenth century in the form *Pantaloun*, designating a figure of the Italian commedia dell'arte (Italian: *Pantalone*). The second sense of *pantaloon* is given as "Trousers, breeches, or drawers" (2021) and the different sub-entry senses distinguish different styles, most of which are now historical, rare or obsolete. Sense 2 d. is important because it defines the sense that is related to the time period relevant for the present study:

Tight trousers fastened with ribbons or buttons below the calf or (later) with straps passing under the boots, which superseded knee breeches and became fashionable amongst men in the late 18th and early 19th cent. Now hist. exc. where retained as part of a livery or military uniform.

This sense may have been the basis for the sense 2 f., which indicates a regional differentiation between the United States and England: It is defined as "Trousers (used generally, without indicating a particular style)" and said to be chiefly used in the U.S., with its first attestation from 1834. As it was the case for *luggage* and *baggage*, the use of *trousers* to define *pantaloons* shows that their meaning is basically synonymous. Not surprisingly, *trousers* is also used in the definition of *pants*. Sense 1 a. defines it as "Originally (*colloquial*): pantaloons. Later: trousers of any kind (in early use applied to men's trousers, but in the 20th cent. extended to include those worn by men and women)". Its first attestation is from 1835 and it is found to have originated in the United States and to be used in this sense chiefly in North America (as well as in New Zealand, Australia and South Africa). By contrast, sense 3 is chiefly British: "(Men's or women's) underpants", which shows that even though *pants* is used in North America and in Britain, the meaning of the form is different in the different regions.

*Trousers*, on the other hand, is described by the *OED* (2021) as an extended form of the noun *trouse*, which was most likely derived from the Irish and Scottish Gaelic word *triubhas* (and not, as it has often been assumed, from Old French *trousse*). The meaning of *trousers* today goes back to the seventeenth century. Sense 2 a. specifies this meaning as follows:

### 3 Tracing enregisterment processes of American English

An outer garment covering the body from the waist to the ankles, with a separate part for each leg; (originally) a loose-fitting cloth garment of this type worn by men, sometimes over close-fitting breeches or drawers; (now more generally) any of various garments of this type worn by either sex (though traditionally more closely associated with men)

The first attestation of *trousers* with this meaning is from 1681, but the *OED* notes that its rise in frequency happened more than a hundred years later, from the beginning of the nineteenth century. In this process, the meaning of *trousers* became more clearly distinct from the meaning of *breeches*: Trousers covered the whole leg while breeches only came just below the knee, and trousers were not as tightly fit as breeches.

Using *pants* and *trousers* as lexical forms in the present study is promising for several reasons: First of all, it provides a good case for studying the interaction between different kinds of perceivable signs, because changes in the meaning of the terms were also linked to changes in the items of clothing that they designated. This can be illustrated by the quote illustrating sense 2 d. of *pantaloons*, which is from the *Chambers's Information for People*, published in 1857: "Pantaloons, which fitted close to the leg, remained in very common use by those persons who had adopted them till about the year 1814, when the wearing of trousers, already introduced into the army, became fashionable". This indicates that changes in fashion, described here as mainly a change from a tight to a loose fit, were related to a change in the use of terms, at least in Great Britain. Yet, in North America the development seems to have been different, with *pantaloons* remaining the lexical form being applied to the garment covering the legs and extending its meaning to all styles of fashion – the main change was the clipping of the form to *pants*, which was apparently first colloquial but then became the general form. The second reason for studying this form is therefore that it provides interesting insights into which role this differentiation on the lexical level used to play in the discourses on language, that is, in the construction of a discursive variety. The third reason for choosing the form is that the analysis of articles containing *baggage* and *luggage* has already revealed that the difference between *pants* and *trousers* is subject to explicit and implicit metapragmatic activities. For example, several newspaper articles listed important differences between the use of lexical items in England and in America, and they often included *trousers* and *pants*. An example is the article entitled "A Few Verbal Errors" published on February 9, 1882[47] in the *Glendive Times* (Montana). It references a prescriptivist work by Alfred Ayres, *The Verbalist*, and gives examples of "the more common errors in the use of words", among them "Pants, for pantaloons, or

### 3.5 Research questions

(better still) trousers". This indicates that, by 1882, *pants* was used in the United States, but that it attracted criticism by prescriptivists, which then circulated in newspaper discourse. A closer analysis of metadiscursive activity surrounding these lexical items thus seems fruitful to investigate if and how *pants*, which is the common form for the item of clothing covering the legs in the United States today, became enregistered as an 'American' form in the nineteenth century.

To conclude this discussion of previous research on the linguistic forms chosen for the present study, it becomes clear that their historical development deserves further attention, not only on the structural but also on the discursive level. This study focuses on the discursive level by investigating their role in the enregisterment of American English in the nineteenth century. Before presenting the results of the analysis in Chapter 4, I will specify the precise research questions guiding the analysis in the following section.

# **3.5 Research questions**

The aim of this study is to investigate enregisterment processes of American English, or, in other words, a register that is indexically linked to the value 'American' in the United States in the nineteenth century. Based on the methodological considerations described in this chapter, the analysis will be focused on the following research questions:


### 3 Tracing enregisterment processes of American English

The research questions subsumed under 1) will be answered by means of quantitative analyses because they aim at discovering the frequency of occurrence of articles containing the search terms. The findings obtained in this part provide an important basis for the subsequent analyses because they reveal when, where and to what extent the phonological forms were part of metadiscursive activity in the first place. It is of course possible and even very likely that the linguistic forms have been subject to discussion in other articles which do not use pronunciation respellings (at least not those selected for the analysis here) or which discuss one of the lexical items without using the other variant near to it, but, as I have already pointed out, using specific search terms makes it possible to create a restricted set of texts and to investigate the articles of this set systematically to identify patterns and trace general developments. In the case of the lexical forms, I will investigate more specifically how often the two variants occur within ten words of each other and identify the frequency of occurrence of articles in which the variants are implicitly or explicitly marked as different (be it in terms of semantic meaning or in terms of indexical values associated with the variants). It is these cases which are particularly interesting because, as in the case of pronunciation respellings, attention is drawn to the difference between linguistic forms, which creates the potential for discovering social differences which are linked to these linguistic differences. The general aim of this first part is to assess the salience of the linguistic forms in metadiscursive activity in quantitative terms because this offers an insight into the questions of when the forms have been enregistered, where this enregisterment took place and the extent to which the form became salient register shibboleths.

The questions subsumed under 2) and 3) will be answered by means of qualitative analyses which focus mostly on the intratextual layer. I will analyze specific examples to show the different ways in which indexical links are created in the articles. This includes the crucial question of how the value 'American' is constructed (Who is American? What characteristics does 'being American' entail?). Furthermore, a central part of the analysis is to identify other discourses and ideologies that are drawn on to establish the indexical links between linguistic forms and social values. In this part, the salience of the form–meaning links is assessed qualitatively by showing how different strategies in the creation of form–meaning links have different effects on the salience of these links. In addition, the analysis also aims at identifying other linguistic forms which are linked to the same social values and characterological figures in order to get a fuller picture of the linguistic repertoire of the registers. Finally, the questions subsumed under 4) focus on the transtextual layer. By adding quantitative elements to the prior qualitative analyses it is possible to identify larger patterns in the creation

### 3.5 Research questions

of links and their role in the enregisterment of American English. This is crucial because, as pointed out in §2.2.1, enregisterment occurs when the typifications occur repeatedly and form-value links become a social regularity. Moreover, it will be possible to trace developments in the social evaluation of the forms under investigation to discover how stable the form-value links were. The following chapter contains the results of the analysis. It is divided into two parts, the first focusing on the phonological forms and the second on the lexical forms. The analysis will be followed by an interpretation of what the results reveal about historical enregisterment processes of American English in Chapter 5.

# **4 Results: metadiscursive activity in nineteenth-century U.S. newspapers**

# **4.1 Metadiscourses on phonological forms**

# **4.1.1 Frequency and temporal and regional distribution of newspaper articles**

The primary aim of the quantitative analysis at this point is to determine the extent to which the linguistic forms which are the focus of the present study are part of metadiscursive activities in newspaper articles. As a first step, I therefore determined the overall frequency of occurrence of articles containing the search terms whose spelling represents a particular phonological form. The result of this analysis, which is shown in Table 4.1, shows that all search terms appear in a number of articles, with *dawnce* occurring least often (in 73 articles) and *bettah* occurring most often (in 374 articles). This already indicates that the pronunciation respellings chosen for this analysis were not just the product of one or two random choices by a few individuals, but they had become reliable indicators of the phonological forms and were used by writers because they expected their readers to recognize which pronunciation they aimed to represent. As can be seen in Table 4.1, I not only identified the number of articles containing the search terms (token frequency), but I also counted how many different articles containing each search term appeared (type frequency) because of the common practice to reprint the same article in different newspapers.<sup>1</sup> Measuring these frequencies provides insights into two important questions: the first one asking how widely the search terms circulated in discourses on language in newspapers and the second one asking the more difficult question of which inferences can be drawn with regard to the salience of the phonological forms represented by the search terms in metadiscursive activities. The first question can be answered by using the token frequencies (and if not indicated otherwise,

<sup>1</sup> If the same article published in the same newspaper appeared in both databases (because the two databases overlap to a small extent), I considered it a duplicate and counted it as only one token.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

all analyses conducted in the present study will be based on the token frequencies). Table 4.1 shows that they vary considerably, with *bettah* occurring in over five times more articles than *dawnce*. This suggests that *bettah* and more generally the representation of non-rhoticity had a wider circulation than that of the other search terms and phonological forms. Articles containing *deah* AND *fellah* as well as *noospaper*/*s* appear less than half as often as articles containing *bettah*, and those containing *hinglish* and twousers appear less than a third as often. While it is possible to argue that a higher circulation also contributes to a higher salience of the pronunciation respellings and the phonological forms that they represent, other factors have to be considered as well to answer the second question.

The first factor that obviously needs to be taken into consideration is the frequency of the lexical item that is used to represent the phonological form by means of a non-standard spelling. If a lexical item is generally more frequent (in standard spelling), it can be expected to occur more frequently in non-standard spelling as well, which means that a higher number of articles containing the search term could also be due to the higher frequency of the lexical item and not due to a higher salience of the phonological form to be represented. Consequently, I have extracted the number of articles containing the lexical item in standard spelling (ASS) and calculated the respective ratios to articles containing the lexical item in non-standard spelling (ANSS), which are shown in Table 4.2. When comparing the ratio of twousers to *trousers* with the ratio of *hinglish* to *English*, for example, it is striking that even though the absolute frequency of ANSSs is almost the same, the ratio of ANSSs to ASSs is considerably different, suggesting that the lexical item *trousers* is much more often respelled to suggest an alternative pronunciation than the item *English*. The ratio of *deah* AND *fellah* to *dear* AND *fellow* is the lowest of all search terms, while the ratio of *bettah* to *better* is the second highest. Therefore, even though both forms indicate a nonrhotic pronunciation, this pronunciation is represented much more often in *dear* AND *fellow* than in *better*. This finding is important in two respects: First, it indicates that the token frequencies of ANSSs are influenced by the lexical items and it is therefore hardly possible to deduce which of the phonological forms is more salient in metadiscursive activities based on the limited number of search terms. (Note that salience at this point is defined as having the quality of being highly frequent.) It is, for example, possible that if a more frequent lexical item containing the bath-vowel had been chosen for the analysis, the search would have yielded a significantly higher number of articles. Nevertheless, as stated above, the frequencies show that the pronunciation respellings drawing attention to phonological forms were not isolated cases but occurred consistently enough to

### 4.1 Metadiscourses on phonological forms

constitute a regular pattern which is worth further investigation. Secondly, the ratios in Table 4.2 serve as an indicator of how closely the lexical item is linked to the phonological form(s) represented by the spelling. Most notably, *trousers* and the combination of *dear* and *fellow* are relatively often represented as being pronounced without the post-vocalic /r/ and, additionally or alternatively, with a labiodental approximant in the case of *trousers*. This suggests that in these cases the indexical links are not just found on the phonological but also on the lexical level. A close connection between a phonological form and a lexical item can contribute to the salience of both forms, lexical and phonological, in other ways than by sheer frequency, as I will show in the second part of the analysis in §4.1.2.

The second factor that can be considered when assessing the salience of the phonological forms in metadiscursive activities in quantitative terms is the relation between the token and the type frequency of the articles. A low TTR-value indicates that the diversity of articles containing a search term is lower, which in turn means that there is a higher number of articles which were reprinted one or more times. Table 4.1 shows that the search terms clearly fall into three groups with *hinglish*, *noospaper*/*s* and *dawnce* having a TTR-value of 71-73%, *deah* AND *fellah* and *bettah* having the same TTR-value of 66% and lastly twousers having the lowest TTR-value of 53%.

Table 4.1: Number of articles containing the search terms in both databases (tokens and types).


How can these values be interpreted with regard to the salience of the phonological forms? It can be argued that a low TTR-value indicates that there is a higher number of popular articles. That is, articles which editors found so appealing that they decided to reprint them in their newspapers. This popularity can have several causes, of course, but since the pronunciation respellings are

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

quite unusual, it is plausible that they at least affected the popularity of the articles in some way. It is striking that both search terms representing non-rhoticity have the exact same TTR-value even though the search term *deah* AND *fellah* appears in less than half the number of articles than *bettah*, which can be taken as a confirmation that the TTR-value can serve as an indicator of the popularity of the phonological form and not just a specific search term (or something else entirely). If a search term appears in a high number of popular articles which have been reprinted several times, this could be interpreted as contributing to the salience of the phonological form – editors notice it and find that the article with the pronunciation respelling might appeal to the newspaper's readers. It is, however, impossible to determine what exactly the editors found appealing, which again underlines the necessity of qualitative analyses to provide a full picture.



Before turning to the qualitative part of the analysis, two important questions remain: When did articles containing the search terms first appear and how did the number of articles change over time? Even though the absolute token frequencies of the articles cannot be compared, it is possible to compare the temporal development of articles containing the search terms over time. Figure 4.1 shows the frequency of articles (tokens) per million and decade for each search term and a clear pattern emerges. On the one hand, articles containing *hinglish* and *noospaper*/*s* appear already early in the century and then keep appearing at a fairly steady rate. There are two obvious exceptions to this steadiness, which are the 1830s for *hinglish* and the 1860s for *noospaper*/*s*, which will need to be investigated in the qualitative part. On the other hand, articles containing *bettah*, *deah* AND *fellah*, twousers and *dawnce* appear earliest in the 1840s and only at low frequencies until their occurrence rises dramatically in the 1880s. Most

### 4.1 Metadiscourses on phonological forms

notably *bettah* and *deah* AND *fellah* show a very similar development, which indicates that it is indeed the phonological forms and not the lexical items which draw more and more attention in metadiscursive activity. Articles containing twousers and *dawnce* appear hardly at all before the 1880s. Based on the selected search terms, it therefore seems as if /h/-dropping and -insertion and yod-dropping played a role in metadiscursive activity throughout the nineteenth century, while the other phonological forms came into focus largely in the last decades of the nineteenth century.

As a means to ensure the reliability of the temporal development found in both databases combined, I compared the development in both databases separately. Figure 4.2 shows the number of articles in the database AHN and Figure 4.3 in the database NCNP. It is striking that the same general trends can also be observed in both databases separately, which indicates that the findings are fairly robust and that the databases are indeed comparable. Moreover, for three of the phonological forms I carried out additional searches in one of the databases (NCNP) to confirm that the observed patterns are related primarily to the phonological forms and not to the lexical items chosen to represent them. Figure 4.4 shows the development of articles containing *Hingland*, *cawnt* and *weally* ('really', the <w> signaling the realization of /r/ as a labiodental approximant) to the ones containing *hinglish*, *dawnce* and twousers.

How does the temporal pattern established above relate to the regional distribution of articles containing the search terms? Figure 4.5 and Figure 4.6 display how often the articles appeared in each state – the darker the color, the higher the number of articles in a particular state. The first general result is that none of the search terms is restricted to articles published only in one state or a specific region, but they are all more or less distributed across the whole country. A second observation is that the two search terms which appeared in articles throughout most of the century, *hinglish* and *noospaper/s*, are also similar with respect to the regional distribution. This might not be apparent at first sight because of the high number of articles containing *noospaper/s* published in Wisconsin (21 articles), but for both search terms there is a relatively high number of articles published in the northeast of the country. For *noospaper/s*, the secondand third-highest number of articles appeared in Massachusetts (12) and New Hampshire (11). Maine (7) and New York (7) also occupy prominent positions on the map. For *hinglish*, the four states with the highest number of articles are Massachusetts (10), Pennsylvania (10), Connecticut (8) and New York (8). In contrast, the other four search terms appear in a relatively small number of articles in the northeast. They are mostly found in articles published in the midwest, west and

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Figure 4.1: The number of articles containing the search term(s) in the databases AHN and NCNP per million and decade

Figure 4.2: The number of articles containing the search term(s) in the database AHN per million and decade

Figure 4.3: The number of articles containing the search term(s) in the database NCNP per million and decade

north of the country, with Colorado and Kansas occupying particularly prominent positions. There is only one eastern state with a relatively high number of articles containing all of the four search terms: Pennsylvania.

There are some dissimilarities between the four search terms as well. Wisconsin stands out among other states for having a high number of articles containing *deah* AND *fellah*, *bettah* and twousers appeared, but the number of articles containing *dawnce* is not particularly high. The high number of articles in Wisconsin is actually something that the three search terms have in common with the search term *noospaper/s*. The search terms representing non-rhoticity also show a very similar distribution. They contrast with the other search terms in being part of a relatively high number of articles published in Missouri (24 articles containing *bettah* and nine articles containing *deah* AND *fellah*). They differ, however, in that *deah* AND *fellah* occurs relatively often in the north (13 articles in North Dakota, 10 articles in Minnesota). The only other term with a relatively high number of articles in North Dakota is *dawnce*. Regarding the southeastern region, another striking difference is that many articles containing *deah* AND *fellah* appear in Georgia (6), which is noticeable because *noospaper*/*s* is the only other search term with a relatively high number of articles in Georgia (8). *Bettah*, on the other hand, appears relatively often in Massachusetts and is similar to

Figure 4.4: The number of articles containing the search term(s) in the database NCNP per million and decade

both *hinglish* and *noospaper/s* in that respect. Finally, Texas and California also appear in the list of the ten states with the highest number of articles containing the search terms *deah* AND *fellah*, *bettah*, *dawnce* and *hinglish*.

Although the regional distribution patterns described above are clearly based on a fairly low number of articles, it is still striking that the differences between the search terms with regard to the development over time are reflected in their distribution across the United States and that the distribution of the search terms representing non-rhoticity is fairly similar. If the distribution was just a reflection of the overall number of articles published in the states, most articles would have to appear in the northeast, which is clearly not the case for all search terms. This suggests that it is likely that the linguistic forms in general circulate more widely in metadiscourses in the respective areas and are highly salient based on their quantity: /h/-dropping and -insertion as well as yod-dropping seem to be part of metadiscursive activity in the northeast more often than in the rest of the country, while the other phonological forms are more often involved in metadis-

Figure 4.5: Number of articles containing *hinglish*, *twousers* or *deah* AND *fellah* in both databases (AHN and NCNP) per state

Figure 4.6: Number of articles containing *noospaper/s*, *dawnce* or *bettah* in both databases (AHN and NCNP) per state

### 4.1 Metadiscourses on phonological forms

courses in the northern and (mid)western parts of the United States. The lower the number of articles containing particular search terms in a state, the more difficult it is to ascertain whether they can be taken as an indicator of the extent to which the phonological form is part of metadiscursive activity in that state. Again, it is only in combination with a qualitative analysis that further light can be shed on the role of region in the enregisterment of these forms. The following section will therefore address the second research question asking which social values and social personae (characterological figures) are indexically linked to the different linguistic forms and the third research question asking how these indexical links are created and how the strategies employed contribute to the salience of the form–meaning links. Section 4.1.3 will then combine the insights of the quantitative and qualitative analyses to provide a comprehensive picture of how the enregisterment of the phonological forms proceeded.

# **4.1.2 Indexical values and social personae: a qualitative analysis**

The main aim of the qualitative analyses in this section is to identify how indexical links between the phonological forms and social values as well as social personae are created. This requires an analysis of the strategies which are employed to typify the linguistic form under investigation and other linguistic and non-linguistic signs by making them point to aspects of the context in which the forms are used. The articles I selected for the following qualitative analyses show the most common indexical links and illustrate different strategies which were used to establish these links. To find out whether the links are stable or changing, the articles were also taken from different points in time.

# **4.1.2.1** *hinglish*

In the databases, the first article containing the search term *hinglish* was published in 1825 in three different newspapers: the *Middlesex Gazette* (Middletown, Connecticut) on July 20, 1825[3] , the *Boston Commercial Gazette* (Boston, Massachusetts) on July 21, 1825[4] and the *Arkansas Weekly Gazette* (Little Rock, Arkansas) on September 20, 1825[5] . As indicated at the end of the article, the newspapers took the article from the *Georgetown Metropolitan*, a newspaper published in Washington, D.C.

The article is a type of text which occurs very frequently in my sample of articles containing pronunciation respellings. I classify such texts as anecdotes because they contain typical characteristics of that genre. According to Bauman (2005: 22), an anecdote is "a short, humorous narrative, purporting to recount a 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Figure 4.7: An article representing the speech of an Englishman, published in the *Middlesex Gazette* (Middletown, Connecticut) on July 20, 1825[3] , retrieved from America's Historical Newspapers

true incident involving real people". Furthermore, there is usually a "focus on a single scene and a tendency to limit attention to two actors" who engage with each other, so anecdotes "tend to be heavily dialogic in construction" (Bauman 2005: 22). In the anecdote above, the truth-claim is established by a first-person narrator who claims that he or she has actually heard an Englishman complaining about Americans. Anecdotes usually focus on actors who can be seen as "representative […] regarding named individuals or groups" (Nicolaisen 2011: 73), so the reader is invited to regard the opinions and behaviors of the Englishman described as being representative of the opinions and behaviors of English people in general. This particular example does not entail any dialogue, but it is an account of reported speech. The American reproduces the voice of the Englishman for the greatest part of the anecdote, thereby providing a representation of the Englishman's accent. The last line of the text is not part of the anecdote. The introduction *Quere* sets it apart from the main text and indicates that the following question is a reaction to the anecdote on the part of the editor publishing the article.

The first line of the text establishes that the main focus of the anecdote is language, particularly the language used by the Englishman, which the narrator la-

### 4.1 Metadiscourses on phonological forms

bels as "a specimen of *pure English*", and the relationship between the English spoken by Englishmen and the English spoken by Americans. The value distinctions between 'good' and 'bad' as well as 'pure' and 'impure' play an important role, which is underlined typographically by italicizing the phrase *pure English*. The evaluation of the Englishman's speech as 'pure' alludes to discourses in England, where, as Mugglestone (2003) shows, purity comes to be associated with speech that does not contain any regionally restricted forms (so-called *provincialisms*). As early as 1687, Cooper described "the best dialect" of London speech as "the most pure and correct" (cited in Mugglestone 2003: 15), which illustrates that links between a repertoire of forms and a place (London) become backgrounded in favor of links between the forms and social values like 'purity' and 'correctness', with the consequence that any use of forms which are associated with different places evokes the lesser values 'impure' and 'incorrect'. Labeling the Englishman's speech as 'pure' in an American context creates two effects: It emphasizes that the speech is purely English, without any American influence, and it also highlights the belief that a form of English speech is associated with correctness and purity, which implies that using any American forms would index impurity and incorrectness. The description of the Englishman's attitude and behavior constructs him as a character who positions himself as linguistically superior: He "was abusing the Americans for many things, and among the rest for speaking such bad *Hinglish* that he could not *hunderstand* them". By setting *Hinglish* and *hunderstand* apart from the rest of the sentence through italics, the author distinguishes the voice of the narrator and the voice of the Englishman. In doing so, he creates an indexical link between /h/-insertion and the Englishman – a link which is reinforced to a great extent by the following mix of direct and indirect representation of the Englishman's speech because every word which would normally be spelled with an initial <h> is spelled without it and vice versa, which creates a picture of a person consistently dropping and inserting /h/ at the beginning of words. The sheer frequency of words representing /h/-dropping and -insertion, the highlighting of these words by means of italics and the almost complete absence of any other markedly differential linguistic forms make the phonological form and its link to Englishness (in contrast to Americanness) extremely salient. Establishing this link serves an important function: It reverses the relationship between English and American speech. The dropping and insertion of /h/ by the Englishman is clearly marked as deviant and non-standard through the spelling, indicating that the phonological form is 'incorrect'. This makes it clear that the aim of the anecdote is to counter the prevailing ideology of English speech being superior to American speech by providing an example of English speech which is incorrect and not pure. The contrast between the la-

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

bel "pure English" and the Englishman's actual speech, marked by the 'incorrect' insertion and dropping of /h/, is a source of irony and humor. The Englishman's arrogance is expressed through his criticism not only of the language used in America but also of the climate ("it is too *ot* in *zummer*"), of the activities ("I *ave* to *habstain* from *hall* the *hinjoyments* of *hexercise* on *orseback* or *hotherwise*") and of the food ("*Hindian* bread" and "*am* and *heggs* [...] with *zumtimes* a fried *en* instead of a chicken"). It culminates in the ridiculous claim that the "*Hamerican orses*, *hoxen* and *ogs* are not equal to the *Hinglish*". The claims of superiority therefore extend beyond language, but it is primarily based on linguistic grounds that these claims are refuted and ridiculed. What is important is that the anecdote does not question the model itself, but it questions the ability of Englishmen to conform to this model while highlighting the Americans' ability to do so.

The question "Was e heducated hat Hoxford?" in the last line adds to the ridicule. As the voice is that of an American asking about the Englishman's education, the use of /h/-dropping and -insertion is marked as mockery, which underlines the irony of the question: It is hardly imaginable that an Englishman deviating from the model of pure speech through the insertion and dropping of /h/ could have attended this prestigious English university and received a good education. This creates an additional indexical link between the phonological form and a lack of education and, by implication, emphasizes the educatedness of Americans. Overall, the strength of the anecdote is that it creates an anecdotal exemplar of an incident – an incident that is to be understood as representative. This impression of representativeness is created through the contrast between the truth-claim and the unrealistic and exaggerated nature of the incident: It is very unlikely for so many words beginning with /h/ or a vowel to co-occur in such a short text. Nevertheless, the reader is called to believe that the general nature of the incident is true, namely that English people think of themselves and of their speech as superior and that they look down on Americans, on their speech and way of life, while they are "in fact" inferior.

The anecdote is a very common text type in the collection of articles containing *hinglish* and anecdotes about similar incidents to the one described above appear repeatedly. An anecdote published on September 14, 1830[7] , in the *Baltimore Gazette and Daily Advertiser* (Maryland) and reprinted in the *Mechanics' Free Press* (Pennsylvania) on October 9, 1830[8] , as well as in *The Arkansas Gazette* (Arkansas) on November 3, 1830[9] , also features an Englishman: "An English *traveller*, who had never been out of the sound of the Bow-bells until he took it into his head to cross the "Hatlantic to take notes"". The encounter between the Englishman and (southern) Americans also includes funny misunderstandings caused by linguistic differences, culminating in the following reaction by the

### 4.1 Metadiscourses on phonological forms

Englishman: "The cockney only half muttered to himself, "vell I do think they murder the king's hinglish most hinumanly in Hamerica." Again, /h/-insertion (and also /h/-dropping) are linked to the speech of the Englishman, who is more specifically characterized as a Cockney, which indicates that there must be some familiarity with a stereotypical Cockney figure, including its linguistic repertoire, which also includes the second prominent linguistic form: the replacement of /w/ by /v/ (as in *vell* 'well'). The humor of the anecdote rests again on the irony created by the contrast between the Cockney's assertion that the English in America is not as 'good' as in England while his own speech exhibits phonological forms which are associated with 'incorrect', 'bad', and 'uneducated' speech.

Another example of an anecdote focusing on /h/-dropping and -insertion to negotiate claims to (linguistic) superiority is the exchange published on July 22, 1856[17] between "a rather waggish New York Judge" and "an overfed John Bull" in a New York Central Railroad car, in which the latter criticizes American pronunciation by stating "It is most hastonishing, sir, to a Hinglish gentleman, to find the pronunciation of the Hinglish lendwidg so defective in this kentry".<sup>2</sup> The negative evaluation of /h/-dropping and -insertion as incorrect, which is the basis for the irony and humor created by this sentence, is reinforced by an explicit metalinguistic comment by the New York Judge, who replies "In this country we call a horse a *horse*, but you call it 'a *nors*', and you think that a man who don't know what a *nors* is must be a *hass*!" In this article, the label "John Bull" is used to indicate that the Englishman is again a representative of Englishmen in general: John Bull is a figure which appeared in eighteenth-century England as a symbolic representation of the English people (see Hunt 2003 for details). As in the first anecdote, the two character traits highlighted are arrogance coupled with ignorance, a combination which creates a source of humor. The characterization of the American figure, the judge, as a wag, portrays Americans as smart, witty, humorous and superior without being arrogant. The final sentence concluding the anecdote asserts that the Englishman has been put in his place and accepted his (linguistic) inferiority: "A laugh, 'like the neighing of all Tattersall's' at this sally, rang through the cars, and our Hinglishman suddenly 'dried up,' and never opened his lips until the train arrived, late at night, at Albany".

<sup>2</sup>This anecdote was published in five newspapers in slightly different versions. The article cited above is the first one. It was published in *The Boston Daily Atlas* in Boston, Massachusetts, on July 22, 1856[17] , under the headline "Orthoepy". The focus on /h/-dropping and -insertion as a linguistic form marking the Englishman's speech as incorrect and inferior is underlined by a change of headline in the second article containing the same anecdote, which was published four days later, on July 26, 1856[18] , in the *Lake Superior Miner*, in Ontonagon, Michigan. It reads "Johnny Bull's Idea of the "Hinglish" Language".

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

A third anecdote, published in the *Bangor Daily Whig & Courier* (Maine) on May 23, 1857[19] (a reprint from the *Boston Herald*), describes the encounter between an Englishman (called Thomas Brown) and an American judge during court proceedings. The Englishman is described as being originally from Liverpool and as having recently come to America with "a confused idea of the [American] nation" because he expects England to be still in control of the country. He is characterized as ignorant and naive, which was the cause of the events leading to the trial: He was exploited and made drunk by some Bostonian "wags" and had to be taken in by the police. The narrator emphasizes that the Englishman was treated in a very friendly manner by the American police - a treatment which is in stark contrast to the behavior of the Englishman during the ensuing dialogue with the judge during the legal proceedings. The Englishman's loud and arrogant manner is highlighted by contrasting it with the calm and self-confident nature of the judge:

"My Lud," cried Thomas, fixing his eye on the Judge, "I'm an Hinglish gentleman, and wishes to no why the hofficer took me in charge last night. My Lud, if I writes to his Highness, minister at Vashington and tells him of this affair, I vould not be responsible for the rage of my Lud Palmeston. I'm 'is friend, and he knows 'ow to protect the 'onor of us Britons." "We have no Luds in this country," said the Judge mildly; "if you wish to say whether you are guilty or not, the Court is ready to hear you." "No Luds in this country, my Lud—no haristocracy, nothing but hingins and histers.—No vonder you 'ave books ritten abont [sic] your nation. Vy the last time I dined with my Lud Palmeston ve vere speaking of 'merica and he told me it was just the place for a man of my talent, and I leave a good business to come here to teach you, and vot is the konsequence?"

# [...]

"[...] I was told of a dozen different situvations, but vhen I vent to get 'em I vas laughed at. My Lud, I,ve [sic] been shamefully treated, and I vants justice, or the valls of this 'ere building will shake with Hingland's rage when she 'ears of it. Remember the fate of Sebastopol, and tremble for your 'omes." The Judge intimated that he was not alarmed, as he knew that the revenue cutter was in port and could afford ample protection to the city against all the fleets which the English would send to revenge his cause. A fine of \$3 and cost was imposed, and as he was being taken down stairs he was begging the officers to let him have pen and paper for the purpose of "riting to the Hinglish Minister."

The anecdote creates a representative incident illustrating the struggle for superiority, with the better end for the Americans. The focus here is not explicitly on language but on perceived differences between the two countries regarding politics and society. The hierarchical English society with a system of nepotism and favoritism is contrasted humorously with the American egalitarian society with its just legal system, which puts the Englishman in his place. As in the anecdote above, the Englishman's arrogance is countered by confidence, humor and wit, thus portraying the Americans not only as superior but also as more friendly and social. Language comes in more implicitly; that is, the speech of the Englishman is clearly marked as different from that of the Americans. The

### 4.1 Metadiscourses on phonological forms

representation of the Englishman's pronunciation focuses on two salient forms: /h/-dropping and -insertion and, as in the first anecdote, /v/-/w/ interchange. In addition, *Lord* is consistently spelled <Lud>, which indicates a non-rhotic form, but as it is the only lexical item, it is not the case that non-rhoticity is associated with the Englishman more generally. Apart from the pronunciation respellings, there are also a few cases of eye dialect: *ritten* 'written', *riting* 'writing', *no* 'know', *konsequence* 'consequence'. With regard to grammatical forms, the marking of the first-person singular by the suffix -*s* is striking (note, however, the exception in *I leave*). This illustrates that the Englishman is characterized as ignorant both explicitly and implicitly. The dialogue demonstrates that the Englishman knows nothing about American politics, society and the legal system, even though he claims to be a businessman with connections to the highest political circles, and this characterization is underlined implicitly through his use of linguistic forms indexing uneducatedness and ignorance.

Next to anecdotes, there are articles containing *hinglish* in which the author discusses the differences between English and American speech explicitly. These articles are few but nevertheless very important as they confirm links created more implicitly in anecdotes. One such article deals with /h/-dropping and insertion. It was published in the *Cincinnati Daily Gazette* (Ohio) on October 18, 1873[27] , and carries the headline "The Hex-Hasperating Hinglish". By using this headline, the author already creates a link between /h/-dropping and -insertion and Englishness and the characteristic of being exasperating and in the article he explains why this is the case. He reacts to a letter to the editor written by "an Englishman" to the *New York Tribune* in which the Englishman denies "the statement that Americans in general have a more correct use of "h" than most of the educated Englishmen". The author first cites the Englishman's argument:

The author's response is full of irony: "Unhappily the h can not serve us; for common schooling and much intercourse have deplorably diffused the faculty of respectable speech. How happy is English society in having a dividing test which everyone aspirates!" He continues to argue that it is precisely that class of tradesmen, shopkeepers, mechanics and laboring men who are the "very substance of the kingdom" and while the Englishman arrogantly disregards them as not respectable, they are embraced by Americans, who recognize that every person plays an important part in the "great political, religious, benevolent, and

In England the proper use of the letter h constitutes a distinguishing line between the upper educated classes and that class composed of tradesmen, shopkeepers, mechanics, and laboring men, the latter never pronouncing it, while its omission by the former would exclude the offender from respectable society.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

social operations" of the American nation. In sum, the absence of /h/-dropping and -insertion is linked to the absence of social hierarchies in America, which is clearly evaluated positively. As a consequence, the argument by the Englishman is not deemed valid. As it is in fact the class whose members drop and insert /h/ which "embraces the very substance of the kingdom", it is concluded in the article that the Englishmen generally speak less correctly. The author underlines this argument by claiming that /h/-dropping and -insertion is "a fault of speech which we verily believes is worse than all the Americans are guilty of". This shows how salient the phonological form is in not only distinguishing English from American people but also as an index of American social and linguistic superiority. Furthermore, the articles show how metadiscursive activity in America is linked to metadiscourses in England. Through articles like these, Americans learn about associations between the phonological forms and social standing ('lower class'), standardness ('incorrect according to the standard'), degree of education ('uneducated', 'ignorant'), and transform these evaluations in their context to fit their purpose of establishing superiority over England.

Another article which explicitly comments on differences focuses not so much on society or politics, but more on manners, behaviors and fashion. It is a letter written by Mr. Bailey, a "Danbury man abroad", and printed in the *Inter Ocean* (Chicago, Illinois) on October 31, 1874[29] . A small part of it was reprinted by the *Lowell Daily Citizen and News* (Massachusetts) on January 30, 1875[30] . Mr. Bailey writes from England and compares English customs and fashion to his American (more specifically New England) home. The part of the letter reprinted in 1875 is headed "The Hinglish Hi-Glass". The introductory sentence states "We exceed the English in building cars, but they completely distance us in wearing an eyeglass". This not only describes a difference, but it also creates irony by comparing an enormous technological achievement to a simple fashion item. That the author evaluates this fashion item and the importance attributed to it as ridiculous becomes clear in the following paragraph:

It is worn only by the English exquisite, and he generally dons it as he asks a question, or on entering a room where there is anybody to see him. Sometimes it is suddenly put up without any apparent provocation. I imagine that it is worms. The wearer has a baggy costume, parts his hair in the middle, and has in his face an expression of mild idiocy, which is much strengthened by the glass.

The writer creates a vivid picture of a typical English upper-class figure ("the English exquisite") by describing his hair style, style of clothing and his inner qualities (or lack thereof). The focus on the eyeglass serves an important function: It illustrates the contrast between how the upper-class man wants to appear

### 4.1 Metadiscourses on phonological forms

and the impression he actually makes on people. By emphasizing that "he generally dons it as he asks a question, or on entering a room where there is anybody to see him", the author suggests that the eyeglass is used by the upper-class man to appear smart and educated. This is contrasted with the "expression of mild idiocy, which is much strengthened by the glass", which suggests that in fact the opposite effect is created. This effect is strengthened by the link between /h/-dropping and -insertion and the lexical item *eyeglass* ("hi-glass") because the social meanings indexed by the phonological form (in particular stupidity) are linked to the object and to the person wearing it. Consequently, the man with the eyeglass appears vain, arrogant and idiotic (because he thinks more highly of himself than is actually justified). In this way, the form serves to characterize a group of people even though they are *not* described as using the form themselves – usually /h/-dropping and -insertion is rather associated with the speech of lower classes, but through this indirect link the qualities of uneducatedness and ignorance, associated with the lower classes, are transferred to the upper classes. In this process, the negative evaluation of the form is in turn strengthened, particularly the qualities of being English, arrogant and ignorant.

That the social meaning indexed by /h/-dropping and -insertion is very wellestablished at least by the second half of the century can be shown by analyzing a particular text type: very short articles, which often consist of just one line, one paragraph or dialogue, and are often humorous in nature. They are normally published together in sections labeled for example "Multiple News Items" or "General and Personal". The following article was published on July 10, 1878[36] , in the *Daily Rocky Mountain News* (Colorado):

Punch puts these words into the mouth of an American lady when viewing Swiss scenery: "O, my! ain't it rustic!" Were she "Hinglish" she would say: "Ho, my hi's! Hisn't hit hawful 'igh!"

Even in this very short text, English condescension towards American speech is countered by the representation of English speech as being marked by /h/ dropping and -insertion. The representation of the phonological form suffices to portray English speech as inferior, contrary to all claims made in England, as in this case by the magazine *Punch*.

Another example of a short article shows how the negative evaluation of /h/ dropping and -insertion is used to mark other linguistic forms used by English people as 'English' and decidedly 'non-American' and to indicate that they carry no prestige in an American context at all. It appeared on April 19, 1881[45] , in the *New Haven Evening Register* (Connecticut) and contains three sentences:

The London *World* tells a story about an American banker, in which he is made to speak of his "top coat." A reward of "4 pun, 6" is offered if the *World* will put its finger on a straight haired American that speaks of his extensive outer garment as a "top coat." That's a "blarsted" Hinglishism.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

The lexical item marked as English and not American is *top coat* (in America, *overcoat* is used instead). This marking is done explicitly by calling it a "Hinglishism" ('Englishism'), in analogy to the terms "Scotticism" or "Americanism" commonly used to mark words as being tied to a particular region and therefore deviant from 'common' and 'proper' English. Here, this view is reversed by marking the English form as deviant from common American English. The use of /h/-dropping and -insertion additionally links the lexical item *top coat* to the social values 'incorrect' and 'improper', an evaluation which is reinforced by the adjective *blarsted* (a respelling of *blasted*, meaning 'damned'), which is itself marked as an Englishism through the use of quotation marks. In addition, the spelling <blarsted> indicates a lower and back bath vowel, which is therefore constructed as yet another English form not evaluated as correct and desirable in an American context.<sup>3</sup>

Even though /h/-dropping and -insertion is only one of several forms which are evaluated as 'English' and 'not American', it is nevertheless a very salient one. This salience is illustrated by the use of the form in articles which do not address linguistic or non-linguistic differences between England and America at all, but have a completely different topic. One such case is an article in which /h/-dropping and -insertion mainly serves to criticize an English person. It was published on May 27, 1882[51] , by the *Salt Lake Tribune* as a reaction to an article written by an English journalist for the New York *World*. The English journalist expressed a favorable attitude towards Mormons in his article, for example finding polygamy "neither unnatural, wicked nor licentious". By labeling the journalist "Han Hinglish Hass" and using this designation as the title of the article, the author reacting to the article not only emphasizes that the author of the original article is English, but he also characterizes him as stupid, a trait evoked through the use of /h/-dropping and -insertion. From the very beginning, it is therefore clear that it is the aim of the article to counter the journalist's view by questioning his character and his abilities, which is also done quite explicitly in the article itself by describing him as "a direct cross between a characterless knave and a pitiful fool" and as being "simply a cockney ass".

A very extreme case of foregrounding the social meaning of the phonological form can be observed in instances where the social meaning indexed by the phonological form becomes part of or even replaces the semantic meaning of the lexical item. An example of this kind is an article published on October 20, 1892[123] , in *The Galveston Daily News* (Texas), which reports the visit of Governor Jim Hogg to Farmersville, Texas, where he spoke to an audience of 2,000 people. The governor is quoted as saying:

<sup>3</sup>A lower and back bath vowel can be indicated by several spellings: <ar>, <ah> and <aw>.

### 4.1 Metadiscourses on phonological forms

They charge me with driving prosperity from Texas. That's Hinglish, you know. All know that we have our flush times and dull times just as we have hot and cold spells. The man who believes the people to be suckers is himself worse than a sucker - he is a mud cat.

In this context, he uses the phrase *That's Hinglish, you know* to deny the charge that he drives prosperity from Texas, which makes the meaning of *Hinglish* equivalent to the meaning 'incorrect' or 'stupid', which are exactly those social values which have come to be indexed by /h/-dropping and -insertion. The actual semantic meaning 'English' is backgrounded, if not completely replaced by the meanings indexed by the phonological form.

The cases discussed so far reveal how indexical values linked to a phonological form are used in the articles and how they are strengthened in each instance they are used. While there is always a potential of change inherent in each use, the social meaning of /h/-dropping and -insertion seems to be rather stable. One of its core indexical meaning is 'Englishness', and the stability of this meaning becomes clear in articles in which Americans are described as trying to imitate English manners, fashion and speech. In the article "The Danbury Man Abroad", already discussed above, the author describes the following case:

There was one young man from Marlborough, Mass., stopping in London last summer, who devoted three whole months, but in vain, to make an eye-glass stay in his eye. I could always tell when he failed by hearing him howl and swear and kick the furniture. At the end of the three months he went home, as both his time and money were exhausted. When his room was cleaned, two full quarts of damaged eye-glasses were gathered up.

This anecdote ridicules the attempts of Americans to imitate English fashion, in this case wearing an eyeglass, a fashion item that was linked to the use of /h/-insertion in articles cited above ("The Hinglish Hi-Glass"). The obvious exaggerations create humor and lead the readers to distance themselves from the American and his attempt to imitate something not desirable.

Whereas, in the case described above, the American imitates English fashion while he is in England, the following article describes this phenomenon happening in America as well. It represents yet another text type, a poem, which can be grouped with the short articles collected in rubrics like "All Sorts of Items". The poem was published originally in the magazine *Life* and was reprinted in the *Daily Evening Bulletin* (San Francisco, California) on January 10, 1888[84] .

A THING TO BE EXPORTED. Oh, why is the Anglo-American proud?– His style is imported, you know. But why is his manner insuff'rably loud?– That's also imported, you know.

With "Lunnon made" raiment he cuts a great dash; For everything "Hinglish" he shells out his cash; No matter the value, to him all is trash That is not imported, you know.

His wines and cigars are the best to be had– That's freshly imported, you know. He makes it a point to adopt the latest "fad" That has been imported, you know. With a little round window stuck into his eye, He ogles humanity as from on high, An asinine figure to cut he doth try– The notion's imported, you know.

It makes a plain Yankee excessively tired To see things imported, you know; Placed up on a pedestal to be admired, Because they're imported, you know. And this Anglomaniac with his odd ways, Who spends time and wealth on some imported craze, Assuredly shou'd, for the rest of his days, Be quickly exported, you know.

The poem contains a strong and explicit criticism of Americans who value English goods, fashion and manners more highly than American ones. The foreignness and non-nativeness of the described products and behaviors are emphasized through the regular repetition of "imported, you know" in the second, fourth and last line of every stanza, except in the very last line where the replacement of "imported" by "exported" creates a strong contrast highlighting the central message of the poem that these Anglo-Americans should leave the country. Even though linguistic differences are not addressed explicitly (except for the expression *to cut a figure*, which is labeled "imported"), they are used to underline this criticism: By designating that what is imported "everything 'Hinglish'", the indexical values associated with /h/-dropping and -insertion are used to mark everything English as inferior to everything American. It alludes to the distinction between the values 'inferior' and 'superior' that also played a great role in the anecdotes discussed above. The American who puts imported English goods on a pedestal is labeled "Anglo-American" and "Anglomaniac" and he is characterized as feeling superior ("ogling humanity as from on high"), while, at the same time, he is also ridiculed and shown to be inferior, for example by describing him as wearing an eyeglass ("a little round window stuck into his eye"). By using English things and imitating English manners, the Anglo-American therefore also occupies the same role as the Englishmen in the anecdotes discussed above. Furthermore, a contrast is created between the apparent insanity and irrationality of the Anglo-American (his admiration of everything English is compared to a mania and a

### 4.1 Metadiscourses on phonological forms

craze), and the 'normality' of the figure of the "plain Yankee", who does not (need to) feel superior but who is simply sane, rational, and proud of American goods and ways of life and calls for an "exportation" of the "Anglomaniac".

The last example I would like to discuss here also contains the figure of the Anglo-American, labeled "American Dude" (this figure will be discussed in more detail in later analyses). It is a short article containing a fictional dialogue between the dude and a "Tramp". It was published on January 12, 1889[90] , in the *Daily Inter Ocean* (Chicago, Illinois), and, like the poem above, taken from the magazine *Life*.

# SWALLOWED IT WHOLE.

*Life*: Tramp—Hi say, sir! Cahn't you 'elp me a bit? Hi'm Hinglish meself, sir.

American Dude (pleased)—Aw—what's that, me good fellow (takes out a bill), and—aw—why d'you think I'm English, y'know?

Tramp—Hoh, sir, henny one could see that! I beg parding; harn't you the Duke of Southampton, sir—Your Grace I mean?

American Dude (sick with bliss)—There, there, me good fellow, take that to help you back to Lunnon (walks haughtily on.)

In this scene, the tramp gets the American dude to give him money by flattering him: He pretends to identify him as an Englishman and establishes common ground by emphasizing his own Englishness (which might as well be put on). The representation of /h/-dropping and -insertion in this anecdote is particularly important because even though both characters want to signal their Englishness, only the tramp uses /h/-dropping and -insertion and he does so extensively. The American dude, on the other hand, is shown to use the filler *aw* several times as well as the possessive pronoun *me* and the discourse marker *y'know*, which, as I will show in later analyses as well, mark him as an imitator of English speech. The character ridiculed in this article is clearly the American dude. He is described to be "sick with bliss" and haughty after he supposedly passed for an Englishman, and not just for any Englishman, but for a Duke. The allusion to sickness indicates that the American dude has lost his capacity to think straight, leading him to fall for the insincere flattery by the tramp who is just interested in getting money. The tramp is portrayed as smart and witty because he takes advantage of the dude's vanity. Dropping and inserting /h/ enables him to index Englishness as well as a position of social inferiority, which he needs to compliment the American dude. The fact that the American is not portrayed as using the form indicates that it still serves as an index of British speech (in contrast to American speech) and a social position that the American does not aim at. His use of other linguistic forms to sound English and to position himself among the upper

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

classes is ridiculed and criticized by showing how it leads him to be exploited by someone who is (supposedly) at the other end of the social ladder.

To summarize, the qualitative analyses of the collection of newspaper articles containing *hinglish* show that /h/-dropping and -insertion is a very salient and stable index of Englishness and social and linguistic inferiority throughout the nineteenth century. It is used to counter perceived arrogant English claims of superiority and mark Englishmen as ignorant and unrefined. The social meaning was so well established that it could be shown to be part of or even replace the semantic meaning of the word (*That's Hinglish* meaning 'That's stupid'). Moreover, it was used to criticize the admiration of English goods, manners and speech on the part of a group of Americans and to ridicule their attempts to pass for Englishmen. In these attempts, the Americans never drop or insert /h/, which underlines that the form remains a stable index of Englishness and low social standing. The qualitative analyses also demonstrate that the forms occur in several text types and that there are several strategies by which the phonological forms are linked to their indexical values. A very effective strategy is to either describe characters and their speech as well as their behavior or to make use of these characters in anecdotes or short humorous dialogues and illustrate their speech and behaviors in concrete situations. In the last two text types, the reader is given the impression that he or she gets an immediate and truthful rendering of an event that actually happened (especially in the case of the anecdote which establishes a truth-claim) or that could have happened (in the case of the humorous dialogues). Regarding the effect of anecdotes, Nicolaisen (2011: 73–74) writes:

Whether anecdotes told about […] individuals are intended to enhance or to denigrate them, the anecdotes relate incidents seen as typical of the individuals' actions or other qualities. In performance, an anecdote is often used to underpin or confirm in its pointedness a characteristic previously ascribed to an individual ("To show you what I mean, let me tell you a story I heard about so-and-so"). Whether they are believed to be true or known to be apocryphal, anecdotes can be powerful rhetorical tools thinly disguised as narrative entertainment.

This quotation illustrates how anecdotes contribute to the creation of characterological figures which are then used in shorter humorous dialogues and texts. These figures embody character traits seen as typical and it is shown how these traits co-occur with a linguistic repertoire and how this plays out in concrete situations. These figures therefore represent models of behavior and speech – usually both positive and negative ones. This supports Agha's (2007: 177) argument that

### 4.1 Metadiscourses on phonological forms

"characterological figures [...] motivate patterns of role alignment in interaction". The readers of the article are invited to align themselves with the normal, intelligent and educated American, either in opposition to the arrogant and stupid Englishman (sometimes Cockney) or in opposition to the manic Anglo-American.

In the following section I will continue the qualitative analyses of newspaper articles by focusing on three search terms: *dawnce, deah* AND *fellah* and twousers. I analyze them in one section because they not only show a very similar temporal development (appearing at the end of the nineteenth century, with an enormous increase in the 1880s), but also because the indexical values and characterological figures associated with the phonological forms are very similar. I will discuss *bettah* in a separate section, however, because even though the search term represents non-rhoticity as well, there are additional social meanings linked to it which cannot be identified for *deah* AND *fellah*.

# **4.1.2.2** *dawnce, deah* **AND** *fellah,* **twousers**

The first result of the qualitative analysis of articles containing *dawnce*, *deah* AND *fellah* and twousers is that the linguistic forms are linked to Englishmen as well. The first article in the databases containing *dawnce* is a short humorous paragraph describing the fate of an Englishman who had been traveling to America and had criticized many aspects of American life. Upon his return to England, he was supposedly arrested and "put in jail by his employers". The paragraph was taken from the magazine *The Knickerbocker* (from "the 5th number of the droll papers") and published in the *Boston Courier* (Massachusetts) on November 12, 1849[14] . It reads:

*Summary*. A. Jarroldy, travellink Hinglishman, who found such faults with our otels, and could n't get any think to suit him, and druv a fine teem of osses while he was in this ked'ntry, and sported a white choke at our principal balls, has landed in Hingland, and been put in jail by his employers. Ha! ha! ha! "Oi loike the monner they dawnce in Frawnce!"

There is an interesting division into two parts. First, the narrator tells the story by using pronunciation respellings indicating /h/-dropping and -insertion (*Hinglishman*, *otel*, *osses*), non-rhoticity, which is restricted to the lexical item *osses*, the realization of -*ing* as /ɪŋk/ (*travellink*, *any think*) and a differential pronunciation of *country* (three syllables and a higher strut vowel). There is also one case of eye dialect (<teem> 'team'). In addition to the phonological forms, there is a grammatical form: the past tense form of *drive* used by the narrator is *druv*. This part of the text can be interpreted as the narrator's voice which exhibits linguistic forms commonly associated with English speech to mock the

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Englishman. The characterization of the Englishman takes up established stereotypes: As in many articles analyzed above, the Englishman is depicted as arrogant and as claiming to be superior to Americans, which becomes visible through his repeated criticism of everything American ("who found such faults with our otels and could n't get any think to suit him"). The humor rests on the contrast between the Englishman's assumed superiority and his actual inferiority, which is pointed out in two ways: implicitly by representing his language and explicitly by describing his social downfall. The social values which I have shown to be indexically linked to /h/-dropping and -insertion mark him not only as English, but also as ignorant, uneducated, and unrefined. In America, he pretends to be a man of high social standing by attending balls, dressing elegantly and driving "a fine teem of osses", but in England he is put into jail, which means that his use of language and his social position are aligned. That the Englishman's social (and linguistic) inferiority is enjoyed by Americans is expressed through the laughter and schadenfreude ("Ha! ha! ha!") on the part of the narrator. After this part, the quotation marks indicate a change of voice to that of the Englishman himself. In this last line of the story, new phonological forms are represented: the price vowel with a back and rounded onset (*Oi*, *loike*), a lowered and backed bath vowel (*dawnce*, *Frawnce*), and also a lowered and backed trap vowel (*monner*). A likely interpretation of this exclamation is that the narrator aims at ironically contrasting the Englishman's preference for fancy balls and upper-class activities with his speech, which clearly marks him as not belonging to these social circles. This interpretation suggests that the back realization of all of these vowels is linked to vulgar speech, which would fit the findings suggesting a negative social evaluation of the back bath-vowel in England in the early nineteenth century (see §3.3.4). As there are no other articles containing *dawnce* which appear as early as the one cited above (the next ones were published in 1877), this interpretation cannot be supported by further evidence, but it remains tentative. The fact that the bath-vowel is only represented in the direct quotation could indicate that the author considers it a new linguistic form which is not wellestablished enough to be used in the mocking description of the Englishman in the first part. The repetition of the vowel in *dawnce* and *Frawnce* and their position at the very end of the article puts a lot of emphasis on the vowel and ensures that it is noticed by the reader. All in all, what can be deduced from the article is that the form was associated with British English speech and that it constitutes tentative evidence of an early stigmatization of a back vowel in bath (and also in trap).

Whereas this early article links the form to the figure of the arrogant but ultimately low-life Englishman that is typically associated with /h/-dropping and

### 4.1 Metadiscourses on phonological forms


A young Oxford swell once described his stock in trade after this fashion: "I have my walking twousers, my standing up and my sitting down twousers, my morning and my evening, my dining and my dancing twousers". He was all twousers. We trouble to think what might have happened had he put on his standing trousers when he was sitting down, or his sitting when he was standing.

It is noticeable that the representation of the speech of the Oxford swell focuses on one phonological form: the realization of pre-vocalic /r/ as [ʋ]. It becomes highly salient because it occurs in the word *trousers*, which is used frequently as it refers to the object discussed in the article. Following the direct quotation, the author comments on the swell and by using the labiodental /r/ ("He was all twousers"), he not only puts emphasis on the form, but he also mockingly distances himself from its use and from people like the swell, who is portrayed as a person of a high social standing (he must have money to afford all these different kinds of trousers), who is educated (as the association with Oxford suggests), young and obviously very much concerned with dressing fashionably. The comment "We trouble to think what might have happened had he put on his standing trousers when he was sitting down, or his sitting when he was standing" is full of irony and illustrates that the American author finds this way of dressing ridiculous and impractical. As in the article "The Hinglish Hi-Glass", published in the January of the same year and which I have analyzed above in relation to /h/ dropping and -insertion, the English upper-class is the subject of ridicule in this article. In "The Hinglish Hi-Glass", I have shown that the negative social meanings indexed by /h/-dropping and -insertion are used to mark the upper-class figures as stupid, even though they are not shown to use the form themselves. This article portraying the Oxford swell confirms that /h/-dropping and -insertion is not part of the linguistic repertoire of upper-class English swells. In contrast to the article "The Hinglish Hi-Glass", however, this article does specify a linguistic form that differentiates the repertoire of the swell from that of other (English) people: the labiodental /r/.

An article which links more linguistic forms to the repertoire of the English swell is a travelogue written by Prentice Mulford for the *Daily Evening Bulletin* (San Francisco, California) and published on April 3, 1873[25] . It is titled "Gossip from London" and contains accounts and comments on the author's experiences

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

when traveling in England. Among other things, he describes how he attended a literary reception in "England's aristocratic halls" and encounters a particular type of young man in this context:

There were present many young gentlemen in full evening dress—tail coat, expansive bosom, white neck-tie, eye-glasses. They could say, "bah jove!" and "mah deah fellah". They would give their eyeglasses the correct twirl. They stood up during the entire evening. As social perpendicularities, they were successes every one. I was delighted with them. I had read of this sort of thing. I had seen it portrayed on the stage but never before in actual reality. In these ran the genuine Simon-pure blood of the Dundreary's.

Again, the men are characterized as young and dressed fashionably. Dress and manner are linked through the description of the twirl given to the eyeglass and the emphasis on the "correct" twirl indicates that dress and manner are constitutive elements of this group of young men. The two examples given of his speech link not only non-rhoticity but also price-monophthongization (*bah* 'by', *mah* 'my') and the phrases *by jove* and *my dear fellow* to the figure of the young English swell. The author's evaluation of this figure is conveyed through irony and humor. His description of them being "social perpendicularities" and "successes" seems to indicate the author's admiration of their success in climbing the social ladder, but the exaggerated nature of this admiration ("I was delighted with them") already hints at irony, which is supported by the contrast to the description of the rather funny looking appearance and manner of the swell and his emphasis on small gestures like twirling the eyeglass. The seemingly innocent comment "They stood up during the entire evening" can be read as an iconic representation of the swell's uptight nature and his anxiety about the social position that he has reached, which does not allow him to relax and sit down. Against this background, the author's delight in them can also be understood as his delight in laughing about them. The last lines of the quotation link his own experience to that of other Americans: They have read about such figures or seen them portrayed on stage. Lord Dundreary is a character in the play *Our American Cousin*, written by the English playwright Tom Taylor and first staged in New York in 1858. Adams (2012: 206) describes Lord Dundreary as a "buffoonish English aristocrat" who balances the main American character in the play: Asa Trenchard, a "loud and often boorish" Vermont backwoodsman. The play was very successful and Lord Dundreary "became one of the great comic turns of the latter half of the century" (Adams 2012: 206) and influenced the image that Americans had of English aristocrats. Mulford's assertion that upper-class people behaving, dressing and speaking like Lord Dundreary are not just fictional constructs, but really exist in England, serves to consolidate this image.

### 4.1 Metadiscourses on phonological forms

Other English people who appear in the articles are also members of the upperclass. An article published in the *Grand Forks Daily Herald* (North Dakota) on September 18, 1883[61] , is a comment on the rumor that the Hon. Lionel Sackville West, a British diplomat in the U.S., gave an interview to advertise the American Northwest to Americans in order to keep them out of Canada and protect the interests of the British aristocracy in Canada. The author of the article writes:

In this article all phonological forms under investigation are represented: nonrhoticity (also in *aw*, 'are', *weah* 'were', *theah* 'there', *doah* 'door', *pewmit* 'permit' and *mawning* 'morning'), labiodental /r/ (*weally*, *vewy*, *pwecedent*, *sorwy*) and the back bath-vowel (*cawn't*). In addition, *aw* is also represented as a frequent filler item. Two further phonological forms are a wider mouth-vowel (*naow* 'now') and a close, monophthongal face-vowel in the name *James*. What is as noticeable as the phonological forms is the extensive use of discourse markers: *there you are* and *you know* occur multiple times. Because they occur so often, they disrupt the coherence of discourse rather than adding to it. In addition, the language used by the diplomat is highly formulaic – the only semantic content expressed is that he cannot do anything because this would create a problematic precedent. This creates the impression of indirectness and evasiveness, which can be considered typical of polite and diplomatic English speech, but which is also portrayed as unsatisfactory because of its low informational content and as annoying because of its incoherence. The exaggerated nature of the representation also shows that this type of speech is rather ridiculed and criticized in this context. The explicit comment that the quotation is based on "probability" and therefore imaginary indicates that the author aims to represent not just the speech of the Hon. Lionel Sackville West, but that of English upper-class diplomats and very likely upperclass men more generally. This has the effect of attributing universality to the form–meaning links because it indicates that there is a model available to create the image of incoherent, indirect upper-class English speech, which is full of polite, but ultimately meaningless phrases. Non-rhoticity, labiodental /r/ and a back realization of bath are evidently part of this model as well. Drawing on this model to create an instance of typical English upper-class speech, which is attributed to a real English upper-class diplomat, serves to transmit the model to a broad audience.

One anecdote that is told or alluded to several times also relates to a real person: a visitor from the other side of the Atlantic who is well-known in America,

The probability is, however, that what he really did say to the reporter was: "Aw - my deah fellah - I weally cawn't do it, you naow, and theah you aw, you naow; if I weah to pewmit this pwecedent, it would be no end of twouble, and theah you, aw again, you naow. Vewy sorwy, that I weally cawn't oblige you - Jeames, kindly attend the - aw - gentleman to the doah - good mawning".

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

the author Oscar Wilde. It was published in two newspapers on February 16, 1882: in *The Lynchburg Virginian* (Virginia, February 16, 1882[48] ) and in the *Boston Daily Journal* (Massachusetts, February 16, 1882[49] ).

One of Mr. Wilde's remarks made in Washington, has become a popular phrase in that city. The correspondent of the *Courier-Journal* says that in response to an invitation to dance at the Bachelors' German, which he attended, escorted by Mrs. Robeson, he said: "I have dined and don't dawnce; those who dawnce don't dine". And now this speech is repeated, with all manner of jeers and jokes, on all occasions.

This shows how newspapers are not only part of speech-chain linkages, but they are also sources which describe how they work. In this case, the back bathvowel is part of a phrase which is attributed to Oscar Wilde, who is perceived as an Englishman of a high social standing even though he was actually Irish.<sup>4</sup> As the phrase is deemed funny it is often repeated, so that the form–meaning link ultimately becomes recognized by more and more people in America.

The articles describing the English swell discussed above demonstrate that /h/ dropping and -insertion is not a form associated with this figure. However, other English figures use /h/-dropping and -insertion in combination with non-rhotic forms and a back bath vowel, but they do not use labiodental realizations of /r/. One such figure is part of a popular anecdote taken from the magazine *Texas Siftings* and published first in the *Atchison Daily Globe* (Kansas) on March 26, 1892[117] , and then in the *St. Paul Daily News* (Minnesota) on March 31, 1892[118] and finally in the *New Mexican* (Santa Fe, New Mexico) on April 30, 1892[120] . 5

# **He Wanted an 'Orse.**

An English visitor, stopping at a prominent New York hotel, sauntered up to the genial clerk during the recent cold snap and, adjusting his eyeglasses, said:

"My deah fellah, cawn't you let me have a sledge?"

"A sledge?"

"Yas."

"John," said the clerk to the porter, "go to a blacksmith shop and get a sledge hammer for this gentleman."

"No, my deah fellah, I don't want a sledgehammer. I want one of those vehicles, you know—a sledge."

<sup>4</sup> In a newspaper article published on February 25, 1882, in the *St. Louis Dispatch*, Wilde is described as looking "like a stout, well-fed, active young Englishman" and his "English accent" was described as "very noticeable". The article was published in a collection of American newspaper articles containing interviews with Oscar Wilde (Hofer & Scharnhorst 2010).

<sup>5</sup>*Texas Siftings* was an illustrated weekly humor magazine, which was first published in Austin, but then later also in New York (by 1884) and in London (by 1887) (Kelsey & Hutchison 2005: 412).

### 4.1 Metadiscourses on phonological forms

"You mean a sleigh. Why, certainly. John, go around to the stables and get a sleigh. Put in a couple of buffaloes."

"Biffaloes? But, my deah fellah, I carn't drive a biffalo, ye know. Cawn't ye let me 'ave an 'orse?"— Texas Siftings

The humor of the anecdote rests mainly on two aspects: first, the misunderstanding caused by the difference between the word used for the vehicle in question, which is designated *sledge* in England and *sleigh* in the United States, and secondly the English visitor's idea of getting around a city like New York City in a sleigh drawn by horses. The clerk responds to this idea with irony, which is designed to be apparent to the American reader, but which is apparently not understood by the English visitor. By asking the porter to get buffaloes from the stables, he plays on the stereotypical English obsession with activities involving horses, suggesting that Americans use buffaloes instead. The fact that the English visitor responds to this seriously makes him subject to ridicule – in the version published in the *St. Paul Daily News*, the headline was changed to "How the English understand us" to emphasize the ignorance of English people who know so little about America that they believe that they use sleighs in cities and that they would use buffaloes to draw them instead of horses. This ignorance is also indexed linguistically by using /h/-dropping and -insertion, a feature which is highlighted in the version of the article cited above through its representation in the heading. That the English visitor is using non-rhotic forms and a back bath-vowel in addition to /h/-dropping and -insertion shows that the linguistic repertoire associated with English visitors has changed over time: English figures appearing in earlier articles do not use these forms (e.g. the Englishman abusing Americans in 1825, the English traveller in 1830, the "overfed John Bull" in the New York railroad car in 1856 and Thomas Brown in the Boston court in 1857), while English visitors in articles appearing towards the end of the century do.

Another example of the "new" English visitor figure is illustrated in a cartoon taken from the magazine *Puck*, an American equivalent to the English magazine *Punch*, and published in the *St. Louis Republic* (Missouri) on May 1, 1892[121] (see Figure 4.8).

Its heading, "Proven", refers to the aim of the two English tourists who are depicted in the two images: They want proof that assumptions circulating in England about Americans are true. The assumption tested here is expressed by the "First English Tourist" who says "Say, me deah fellah, they say an Hamerican halways ahnswers a question by ahsking one. Let's try it and see for ourselves". The "Second English Tourist" therefore suggests "Try the fellah at the windah, me deah boy". The second image depicts the tourist talking to the "Official" at

Figure 4.8: A cartoon depicting two English tourists, published in the *St. Louis Republic* (St. Louis, Missouri) on May 1, 1892, retrieved from America's Historical Newspapers

### 4.1 Metadiscourses on phonological forms

the ticket office asking him "Aw, me good fellah, what time does the next train leave?" to which the "Official" replies "Where to?" While this question might be taken by the Englishmen as a proof that their assumption is correct, it becomes clear to the reader that it cannot be taken as evidence of Americans exhibiting a general tendency of answering a question with a question because the question posed by the English tourist requires clarification, which makes the question by the ticket seller the only sensible reaction. This shows that while the Englishmen fail to prove assumptions about Americans, they do in fact prove assumptions about Englishmen travelling to America, namely that they are arrogant and ignorant. Their ignorance is also indexed by their use of /h/-dropping and -insertion (*Hamerican*, *halways*). As the English visitor figure who "wanted an 'orse", the tourists in this cartoon also use non-rhotic forms (*deah, fellah*, *windah*) and a back vowel in bath (*ahnswers*, *ahsking*), which confirms that these forms have come to index Englishness.

The visual illustration links the form to other non-linguistic signs. First of all, the English tourists are depicted as wearing eyeglasses. The eyeglass has been described as an English item in several articles analyzed above, e.g. in "The Hinglish Hi-Glass" (January 30, 1875[30] ) and in "The Gossip from London" (April 3, 1873[25] ). In these articles, the eyeglass has been linked to members of the upper class, the "English exquisite" and "young gentlemen" in "England's aristocratic halls". In this article and the article "He Wanted an 'Orse", both published in 1892, the eyeglass is worn by an average English visitor, which indicates that the indexical link between the eyeglass and a high social position in society has weakened, while the link between the eyeglass and Englishness has become stronger, now extending to at least the upper-middle-class (English visitors to America must have some financial resources to afford travel). The fact that the English visitors use /h/-dropping and -insertion also indicates that they are not part of England's social elite. The absence of the labiodental realization of /r/ in their speech suggests that this phonological form remains tied to the figure of the upper-class English swell. Further non-linguistic signs marking the English tourists as English are highlighted in the cartoon by depicting the tourists as being dressed completely alike. This makes it clear to the reader that the dress and the accessories chosen by the Englishmen are not based on personal preference, but on English ideas of what is fashionable. The style of the coat and the pants is identical, and a particular focus is put on the pant legs, which are turned up at the bottom (this will play a role in later analyses). The hats have the same shape and pattern, they both wear a small bag strapped across the body, and both tourists hold a cane – it is noticeable that they do not use the cane to get more stability in walking or standing, but that they hold it either in the middle or put it in

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

the coat pocket, which suggests that the canes rather serve a decorative purpose. The cartoon therefore illustrates well how linguistic and non-linguistic signs are combined in creating and transmitting the stereotype of an English visitor to America.

The last article which focuses on the representation of English people that I analyze here is important because it illustrates that the back bath vowel is not only associated with English swells and ignorant English visitors but also with lower class English people. It therefore takes up the form–meaning link that occurred in the very first article containing *dawnce*. It is a travelogue written by Kenneth Lamar and published in four newspapers contained in the databases: in the *Worcester Daily Spy* (Massachusetts) on September 12, 1893[128] , in *The North American* (Pennsylvania) on September 13, 1893[129] , in the *Bismarck Daily Tribune* (North Dakota) on September 14, 1893[130] and in the *Atchison Daily Globe* (Kansas) on September 15, 1893[131] . The author describes the town Escanaba on the Upper Michigan Peninsula, which he labels "a town of aliens" – a label which is used as a heading for the article. The subheadings also indicate the main characteristics of the town: "American Soil, but Largely Controlled by British Capital", "Dragooned by a Motley Crowd—A Halfway Unknown Region of the Union". This indicates why the town attracts the interest of American newspapers: It raises the question of what makes a town American other than its geographical position. The author's view becomes clear when he describes the people who live there as "a jabbering crowd of aliens, with only here and there an American" and a "motley mob of miners, lumbermen, dock wallopers and human drift in general". He describes a particular encounter with a man, who he later explicitly calls an "Englishman", while walking down to the docks:

"Now, listen to that, lads, will ye? Well, he'll dawnce to our music 'fore he's out o' this, unless he let's us 'awve a drink," and he laughed a low, brutal laugh that made me shudder and look round for a possible policeman.

This scene characterizes the Englishman as a dangerous, threatening figure and the impulse of the American to look for the police emphasizes his need for government authority in a place where people have apparently established their own rules, expressed by the Englishman telling him that "he'll dawnce to our music 'fore he's out o' this, unless he let's us 'awve a drink". Linguistically, the two most prominent forms linked to this figure are /h/-dropping and -insertion and not only a back bath vowel but also a back trap vowel (in *'awve* 'have' and *'awving* 'having'). They are foregrounded through the American's inability

<sup>&</sup>quot;'Ello, pard!" cried a great beefy brute, coming toward me. "Hit's a 'ard time we're a-'awvin. Cawn't ye set up an arf an arf?"

<sup>&</sup>quot;What is an arf an arf?" I queried sharply.

### 4.1 Metadiscourses on phonological forms

to understand *arf an arf* – even though it might be the case that he just does not know the lexical item (because he is not depicted as having difficulties understanding the Englishman in general). The differences in pronunciation are highlighted as well because he repeats the item in the question using the same phonological form as the Englishman. This shows that the back bath vowel can be linked to a whole range of English figures – what is important is that they are in opposition to American values and ideals. This is emphasized in this article by connecting the Englishman to a "wild, dismal, half way unknown region of the Union—a region mainly controlled by British capital—a region that has hardly a thing in common with the distinctive march forward of American ideas". The conclusion by the author at the end of the article represents an urgent call for the Americanization of the region and even though it is not stated explicitly it becomes clear that this claim comprises language as well: "It is a duty we owe to our money and muscle to conquer and control it, or a few years hence it will be in utterly foreign hands".

The second result of the analysis is that the search terms *dawnce*, *deah* AND *fellah* and twousers are not only linked to English people, but that they also came to index a particular group of Americans: the Anglo-Americans that have already been subject to analysis in the section on /h/-dropping and -insertion. While Anglo-Americans never insert or drop /h/, they are shown to use a back bath vowel, non-rhotic forms, and a labiodental approximant as a realization of /r/. In the following analysis, I will take a closer look at different types of Anglo-Americans and the historical development of the figures by paying particular attention again to the strategies used to link language and social attributes of speakers.

The earliest articles featuring Anglo-American figures occur in the 1870s. The following short paragraph was published by the *Trenton State Gazette* (New Jersey) on September 12, 1877[35] :

By labeling the man a "Trenton Dundreary", the author takes the English figure of Lord Dundreary and transposes it to an American context, claiming that there are men in America who behave and speak like the English Lord. The figure is portrayed as being as ridiculous as the English one – his childishness (asking his mother for permission, enjoying "ramping", i.e. roaming wildly, in the woods)

A Trenton Dundreary at the Water Gap—The following lament was uttered during a long rain at the Water Gap, by a Trenton Dundreary, who is suspected of painting and wearing corsets: He said, "Nawthing but wain–cawn't have any fun. Maw maw won't let me dawnce, and it is too wet to wamp in the woods, so I cawnt have any fun–cawnt wamp nor nawthin'." ---*Water Gap Observer*

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

and his (suspected) effeminate traits (wearing corsets) emphasize his lack of masculinity. The two linguistic features highlighted here are the back bath vowel and labiodental /r/ (non-rhoticity is not represented).

The back bath vowel is also the most prominent feature in a dialogue that occurred in all sorts of variations and contexts in several articles. It appeared first in the *Lynchburg Virginian* on February 1, 1881[40] , and was embedded in a short paragraph in the *Worcester Daily Spy* (Massachusetts) on February 5, 1881[41] , four days later:

When the society idiot asks, "Do you dawnce the lawncers?" the proper reply is said to be: "No, I don't dawnce the lawncers, but my sister Frawnces dawnces the lawncers and several fawncy dawnces." The management of this sentence assures entrance into the highest circles.

This paragraph shows that back realization of the bath vowel has come to index belonging to high social circles and that its use can even make access to these circles possible. By "management of this sentence" the commentator refers particularly to the pronunciation of the words, which is made obvious through the inclusion of nine words containing the vowel in the short exchange, one of which is not even a member of the bath set (*fancy*). This makes the back realization of the vowel highly salient and ensures that people recognize the link between the vowel and the social position that it indexes. At the same time, however, the paragraph also contributes to the creation of further indexical links: By calling the person who is asking the question a "society idiot", the form is evaluated negatively and associated with idiotic people in higher social circles and therefore questions whether it is even desirable to use the form to be part of that group of 'idiotic' people. In another article, the origin of this dialogue, which is described as very fashionable ("the very latest") is attributed to an "Albany genius", thus creating a link between the vowel and urban centers in the northeast of the United States.<sup>6</sup> How such evaluations are transmitted can be illustrated by an February 5, 1881[42] article published in *The Galveston Daily News* (Texas), which is a letter from a correspondent of the newspaper in Washington who quotes the short paragraph and comments "I mention this for the information of some of our young society folks in Texas, particularly at Tyler and Mineola".

These early articles show that the figure of the English swell has been transferred to American figures with similar characteristics. While the articles illustrate and transmit the high social position indexically linked to a back bath vowel and labiodental /r/, they also criticize and ridicule people making use of

<sup>6</sup>The article appears twice in my data collection: The first one was published on February 7, 1881[43] , in the *Daily Evening Bulletin* (San Francisco, California) and the second one on February 10, 1881[44] , in the *St. Louis Globe-Democrat* (Missouri).

### 4.1 Metadiscourses on phonological forms

these indexical links to achieve or ensure their social standing. This figure of the American swell is developed in much more detail in the following years and it is given a new label: "the dude". In one of my initial exploratory searches, I came across an article which reveals the point in time when the label "dude" was coined and which type of person it designates. Its title is "The Dude" and it was originally published in the *Brooklyn Eagle*, a New York newspaper, and reprinted in the *Milwaukee Sentinel* (Wisconsin) on March 11, 1883[57] :

# **THE DUDE.**

# **A Full Description of the Newly-Discovered Animal.**

A new word has been coined. It is d-u-d-e or d-o-o-d. The spelling does not seem to be distinctly settled yet, but custom will soon regulate it. Just where the word came from nobody knows, but it has sprung into popularity within the last two weeks, so that now everybody is using it. It means a masher and yet it means something more than a masher. For instance, a masher may be young or old, or he may mash by virtue of his politeness, of his accomplishments, of his wealth, beauty, eyes, nose or fame; he may be a man of mature years, an old man, a young man or a boy. In speaking of mashers, one is never sure exactly what sort of a man is meant. There is a class of mashers in New York who will now have a definite place in the language of the town as dudes. A dude cannot be old; he must be young, and to be properly termed a dude he should be of a certain class who affect the Metropolitan theaters. The dude is from 19 to 28 years of age, wears trousers of extreme tightness, is hollow-chested, effeminate in his ways, apes the English and distinguishes himself among his fellow men as a lover of actresses. […] The word dude is a valuable addition to the slang of the day.

The author of the article highlights many important characteristics of the dude: his age (young adult), the place where he lives (New York), his body (hollowchested), his manner (effeminate), his behavior (an imitation of the English) and his clothing style (tight pants). His interest in theater and actresses as well as the description as hollow-chested also hints at a lack of sincerity and of substance – he is characterized as being more interested in a show than in real people and real life.

The article above does not link the dude figure to linguistic forms, but the articles containing the search terms *deah* AND *fellah*, *dawnce* and twousers show that these terms and the phonological forms represented by the spellings are highly salient indexes of the American dude. An early example linking these forms to the dude is the following anecdote, published in the *Rocky Mountain News* (Denver, Colorado) on September 17, 1883[60] , that is, half a year after the description of the dude in the article above.

# **FOUND AT LAST.**

# **He Makes His Appearance and Creates a Lasting Impression.**

Yesterday a dude walked to the Burlington ticket office. On his closely cropped, bullet head was perched a mammoth white beaver, the wide brim of which curled up like a scoop, and formed a roof for the protection of a large pair of pigeon-toed ears. The tail of his delicate coat flirted around his suspender buttons, and entirely failed to cover the southern exposure of his tight-fitting pants, which were so small in the legs that they looked like two umbrella covers. Below the bottoms of the pants, which clasped the ankles like a Langtry glove, were a pair of feet which the Smithsonian Institute has been trying to find for a long time. They simply stood straight out from the rest of the dude as if they belonged to another family. While ticket clerks were wondering how the dude managed to slip his feet through his pants' legs without using a shoe-horn, he leaned both elbows on the counter, and, with a gentle smile, said: "Cawn't you sell me a sleeping caw ticket to Kansaw Citaw?" Yes." responded the ticket clerk. "Aw! well do so, me deah fellaw, please. Aw! but I forgot to ask the price, you knaw." He was told the price of the ticket, and handing out the money he began drumming on the counter with his fingers and singing:

> We nevah speak as we pass by, Although a teah drop–

"Aw, but I say, are you quite suah the sleepaw will go right though to Kansaw-Citaw?"

"Yes, it goes right there without fail."

"Well, you knaw, that's what I look out for when I take a sleepaw. I hate a beastly change, you knaw–

We nevah speak as–

"Ah, my deah fellah, tell me, there is no danger of a lay ovah?"

[He continues to ask questions and to sing, until the clerk becomes very annoyed. He then announces that he has to leave to get his things to the car.]

As he glided down the street the people stopped, and looked, and wondered if they could buy the thing to play with.

The incident narrated in this anecdote is that of 'normal', 'average' Americans encountering the dude for the first time. The setting is Burlington, which is probably Burlington, Colorado, as the newspaper is a Colorado one. The description of the dude in the beginning is similar to the one in the article before but much more humorous because it is full of similes and hyperboles (e.g. the pants looking like two umbrella covers). The conversation between the dude and the ticket clerk establishes a contrast between these figures. The dude is described as being annoying by asking silly questions and singing and drumming his fingers on the counter, whereas the ticket clerk is depicted as being calm and reasonable until the very end, when he reaches for a paperweight to throw at the dude. The dude is portrayed as an exotic figure which stands out in the normal, everyday life of a western city. The ticket clerk, on the other hand, represents a man of an average profession; he seems respectable, but in no way remarkable. The last sentence emphasizes not only the otherness of the dude, but also the fact that he is not taken seriously: He is perceived as more of a thing than a human and as something that people would play with (like a toy or a puppet) rather than interact with on an equal level. Linguistically, the form that stands out most in the dude's repertoire and distinguishes it from the speech of the clerk is non-rhoticity. Even though it is not represented in every instance (e.g. *forgot*, *for*, *there*, *danger*), a spelling using <ah> or <aw> to indicate a non-rhotic form is found in nearly every sentence (e.g. *caw*, *nevah*, *suah*, *teah*, *ovah*). A back vowel is represented not only in

### 4.1 Metadiscourses on phonological forms

bath words (*cawn't*), but also in several instances where a back vowel would not be expected (e.g. *Kansaw Citaw*, *sleepaw*, *fellaw*). In addition, the goat vowel is also represented by <aw> in *knaw* ('know') which indicates a mid to mid-close monophthong, but in initial or medial position the vowel is not represented as being different. This creates the impression that the dude makes use of long back vowels rather inappropriately and too frequently in general, which signals hypercorrection and, connected to that, linguistic insecurity. Next to phonological forms, the phrases *you know* and *I say* also mark the dude's speech as different from that of the ticket clerk.

That the speech of the dude is a deliberate imitation of English norms is explicitly described in an article written by George Salisbury for *The Epoch* and reprinted in the *Atchison Daily Champion* (Kansas) on October 8, 1887[82] .

AS AN ENGLISHMAN. THE EFFORT YOUNG AMERICA MAKES TO POSE AS A BRITON. A Long and Tedious Process of Preparation–The "English" Method of Speech, Oddities of Dress–A Distinctive Walk. The Eyeglass.

There is a large number of young men in these free states whose chief object in life is to be taken for Englishmen.

The youth who wants to pass as an Englishman is obliged to put himself through a long and tedious process of preparation. He usually commences with a study of the "English" method of speech. The first task is to learn how to talk "away down the chest," and the phrase chosen to experiment upon is, invariably, "By Jove." When he can say this with the proper accent he next ventures upon "You don't say so?" He then passes on to such sentences as "How awfully jolly. I cawn't believe it, you know" and so on.

[...]

If you live in the same house with him you can hear him up to a late hour of the night repeating over and over such words as "dawnce," "cawn't," "pawth," "chawnce," "rathaw," "fathaw" and "aw."

[...]

This article shows that rather than emphasizing the use of a low back vowel in a specific set of words, the dude's speech is generally characterized as "to talk 'away down the chest'", which is a description of the impression that the use of the vowel creates for listeners rather than an accurate linguistic analysis by experts. This impression is foregrounded here, but non-rhoticity is also represented (in *fathaw* and *rathaw*) and connected to it (not least through the same spelling pronunciation). Again, a particular emphasis is also put on discourse markers (*by Jove*, *you know*) and lexical items (*awfully*, *jolly*). By focusing specifically on the process of acquiring the accent, the author stresses the unnaturalness of the pronunciation, which can only be attained through tedious practice and comes at the price of not sleeping at night. This alone makes it clear for the reader that it is not a desirable task, but the main reason why the phonological forms are

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

evaluated negatively is that they are marked as English and not as American. What is constructed as being natural for an Englishman is at the same time constructed as not natural and impossible to acquire for an American and the article points out that young Americans who attempt to imitate English pronunciation make themselves subject to ridicule and derision. By describing a particular order which is followed in learning "the English method of speech", the author suggests that the result is a uniform way of talking which comes at the expense of giving up the possibility of expressing one's individuality.

In the late 1880s and 1890s, the stereotype of the American dude is strengthened continually through the appearance of this characterological figure in many humorous articles – comic dialogues, jokes, cartoons and anecdotes. I will give several examples here to show which traits of the dude were highlighted in these articles, which figures he was contrasted with and to what extent the figure was also subject to change.

The first article is an anecdote published in the *Milwaukee Daily Journal* (Wisconsin) on June 29, 1889[92] . Its title and subtitle introduce the main characters as well as their social relationship and they also summarize the action: "Stunned. How a Montana Girl Paralyzed a New York Dude". It consists only of direct speech, starting with that of a hostess of a social event in Boston, who introduces the New York dude, Mr. Chester de Montague, to the girl from Montana, Miss Sharpe. The remaining part of the anecdote consists of a conversation between these two characters: He is interested in her and asks her to dance with him, while she has trouble understanding him, regards his behavior as stupid and idiotic and rejects and insults him:

```
"Do you dawnce?"
"Do I what?"
"Dawnce – dawnce."
"I – I – oh, do I dance? Is that it?"
"Aw, yaas, yaas – to be sure."
"Well, I dance sometimes, but I'm not going to dance any more tonight."
"Oh, weally, weally, me deah Miss Sharpe, you are too, too cruel. Mah I not have the honah of just
one waltz – ah?"
"No, not to-night."
"Naw, now, weally? Naw? You are vewy, vewy unkind – you weally are!"
"I guess you'll live through it."
"Beg pahdon?"
"What?"
```

```
"Oh – aw – I merely begged your pahdon."
```
"What for?"

"Aw, weally, Miss Sharpe, you oughtn't to guy a fellah so, you weally oughtn't. Naw, be Jove; hanged if you ought."

"Look here, young man, ain't you a little off? And don't this State make any provision for its idiots? And if it don't, you come out to Montana and you'll be taken care of without any expense to your friends. Good-bye!"

By contrasting the dude with a girl in this way, several attributes and character traits are highlighted. First of all, there is the gender difference: The dude is the male figure who tries to engage in a courting scenario with a female figure by following social conventions requiring the man to take the initiative and ask the girl to dance with him. In the course of the conversation it becomes clear, however, that the character traits typically associated with male and female figures are reversed in this case. The girl is very direct in her refusal to dance and appears tough, bold, practical and sane. The dude, on the other hand, is portrayed as indirect, hesitant, insecure and whiny. The anecdote therefore illustrates the effeminate behavior of the dude that has been explicitly mentioned in earlier articles and that is even more prominent through his juxtaposition to a girl (not a woman) who is more manly in her behavior than he is.

Secondly, the character traits are linked to different places. The dude comes from New York, a northeastern city, while the girl comes from Montana, a northwestern rural state. Their manner and behavior is therefore constructed as representative of people living in these places. The Montana girl is not interested much in social conventions, but appears natural, authentic and true to herself, while the dude is presented as affected and unnatural. This culminates in her assumption that he might be "a little off" and her practical suggestion to take him to Montana to become sane again.

Thirdly, the two figures are contrasted on a linguistic level. Their inability to understand each other puts an emphasis on the back vowel in *dance* and the nonrhotic pronunciation of *pardon* (which present difficulties to the Montana girl) and on the expression *I guess you'll live through it* (whose meaning is not clear to the dude). In general, the Montana girl's use of *ain't* and the absence of thirdperson singular marking in *it don't* signal a lack of formal education, but her name (Miss Sharpe) indicates that she is intelligent and witty nevertheless. The use of these forms also underlines that she does not change her speech to appear educated, but that she remains authentic and down-to-earth. It is in fact the dude who, despite his education and high social position, is presented as inferior to the girl because of his affected pronunciation (the labiodental realization of /r/ is also indicated by spelling several times) and the discourse markers that he uses (*by Jove*, *hanged if you oughtn't*), which make his speech appear incoherent and

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

meaningless. The formulaic nature of his speech is reminiscent of the speech of the English diplomat represented in the article "Note and Comment" (September 18, 1883[61] ) and in both cases it is ridiculed because of its emphasis on empty form over function (conveying meaning).

The anecdote illustrates very well how different types of performable signs are not only linked to different places, social personae and social relationships, but also to stereotypic pragmatic effects: The dude is stunned and paralyzed by the girl and therefore unsuccessful in his courtship attempts. Taking into consideration the representative nature of the anecdote, this specific effect hints at more general effects, suggesting that the manners and behavior of young men in eastern cities will lead to immobility and stand in the way of growth and progress, while the activity and practicality of westerners will advance the country's development. This shows that the social meanings indexed by the linguistic forms are embedded in broader discourses creating stereotypical differences between easterners and westerners and I will discuss three articles which also draw on these differences and serve to reinforce them.

The first one is a cartoon which contrasts the dude with a farmer and therefore emphasizes the opposition between city and country life and people (see Figure 4.9). It originally appeared in the magazine *Truth* and was published in two newspapers on September 16, 1894 (in Nebraska and in Massachusetts).<sup>7</sup> It depicts the two figures standing across from each other on a field next to a farm, divided by a wooden fence. The dude is leaning on the fence because, as the caption explains, he wants to get over the fence, but he is worried that his "twowsahs" might get bagged at the knees in the process. The farmer on the other side stands a few feet away from the fence and suggests that the dude should simply take his pants off. The visual illustration of the two figures highlights their differences and illustrates their relationship in several ways. First, there is an obvious age difference, with the dude being much younger than the farmer. The farmer's long beard, which is contrasted with the dude's shaven face, underlines this difference in age and also in experience. Secondly, their style of clothing is very different: The dude is dressed elegantly, but also uncomfortably (e.g. the high and stiff collar) and the clothes are not suitable in a rural farming context. The farmer, on the other hand, is depicted as wearing clothes which are not fashionable, but comfortable and suitable for working on a farm. Thirdly, the difference is underlined by the accessories they bring to the scene: The farmer holds a rake in his hand, a useful tool for his work, while the dude holds a cane and wears an

<sup>7</sup>The dialogue without the cartoon appeared additionally in five newspapers in Colorado, Nebraska, Oregon and New Mexico.

4.1 Metadiscourses on phonological forms

Figure 4.9: A cartoon depicting a dude and a farmer, published in the *Omaha World Herald* (Omaha, Nebraska) on September 16, 1894[145] , retrieved from America's Historical Newspapers

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

eyeglass, two items which, in this context, serve decorative purposes only. The cane and the eyeglass also serve to mark the figure as a dude because they are linked to English urban fashion, which is illustrated for example in the cartoon "Proven" published a year earlier, which depicts two English tourists wearing eyeglasses and holding canes in the same manner as the dude in this cartoon. Another important visual element is the fence between the two figures, a physical barrier that alludes to the social barrier that also exists between them. This impression of a barrier between the figures is reinforced by the farmer holding the rake upright in front of him, as if to make sure that the dude, who is leaning towards him, would not come near him. The textual elements add to the visual elements by providing a heading, labels for the figures, explicit descriptions of the emotional states of the figures and a dialogue. The heading "A City Boarder" marks the dude as a person living in a city and coming to the country temporarily as a boarder. His emotional state ("angrily") and his question ("How the deuce can I get over this blamed fence without bagging me twowsahs at the knees?") reveal not only his contempt for and impatience with the life in the country, but also his concern with dressing and speaking fashionably. The farmer's answer reflects his emotional state: He "laconically" replies: "Take 'em off". This reply in combination with his relaxed posture indicates that he is neither intimidated nor impressed by the dude, but down-to-earth and used to finding practical solutions to problems. All in all, the dude seems entirely out of place in the country and his extreme concern with fashion, his uptightness and his affected and impractical manner are presented as ridiculous in comparison to the concerns of a hard-working American farmer. Finally, his inability to get over the fence can also be interpreted as an inability to overcome the social barrier dividing him and the farmer and consequently his inability to acquire characteristics that are regarded highly in (especially rural) America.

A cartoon which draws on and contributes to the same stereotypical differences is headed "Trees and Trees" (see Figure 4.10) and was published in the *Philadelphia Inquirer* (Pennsylvania) on September 19, 1897[166] and in *The Wheeling Register* (West Virginia) on September 22, 1897[167] . It juxtaposes the dude with the figure of the American cowboy. Even though the label *dude* is not used here, the characterological figure is clearly recognizable due to his clothing and his language: The eyeglass, the cane, the hat and the pants which are extremely tight at the bottom suggest again a concern for English fashion. The cigarette in his mouth is also a characteristic feature of the dude which is mentioned in other articles. He is given the name "Tender Foot" and is correspondingly depicted as a small and weak figure that has to look up to the big and tough cowboy on his horse. The pun of the dialogue rests on the polysemy of the word *tree*. While the

### 4.1 Metadiscourses on phonological forms

Figure 4.10: A cartoon depicting a dude and a cowboy, published in the *Philadelphia Inquirer* (Philadelphia, Pennsylvania) on September 19, 1897[166] , retrieved from America's Historical Newspapers

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

dude tries to impress the cowboy with his aristocratic family tree, the cowboy counters it with a more literal meaning of *family tree*: "Well, by thunder there's our family tree. Father and my grandad were both hung on it". This illustrates a difference in values that are regarded as important: on the one hand, fashion, refinement and a system of power and social hierarchy based on ancestry (as in aristocratic England), and on the other hand masculinity, toughness, and a system of power based on sheer physical characteristics. Linguistic differences are linked to these differences in values: The dude's imitation of English values is also reflected in his speech through the use of non-rhoticity, possessive *me* and the discourse marker *don che know* ('don't you know'). The cowboy, on the other hand, uses the discourse marker *by thunder*, which creates a contrast to the marker *by Jove*, which is attributed to the dude's linguistic repertoire in other articles. This marks the cowboy as calling on forces of nature rather than religion and links the figures to the extreme positions of culture versus nature.

How region is linked to the figure of the dude and contrasting figures, such as the cowboy, becomes clear in a cartoon which was published in the *Grand Forks Daily Herald* (North Dakota), on April 17, 1898[168] (see Figure 4.11).<sup>8</sup> The figure shown is a man who appears to be strong, stout, sturdy and determined. His clothes and shoes are simple, but practical and even though they are worn out, they seem comfortable as well. His hat has a wide brim, which, together with his beard, protects the man from the harsh conditions of outdoor labor. He is depicted as being on the move with bags, boxes and a sleeping mat. The labels on these items show that he is from Dunkardville, Indiana, and is headed for North Dakota. They also serve to characterize the man explicitly as having a "good character", "good habits" and as being hard-working and thrifty so that he has accumulated "hard earned savings". The headline explicitly puts the figure in opposition to the dude: "He is No Dude". The caption under the cartoon reads "But he is a good citizen, and he is welcomed to North Dakota—thousands of him". The word *but* indicates that the creator of the cartoon recognizes that some of the values associated with the dude are indeed positive – perhaps refinement, being part of higher social circles, being rich and having an interest in art and fashion. However, the cartoon points out the negative characteristics of the dude and portrays the opposing values as desirable, especially in North Dakota, a very rural state, and uses this positive characterization of rural workers to attract men to move there.

<sup>8</sup> I found this article in a separate search, which aimed at finding cartoons or other visual illustrations of the dude. Even though it does not contain any of the search terms and is therefore not part of metadiscursive activities, it nevertheless illustrates that discourses on language are inextricably tied to other discourses – in this case to the figure of the dude and its opposite.

### 4.1 Metadiscourses on phonological forms

Figure 4.11: Cartoon "He is No Dude", published in the *Grand Forks*

*Daily Herald* (Grand Forks, North Dakota) on April 17, 1898[168] , retrieved from Nineteenth-Century U.S. Newspapers

Although the contrast between east and west is very often tied to the contrast between urban and rural life, there is a comic dialogue in my collection of articles which illustrates that the differences between east and west extend to cities as well, in this case to New York and Chicago. It was published in the *Atchison Daily Globe* on July 1, 1889[93] .

# **Only an Overgrown Village.**

In the window of a New York club:

Doolittle—How long have you been heah, Idlewild?

Idlewild—'Bout three houahs. And you?

Doolittle—I was heah an houah befaw you came in. Lots of girls pawsing. By the way, the fellaws tell me you have been to Chicago lately.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Idelwild—Ya-as, went out theah to look aftah some pwoperty left to me by my uncle. Queah place, Chicago. Some fellahs work out theah.

Doolittle—Naw.

Idlewild—Fact, deah chappie, weally. And you'll nevah believe me when I tell you that all the time I was theah I nevah saw a solitarwy fellah sit in the club window and watch the girls foh moah than a half houah.

Doolittle—Ah, well; Chicago is only an overgrown village, after all. —America.

Doolittle and Idlewild are both telling names which link the dude to the character trait that is highlighted in this comic dialogue: laziness. The dudes are portrayed as spending their days sitting in clubs at the window, watching girls and finding this activity not only normal but typical of urban life, which leads them to conclude that Chicago can only be a village and a "queah place" because men actually "work out theah". By ridiculing this attitude, the comic dialogue conveys a very positive evaluation of hard work, which connects the rural and the urban parts of the west and differentiates them from eastern cities like New York where idleness stands in the way of progress. This connects the article to the cartoon "He is no Dude", which also places an emphasis on the fact that the savings by the rural worker are "hard earned", which implies that he has worked for his money while the dude's wealth is based on inheritance, as indicated in this dialogue when Idlewild tells Doolittle that the property he owns in Chicago was "left to me by my uncle".

The laziness of the dude is also the source of humor in other comic dialogues and cartoons. The following article (see Figure 4.12) was taken from the magazine *Life* and published on March 7, 1888[86] , in the *Atchison Daily Globe* (Kansas). The two dudes are named Gus and Cholly, which are shortened and diminutive forms of the traditional names Augustus and Charles. These names were often used by emperors and kings throughout Europe so that they may evoke associations with royalty and a superior social position. However, using the nicknames instead of the full names rather creates links to familiarity, equality, closeness and, in the case of the diminutive form, even to inferiority and infantility.<sup>9</sup> Gus inviting Cholly to have a "glass of sodah" alludes to the dudes' favorite activity: sitting in a club and having a non-alcoholic drink. Cholly's reply is therefore surprising because the dude normally does not have important business to attend to. The humor rests on the description of the business that Cholly deems important:

<sup>9</sup>Even though especially the name Charles was and still is a popular name that could be seen as ordinary and unremarkable, the choice of names can also be interpreted as symbolizing the contradiction inherent in the dude figure: He aspires to greatness but usually achieves the opposite.

### 4.1 Metadiscourses on phonological forms

Figure 4.12: A cartoon depicting two dudes, published in the *Atchison Daily Globe* (Atchison, Kansas) on March 7, 1888[86] , retrieved from Nineteenth-Century U.S. Newspapers

the purchase of stamps and a new pair of pants. That he buys the stamps for his mother makes him seem like a little boy and reinforces associations with weakness, dependence and unmanliness evoked by his name. That he needs to buy a new pair of pants alludes to his preoccupation with fashion and outer appearance, a characteristic illustrated by the visual elements of the cartoon. The two dudes look almost completely alike: They wear the same hats, they hold a cane in exactly the same manner and they wear the same type of shirt (with a high collar), overcoat and pants. The pants fashion seems to have changed: In earlier articles the pants have been described as extremely tight (like umbrella covers), while they are now depicted as rather wide. The overall message of the cartoon becomes clear through the heading "It's Worry That Kills", an ironic comment which makes the dude's worries appear ridiculous.

The dude's preoccupation with dressing fashionably is also ridiculed in a comic dialogue published in the *Emporia Daily Gazette* (Kansas) on January 31, 1893[124] , and which also highlights another characteristic of the dude: his unattractiveness to women.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

# **Not Quite Desperate.**

Cholly (disconsolately)—Yaas, she wefused me, and she lawfed at me, too. If it wasn't foh one thing, I'd drown myself.

Friend—You still hope?

Cholly—No, but the watah would take the cweases out of my twousers, you know.

Even though the dude seems desperate after having been refused by a woman, his fear of not being dressed according to the latest (English) fashion, which in this case are creases at the front of the pants, keeps him from killing himself. By presenting the interest in fashion as a matter of life and death, the author of the dialogue creates the impression of absurdity, suggesting that the dude is stupid and cannot be taken seriously.

The stupidity of the dude is also targeted in a cartoon taken from the magazine *Judge* and published in the *Philadelphia Inquirer* (Pennsylvania) on July 9, 1893[127] (see Figure 4.13). The figures are immediately recognizable as dudes based on their dress and the canes they carry. The dude on the right is wearing clothes with a check pattern, which is described as being fashionable in England in other articles. The dude on the left is depicted as sucking his cane and this behavior in combination with the heading and the dialogue indicate that the title "lack of space" refers to the dude's lack of brain capacity.

Figure 4.13: A cartoon depicting two dudes, one of them sucking his cane, published in the *Philadelphia Inquirer* (Philadelphia, Pennsylvania) on July 9, 1893[127] , retrieved from America's Historical Newspapers

### 4.1 Metadiscourses on phonological forms

A characteristic which is absolutely crucial for the argument developed in the present study is the dude's lack of authenticity and Americanness that is constructed in several articles, including the following comic dialogues. The first one was taken from the magazine *Harper's Bazar* and published in the *Milwaukee Sentinel* (Wisconsin) on June 6, 1899[172] .

# **Smuggled.**

Harper's Bazar: Chappie had just returned from a visit to England.

"Now, my dear boy," said his friend who met him on the pier, "keep your mouth shut. Don't say a word to the custom house people."

"Fawncy, now!" said Chappie. "And why, me deah fellah?"

"Because they'll make you pay duty on that new English accent of yours."

"Quite so!" said Chappie.

And he smuggled it in.

The article explicitly discusses the dude's linguistic behavior and marks his accent as English and therefore not American. The use of the smuggling metaphor evokes associations with illegal activities and sneakiness, which contributes to the negative evaluation of the accent. It also emphasizes that the accent is something foreign that has not been granted official acceptance in America.

The second example is an article which was taken from the *Chicago News* and published in *The Galveston Daily News* (Texas) on November 23, 1884[66] . Even though the figures speaking are not explicitly named, it becomes clear that an American dude is speaking to an Englishman. The irony of the dialogue is created by the dude denying the Englishman his English identity and authentic English behavior, while it becomes clear that he is in fact the person lacking authenticity and struggling with constructing a genuine identity. By adopting "English ways", the dude becomes un-American; in addition, even these supposedly English ways are explicitly marked as not genuinely English, which makes the dude appear even less authentic. His attempts to be English fail as completely as his attempts to be accepted by American society.

# **He Was the Genuine Imported Article.** [Chicago News]

"Aw, my dear fellah, I notice you cawwy youah cane by the handle."

"Yes, sir; that is what the handle is made for."

"But, you knaw, that is not the pwopah capah; it's not English, you knaw."

"I don't care."

"But tell me, deah fellah, why you do not assume English ways as the west of us do? You are so deuced odd, ye knaw."

"I don't know, unless it is that I am English and have lived in England all my life."

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

The third example is a comic dialogue which was taken from the *Washington Star* and published in two newspapers contained in the databases: first in the *Daily Inter Ocean* (Chicago, Illinois) on January 25, 1896[156] , and secondly in the *Morning Oregonian* (Portland, Oregon) on February 10, 1896[157] . It addresses the question of what makes a person a true and authentic American in a humorous way:

"Chawles is what I call a twue patwiot," remarked Willie Wibbles. "He's Amerwican to the back bone."

"He keeps his twousahs tuhned down now, whethah it is waining in London or not."—Washington Star.

The performable sign that is the main subject of the anecdote is the way that the bottom parts of the pants are worn. As I have shown above, the cartoon "Proven" (1892), which depicts two English tourists, links turned up pant legs to stereotypical Englishmen. The dialogue here makes use of this link by establishing that an American who turns up the bottom part of his pant legs is un-American and unpatriotic, while an American who keeps them turned down is an American "to the back bone". The addition that Charles now wears them turned down "whethah it is waining in London or not" ridicules the imitation of English fashion on the grounds that it appears completely unreasonable and stupid to dress according to the weather in another country. By implication, speaking according to the fashion in another country is ridiculed as well and it marks the speaker, in this case Willie Wibbles, who is shown to use non-rhotic forms and labiodental realizations of /r/, as un-American and unpatriotic.

The question of what motivates the dude to imitate English manners and speech is also addressed humorously in the articles, for example in the following comic dialogue, published in the *Milwaukee Journal* (Wisconsin) on October 12, 1894[146] .

# **The Social Test.**

Algernon (employed in extracting nourishment from his cane)—I say, Chawles, me boy, why does a fellah have to suck his cane, don't you know? Why cawn't a fellah do without doing it, don't you know?

Charles (similarly employed)—Deah me! cawn't you see, me boy, why it is? It's the social test, don't you know!

Algernon—Cawn't any fellah do what he likes with his cane?

Charles—P'whaps so, Algy, but they don't count in society, me boy. It's only intellect, culture and wefinement that count, don't you know.

Algernon—Haw! why so, Chawles?

<sup>&</sup>quot;How do you know?"

### 4.1 Metadiscourses on phonological forms

Charles—Pon me life, Algy, cawn't you see, me deah boy! Society won't tolerate those who haven't bwains enough to suck their canes, don't you know?

Algernon—Haw!—*Truth*.

Sucking the cane is presented here as a behavior that is regarded by the dude as indexical of "intellect, culture and wefinement", qualities which are needed to be accepted as a member of high society. The dialogue illustrates that by imitating English speech (by using non-rhotic forms, a back bath vowel, labiodental /r/ and phrases like *don't you know*) and English manners (carrying a cane), American dudes seek to index these qualities, hoping to achieve a good position in society. At the same time, it ridicules this indexical relationship by showing that the dude not only regards wearing a cane as a "social test", but also the activity of sucking it. Sucking the cane is reminiscient of childish behavior (e.g. children sucking their thumb) and as such in complete opposition to refined behavior (of adults). That the dude does not understand this marks him as stupid: The irony created by Charles' statement "Society won't tolerate those who haven't bwains enough to suck their canes, don't you know?" makes it clear to the readers that it is in fact the dude who lacks intellect and does not understand that the very activity of sucking the cane shows that he lacks the qualities of being part of upper-class society.

The following article sheds more light on the dude's motivation and his position in society. It is a comic dialogue, taken from the magazine *Truth* and published in *The Duluth News Tribune* (Minnesota) on April 18, 1896[160] .

# **HE GAVE UP.**

Why Chappie Was Baffled in His Efforts to be English.

Chappie—"Aw, there, deah chappie; I hardly expected to find you at the club today. What's up?"

Algie—"Everything. I've given up. That's what's the matter."

Chappie—"Given up? Good gwacious, deah boy, you don't mean to say that you're going to quit us?"

Algie—"That's just it."

Chappie—"Why, you've been the greatest monochromic-maniac of us all. What will we do for a leader without the white plume of Navarre and all that sort of thing we used to hear about at college?"

Algie—"Can't help it; I'm done for, old fellah."

Chappie—"Why, what do you mean?"

Algie—"Why, just this. Haven't I bought all my clothes in London?"

Chappie—"Yes; that's English, you know."

Algie—"And not paid for them?"

Chappie—"Yes; that's English y'know."

Algie—"And turned up my trousers and played golf and yelled for the Valkyrie III and the Cambridge athletes and all that sort of thing?"

Chappie—"Yes, that was correct English, y'know."

Algie—"Well, just at the end I have come to the limit of my resources."

Chappie—"Aw, you don't mean it, deah boy?"

Algie—"I do. I have discovered that I cannot marry a daughter of the Vanderbilts."

Chappie—"Poor boy!"

Algie—"Yes—and I've got to remain poor. That's just what's the matter. —Tom Hall in *Truth*.

The dialogue shows clearly that "Algie" is a poor young man who aims at social advancement: By marrying the daughter of the Vanderbilts he hopes to gain access to the highest social circles in New York City. All his "efforts to be English" are part of his strategy to become accepted by high society (and ultimately chosen by the daughter of the Vanderbilts as a husband). That he is not alone in this behavior, but that the imitation of English ways is constitutive of the group of dudes is emphasized by Chappie asking him whether he is "going to quit us". This reveals a development of the figure of the dude. Initially, the group of dudes comprised young men who were already part of the wealthy upper class, as I have shown in the analyses of articles above in which they are described as being interested in theaters and actresses, as not having to work because they inherited money and property (Doolittle and Idlewild) and as attending important social events like the New York dude meeting the Montana girl at a ball in Boston. Increasingly, the figure of the dude is extended to young men who are not part of these circles, but for whom the imitation of English speech and fashion becomes an instrument for social advancement. The negative evaluation of this behavior is shown in this article – not only is the goal of marrying the daughter of the Vanderbilts a completely unrealistic one for a poor man in the first place, but his realization that all his efforts were futile and his decision to give up trying to be English and quit the group of dudes show that social advancement through the imitation of English manners and speech is not possible. Algie's ability to simply "give it up" is demonstrated in the dialogue through the absence of any linguistic forms indexing Englishness in his speech: He does not use non-rhotic forms, no back vowel in bath and no labiodental realizations of /r/ anymore. It also emphasizes again that the dude's behavior and speech are inauthentic and affected. All in all, the dialogue therefore functions as an appeal to young American men to be authentic and achieve their goals by other means.

The negative effects of the "Anglomania" have also been described for the upper-class dudes. The following article combines an explicit description and

### 4.1 Metadiscourses on phonological forms

comment on the development of the "Anglomania in New York" with an anecdote about a "young man about town of the name of Lamm" who talks to a friend about a "loan of twenty pounds" and how they draw money from the bank. It was originally published in the *New York Tribune* only three years after the dude figure had been described first. It was reprinted in the *Wheeling Register* (Virginia) on December 12, 1886[74] .

# ON THE DECREASE. ANGLOMANIA IN NEW YORK SAID TO BE ON THE WANE.

# One of the Most Curious Examples of Unrestorable Anglomaniac–A Unique Case–Making a Loan of "Twenty Pounds."

Anglomania in New York is on the wane. The influence of the more American clubs, such as the Union League, the Lotos and the University, has been directed against it so steadily and with such effect that its manifestation is now mercilessly guyed. Nothing cures a weakness so quickly as ridicule. Among wealthy young club men, however, some advanced and hopeless cases remain. They have been proof against all the shafts of wit, against the ostracism of sensible men, against the contempt of women, against all the influences usually successful in restoring a mental equilibrium which some absurd charge of society has disturbed.

Among the most curious examples of the unrestorable Anglomaniac is a young man about town of the name of Lamm. At least, if his name is not Lamm, it may as well be for all present purposes. He is wealthy, a college graduate, and a fellow of real intelligence. If his mind had not been wrecked by this unhappy mania, he would be an agreeable companion socially and perhaps useful in an honorable profession. But he is completely given over to Anglicisms.

# [...]

This beginning of the article shows that the main metaphor used to discuss the imitation of English manners and fashion is illness. The admiration of "Anglicisms" is described as affecting the mental health of individuals and spreading uncontrollably so that quite some effort is required to contain and to cure it. The effects of the "illness" are first of all detrimental to the individual's well-being: Even though the young men affected by the "illness" may possess many desirable characteristics (intelligence, high education, wealth), they become irrational, wrecked and weak and consequently subject to ridicule and contempt, which ultimately leads to their social isolation. In addition, the effects are detrimental to society because the equilibrium is disturbed and needs to be restored. This illness metaphor creates the impression that the individual is not in control and that a common effort by society as a whole is required to cure it to "restore" the individual to his original, healthy state, and the society to an equilibrium. That the anecdote serves the purpose of illustrating a representative case of an "unrestorable Anglomaniac" is made explicit when the narrator writes "At least, if his name is not Lamm, it may well be for the present purposes". Even though the fictitiousness of the person and of the ensuing dialogue between him and his

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

friend is made transparent, the message created through the anecdote adds to the efforts described explicitly before: to cure the Anglomania through ridicule.

The last article relating to the dude figure which I analyze here is interesting because it draws on the overwhelmingly negative stereotypical traits of the dude to ultimately subvert them and highlight positive characteristics of the figure. It was taken from the *Omaha Bee* (Nebraska) and was reprinted in the *St. Louis Globe-Democrat* on August 29, 1884[65] .

# **A MUSCULAR DUDE.**

# **He Wouldn't Drink Whisky but Mopped Up the Floor with a Cowboy.**

"What's that?"

The question came from a long-haired, big-hatted, leather-coated, wild-eyed specimen of Montana cowboy, who carried a belt full of revolvers and an odor of bovine impurity about him. He stood up before the bar in a Broadway saloon the other day, and as he spoke tipped his head in the direction of a pale-faced, hollow-eyed chap, who was leaning against the other end of the bar, tapping the pointed toe of his patent-leather shoe carelessly with a light-complexioned rattan cone [sic].

"That's a dude," replied the bartender.

**"A dude,"** repeated the cowboy, while a peculiar grin began to dispense itself over his sunburned features; "**a dude, is it? I've heard o' them things, but I never got my lookers on one o' them before; it's a dandy, ain't it? Make a good toothpick, wouldn't it? Jest size ther critter up, will yer? Look at them ar legs o' his; I'm a slum gullion ef I don't spit terbacker juice all over 'em lookin' glass shoes.**" With this the cowboy expectorated a deluge of nicotine humidity in the direction of the dude's pedals that made them look as if they had just been pulled out of the manhole of a sewer. The dude slowly changed positions and lighted a cigarette, which he carelessly puffed without looking in the direction of the cowboy.

[The cowboy tries to talk the dude into drinking whisky with him and when the dude politely declines, the cowboy insults the dude and attacks him. Surprisingly, the dude fights back.]

The collision lasted about three minutes, when the remnants of the cowboy were jammed down into a corner in an unconscious state. Readjusting his cravat, the dude laid a chunk of chewing gum down on the bar and said: **"Now, deah fellah, give me a glass of seltzah with a drop of lemon in it. That is weally the hardest work I've done since I played first base in the Yale nine."**

# [emphasis mine]

As in anecdotes and cartoons analyzed above, the dude is contrasted with a cowboy figure. The setting is a Broadway saloon in New York City, that is, an urban place in the northeast. The cowboy is from Montana, a prototypically rural and northwestern state. He is obviously a visitor because he encounters a dude for the first time. However, his remark that he has heard of dudes before shows that the stereotype of the dude has already been transmitted even to remote regions. In contrast to the articles already analyzed, the cowboy is not evaluated positively here. He appears dangerous, crazy, wild, unpredictable (not the least because of his whisky drinking), uncultivated (chewing tobacco, spitting tobacco juice), animalistic and dirty ("a bovine impurity") and eager to pick a fight with

a dude who is depicted as calm, polite, controlled (because he is only drinking lemonade and water), elegant (smoking cigarettes, well-dressed, clean) and unimpressed by the cowboy's aggressive behavior. The anecdote therefore stresses the negative characteristics of the cowboy and the positive ones of the dude and culminates in the unexpected reversal of inferiority and superiority. While the reader might expect and accept that the dude is more cultivated than the cowboy and thus more powerful on this dimension, the cowboy is clearly expected to be physically more powerful – an expectation which is violated here. This is already hinted at through the heading "A Muscular Dude" and it is confirmed when the dude beats the cowboy unconscious in the end and proves to be superior to him on all levels.

This anecdote serves an important function: It takes two stereotypical figures which mark extreme positions in society – the very end points on a continuum between nature and culture – and uses the fact that it is the cowboy figure which tends to be evaluated positively (as authentic, natural, masculine and physically powerful) to create humor. By reversing the roles and making the dude physically powerful, the anecdote subverts the stereotypes and emphasizes that physical power and cultural refinement are not opposites but can be embodied in one person. However, it is not the case that the dude is presented as a role model because other negative stereotypical character traits are still present. As in other anecdotes, the dude is depicted as unhealthy, feeling aloof and his refusal to drink any alcohol at all but only "seltzah with a drop of lemon" makes him seem childish and eccentric. Rather, the anecdote creates a positive evaluation of the middleground between the two poles by combining the positive characteristics of both sides (e.g. elegance, cultivatedness and calmness on the one hand, and authenticity, physical strength and a hard-working character on the other hand) and by avoiding the negative ones (e.g. lack of cultivation, lack of control, arrogance, affectedness, childishness). The representation of language supports this interpretation. The voices of the dude and the cowboy are clearly marked as different and both differ from the narrator's voice, which serves as the representative of the 'neutral' middle position. The cowboy's deviation from both the narrator and the dude on the grammatical and lexical level (his use of *ain't* and demonstrative *them* as well as the use of *lookers* 'eyes',*critter* 'creature' and *looking glass shoes*) indexes his lack of cultivation and education, especially in contrast with the narrator's frequent use of Latinate words (e.g. *expectorated*, *deluge*, *nicotine humidity*). But the three-way division becomes most noticeable on the phonological level, in particular with reference to the realization of /r/. The speech of the cowboy exhibits hyper-rhoticity (e.g. *ther* 'the', *yer* 'you', *terbacker* 'tobacco'), whereas the dude is shown to use non-rhotic forms (*deah*, *fellah*, *seltzah*) and real-

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

izes pre-vocalic /r/ as a labiodental approximant (*weally*).<sup>10</sup> Both hyper-rhoticity and non-rhoticity are therefore linked to the negative extreme positions, which creates an association between rhoticity and the positive middle position occupied by the 'neutral' narrator. Overall, this analysis illustrates again the general importance of the anecdote: It creates a representative instance of a physical fight symbolizing the ongoing fight over social and linguistic values and invites the readers to align themselves with the middle position between the dude and the cowboy.

Apart from the dude figure which represents the typical male Anglomaniac, there are also a few articles which present female figures that admire and imitate English manners and behavior. One of these article appeared before the first mention of the dude, on November 19, 1882[54] , in the *Daily Republican-Sentinel* (Milwaukee, Wisconsin). It contains a description and a comment on "The Girls in Gotham" or, in other words, on "society girls" in New York City. In general, the author compliments the society girls on their "charmingly frank and earnest manner", which he calls "a big improvement over the maidenly simper that formerly prevailed". Despite all this praise, he also finds points of criticism, which he explicitly links to language use:

It is the craze for the English which does her the greatest harm. In her struggle to get the English accent she lays herself open to ridicule. She is guilty of calling street-cars "trams", and says such things as "I cawn't dance any more", or "I can't dawnce any more", combining the American and English in a most hybrid and enervating way.

With regard to pronunciation, this article focuses on the back vowel in bath. By comparing the girls' attempts to speak with an English accent to a struggle, the author emphasizes how difficult it is for the American girls to acquire the lexical distribution of the back vowel. The resulting "hybrid" pronunciation is evaluated negatively because it carries connotations of being neither authentically American nor English. Given that the author praises the girls for their unaffected manners, this reproach of lacking authenticity carries even more weight. Moreover, calling the girls *guilty* indicates a violation of social norms which should not be accepted. Finally, the author clarifies that the reaction triggered by this way of talking is not positive: The word *enervating* rather suggests that it puts a strain on other people.

A second article also links the back vowel in bath to American society girls. It was published in *The Philadelphia Inquirer* (Pennsylvania) on September 14, 1895[153] , and contains a report on "an international cricket match between Oxford and Cambridge and the University of Pennsylvania". The author not only

<sup>10</sup>For a discussion of hyper-rhoticity and non-rhoticity see §3.3.5.

### 4.1 Metadiscourses on phonological forms

describes the results of the game, but also aims to give a picture of the atmosphere of the event that took place at the University of Pennsylvania. He locates the people attending the event on the upper part of the social scale and mainly distinguishes three groups: English boys from Oxford and Cambridge, Philadelphian boys and American society girls. The group of society girls is characterized by means of an alliteration, which also highlights their relation to the other groups: "pretty, lisping lassies, who lavished lovely looks upon them [the "English lads"]". While they do not neglect the Philadelphian boys, the English boys get most of their attention. This is illustrated by means of an anecdote about a group of four people, which is embedded in the report: an "Oxford boy", "a Philadelphia layman" and two well-known society girls, who admire the former and neglect the latter. Language plays an important role in the girls' attempt to get the attention of the English boy: The author reports that the girls pronounce *dance* with a back vowel and use the phrase *y'know*, and he comments that "you could almost catch the accent a square away" and that "accent was laid on as with a mortar trowel". That the accent is English is made clear through the heading of the part of the report, which is "English, y'know", and other explicit descriptions such as "the bonnie accent of old England was a foot thick everywhere". The girls' aim to impress the English boy through their imitation of an English accent is characterized as a "little affectation" which "was charming indeed–sometimes". These descriptions of the accent emphasize that it is not authentic and not American, and even though the author finds it charming, the restriction to "sometimes" hints at a critical stance. The anecdote can also be read as a reproach: that the superficial American girls, who fall for the vain Oxford boy's game, neglect their own countrymen.

The lack of authenticity of the accent is also highlighted by means of a second anecdote that is told in the article, which involves the English boy and a group of "Wissahickon waifs", creating a stark contrast between the upper class people and the young straying men from the Philadelphia neighborhood, who are at the very bottom of society. Even though they are poor, the boys are depicted as being full of spirit and humor and having a carefree attitude to life. They are fully themselves and their sincerity is for example indexed by the expression *on the level* in Jimmy's request to Rock to tell him whether the boy is "a real British mug". That their sincerity and authenticity makes them superior to the Oxford boy is established through their mockery: The "Dun-raven" joke uses the homophony of the name *Dunraven* (the name of an English Lord whose possible attendance is discussed in the article) and *done raven* (*raven* probably used in the sense "to prowl ravenously after prey" given in the *OED* 2021) to play on the English boy's preying on the American girls. This reverses their roles: Even though the boys are

intruders on the cricket grounds, they construct the English boy as an intruder who attempts to take American girls away from American boys. The primary target of their mockery is the English boy's accent and by calling it a *haccent*, they evoke the negative social meanings associated with /h/-dropping and -insertion and not only establish their superiority, but also mock the society girls' admiration of that accent. Their own linguistic repertoire is also marked as different: Lexical items like *mud*, *queer* and *nob* are colloquial (not used by the author in the rest of the article), the second person plural pronoun *you's* marks a grammatical difference, and on the phonological level final -*ing* is realized as an alveolar /ɪn/ (*lookin'*, *bloomin'*), a final /ən/ is hyper-corrected to /ɪŋ/ (*parding*), the dress vowel is raised (*git*), there is TH-stopping (*dat*, *de*) and there are hyper-rhotic forms (*yer* 'you', *onter* 'onto'). This anecdote therefore creates a similar effect to the article "The Muscular Dude": It defines extreme positions, both socially and linguistically. While the anecdote about the dude subverts the stereotypes of the strong cowboy and the weak dude, this anecdote challenges the prestige associated with English accents in upper class circles by opposing their affectedness to the authenticity, sincerity and humor of the lower-class Philadelphian boys. Nevertheless, as in the case of the dude and the cowboy, neither the Oxford swell nor the Philadelphian "waifs" are constructed as role models – they are rather used to delimit a middle ground and clarify which linguistic forms are *not* part of 'normal' and 'neutral' middle-class speech.

# **THE BIG CRICKET GAME**

[…]

Cricket is a "society game," and while it does not partake of the excitement of football nor the interest of the national game, yet it draws a patronage from exclusive circles and is recognized as the favored sport of upper tendom.

# […] English, y'know. […]

But the bonnie accent of old England was a foot thick everywhere. The dainty American girls, society belles and heiresses, some light, others dark, some short and plump and others tall and most divinely fair, all used it. Their little affectation was charming indeed—sometimes. When play was off and during the intermission between the first and second half, the English lads were always the centres of admiring feminine groups. They talked in broad accents and were answered back in "y'knows" by pretty, lisping lassies, who lavished lovely looks upon them. Of course, the Philadelphia boys were not neglected, for there were plenty of bright and pretty girls, enough to go around. But one group was particularly noticeable, probably not against its will. There were **two girls, both society belles**, quite well known, of whom the public often reads, and there were two young men, one **an Oxford boy**, the other **a Philadelphia layman**. And in the patois of the street, the latter was like the driver of the hearse in the funeral cortege—he simply wasn't in it. Everything went the Englishman's way, and he held one of the girls' parasol over his curly head with an air of nonchalance that seemed to say, "**Look at me; I'm a real English cricket swell.**" And they all did look, the girls anyway. Both belles, one of them is a noted blonde beauty, were telling the cricketer how much she'd like him to come to the dance. **She called it "dawnce," and you could almost catch the accent a square away**. In fact, accent was laid on as with a mortar trowel. Then the other girl wanted him to come and "**dine with us, y'know**," while the **hero of**

### 4.1 Metadiscourses on phonological forms

**it all** twirled the parasol, uttered laconisms and **looked superior**. The **poor Philadelphia boy looked miserable** and seemed to feel that for the nonce, at least, he was lost to mind, and fully realized that it was not his day.

Across from this group were two **Wissahickon waifs**, who had somehow eluded the vigilance of the police officers and the club's employes and surreptitiously stolen into the grounds. Said one of them: **"Hi, Jimmy! Git onter his nobs, de Englisher. Wonder if he's Dun-raven yet? Hah!"** The point of the joke was evidently relished by "Jimmy," for he roared with exuberance of spirit. **"Say, on de level, Rocks,"** he replied. "Is dat a real British mug? My eye, wa[?] a lookin' guy. See me queer 'im," and he walked up to the group of four.

**"Say, mister,"** he said, with a scrape and a bow, **"will you's parding me fer a minnit?"**

"Well, what is it, Bobbie?" asked the cricketer.

**"I just wanted to ask yer wat yer'd take"**—here he got ready to run—"fer yer bloody, bloomin' haccent," and off he went with a mocking laugh. […]

[emphasis mine]

The last article that I analyze in this part involves a female figure who not only uses a back bath vowel but also non-rhotic forms and a labiodental realization of /r/. It is an anecdote which was published on December 29, 1885[70] , in the *Rocky Mountain News* (Denver, Colorado) and which consists almost exclusively of a dialogue between two salesladies. The introductory sentence establishes the truth-claim explicitly and it is striking that the reporter who wrote the anecdote emphasizes that he did not influence the women's conversation in any way (they were "wholly oblivious" to his presence) and that he did not intend to overhear them (the conversation was "forced" on him). This reinforces the impression that the dialogue has indeed taken place exactly in the way that it is rendered by the writer. The main topic of the dialogue is the language use of one of the salesladies who has changed her style after changing from the department of "plain goods" to the "ribbons". The ribbons department seems to attract more upper-class customers and it is described by the saleslady as more cultured, which has motivated her to change her language style to the style used by her customers (described as "some of the highest-toned and dwessiest lady's") as well as by the other salesladies in the department ("neahly all of us in the 'wibbons' are at it togetha"). Her friend compliments her on her progress, which she deems remarkable because of her lack of education (which she points out repeatedly), but she is reluctant to change her speech herself. The 'cultured' speech style used by the saleslady is marked most prominently by non-rhoticity and the labiodental /r/ (e.g. in *wibbons depawtment*), but the back vowel in bath is represented as well (e.g. in *cawnt*).

The evaluation of the speech style conveyed by the anecdote is very negative. First of all, it is presented as unnatural and affected: The saleslady repeats several times that she needs to practice the style and that she has to take lessons.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Moreover, already in the first sentence she corrects herself ("all that sort—soath of thing"), which implies that her true natural speech still comes through. Secondly, the saleslady who changed her speech is characterized as unintelligent and uneducated. It is a source of humor in the article that she wants to convince her friend to change her speech because she does not understand that her friend is too smart, sensible and authentic to fall for the new fashion. Statements made by the friend, like "You see being at the desk sort of dulls one", are full of irony because the dialogue makes it clear that it is indeed her friend who is the dull person. The saleslady's lack of education is also underlined by the use of eye dialect to represent her speech (e.g. in *twubble* 'trouble') and by her 'incorrect' pronunciation of words like *certainly* (*certahainly*). This shows that the subheading "Remarkable Proficiency Manifested by Her for a Beginner" is full of irony because the article does not portray her in a positive light. Thirdly, the speech style is linked to pressure and force. The saleslady is explicitly labeled a "victim" through the subheading and the dialogue establishes that she experiences pressure from the upper-class customer (she is afraid of not being able to communicate with them) and from her fellow peers (the salesladies at the ribbons department who have decided that they have to pay a fine for every mistake). The force exerted on her is not direct – she voluntarily adopts the speech style and seems very proud of it – but the author of the anecdote conveys that the pressure works indirectly and he compliments people like the friend of the saleslady who resist the pressure and stay authentic and true to themselves.

The strategy used to convey the evaluation of the speech forms is very powerful. The author does not condemn "the latest lingo" explicitly; on the contrary, he emphasizes that he is only an observer without an agenda – someone who lets the "facts" speak for themselves. The basic messages conveyed by the salesladies' conversation, namely that there are linguistic differences between saleswomen working in different departments, and that these differences correlate with the social class of the customers who shop at the department, are realistic. In fact, Labov's Department Store Study, conducted in 1962 (Labov 2006), presented evidence of this correlation by comparing the use of rhotic forms by salespeople working in three department stores in New York City, and this correlation was confirmed in two replica studies (Fowler 1986 and Mather 2012). The striking difference is that in the anecdote, non-rhoticity is presented as the upper-class prestige variant that the saleswomen working in the ribbon apartment use to accommodate to their customers, while in the New York Department Stores in the twentieth and twenty-first centuries salespeople in the upper-class stores use a higher percentage of *rhotic* forms compared to the other stores. (The implications of this difference will be discussed further in §5.3.) By presenting himself

as a neutral observer giving a realistic account of the salesladies' conversation, the author of the article appeals to the readers' intelligence and common sense to form a judgement. He refrains from any explicit disapproval of the saleslady's "latest lingo" because this might have provoked objections to his view, whereas the humorous nature of his account makes the issue appear light and not serious, so that anyone objecting to the argument presented would risk appearing like a humorless and unpleasant person.

# **CULCHAWED CLEHKS.**

# **The Latest Lingo and How it is Acquired by Victims.** A Saleslady Tells a Companion How She Got it So Well. **Remarkable Proficiency Manifested by Her for a Beginner.**

The following conversation was forced on a News reporter yesterday in a street car by two young "salesladies," who appeared to be wholly oblivious to the presence of the scribe:

"Oh, deah, I'm so tiahed of Quissmiss and holidays and all that sort—soath of thing, you knoah."

"Why, you do that remarkably well for the time you are at it."

"Doah whaat?"

"Oh, talk that way, of course. Why, you remember when you went into 'ribbons' three months ago you talked quite plainly considering your lack of education."

"Oh, deah, you flattah me."

"Not in the least, I assure you. You must have applied yourself very steadily."

"Why, weally, I cawnt say os I have, but you see the 'ribbons' is vewy culchawed, an' one gets on fostah theah than in 'plain goods.'"

"I judge so."

"But of coahs I didn't get all my culchaw theah!"

"Did you take lessons?"

"Why, cetaihanly. I've been an' taken pwivate lessons and I've joined a litwy club foh exesize."

"This must benefit you greatly."

"Oh! deah, yes; I could nevah get along without pwactice. You see it's easy enough to get onto the way of doing it but the twubble comes in putting it in pwactice."

"It must be very difficult."

"Yes, without pwactice impossible, I should expect. But you see neahly all of us in the 'wibbons' are at it togetha and that helps us wondahfully. We've agreed upon a fine foh any one that makes a mistake. Won't you guess whaat it is?"

"I'm sure that I don't think I would be able to guess. You see being at the desk sort of dulls one."

[…]

[The friend tells her that she is afraid of trying to pick up the style.]

"Oh! deah no, you mustn't be afwaid. Why yove had an education, and ought to be able to pick up weal fast."

"I fear not. As near as I can judge from the people who talk in this stylish manner the absence of education is an assistance rather than a drawback."

"Oh! deah, no, why some of the highest toned and dwessiest lady's can talk it just as good and even bettah than us in the wibbons depawtment."

To summarize the analysis of articles containing *dawnce*, *deah* AND *fellah* and twousers, I have shown that the phonological forms they represent are, first of all, linked to English people. The range of indexical meanings is greatest for the back bath vowel, which comprises links to English people of all social classes. The labiodental realization of /r/, on the contrary, is restricted to the figure of the upper-class English swell. Non-rhoticity can index the swell, but also more average English people, often embodied in the figure of the English visitor or tourist. Next to these English figures, there are also American figures associated with the forms: people who imitate English manners, fashion and behavior. They are either part of the upper class or they seek access to high society, which shows that the forms must have acquired the indexical meanings 'refined', 'cultivated' and 'of high social position'. The most prominent figure is the American dude, who is originally an American version of the upper-class swell, but the figure is developed further to comprise also poor young men who hope to advance socially. The dude is usually portrayed as weak, effeminate, childish, affected, lazy, sickly, unintelligent and overly concerned with his outer appearance and with impressing females.

Female figures linked to the use of the phonological forms are the society girls (in both articles, however, only the back vowel in bath was represented) and the saleslady, who wants to become more 'cultured' and be able to interact with upper class ladies. The females' language use is also evaluated as affected and the female figures were characterized as rather unintelligent, easy to impress and insecure in their eagerness to please others. The affectation and unnaturalness indexed by the forms is underlined by opposing these figures to American figures with a lower-class and/or a rural background, such as the cowboy, the Montana girl or the Philadelphian "waifs". These appear as authentic, natural and downto-earth, the cowboy also being tough, strong and hard-working, the Montana girl also being sane, practical and frank and the Philadelphian "waifs" also being witty, courageous and street-wise. The Montana cowboy and the Philadelphian "waifs" have been shown to also mark the other end of the extreme in two anecdotes: Rather than serving as figures to identify with, they are used to define the middle-ground, which is particularly well illustrated in the case of rhoticity, where the realization of post-vocalic /r/ marks the neutral middle position between hyper-rhoticity on the one extreme and non-rhoticity on the other. Rhoticity will also be the focus of the analysis in the next section, which deals with articles containing the search term *bettah*.

### 4.1 Metadiscourses on phonological forms

# **4.1.2.3** *bettah*

As the search term *bettah* represents non-rhoticity, it is not surprising that there are a number of articles linking *bettah* to the same social values and social personae as *deah* AND *fellah*: Englishmen and Americans who imitate English fashion and speech, in particular the dude. However, there are three additional broad groups of people that become indexically linked to non-rhoticity in the articles containing *bettah*, which is why I analyze these articles here separately. The three groups can roughly be categorized as Black Americans, white southern Americans and mountaineers. The fact that these three groups are not indexically linked to *deah* AND *fellah* suggests that the phonological and the lexical levels are connected here: The phrase *dear fellow* is not constructed as being part of their linguistic repertoire.

Before discussing the three groups in more detail, however, I will analyze the first article containing *bettah* in the databases because it constitutes a very good example of how non-rhotic forms were linked to British English speech. It was published relatively early, on April 17, 1844[11] , in the *Weekly Ohio Statesman*, in an article which quoted a speech by John Teesdale, a British immigrant who was born in York, England, and who had arrived in Philadelphia with his parents in 1818 and became the editor of several newspapers, for example the *Ohio State Journal* (Snodgrass 2008: 526–527), which was the main competitor of the newspaper in which the article appeared. It is therefore not surprising that John Teesdale is portrayed in a very negative light. The representation of his speech is part of this negative characterization: His English origin is emphasized to discredit him and his political opinions. Linguistic forms represented in his speech not only comprise non-rhoticity, but also several others, as the following sentence shows: "A high Tawiff, sah, enhables owah manufactuahs to get bettah prices foh thah goods, and, sah, what is pekooliahly remarkable, it also enables the people to buy the same goods at low prices!" In this example sentence, non-rhoticity is represented in all instances except one (*remarkable*) and there is one instance of a labiodental realization of /r/ (*Tawiff* ). There is also one instance of /h/-insertion (*enhable*), but it is noticeable that the /h/ is not inserted in the second instance of *enable* and it is not dropped in *high*. Furthermore, there is also an instance of yod-dropping (*pekooliahly*), a form which I will discuss in more detail in the next section. Given the results of the analysis of articles containing representations of /h/-dropping and -insertion, non-rhoticity and labiodental realizations of /r/, it is striking that the labiodental approximant [ʋ], which comes to be indexically linked to the English upper-class swell, co-occurs with /h/-dropping and -insertion here, which usually indexes that the speaker is uneducated and

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

not part of the upper class. As the social meaning of /h/-dropping and -insertion was already fairly established in the 1840s, this suggests that labiodental /r/ (and probably non-rhoticity as well) signaled first and foremost that a speaker was English. Secondly, they probably signaled incorrectness (like /h/-dropping and -insertion) and were therefore used to question the competence and intelligence of the speaker and to ridicule his argument. This interpretation is strengthened by taking an article into consideration which was also published in the *Ohio Statesman* (on May 21, 1845[12] ) and in which John Teesdale's English is discussed in more detail:<sup>11</sup>

# **'Orrible.**

The editor of the Journal, it grieves us to say, does not like our mode of writing the English language, and, fancying himself a "schoolmaster abroad," he has taken us to task for false grammar. Not being a "*native* Hinglish cockney," as is Mr. Teesdale, the English cannot be strictly called our mother tongue. Mr. Teesdale not only speaks the English language with fluency, but he speaks the *language of England*, and that of her tory Ministers, on all questions at issue between Great Britain and the United States.

The author of the article reacts to John Teesdale's criticism of the language used in articles published in the *Ohio Statesman* by pointing out that Teesdale's arrogant claims to linguistic superiority are mainly based on his English origins and the underlying belief that British English is better and more correct than American English. The strategy used to criticize Teesdale is irony: The author pretends to admit that "the English cannot be strictly called our mother tongue", but by calling Teesdale a "*native* Hinglish cockney" (the word *native* being emphasized through the use of italics), he implicitly conveys his real position, namely that even if someone was born in England, he or she did not necessarily speak English more correctly than an American. Calling Teesdale a "cockney" and linking him to /h/-dropping and -insertion (prominently in the heading *'Orrible* and in the word *Hinglish*) serves to underline the argument that he does not speak correctly and that he is neither a role model for Americans nor entitled to criticism. This suggests that non-rhoticity and labiodental /r/ were primarily markers of incorrect speech in the 1844 article as well, although the conclusion has to remain tentative and needs further support.

While the first article links *bettah* to English speech, the majority of articles link it to the three groups mentioned above, to which I will now turn. The second article in the set of articles containing *bettah* in the databases already connects

<sup>11</sup>I found the article by searching for articles about John Teesdale in order to find more information about him. The fact that it also contains the search term *hinglish*, but is not identified in the search for *hinglish*, illustrates the methodological problem of the study that the search software very likely did not succeed in identifying *all* instances of the search terms.

### 4.1 Metadiscourses on phonological forms

the form to the group of Black Americans. It was published six years after the first article, on August 02, 1850[15] , in the *Scotio Gazette*, which is also an Ohioan newspaper (of the city Chillicothe). The article reports on the state of health in the city and discusses the fear of cholera, which had presumably caused the death of several "colored persons". The author concludes the report with a humorous anecdote involving a "real" person, "a black servant in our family", whose sister was expected to die of cholera. She went to see her and reported, when she came back, that her sister was getting better despite having been infected either with cholera or with whooping cough.

# **Health of Chillicothe.**

Our city yet continues healthy—as we trust, by the favor of Providence, it may throughout the season. Reports to the contrary have spread about the country, which had their origin in the fact that nine colored persons, old and young, and of both sexes, have died within the last 48 hours, several of them suddenly. Of these, three died of choleratic symptoms, one of old age, and another of old corn whisky. We have no doubt the real Asiatic cholera carried off some of them—and there is much alarm among the colored population in consequence. White folks, too, had as well be careful of their diet, for there is no telling how soon the erratic disease may appear among us.

This morning, a black servant in our family, was sent for, "to see her sister die of cholera." She was absent till about dinner-time, when she returned, with a thicker

> "Pout on her lip, and Smile in her eye!"

On being inquired about her sister's health, she replied:

"I dozzant know wedder 'twas de kolra or de hoopin'-koff—but she's gittin bettah!"

The anecdote aims at ridiculing the Black female servant who is smiling, carefree and optimistic even in the face of serious and dangerous illnesses because she does not know any better. Her inability to distinguish cholera and whooping cough characterizes her as ignorant because whooping cough can be recognized easily by one of its symptoms: the severe cough which is indicated by the name of the disease. Her ignorance is also signaled linguistically through the use of eye dialect: spelling *cholera* <kolra>, *whooping cough* <hoopin'-koff> and *doesn't* <dozzant> creates the impression of illiteracy. Her speech is additionally marked as different on a grammatical level (first-person singular -*s* in *I doesn't*) and on the phonological level (TH-stopping in *wedder* and *de*, a raised dress vowel in *gittin'*, alveolar -*ing* in *hoopin'* and *gittin'*, initial /h/ instead of /hw/ in *hoopin'* and initial unstressed syllable-deletion in *'twas*). She is also described as having a "thicker pout on her lip" (indicating that she has full lips) to emphasize her difference on a physical level. All in all, the anecdote functions as comic relief for the readers at the end of a serious article, relieving their tension at least temporarily after reading about a potentially serious threat to their lives, at the expense of the nameless Black American servant who becomes the object of ridicule.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

There are several anecdotes and short paragraphs which link non-rhoticity to Black Americans and social characteristics like being poor, lacking education and having jobs in which they serve other people (mainly whites), like household servants, washerwomen, bootblacks, waiters and porters. Sometimes they are depicted as criminals (often thieves). The following three examples illustrate how these links are created.

One example is an anecdote with the heading "Rich", which was published in the *Atchison Daily Champion* (Kansas) on August 20, 1889[95] , but taken from the magazine *Youth's Companion*. It depicts the conversation between "a lady" and "her colored washer-woman" by using direct quotations to construct the voices of the two figures. The narrator describes what the anecdote serves to illustrate in the first line: "People have widely different ideas of what constitutes wealth". In the conversation, the Black woman explains to her lady that her daughter has married and that this daughter and her husband are so rich that they do not have to worry about money. The humor rests on the sum of money that the woman assumes to be necessary to be rich like that, namely 139 dollars, and the readers' knowledge that this sum does not constitute wealth. On the contrary, the washerwoman's expectation that they will live a wealthy life emphasizes her own poverty. It also creates a stark contrast between her own way of life and that of the lady who she is talking to. This contrast is linked to the contrast between their voices. The lady's voice takes up less space than the Black woman's voice and does not exhibit any differences to the 'neutral' voice of the narrator. The constructed dialogue consists mostly of the Black woman's speech, which is marked by a large number of differences on all linguistic levels. As in the article "Health in Chillicothe", the Black female speaker is marked as non-rhotic, but non-rhoticity is not represented consistently (not in *her* and *dollars*). In contrast to the first article, the absence of /r/ extends here even to intervocalic contexts (*mah'ied* 'married). The counterpart to /r/-loss is also present in her speech as there are three cases of /r/-insertion or hyper-rhoticity (*ter* 'to', *erlong* 'along', *par* 'pa'). Like the Black servant in the first article, the washer-woman exhibits fricative-stopping (*de*, *deyselves*), but the stopping comprises not only voiced interdental fricatives here, but also voiced labiodental fricatives (*hab* 'have', *nebbah* 'never'). A raised vowel in *get* (*git*) is also represented again here, as well as initial unstressed syllable deletion (*'deed* 'indeed', *'bout* 'about'). A phonological form that did not occur in the first article but that is prominently represented in this one is the elision of the final consonant in syllable-final consonant clusters (*ole* 'old', *hol'* 'hold', *han's* 'hands', *las'* 'last', *lef'* 'left', *jess* 'just'). The weakening and reduction of *and* to *en* or *'n* is represented orthographically as well. Considering the large time gap of almost forty years between "The Health of Chillicothe" and

### 4.1 Metadiscourses on phonological forms

this anecdote, it is striking that the repertoire of phonological forms linked to Black American speech has remained fairly stable over time.

On a grammatical level, there is first-person singular -*s* marking, as in the first article, but also third-person plural -*s* marking (*I'se wu'ked*, *dey has no need*). What is new in this article is the representation of perfective *done*, once with a contracted form of *be* or *have* (*she's done mah'ied*) and once without an auxiliary verb (*his par done died*). The third-person plural reflexive pronoun *deyselves* is also linked to the Black washer-woman's speech. Lexically, the use of the phrase *I reckon* and the creole form *gwine* 'going' are linked to the Black female figure. What is also highly relevant are the different labels used to designate her. The writer of the anecdote labels her a "woman", while he calls the other person a "lady", which reflects a difference in social status, the "lady" being superior to the "woman". Furthermore, the Black woman is represented as labeling herself an "ole mammy", which evokes the mammy stereotype, which is implicitly already present in the first article about the female Black servant in Chillicothe. In a detailed study of the stereotypic cultural representations of the mammy figure, Wallace-Sanders (2008) describes their profound influence on American culture. Her analysis shows that the mammy figure was created in the 1820s and became increasingly popular and widely recognized already by the middle of the nineteenth century. She defines "the standard, most recognizable mammy character" as

a creative combination of extreme behavior and exaggerated features. Mammy's body is grotesquely marked by excess: she is usually extremely overweight, very tall, broad-shouldered; her skin is nearly black. She manages to be a jolly presence – she often sings or tells stories while she works – and a strict disciplinarian at the same time. First as slave, then as a free woman, the mammy is largely associated with the care of white children or depicted with noticeable attachment to white children. Her unprecedented devotion to her white family reflects her racial inferiority. *Mammy* is often both her title and the only name she has ever been given. She may also be a cook or personal maid to her mistress – a classic southern belle – whom she infantilizes. Her clothes are typical of a domestic: headscarf and apron, but she is especially attracted to brightly colored, elaborately tied scarves. (Wallace-Sanders 2008: 5–6)

In this anecdote, several of these characteristics are also present: She is a washerwoman and thus a hard worker, she is inferior to the "lady" and she is depicted as jolly and oblivious to her social and economic position. This is particularly

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

highlighted in her last statement "Some folks jess seem ter be bawn lucky", which is ironic because she refers to her daughter and her husband as the lucky ones, wile it is clear to the reader that all three of them are poor – not least because of their ethnicity, which they were given by birth (so they were in fact *not* born lucky). The implicit statement conveyed by the anecdote is that Black Americans are not lucky because they are rich, but because they are too ignorant to realize how poor they are, and that because of their jolliness they do not need help or pity. This condescending attitude ultimately serves to evoke feelings of superiority on the part of the white readership, so that the superiority/inferiority distinction is also linked to rhoticity/non-rhoticity and the other linguistic differences as well.

# **Rich.**

People have widely different ideas of what constitutes wealth, as the following incident illustrates: "I hear, Amanda, that your daughter is married," said a lady to her colored washer-woman.

"Yes'm," was the reply; "en I tell yo' she's done mah'ied bettah 'n her ole mammy did, en she'll nebbah hab to wu'k like I'se wu'ked. No'm, all she'll hab ter do 'll be to set en hol' her han's." "Is her husband rich?"

"'Deed he is, ma'am; he had fifty dollars in bank de day he got mah'ied, en his par done died las' week en lef' him seventy-five mo', en my Tilly she had fo'teen of her own, so I reckon dey has no need to worry deyselves 'bout dey's gwine ter git erlong. Some folks jess seem ter be bawn lucky."—Youth's Companion

The second example is a short paragraph consisting of a direct quotation of "Uncle Eben". It must have been very popular because it was published in six different newspapers after its first occurrence in the *Washington Star*: in the *Morning Oregonian* (Portland, Oregon) on December 08, 1893[135] , in the *Daily Picayune* (New Orleans, Louisiana) on December 14, 1893[136] , in the *Idaho Statesman* (Boise, Idaho) on December 27, 1893[137] , in the *Atchison Daily Globe* (Kansas) on December 27, 1893[138] , in the *Northern Christian Advocate* (Syracuse, New York) on January 3, 1894[139] , and in the *Boston Investigator* (Massachusetts) on March 28, 1894[141] . The following version is the one that was published in the *Idaho Statesman* and while the paragraph is largely the same as in the other newspapers it has been given the heading "Philosophy" here. This function of this heading is to increase the humor of the paragraph: The term *philosophy* raises the expectation of a sophisticated mental exercise performed by an educated person with time and motivation to deal with the basic questions of human existence. This expectation is violated, however, by the quotation of the paragraph, which expresses a belief about life which is rather simple, namely that one should take care of basic needs (like nutrition) before aiming at improving one's outward appearance.

### 4.1 Metadiscourses on phonological forms

The simplicity of the belief is underlined by the metaphors used to illustrate it: A "cabbage undah yo' wais'-coat" symbolizes the need for nutrition, while the "chrysanthemum in yer button-hole" symbolizes the care for outward appearance. The person uttering the statement is therefore portrayed in an ambivalent way: On the one hand, he is depicted as a character that is down-to-earth and not interested in vanity or luxury, but on the other hand, he is also ridiculed as a simpleton leading a way of life marked by poverty and the need to fulfill his basic needs. His Black ethnicity is marked through the use of the address term *Uncle*, which was commonly given to male Black Americans instead of other address terms, such as *Mister* (Harris 2008: 10).

Regarding the representation of Uncle Eben's voice, non-rhoticity (*bettah*, *undah*, *yoh*) is again a prominent form, but, as in the anecdote "Rich", some words are not marked as non-rhotic (*yer* 'your', *outward*) and hyper-rhotic forms occur as well (*ter* 'to', *er* 'a'). The determiner *your* is represented once as rhotic and once as non-rhotic, which creates the impression of inconsistency in Uncle Eben's speech. Further phonological forms shared with the representation of Black speech in the anecdote "Rich" are final consonant cluster reduction (*Doan* 'don't', *min* 'mind', *wais'* 'waist') and the stopping of voiced interdental (*dan* 'than') and labiodental fricatives (*hab* 'have'). A form that did not occur in the two articles discussed above is *Hit*, a case of /h/-insertion, but as it only occurs in one function word, it is not a salient form here. On a grammatical level, subjectverb agreement is also marked as different in *Hit am bettah*. Overall, the shortness of Uncle Eben's philosophical statement in combination with the multitude of linguistic forms signaling difference from the 'neutral' speech forms used in surrounding newspaper articles underlines the uneducatedness of the figure and the simplicity of his approach to life and it creates the impression of otherness.

# **Philosophy.**

"Doan put yer min too much on outward decorations," said Uncle Eben. "Hit am bettah ter hab er cabbage undah yoh wais'-coat dan er chrysanthemum in yer button-hole."—Washington Star.

The third example is the anecdote "An Unwelcome Fifteenth", published originally by the *Detroit Free Press* and printed in the *Dallas Morning News* (Texas) on August 1 and August 2, 1899. The narrator describes a "couple of tourists" who talk to "an old negro" during their journey "in the rural district of the south". The "negro" lives in "a small log cabin, out of which children of all sizes and age came swarming like bees from a hive" and he sells food to passing tourists. When he sends one of the children, called Judas Iscariot, to catch a chicken, the tourists ask him why he has given the boy this name. He answers

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

"Well, I'll tell yo', sah. Hit's like dis: Yo' see I'd had fo'teen chillum befo' Judas Iscariot was bawn, an' fo'teen chillum is a mighty big fam'ly fo' a po' man ter raise en keer fo', thout habin' no mo', so when Judas Iscariot came erlong I gib 'im dat name caze you know de Bible hit say it'd be bettah fo' Judas Iscariot if he'd nebbah been bawn."

This explanation highlights several supposed characteristics of Black people: their poverty, their big families, their religiousness and their lack of education and intelligence. The last aspects are foregrounded here, as the main aim of the anecdote is to ridicule the way that Black people understand religious aspects and how this influences their lives. To educated religious people, the main characteristic associated with Judas is that he was a traitor whose betrayal led to the crucifixion of Jesus, which is why they would not name a child Judas. Because of his lack of understanding the bible, the Black man only focused on the statement that it would have been better if Judas had never been born and used this as a basis for naming his child. His lack of intelligence and education therefore has negative consequences for the child. The comparison of the family's cabin to a beehive evokes associations with animal-like behavior and adds to the impression that while Black people try to be civilized, for example by being religious, they fail to achieve it.

Linguistically, the forms that mark the Black man's speech as different from that of the narrator are strikingly similar to those used in the anecdote "Rich" and in the short paragraph "Philosophy", which were published ten and six years earlier. Non-rhoticity is combined with hyper-rhoticity, and in one instance (*keer* 'care') non-rhoticity is not marked. In the same word, however, the spelling of the square vowel as <ee> in *care* suggests a differential realization of the vowel as a high front monophthong. Furthermore, there are several instances of interdental and labiodental fricative stopping. Final consonant cluster reduction is not marked in the quotation above because there are no words containing final clusters, but it is marked in other parts of his speech represented in the anecdote. The third-person singular personal pronoun *it* is pronounced *hit*, as in Uncle Eben's speech, and it becomes clear here that this does not represent a general process of /h/-insertion, but that the insertion is lexically restricted to the pronoun because if the author had wanted to mark the speaker as exhibiting /h/-insertion on a phonological level, it is very likely that he would have spelled the name Iscariot with an initial <H>. Furthermore, alveolar -*ing* and initial unstressed syllable deletion are forms marking the Black man's speech as well (*thout habin'* 'without having'). Additional forms, which have not been part of representations analyzed above, are the result of phonological reduction processes affecting lexical items: *chillum* 'children' and *fam'ly* 'family'.

### 4.1 Metadiscourses on phonological forms

On a grammatical level, differences in subject-verb agreement are also represented, but this time it is the absence of -*s* marking the third-person singular (*it say*). There are also grammatical forms not found in the articles above. In one case, a past tense form is not marked (*when Judas Iscariot came erlong I gib' im dat name*). Furthermore, *de Bible hit say* is a case of left dislocation, where a noun phrase is repeated and replaced by a pronoun in its second occurrence (Schneider 2015: 196). On a pragmatic level, the address term *sah* is used frequently by the Black man (at least once in every contribution to the dialogue) and he also uses the address term *gemmen* 'gentlemen' twice. As the tourists are depicted as not using any terms of address when talking to the Black man, their higher social position is underlined by linguistic means. The "gentlemen" are not required to be polite to the poor, uneducated Black man, but they may simply use his help to fulfill their basic need for food.

The three examples analyzed so far not only illustrate typical characteristics associated with Black people, but they also show that in the last two decades of the nineteenth century a fairly stable repertoire of linguistic forms was used to represent their speech. Before turning to the analysis of other text types, which exhibit a number of different strategies to establish indexical form–meaning links, I will analyze one last anecdote here because it aims at illustrating how stereotypes are formed and that relying on them can have negative consequences.

The anecdote tells a story about a white man who jumps to conclusions too quickly. It is entitled "Too Hasty" and it must have been a very popular anecdote because it was published four times in 1891 (first in the *Dallas Morning News* (Texas) on December 20, 1891[110] , then in the *Atchison Daily Globe* (Kansas) on December 22, 1891[111] (this version is cited below), then in *The Galveston Daily News* (Texas) on December 24, 1891[112] , and finally in the *Philadelphia Inquirer* (Pennsylvania) on December 27, 1891[113] ) and three times in 1892 (in *The Emporia Daily Gazette* (Kansas) on January 28, 1892[114] , and in *The Milwaukee Journal* (Wisconsin) on February 29, 1892[115] , and in the *Aberdeen Daily News* (South Dakota) on April 19, 1892[119] ). It was originally published by the magazine *Youth's Companion*. The anecdote exhibits a structure that is typical of anecdotes: In the first sentence, the narrator summarizes the main message that the incident described in the anecdote is supposed to illustrate: "Jumping at conclusions often results in embarrassment to all concerned". The narrator continues by introducing the main figure that is the focus of the anecdote: "a certain clerk in a Court street law office". The following part is mainly characterized by dialogues between the clerk and Black American figures: first an "old Negro" and then "another sable head". Both of them offer to clean the window, but the clerk sends them away. The third person entering the office is described as "a dark face surmounted by

a rather rusty hat". The clerk, who gives the incomer only a quick look, assumes that this is another Black American who wants a window-cleaning job and says "I suppose you want to wash windows, too, don't you?" The humor of the anecdote is created in the last paragraph which shows that the clerk's assumption was wrong and that the man is actually the father of "the senior member of the firm". This is revealed through the senior member's enthusiastic greeting, accompanied by the words "Why, my dear old father, this is the most delightful surprise of my life!" This sentence, representing the voice of the senior member of the firm, indicates that the man in question is a white man because it does not exhibit any of the forms marking the speech of the two Black men who asked for a window cleaning job before. The father must therefore also be white, and the darkness of his face was probably rather a result of the hat casting a shadow. The clerk's hasty conclusion draws attention to the way that stereotypes come into existence: Based on two encounters with Black men who offer to clean the windows, the clerk draws the generalization that all Black men who enter the office want to get a window-cleaning job, so, in other words, he assumes a social regularity based on a recurrent co-occurrence of signs (ethnicity of the speaker and his job/social position). However, the main point made in the anecdote is that perception can be influenced by stereotypes. Based on his prior experience, the clerk immediately assumes that the dark face of the third incomer belongs to a Black man who wants to clean windows. His expectation prevents him from taking a closer look at the man and this causes great embarrassment. It is notable that the anecdote does not call into question stereotypical assumptions about Black people and the jobs they hold, but it warns about applying these assumptions to the wrong people (that is, people who are not Black).

Language plays an important role in differentiating and marking the ethnicity of the speakers. The window cleaners' speech exhibits linguistic forms which are typically associated with Black speech. Non-rhoticity is salient again and highlighted, for example through the use of a parallel structure when the second window cleaner and the clerk exchange greetings: "Good aftahnoon, boss". / "Good afternoon". The repetition of the lexical item *afternoon* draws attention to the phonological difference, that is, the window cleaner's non-rhoticity. The use of the address term *boss* by the window cleaner and the absence of any address term by the clerk signal their unequal social relationship. The Black man also uses incomplete sentences: There are no subjects in the sentences *Want yoh windows cleaned?* and *Do it cheap, boss*, where they would normally be required. The other Black speaker also uses an address term (*cap'n*) and he exhibits TH-stopping (*dey* 'they'), hyper-rhoticity (*kinder*), a longer, higher and fronter vowel in *leetle* 'little' and third-person plural verb forms which are marked by an -*s* (*they looks*, *they*

### 4.1 Metadiscourses on phonological forms

*needs*). Both Black speakers also exhibit elisions (*cap'n*, *'em*, *lemme*) and it is noticeable that they both use the word *better* as part of a modal construction, which van der Auwera et al. (2013) call a better construction. In their framework, the better constructions consist of one of the comparative modals *had better*, *'d better* or *better* followed by a verb, and they usually express the speaker's advice and more rarely the speaker's hope. In the anecdote, both Black speakers use the same construction (the modal *better* followed by the verb phrase *lemme* + V and without an explicit subject) to advise the clerk that the windows need to be cleaned. This parallel structure draws attention to the modal construction and it therefore illustrates how based on the form *bettah* indexical links are created both on the phonological and the grammatical level. In contrast to the Black speakers, the senior member's linguistic repertoire does not contain any forms that are different from the narrator's voice or the clerk's voice, which is a clear indicator of his whiteness. The fact that his whiteness is not explicitly pointed out but implicitly conveyed through his use of language (and also through his high social position in the firm) illustrates how white speakers and their speech are marked as the 'default', whereas the Black speakers are constructed as the 'other'.

# **Too Hasty.**

Jumping at conclusions often results in embarrassment to all concerned. Perhaps nobody knows this better than a certain clerk in a Court street law office. He was sitting at his desk, writing busily, the other afternoon, when the door opened and an old negro put in his head.

"Say, cap'n, don't yoh want yoh windows washed? Dey looks kinder like dey needs it."

"No, not today; they were washed only last week."

"Bettah lemme touch 'em up a leetle, cap'n."

"No, no," replied the clerk, going on with his work; "come around in a couple of weeks."

With another intimation that the windows were susceptible of considerable improvement, the ancient cleaner withdrew.

But window washers were evidently on full force that day; for five minutes had not elapsed when the door opened again and another sable head popped in.

"Good aftahnoon, boss."

"Good afternoon."

"Want yoh windows cleaned?"

"No, not today."

"Do it cheap, boss; bettah lemme clean 'em."

"No; just engaged a man."

Three minutes later the door opened again, and a dark face surmounted by a rather rusty hat peeped in.

"Well?" asked the clerk, looking hastily up, "I suppose you want to wash windows, too, don't you?"

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

It was difficult to tell whose surprise was greater, the newcomer's or the clerk's, when the senior member of the firm hastened forward from his room, and grasping the stranger affectionately by the hand exclaimed, "Why, my dear old father, this is the most delightful surprise of my life!"—Youth's Companion.

Apart from anecdotes and short humorous paragraphs, there are other text types which link Black Americans to non-rhoticity and other linguistic forms and, in doing so, emphasize several social characteristics related to being Black and contribute to the construction of several stereotypes of Black Americans. At first, I will discuss two examples of reports which focus on a specific Black figure: the Black preacher.

The first report that I will analyze here is entitled "A Negro Revival". It was published on May 4, 1875[31] , in the *Georgia Weekly Telegraph* (Macon, Georgia). The author describes a religious meeting of a group of Black people and focuses on the characteristics of the Black preacher and his interaction with his audience. He first addresses the preacher's outward appearance by calling him "shining" and using the image of a "glossy [...] varnished beaver" for comparison. This already creates the impression that the preacher is very interested in impressing his audience (he wants to shine), but that he is also a slick character whose inner qualities and competence do not hold up to the shiny outward appearance. This impression is supported further by the author's depiction of the preacher's sermon, which despite some exaggerations also seems to aim at giving a realistic impression of the preacher's rhetorical style. He presents it as an almost theatrical performance by describing the preacher's quality of voice as "low and reverential" in the beginning but changing over time, when he starts a "wailing chant, with a prolonged sound in a higher key on emphatic words and syllables". It is thus on the level of intonation that the preacher and the audience are described to connect: They "unite with the preacher in a piteous moan, between words, gliding down from the dominant note to the minor third below, and dying through diminuendos into sobs and sighs". This description creates the impression of an extremely emotional, but also unstable person, who has a strong effect on his audience, but not in that he provides strong, reliable guidance, but more in that he creates strong emotions. This impression fits the explicit characterization of the preacher in the sub-heading: He is labeled "A Colored Moody". The author addresses other linguistic aspects as well and connects them to an explicit evaluation: He quotes the preacher as saying "Thou Knoweth" and "You knows" and calls the use of this agreement pattern "indulging in doubtful grammar". Furthermore, the use of "Thou" and "You" to address the Deity is regarded critically by the author because the two forms are used "indiscriminately". Consequently, in both cases, the linguistic competence of the preacher is questioned because even

though he uses forms which are found in the bible and thus associated with formal and old-fashioned religious language (*thou* and *knoweth*), he is not consistent in the use of old and modern forms and chooses an agreement pattern in both cases that is evaluated as incorrect. Next to the description and evaluation of the preacher and the meeting in general, the author also includes a direct quotation of the preacher's sermon, which he calls an "exhortation", thereby constructing the voice of the preacher on several linguistic levels.

The sub-heading already contains two direct quotations which highlight important phonological forms: voiced interdental and labiodental fricative stopping (*wid* 'with', *de* 'the', *dan* 'than', *debble* 'devil'), alveolar -*ing* (*foolin*) and nonrhoticity (*whispah* 'whisper'). However, it also shows that non-rhoticity is not consistently marked because *Lord*, *better* and *holler* are not represented as being pronounced without post-vocalic /r/. While *Lord* is also not represented as nonrhotic throughout the direct quotation, *bettah* is marked as non-rhotic there (*you bettah stay away*). The other phonological forms represented in the long direct quotation are initial unstressed syllable deletion (*foah* 'before', *thout* 'without', *roun* 'around', *'tirely* 'entirely'), final consonant cluster reduction (*bes'* 'best') and syllable reduction (*foh'rd* 'forward'), which are all forms that I have shown to be represented as part of Black speakers' linguistic repertoire in later articles as well. Two forms which were not present in those articles are a lower kit vowel in *if* (spelled <ef>) and the fronting of the voiceless interdental fricative in *fru* 'through'.

On a grammatical level, differences in subject-verb agreement are most noticeable. The second-person plural is marked by the inflectional suffix -*s* and this is highlighted through the use of parallelism. Several short clauses starting with *you* and followed by a verb occur right next to each other:

You come foah you's ready. You starts too soon. You don't repent; you's no mounah. Your foolin' wid de Lord. You comes struttin' up to de altah; you flops down on your knees, an' you peeps fru you fingahs dis way, an' you cocks up you eahs to see who's makin' de bes' pray'r. You's 'tirely too peart for peniten's. You's no mounahs.

The section not only draws attention to the inflectional second-person plural -*s*, but it also highlights inconsistencies. *You come* and *you comes* both occur without any indication why the marking with -*s* is absent in the first case. With regard to the second-person plural form of *be*, it is usually *is* (*you's ready*, *you's no mounah*, *you's 'tirely too peart*), but in one case it is also *are* (*your foolin'*). Even though *be* is an auxiliary verb in the second case and not a main verb, it is not clear whether this is a pattern causing the difference in inflection. *You don't* in *you don't repent* also violates the pattern established in the rest of the

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

speech, as the usual marking by an -*s* is absent here as well. As in the first two articles analyzed here ("Health of Chillicothe" and "Rich"), the first-person singular is marked by -*s* (*I knows*). All in all, the deviating pattern of subject-verb agreement, emphasized through the repeated use of short clauses containing a second-person plural subject and verb and combined with inconsistencies in the pattern, adds to the impression of linguistic incompetence. A further strategy to underline this impression is the use of eye dialect in *wan't* 'want', *peniten's* 'penitents' and *your* 'you're'. A form that can be observed here but which is not present in the articles above is *a*-prefixing (*a*+verb+*ing*, see Cukor-Avila 2001: 96), which occurs in the clause *you sinnahs comes foh'rd an' holds your head too high a-comin'*. As it never occurs in the narrator's speech in any of the articles analyzed here, nor in the speech of the (white) people interacting with Black people, it comes to be associated with Black speech in this article. A grammatical form that links this article to the anecdote "Too Hasty" is the better construction: *bettah* is used by the preacher as part of a modal construction to give advice to his audience: "you bettah stay away".

The construction of the speaker as linguistically not competent and therefore also not educated is underlined further by the explicit description of his sermon as "less eloquent", but "certainly very practical". The use of a tool metaphor to describe the preacher's speech underlines this aspect of practicality, and additionally creates an image of directness and forcefulness: "The preacher struck nails square on the head as he hammered away". This is in stark contrast to eloquence and refinement and to an argumentation that is persuasive because of its rhetorical strength and not because of its simple and repetitive structure, emotionally appealing intonation and direct appeals. The repeated criticism leveled by the preacher at the Blacks attending the meeting, namely that they are not ready because they are "too peart" and "no mounahs" adds to the negative impression created of Black people in general. The reproach that they constantly look for models because they do not know how to pray paints a picture of Black people being insecure in religious matters ("you peeps fru you fingahs dis way, an' you cocks up you eahs to see who's makin' de bes' pray'r") and also of being insincere because instead of establishing their own individual connection with God they appear to be mainly interested in seeing what their neighbors are doing and in comparing the quality of the prayers. The message that is conveyed by the report is that Black people might be interested in religion, but that they are unable to practice it and that they are insincere because their preachers value loud displays of emotions and a shiny appearance over quiet and sincere repentance.

4.1 Metadiscourses on phonological forms

# **A NEGRO REVIVAL A Colored Moody who Wants "No Foolin wid de Lord"—"Better Whispah to de Lord dan Holler at de Debble."**

Mississippi Corr. [?] Commercial

We must give the reader a few specimens of a prayer and exhortation we heard in a revival meeting among the colored folks. A shining black preacher, glossy as a varnished beaver, gave us a characteristic article in this line. Beginning his prayer in a low and reverential voice, he addressed the Deity as "Thou" and "You" indiscriminately, and sometimes indulging in the doubtful grammar of "Thou Knoweth" and "You knows." Soon his words were uttered as a kind of wailing chant, with a prolonged sound in a higher key on emphatic words and syllables. The peculiar intonation, especially when the congregation would catch the key from the plaintive sounds, and unite with the preacher in a piteous moan, between words, gliding down from the dominant note to the minor third below, and dying through diminuendos into sobs and sighs. The effect was at times thrilling. Some parts of an exhortation to which we listened, however, while less eloquent, were certainly very practical. The preacher struck nails square on the head as he hammered away. For instance:

"Now, brethren and sisters, we wan't mounahs heah to-night. No foolin'. Ef you can't mouhn for your sins, don't come foolin' roun' dis altah. I knows ye. You's tryin' mighty ha'hd to be convarted 'thout bein' hurt. The Lord 'spises mockery. Sometimes you sinnahs comes foh'rd an' holds your head too high a-comin'. You come foah you's ready. You starts too soon. You don't repent; you's no mounah. Your foolin' wid de Lord. You comes struttin' up to de altah; you flops down on your knees, an' you peeps fru you fingahs dis way, an' you cocks up you eahs to see who's makin' de bes' pray'r. You's 'tirely too peart for peniten's. You's no mounahs. Ef you comes heah to fool, you bettah stay away. […]"

The second example of a report focusing on a Black preacher figure was published on July 4, 1885[67] , so ten years later than the article "A Negro Revival", under the heading "An Old Time 'Fo'th'" in the *Rocky Mountain News* (Denver, Colorado). As the title indicates, the author of the article reports on the Fourth of July celebrations by Black people and in this context also discusses and compares the patriotism of Black and white Americans. From the beginning, the superiority of white Americans is firmly established: Black men are described as imitators of white men and the comment that Black men are "naturally close observers" attributes their reliance on whites for guidance to their nature and thus to something that cannot change. Furthermore, Black men are described as being less patriotic because they care more about "powder, fire-crackers and yells" than about remembering the actual events of fighting (for example the battle of Bunker Hill). The comparison of Black men to small boys makes them seem childish, and the author points out several times that after the Civil War, Black people started celebrating "whenever an opportunity presented itself", which characterizes them as rather lazy and more interested in entertainment and other pleasures like eating good food than in contributing to the progress of the nation. Against this background, the figure of the Black preacher is described in more detail: He is called "Uncle Ephraim" and characterized by the author as "ignorant as a field hand" but nevertheless "shrewd and cunning". Furthermore, he is presented as

a person who wants to appear eloquent and dignified, like "a Roman senator". However, the direct quotation of the speech that he gives, in which he "utter[s] his patriotic thoughts mixed with religious sentiments", is supposed to ridicule his efforts and prove that he is in fact neither eloquent nor able to speak in a dignified manner.

The most salient linguistic form is again non-rhoticity, for the reason that it is already represented in the heading, where the word *fo'th* 'fourth' is separated from the rest of the heading through quotation marks to indicate a change of voice. As the text is about the Fourth of July celebrations and the (lack of) patriotism of Black people, this word prominently links non-rhoticity to the topic of the article and the group of people that are indexically linked to the form. However, non-rhoticity is again not represented consistently and several forms keep the <r> in the spelling (e.g.*'marked*, *fadder*, *wedder*). Another phonological form which is particularly emphasized in this article is the deletion of initial unstressed syllables: The Black preacher says *'dependence* instead of *independence* and therefore changes the meaning of the word to its opposite, so that he unintentionally creates humor when he says "To-day [...] am de declamation of 'dependence". This very sentence also highlights a grammatical difference (third-person singular *am*) and a lexical difference: the use of *declamation* instead of *declaration*. The words differ in only one sound, which invites the likely interpretation that the preacher mixed up the two words. This mix-up signals the linguistic insecurity of the speaker when it comes to Latinate words, which are associated with linguistic and rhetorical competence. The preacher's use of *unneccessarious* instead of *unnecessary* and *dissolutionary* instead of *revolutionary* underlines this impression of a lack of rhetorical skill and therefore creates humor through the contrast between the preacher's intention to appear eloquent and his failure to achieve it. Apart from these linguistic forms, voiced fricative stopping is also represented consistently in his speech (e.g. in *fadders*, *de* and *ob*) and in the song sung by everyone at the very end, final consonant cluster reduction and alveolar -*ing* are highlighted because they occur at the end of the lines (*eas'*, *wes'*, *mo'nin'*). There are several further grammatical forms marked in his speech: the demonstrative *them* (*dem heroes*), the past tense form of *fight*, which is once *fought* and once *fit* (indicating not only a different form, but also inconsistency in use), and the pastparticiple of *frost-bite*, which is *frost-bited* instead of *frost-bitten*. In the fifth line of the song sung in the end, an invariant *be* (derived from *will/would* deletion) occurs as well.

The final comment in the last paragraph emphasizes again the thoroughly negative attitude towards Black people: Their "scramble for the lunch baskets" makes them appear hungry like animals and their fighting also characterizes them as

uncivilized. That the preacher himself was involved in the fighting is the final step in destroying the picture of eloquence and dignity that he was anxious to portray and in providing "proof" of Black inferiority.

# AN OLD TIME "FO'TH" […]

Whatever a white man does a colored man must do. They are naturally close observers, and, being good imitators, never fail to do what they have seen done. It may be that the fires of patriotism do not burn so brightly in their breasts as in the bosoms of white American citizens. They may not grow as enthusiastic as the New Englander when the battle of Bunker Hill is alluded to, but the average colored citizen takes as great delight in a Fourth of July celebration as the small boy, who connects the occasion with powder, fire-crackers and yells.

In the years immediately succeeding the late unpleasantness the negroes, recognizing that they were freemen

# began to celebrate the fourth,

a day which there was never much fuss made over in the South. Christmas time being the holidays of holidays for whites as well as blacks. But after the war things were changed. The negro felt it his bounden duty to celebrate whenever an opportunity presented itself and celebrate he did. The Fourth in a great measure took the place of the barbecue, a place where the dusky brother was wont to disport himself and feast on "eatins' as good as de white folks."

Then there was the colored preacher who, always anxious to display his eloquence and air his oratory, made the Fourth an occasion to utter his patriotic thoughts mixed with his religious sentiments, and a Fourth of July celebration in the hands of the colored people was generally a cross between a camp meeting and a corn shucking.

[…] The orator of the county, the bright particular star in the black firmament, was Ephraim Jenkins,

### "Uncle Ephraim," as he was called.

"Uncle Ephraim" was a portly man, black as ebony and as ignorant as a field hand. But he was shrewd and cunning, had a retentive memory, a facile command of language, not the most choice, and could talk, or rather rhapsodize, for a day at a stretch.

### […]

## The orator of the day

arose with all of the dignity of a Roman senator. "To-day," he said, "am de declamation of 'dependence." After having made that remark, he paused to see what effect the startling announcement had created, and during the pause, wiped the perspiration form his shiny brow. He then resumed: As I 'marked befoah, dis am de declamation of 'dependence and our fadders—yo' fadders as well as my fadders—fought in de dissolutionary wah and marked de snow and ice of Valley Fo'ge wid blood. Some of my heah has had de same 'sperience in de wintah, de ground froze in de co'n field, de toes out'n de shoes and run down at de heel. Some ob yo' I spec is frost bited yet.

It am unneccessarious fo' me to tell de history ob dem heroes who fit at Bunkum hill and died, yes, died, fo' me an' fo' yo', like de blessed Lamb.

I'd like for de audience to sing befoh we 'sperse for 'freshments "You may bury me in de eas'."

[…]

"You may bury me in de eas', You may bury me in de wes', But I'll heah dat trumpet soun' in de mo'nin'. My eahs may change to clay, An' my tongue be waste away,

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

But I'll answer dat trumpet in de mo'nin, In de mo'nin', in de mo'nin', in de mo'nin' ob de Lawd. An' we'll all be togedder in de mo'nin'."

When the song had been sung the audience was dismissed by Brother Euripides Shands, a local light, and a scramble for the lunch baskets ensued. A portion of the day was spent as became patriots, but the balance of it was devoted to fighting, and "Uncle Ephraim," the orator, was one of a number who were slashed with razors. He was cut by the Rev. Euripides Shands in a dispute over a sister.

A further text type linking Black Americans to a specific linguistic repertoire is the short story. The example I will discuss here is entitled "Dat Deceptious Mule". It was written by A. T. Worden, a white author, and published on September 21, 1895[155] in the *Yenowine's News* (Milwaukee, Wisconsin). The difference to the anecdotes and the reports discussed above is that no claim is made that the story is based on real events or that it is to be regarded as a typical incident of events or people occurring in real life. Nevertheless, there are also striking similarities to the anecdote because the story consists almost entirely of direct speech, quoting the conversation between two Black men, one wanting to sell a mule and the other looking to buy one. For this reason, almost the whole text represents the two Black voices, interrupted only very few times by the narrator. Due to this large absence of a narrator, the two Black men are not explicitly described. However, they are characterized implicitly based on what they are saying and they are also portrayed in an accompanying illustration (Figure 4.14). The seller is presented as very optimistic and good-natured. He praises the mule in every possible way. He addresses the potential buyer as "Eldah Pokeberry", which suggests that the man is older and has a social position of authority. The illustration confirms this: The seller is rather young and the way that he leans on the mule to his left and on a stick to his right suggests a laid-back and confident attitude. He is wearing practical, comfortable clothes and a cowboy hat. The Elder, on the contrary, is standing upright, folding his hands in front of him and holding a cane and a briefcase under his arm, which makes him seem genuinely interested and ready for doing business. He wears a frock coat and a top hat and therefore appears more elegant than the seller. Despite these differences, however, the striking similarity between the figures is the way that their ethnicity is caricatured: Their skin color is black and their facial features, big and protruding eyes, ears and lips, are exaggerated to an extent that it makes them subject to ridicule and derision. This fits the story, which also mocks both Black figures. The first part is quoted below and shows how the seller praises the mule to the Elder in a greatly exaggerated manner. He puts special emphasis on the friendliness of the mule, which supposedly makes it an ideal family mule. However, in the second part of the story, the mule suddenly starts kicking aggressively

### 4.1 Metadiscourses on phonological forms

Figure 4.14: Two Black men, the seller of the mule and Elder Pokeberry, in the short story "Dat Deceptious Mule", published in the *Yenowine's News* (Milwaukee, Wisconsin) on September 21, 1895[155] , retrieved from Nineteenth-Century U.S. Newspapers

and proves to be the complete opposite of the description of the seller. The seller explains the behavior of the mule by saying that it "understan's talk" and that it was set off by the Elder's remark that "he kin cyarry double". This unrealistic explanation makes it clear that the seller lied from the beginning about the characteristics of the mule to get the Elder to buy it. In the end, the seller changes his mind about selling the mule and says "Yes sir, good-day. You can't buy dat mule. I'll keep him in de fambly". It is clear, however, that he just tries to save face because it is unlikely that the Elder is still interested in buying the mule after seeing how aggressively it behaves. The Black seller is therefore characterized as someone who is cunning and who cannot be trusted because his optimistic and good-natured manner is only put on to achieve a goal. Ultimately, it is not

the mule that is deceptive, but it is the Black seller of the mule, which reinforces negative stereotypes about Black people.

On a linguistic level, the Black men's voices are represented as very similar and the sheer number of forms marking their speech as different is again striking. Non-rhoticity is highlighted as a phonological form in the text because it is one of the few forms that are contrasted in the text with the 'neutral' voice of the narrator (*Eldah* vs. *Elder*).<sup>12</sup> However, as in other articles discussed above, it is not represented consistently (it is not marked e.g. in *dar* 'there', *burnt* and *air*). There are also instances of hyper-rhoticity in function words, e.g. in *ter* 'to', *er* 'a', *der* 'the' and in the content word *sorrer* 'sorrow' (in the second part of the story). Cases of elision are frequent as well: of final consonants (*da* 'that', *o'* 'of', *le'* 'let'), of initial unstressed syllables (*'count* 'account', *spose* 'suppose'). Final consonant clusters are reduced, as in *understan'*, *wouldn'*, *behine*, *stan'*, *tole* and *jess* 'just'. Stopping of voiced interdental and labiodental fricatives is marked, as is fronting of voiceless interdental fricatives (*da*, *dar*, *de*, *dis*, *dem*, *wid*, *dat*, *dey*, *ob*, *ober*, *mouf*, *fink*). Final -*ing* is realized by [ɪn] (as e.g. in *zoonin'*, *standin'*, *maunin'*, *durin'*, *alludin'*). The dress vowel is raised in *git*, as in representations of Black speech in articles discussed above, but in this story the vowel is also raised in *Giniral*. Three vowel forms which have not been marked in articles above are realizations of choice with a lower and fronter onset (*appintments*, *pint*), of square with a lower and backer onset or even a low monophthong (*dar* 'there', *har* 'hair') and of price with a monophthong, although this last form is restricted to the pronoun *ma* 'my'. A consonantal process that has not been represented in the articles discussed so far is the insertion of a palatal glide after /k/ in *cyarry*.

Regarding grammatical forms, there are numerous differences as well. With regard to subject-verb agreement, the third-person singular form of *be* is sometimes *am* (*dis am a fambly mule*), sometimes *are* (*dis ar de mule*) and sometimes *is* (*He's lonesome, dat mule is*). The suffix -*s* is used to mark first-person and secondperson singular (*I calls*, *yo' falls*) and third-person plural (*De chillun cracks*). Thirdperson singular is not marked, e.g. in *dat mule stan' an' sleep*. Past-tense forms are sometimes marked by inflection and sometimes they are not, e.g. in *Dar was de Stiggins' mule dat kick so fast dat de friction of his legs burnt de hair off an' made de air wam all aroun' dar*. There is a bare infinitive construction (*dat tried kick*), sometimes there is no do-support (*How much you want*), adverbs are not marked

<sup>12</sup>In the other cases, the comparison is implicit, which means that the author of the story assumes that readers compare the linguistic repertoire to their own repertoire or that of other articles in the newspaper.

### 4.1 Metadiscourses on phonological forms

(*win de money easy*), demonstrative *them* (*dem cohns*) and *this here* (*disher mule*) can be found, the negator *ain't* is used (*I ain't gwine to sell*) and negative concord occurs, that is the occurrence of multiple negative elements as in *dey couldn' git nobody to hold de drum*. Moreover, the genitive construction *dat one ob Lawson's* is used and there is an instance of perfective *done* (*you done heard about*). Another interesting grammatical form is the use of periphrastic *do* in an affirmative sentence where it does not seem to be used for emphasis: *I des tell yo'*. <sup>13</sup> Finally, the Black seller sometimes uses reduced sentences, for example *Kick, not him*, which consists only of a verb and the apposition *not him* to give information about the agent of the activity denoted by the verb. Reduced sentences like this evoke the value of simplicity, so that the speaker appears naive.

The combination of phonological and grammatical forms creates the impression of the Black speakers having a linguistic repertoire that deviates enormously from that of white speakers. The two Black speakers are, however, also marked as different from each other on a lexical level: The Elder Pokeberry uses words of Latinate origin and rather infrequent (*curb*, *volubility*) and formulates rather complicated and long phrases (*befo' yo' falls into de habit ob exaggeration*). This creates the impression that he aims to appear educated and that he wishes to emphasize his high social status as an Elder. However, as other phonological and grammatical "deviations" are present in his speech as well, his attempt has a rather comic effect because of the apparent incongruity between the lexical level and the other levels of speech. This effect can be illustrated by considering just one sentence: *yo' bettah curb de volubility ob yo' imagination befo' yo' falls into de habit ob exaggeration an' slops ober* combines phonological forms (non-rhoticity, fricative stopping) and grammatical forms (second-person singular -*s* and the better construction that also occurred in "Too Hasty" and "A Negro Revival"), which are associated with uneducated Black people, with infrequent lexical items of Latinate origin, which are associated with highly educated people. This particular combination also links the figure of the Elder in this story to the figure of the Black preacher in the report "An Old Time Fo'th" because they both occupy high social positions in the Black communities and try to adapt to this position

<sup>13</sup>Schneider (2015) finds "the use of periphrastic *did* with past time reference, in affirmative structures where neither an emphatic nor a habitual reading may apply" in BLUR (The Blues Lyrics Collected at the University of Regensburg), a corpus of African American Southern English containing blues lyrics from 1920 to 1969. He analyzes it as "an analytic, preverbal tense marker" and emphasizes the need for further research on this structure. In the structure *I des tell yo'*, found in the present short story, it is possible that *does* also functions as an analytic marker of present tense. However, as it occurs only once in the story, it is difficult to identify a pattern here.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

linguistically by using specific lexical items. Ridiculing this by showing how the presence of other linguistic deviations makes this attempt absurd and impossible establishes a relationship of superiority of the white authors to the Black figures that they describe or invent.<sup>14</sup>

# **Dat Deceptious Mule.**

"I understan' yo' Eldah Pokeberry da yo' wanter buy a nice peaceable mule ter ride to yo' appintments an' git dar on time an' all in one piece. Dis ar de mule wot you bin zoonin' for. I raised disher mule in ma own fambly. Dis am a fambly mule. Buttah wouldn' melt in dis mule's mouf. I calls him Giniral Grant 'count o'him being so quiet. Kick, not him. I des tell yo' Eldah he had cohns on his behine feet from standin' still so long an' I used ter cut dem cohns ever maunin' wid ma jack-knife an' dat mule stan' an' sleep all de durin' time. Dat wot I call a gentle mule, mon."

"Does you fink dat mule would be kind to de chillum?" asked the Elder.

"Now you's alludin' to de strong pint in dat mule. De chillun cracks pecan nuts on dat mule's heels wid a brick. He's lonesome, dat mule is, onless he got some chillun aroun'. Dar was de Stiggins' mule dat kick so fast dat de friction of his legs burnt de hair off an' made de air wam all aroun' dar. He would go out in de lot an' practice all alone an' Stiggins he made a bet dat de mule could kick de double drag on er base drum to the time of De White Cockade but dey couldn' git nobody to hold de drum else de mule win der money easy."

"Ma frien'," said the Elder slowly, "yo' bettah curb de volubility ob yo' imagination befo' yo' falls into de habit ob exaggeration an' slops ober." [...]

All in all, this example shows that the short story as a text type is similar to the text types analyzed above because it also contains direct representations of speech and thus creates social personae whose behaviors and appearances are linked to a specific set of linguistic forms. It is a particular characteristic of the short story "Dat Deceptious Mule" that it consists almost completely of dialogue and that the narrative voice is reduced to indicating who is speaking ("said the Elder"). The events of the story are instead told through the dialogues and illustrated by the accompanying visual elements and onomatopoeic words ("Rip! ping!! splum!!! bang!!!!"). The dialogic character in particular creates a similarity to the text type of the anecdote. However, the short story is also different because it is marked as entirely fictional: It does not claim to describe encounters between speakers that really happened (as in anecdotes) or to give realistic accounts of speakers (as in reports citing actual speakers). Whereas this reduces the authenticity that readers are invited to attribute to the representation, it increases the readers' readiness to accept exaggerations or unrealistic depictions

<sup>14</sup>In "An Old Time Fo'th", the author is not explicitly named, but the text leaves no doubt that the perspective taken on Black people is not an inside but an outside one – formulations like "The negro felt it his bounden duty to celebrate [...]" show how the author distances himself from the group he describes. The author of this short story is A. T. Worden, a white man who lived in New York and had a career as a writer and as a Baptist minister. He is also the author of several poems using racist imagery and stereotypes (Young 2008: 62–63).

### 4.1 Metadiscourses on phonological forms

(like the mule understanding human speech). Due to its greater length, the short story also has more room to create a fictional world and a story that happens within this world, which allows for more detailed characterizations and much longer representations of speech.

An article type which rests more heavily on illustrations than on text is the cartoon. I will discuss two examples here which illustrate how visual elements are linked to linguistic forms which are constructed as part of Black speakers' voices. The first example is headed "Dangerous" and it was published in the *Milwaukee Journal* (Wisconsin) on February 12, 1897[165] (see Figure 4.15). The heading, which is printed in big capital letters, raises the expectation that something dangerous is depicted below. This has the effect that the reader is invited to associate the two Black people in the illustration with the characteristic of being dangerous and it shows that the cartoon plays on the stereotype of the "brute", a "barbaric black out to raise havoc" (Bogle 1990: 13). The dialogue underneath

Figure 4.15: A cartoon depicting two stereotypical Black figures and a mule, published in the *Milwaukee Journal* (Milwaukee, Wisconsin) on February 12, 1897[165] , retrieved from Nineteenth-Century U.S. Newspapers

the illustration, however, makes it clear that it is in fact the mule, which is rather in the background behind the left figure, that the adjective *dangerous* relates to. The owner of the mule, Erastus, asks the woman, Mrs. Johnsing, what she thinks about the mule that he had just clipped. She answers that it is dangerous to clip a mule in winter time because the mule is likely to get kleptomania. The pun is that while it makes sense that Mrs. Johnsing assumes that the mule might develop an illness because it could get too cold in winter with its clipped mane, she uses the term *kleptomania* for the possible illness because the phonological form of the word is reminiscent of the phrase *clipped mane*. As kleptomania is in fact a psychological condition in which people have the impulse to steal items and cannot resist it, Mrs. Johnsing's unintended suggestion that the mule could develop this condition leads to the absurd image of a mule being dangerous because it is likely to become a thief. After reading the dialogue, it therefore becomes clear to the reader that the heading "Dangerous" mocks the word choice by the Black figure and draws attention to the inability of Black speakers to use specialist vocabulary.

The ridicule of Black Americans is also underlined by the visual elements: As in the illustration accompanying the short story "Dat Deceptious Mule", the depiction of the Black speakers focuses on their physiognomy and highlights the color of their skin and especially the eyes and the lips, which are so big that they appear unnatural and make the two Black people appear less human. The female Black person, Mrs. Johnsing, has a round figure and her simple dress and head scarf are practical clothes, which, in combination with the basket and her worry about the mule getting sick, clearly reference the mammy stereotype again. The man is round and stout and small and his white hair suggests that he is rather old. His clothes are simple as well, with patches on his pants indicating that he is not particularly wealthy.

The mocking portrayal of the Blacks as simple-minded is supported by the representation of their voices in direct quotations, which exhibits several typical features associated with Black speech. Non-rhoticity is marked in *mawning*, *bettah*, *wintah*, *fus'* and the form *bettah* links non-rhoticity to the grammatical level again because as in three of the articles analyzed above it occurs in a modal construction used to give advice (*Bettah look out*). Final consonant cluster reduction (*fus'*), fronting of the voiceless interdental fricative (*fink*, *fing*), stopping of voiced labiodental fricatives (*ob*), a raised dress vowel in *get* (*git*) are represented as well, but the phonological form emphasized the most in this cartoon is the use of alveolar *-ing*. Mrs. Johnsing's speech exhibits this form in *clippin'*, but there is additional attention drawn to the form by giving her the name Mrs. Johnsing, which indicates a hypercorrect form: The final /ən/ in *Johnson* is reinterpreted as

### 4.1 Metadiscourses on phonological forms

a realization of an -*ing* suffix and consequently changed to the supposedly more correct realization /ɪŋ/. This adds to the implicit characterization and mockery of the Black woman as trying to appear more educated than she actually is: She not only fails to use a specialist term for a sickness (a Greek loanword associated with a high degree of education), but she is also shown to fail in her attempt to use a 'correct' phonological form.

A prominent element of the story is therefore the mule – it also creates an intertextual relation between this cartoon and the short story "Dat Deceptious Mule". Ferris (2009: 332) writes:

The lives of African American workers and the mule are intimately linked in every period of southern history. As slaves before the Civil War and as tenant farmers afterward, blacks worked with mules; the ubiquitous team of mule and African American driver was essential to the southern economy.

In the short story, this close relationship is used to create the impression of a similarity in character traits between the mule and Black men. The main twist of the story is that the animal has the human characteristic of understanding language and that it has feelings which can be hurt so that it starts kicking when Elder Pokeberry suggests that it can carry double. By attributing human characteristics to the mule, it is implicitly suggested that characteristics of the mule can apply to humans as well, more particularly to Black men, with whom they are in a close relationship. In the story, the mule is described as sturdy, patient and gentle, but it turns out that it has an aggressive character as well. In the cartoon "Dangerous", the suggestion that the mule can develop kleptomania might be unintended and absurd, but if characteristics of the mule are understood as applying to Black men as well, it can be read as an implied suggestion by the creator of the cartoon that Black people are likely to become thieves, especially when they are stripped of their resources (a metaphorical reading of the clipping of the mane). Both the cartoon and the short story therefore contribute to the construction of a very negative image of Black Americans and to a process of othering, in which they are constructed as physiologically, psychologically, socially and not the least linguistically markedly different from white people.

The second example of a cartoon linking *bettah* to Black Americans was published in the *Philadelphia Inquirer* (Pennsylvania) on August 13, 1899[175] (see Figure 4.16). The cartoon does not have a heading and the caption under the illustration is very small, which has the effect of foregrounding the visual element because the readers' attention is drawn to the illustration first, simply because it

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

requires some effort to read the small print of the text underneath. The illustration shows a Black boy and a white boy who encounter each other while they walk their dogs in a rural area. What is most noticeable is the difference in size between the boys, but especially between the dogs: The Black boy is smaller than the white boy, and while the white boy's dog is as big as the boy himself, the Black boy's dog is tiny and thin. The second striking difference is a difference in posture: The white boy and his dog stand still and upright, while the Black boy's dog is moving towards them and the Black boy is trying to hold him back, with his big eyes and slightly open mouth indicating that he is scared. A difference in social status in indicated through their difference in dress. The white boy is dressed more elegantly than the Black boy: He is wearing well-fitted knee-pants and black stockings and his hat is fancier than the practical and simple hat worn by the Black boy. The impression created by these visual elements only is that the Black boy is intimidated by the white boy and his dog and tries to get away. However, the humor of the cartoon is created by the textual element, which contradicts the impression created by the image. The text reads: "Yo' had bettah get away quick, white boy, 'cause I can't hold dis yer dawg by mine much longer". As the voice addresses the conversational partner with "white boy", it is clear that the Black boy is speaking, but contrary to the assumption that he fears the white boy and feels inferior to him, he is now presented as feeling superior and being in

Figure 4.16: A cartoon depicting a Black and a white boy with their dogs, published in the *Philadelphia Inquirer* (Philadelphia, Pennsylvania) on August 13, 1899[175] , retrieved from America's Historical Newspapers

### 4.1 Metadiscourses on phonological forms

a position in which he can protect the white boy. Given the objective difference in size between the dogs, the assumption that the small dog poses a threat to the white boy and his big dog seems ridiculous, so that the cartoon mainly conveys that Black people, embodied here by the Black boy, have a self-confidence that is unjustified given the situation that they are in, and that they try to cover their actual inferiority with boastful talk.

Linguistically, the Black boy's speech is marked by non-rhoticity in *bettah*, but non-rhoticity is not marked in *yer* and *longer*, suggesting either an inconsistent use or an inconsistent representation of the form. Another phonological form is the insertion of /j/ and dropping of /h/ in *yer* 'here' and, as found frequently in the articles above, the voiced interdental fricative is replaced by a stop (*dis*) and the initial unstressed syllable is elided in *'cause*. As in four of the articles analyzed above, *bettah* is used again in a modal construction, but this time the modal is *had better*, which according to van der Auwera et al. (2013: 131) is the historically older form. Furthermore, there are three grammatical forms which were also present in "Dat Deceptious Mule": a double demonstrative consisting of a demonstrative determiner and a locational adverb (*dis yer dog*),<sup>15</sup> an unmarked adverb (*quick*) and an unusual genitive construction (the phrase *by mine* is used to mark possession of the preceding noun). All in all, the large number of marked forms in the one sentence spoken by the boy underlines his inferiority to the white boy, who does not even need to reply to the Black boy to assert his superiority.

The two cartoons show how powerfully visual and textual elements interact in the process of stereotyping. The short representations of speech highlight particular linguistic forms and the voices created in this way are linked to images, which also highlight particular aspects of the outer appearance, manner and attitudes of the figures. Furthermore, the visual element draws the readers' attention to the cartoon and makes it likely that they will read the dialogue underneath. Another reason for the popularity of the cartoon is that readers expected to be entertained by it and that it even appealed to readers who had no interest in (or who were not capable of) reading longer texts.

The last two text types that I will analyze here to show how linguistic forms, among them non-rhoticity, become indexically linked to Black speakers are a poem and a soliloquy (a dramatic monologue that usually occurs in a play but

<sup>15</sup>Diessel (1999: 74) writes that such forms are a product of a grammaticalization process in which locational deictics which were used adnominally to intensify a demonstrative determiner (as in *this guy here*) become part of the demonstrative form. Diessel's examples are not taken from English (varieties), however, but from Afrikaans, Swedish and French. Nevertheless, the Swedish form *det här hus-et* 'the/this here house-the' is parallel to the form associated with Black American English in the cartoon analyzed here (*this here dog*).

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

appears as an isolated text in the newspaper). Like anecdotes, short paragraphs, short stories and cartoons, these text types are included in newspapers primarily for entertainment and not for information or critical discussion of current affairs. The soliloquy is entitled "Cato's Soliloquy" and it appeared as a part of the section "Denver Postscripts" in the *Denver Evening Post* (Colorado) on July 2, 1899[173] . The title indicates an intertextual relationship with Joseph Addison's tragedy *Cato* (1712), which features a soliloquy in which Cato, the hero, ponders suicide. The soliloquy in Addison's play in turn contains intertextual links to Shakespeare's *Hamlet*. In the poem which appeared in the newspaper, the relationship with *Hamlet* is obvious as well because the opening line is almost the same as the one in Hamlet's soliloquy. The statement "To be, or not to be, that is the question" is turned into "To steal, or not to steal? Dat am de question". This line is crucial: It not only links the soliloquy to *Hamlet*, but through this link it also establishes a reference point for comparison. On the one hand with regard to content: By replacing the verb *be* with the verb *steal*, the existential question of being, of life or death, is turned into the more profane question of stealing. On the other hand with regard to language: By indicating TH-stopping orthographically in *Dat* and *de* and by changing the third-person singular form of be to *am*, differences in the linguistic repertoire between this soliloquy and Hamlet's soliloquy are highlighted. Given that form–meaning links between these two linguistic forms and Black people had been circulating widely for more than twenty years, it is likely that these forms in the first line already evoke the voice of a Black speaker even before his ethnicity is made explicit in the fourth line when he talks about "de white folks' moufs". The content of the soliloquy expresses a negative view on Black Americans: The Black man ponders the question whether he should steal a watermelon from a patch owned by white people. His main argument against stealing is his conscience, which he regrets having, and he makes religion responsible for it: "'Tis conscience dat makes cowards ob us coons, / Especsh'ly dem dat's jes' done jined de church / An' had deir brack souls washed as white as snow". Whiteness is associated with innocence and moral and Christian behavior here and the Black man's evaluation of this behavior as cowardice serves to underline the moral corruption of Black people, the need to educate them, and to introduce them to religion. This undertaking is presented as very difficult, however, as the humor of the soliloquy rests on an argument developed by the Black speaker which justifies stealing the watermelon precisely *because* of religious reasons. He argues that the preacher said that he loved watermelons and since the preacher is "de agent ob de Lawd", pleasing him by sharing the stolen watermelon with him becomes an act of pleasing God. He therefore "confiscates" the watermelon "in de Mastah's name" and plans to "invitate" the

### 4.1 Metadiscourses on phonological forms

minister to share the watermelon with him. What is thus constructed here is an inability of Black people to control their impulses and their behavior. Especially the depiction of their craving of food, emphasized through the use of onomatopoetic words ("Yum! Yum! [...] Tunk-tunk-tunk!") and the verb *devour* in the last line of the soliloquy, is used as a means to characterize them as animalistic and uncivilized. This uncontrollable craving is ultimately presented as the cause for stealing (and not actual hunger). The label used by the Black voice to designate himself and Black people like him is "coons", which establishes a link to stereotype of the coon popularized at the end of the nineteenth century by so-called "coon songs". Cox (2011: 15), in her study on the creation of the south in American popular culture, writes:

The term "coon" was not used to describe blacks until the 1880s and can be attributed to the popularity of coon songs, which became a trend in music publishing that lasted through the first decade of the twentieth-century. Whites associated southern blacks with eating raccoons, and in coon songs they also became known as chicken-thieving, watermelon-eating, razorwielding oafs. These tunes were enormously popular in the decade of the 1890s, not surprisingly during the period of heightened racial violence nationally. In that decade, over 600 coon songs were produced as sheet music and performed in music halls and vaudeville shows around the country. In effect, coon songs expressed American racism and were important to popularizing black stereotypes.

The Black man speaking in the soliloquy conforms to the stereotype of the coon and it clearly marks him as inferior to whites, whose superiority is also underlined here by the fact that the intertextual reference to Addison's *Cato* and Shakespeare's *Hamlet* can only be recognized by educated people who are familiar with British literary classics. The author, who is most likely white, draws on white cultural knowledge to exclude the objects of derision and ridicule from the group of people who can achieve a full understanding of the text and to emphasize the importance of such knowledge as a uniting factor for this group of educated, white Americans.

With regard to the linguistic forms represented as being part of the Black speaker's repertoire, they are largely similar to the forms found in other articles, e.g. in "Dat Deceptious Mule". What is again noticeable is the combination of non-rhoticity with the hyper-rhotic form *shadder* 'shadow'. As in the other articles, non-rhoticity is also not consistently marked here, which is particularly striking in the line *Come to mah a'ms an' lie again mah heart* because it is marked

in *arms* but not in *heart*, which is in close proximity. Two phonological forms which, of all the articles containing *bettah* analyzed so far, have only occurred in "Dat Deceptious Mule", are the monophthongization of price in *mah* 'my' (but not in other lexical items like *moonlight* or *night*) and the lower onset of choice (*jined*). An unusual phonological form, which is not found in any of the articles above, is the replacement of /l/ by /r/ in *brack* 'black' and *bressed* 'blessed'. What is most striking on the lexical level is the phonological form of *melon*, which is represented by *millyun*, a spelling which indicates a raised dress vowel and the insertion of a palatal glide before the unstressed vowel. As the form occurs several times in the soliloquy and as the spelling differs considerably from *melon*, it serves to emphasize the linguistic differences of the Black speaker to white speakers. Another noticeable difference is the malapropism *invitate* 'invite', which is reminiscent of malapropisms found in other articles (e.g. *unneccessarious* in "An Old Time Fo'th") and serves to mark the speaker as uneducated.

On the grammatical level, the soliloquy exhibits several forms next to *am* being highlighted in the first line as a third-person singular form of be. First of all, thirdperson singular *is* occurs as well (*inside him is dat conscience*), and *am* is also used to mark third-person plural (*all de fruits ob earth am jes' de Lawd's*). With regard to marking agreement on other verbs, there is first-, second- and third-person singular -*s* (*I gibs*, *you's a beauty*, *'Tis conscience dat makes*), but unmarked firstand third-person singular forms occur as well (*if I take*, *Ol' Mastah Petah tell me*). Secondly, *bettah* occurs again as part of a modal construction (*he bettah make a sneak*). Thirdly, negation with *ain't* and negative concord is used (*dat ain't no place fo' niggahs*). Fourthly, there is a case of perfective *done* (*dem dat's jes' done jined*). And lastly, *a*-prefixing (*stop a hunchin'*) and the reflexive pronoun *heself* increase the impression of a text which constructs the voice of a Black speaker as deviating enormously from the 'neutral' voice of the surrounding newspaper articles.

Overall, what makes this text stand out is the contrast between the expectations raised by the title and the form of the soliloquy and its actual content and form. The plays *Cato* and *Hamlet*, which are an important part of the literary canon of the time and therefore constitute an element of white cultural knowledge, are used as points of reference, understood only by educated white speakers who have the necessary cultural knowledge. Those white speakers are the implied audience because they are able to understand the humor and to confirm their superiority over Black people based on the text linking linguistic differences to much more fundamental differences in educatedness, degree of civilization and morality.

# CATO'S SOLILOQUY.

To steal or not to steal? Dat am de question; Whedder to let dat watermillyun lie An' slumbeh in de moonlight in de patch An' sabe its sweetness foh de white folks' moufs, Or tell mah conscience to behabe itself An' stop a hunchin' at me dis-a-way, An' sneak 'roun' froo de shadder ob de trees Wha' not an eye can see me but de Lawd's, An' separate dat millyun from de vine. De Lawd! Ah! dar's de rub, fo' mebbe He Would chalk it down again dis niggah, so Dat when de trumpet soun's an' I arise On golden wings an' light beside de gate Ol' Mahstah Petah tell me to go 'long, An' smack me on de trousehs wif his boot, An' say dat ain't no place fo' niggahs dat Doan' propagate deir own ripe millyuns 'stead Ob swipin' dem from other folkses' patch. 'Tis conscience dat makes cowards ob us coons, Especsh'ly dem dat's jes' done jined de church An' had deir brack souls washed as white as snow. Dar lies dat millyun in de moonlight, an' Hyur stands dis niggah hidin' 'hin' de fence, An' hyur inside him is dat conscience dat's A tellin' him he bettah make a sneak Away from hyur an' git down on his knees An' ask de blessed Lawd to pahdon him Fo' eben thinkin' 'bout fo'bidden fruit. But stay! Las' night I hea' de preacher say Jes' afteh prayehs dat nex' to de Lawd He loved a big fat watermillyun wif A heart as red an' temptin' as de lips Dat o'naments de face ob coal brack wench. He am de agent ob de Lawd, an' if I take dat millyun an' divide wif him I gibs it to de Lawd, an' wha's de sin Returnin' him jes' what He grow heself, Fo' all de fruits ob earth am jes' de Lawd's. Come hyur, ol' millyun, go along wif me! I confiscates yo' in de Mastah's name! Yum! yum! but you's a beauty! Tunk-tunk-tunk! An' mighty ripe, too, 'cordin' to de soun'. Come to mah a'ms an' lie again mah heart Twell I can tote yo' home, an' den I'll go An' invitate de ministeh to come An' ask a blessin' on yo', an' den help Devou' yo' in de bressed Mahstah's name!

The last article analyzed here that links non-rhoticity to Black speakers is the poem entitled "When de Co'n Pone's Hot" and it is different from all other articles analyzed so far because it was written by a Black poet, Paul Laurence Dunbar. In the two databases there are ten articles which contain a version of the poem.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

The first one was published in the *Philadelphia Inquirer* on November 12, 1893[134] . It is indicated that it has been published before in the *Chicago Record*, but this newspaper is not part of the databases used for the present study. The following seven articles were published three years later, in 1896, and in four of them the poem is used as a part of the same advertisement (published in the *State*, in Columbia, South Carolina). Two articles were published in 1898, but they do not contain the whole poem, but only parts of it, and in one case, this part is quoted within a longer report. The following analysis will focus on three versions of the complete poem, the first one published in 1893, the second one in the *Penny Press* (Minneapolis, Minnesota) on July 25, 1896[162] , and the third one in the *Omaha World Herald* (Nebraska) on October 11, 1896[163] (see Table 4.3). The third version is introduced by the following comment:

# NEGRO POET IS BORNE **Mr. Paul Lawrence Dunbar Sings of "de Co'n Pone"**

Mr. Paul Lawrence Dunbar has been until recently an elevator boy in Dayton, O. While engaged in the ups and downs of life in that capacity he has cultivated his poetical talents so successfully that his verse has found frequent admission into leading magazines. At last a little collection of these verses reached William Dean Howels [sic], and Mr. Dunbar's star at once became ascendant. He is said to be a full-blooded negro, the son of slave parents, and his best work is in the dialect of his race.

This introduction describes the increasing popularity of Dunbar as a poet, which was particularly helped by a positive review by the white and well-known author William Dean Howells, and his ethnicity is explicitly emphasized through the heading, where he is labeled a "negro poet", and in the text, where he is called "a full-blooded negro". The introduction also draws attention to the linguistic repertoire used in the poem by mentioning that it is an example of a poem "in the dialect of his race". This statement already links the forms of the poem to Black speakers. Added to this is a positive evaluation of the representation of Black speech in the poem because his dialect poems are judged to be "his best work". This evaluation takes up Howells' opinion expressed in the review of Dunbar's collection of poems *Majors and Minors* (published in 1895). The review appeared in the magazine *Harper's Weekly* on June 27, 1896 (Nettels 1988: 80). The introduction therefore shows that the poems published in 1896 were related to a well-known and highly praised poet, whereas the same poet had not been famous when the first version of the poem appeared in 1893.

The content of the poem is the same in all three versions, except for the possessive pronoun in the second to last line, which is *my* in the third version and *yo'* in the first and in the second version. The poem describes the calming and soothing effect of corn pone (a type of corn bread). The stanzas are about feelings of instability ("When de worl' jes' stahts a-spinning"), about being held back

### 4.1 Metadiscourses on phonological forms


Dey is times in life when **nacher** Seems to slip a cog an' go, Jes' a-rattlin' down creation, Lak **a** ocean's **ober**flow; When **the** worl**d** jes' sta**'**ts a-spinnin' Lak a picanniny's top, An' yo' cup **ob** joy is brimmin' Twel**l** it seems about to slop An' yo feel jes' **lak** a **racer**, Dat is trainin' **fo'** to trot— When yo' mammy ses de blessin' An' de co'n pone's hot. When you se**'** down at the table, **Sort** o' weary lak an' sad, An' **youse** jes' a **lettle tired** An' perhaps a **leatle** mad— How yo' gloom tu'ns into gladness, How yo' joy **dribes** out de doubt— When de **oben** do' is opened, An' de smell comes po'in' out; Why, de 'lectric light**s ob** heaven Seems to settle on de spot, When yo' mammy ses de blessin' An' de co'n pone's hot. When de cabbage pot is steamin' An' de bacon good an' fat, When de chittlin**'s** is a **sputtr'in'** So's to show yo' **whar** dey's at— **Tek** away yo' sody biscuit**s**, **Tek** away yo' cake an**d** pie, **Fo'** de glory time is comin', An' **its** 'pr**oa**ching **bery** nigh, An' yo' want to jump an' hollah, **Do' yo'** know **'yo** bettah not, When **yo'** mammy ses de blessin' An' de co'n pone's hot. I **hab heerd ob** lots **ob** sermons, I **hab heerd ob** lots **ob** prayers, An' I've listened to some singin' Dat **hab took** me up de stairs **Ob** de **g**lory lan**d** an' set me Jes' below de Mast**ah**'s **throne** An' **hab** lef my **heart** a-singin' In a happy aftah tone. But dem **words** so **softly** murmured Seem**s** to **touch** de softes' spot, When **yo'** mammy ses de blessin' An' de co'n pone's hot. —Chicago Record. This version was published in the *Philadelphia Inquirer* (Philadelphia, Pennsylvania) on November 12, 1893[134] .

Dey is times in life when **Nature** Seems to slip a cog an' go, Jes' a rattlin' down creation, Lak **an** ocean's **over**flow; When **de** worl**'** jes' sta**h**ts a-spinnin' Lak a picanniny's top, An' yo' cup **o'** joy is brimmin' **'**Twel it seems about to slop An' yo' feel jes' **like** a **racah**, Dat is trainin' **fur** to trot— When yo' mammy ses de blessin' An' de co'n pone's hot. When you se**t** down at **de** table, Kin' o' weary lak an' sad, An' **yo'se** jes' a **little tiahed** An' perhaps a **little** mad; How yo' gloom tu'ns into gladness, How yo' joy **drives** out de doubt When de **oven** do' is opened, An' de smell comes po'in' out; Why, de 'lectric light **o'** heaven Seems to settle on de spot, When yo' mammy ses de blessin' An' de co'n pone's hot. When de cabbage pot is steamin' An' de bacon good an' fat, When de chittlin**s** is a **sputter'n'** So's to show yo' **whah** dey's at; **Take** away yo' sody biscuit, **Take** away yo' cake an**d** pie, **Fur** de glory time is comin', An' **its** 'pr**o**ching **very** nigh, An' **you** want to jump an' hollah, **Do you** know **yo'd** bettah not, When **yo'** mammy ses de blessin' An' de co'n pone's hot. I **have heard o'** lots **o'** sermons, An' I've **heard o'** lots **'** prayers; An' I've listened to some singin' Dat **has tuck** me up de stairs **Of** de **G**lory-**L**an**'** an' set me Jes' below de Mast**ah**'s **th'one** An' **have** lef' my **hawt** a singin' In a happy aftah tone. But dem **words** so **sweetly** murmured Seem to **tech** de softes' spot, When **yo'** mammy ses de blessin' An' de co'n pone's hot. —Lawrence Democrat. This version was published in the *Penny Press* (Minneapolis, Minnesota) on July 25, 1896[162]

WHEN DE CO'N PONE'S HOT WHEN DE CO'N PONE'S HOT Here is one of Mr. Dunbar's dialect poems entitled, "When de Co'n Pone's Hot:" Dey is times in life when **nature** Seems to slip a cog an' go, Jes' a-rattlin' down creation, Lak **an** ocean's **over**flow; When **de** worl' jes' sta**h**ts a-spinnin' Lak a picanniny's top, An' yo' cup **o'** joy is brimmin' **'**Twel it seems about to slop An' yo' feel jes' **lak** a **racah**, Dat is trainin' **fu'** to trot— When yo' mammy ses de blessin' An' de co'n pone's hot. When you se**t** down at **de** table, Kin' o' weary lak an' sad, An' **you'se** jes' a **little tiahed** An' perhaps a **little** mad; How yo' gloom tu'ns into gladness, How yo' joy **drives** out de doubt When de **oven** do' is opened, An' de smell comes po'in' out; Why, de 'lectric light **o'** heaven Seems to settle on de spot, When yo' mammy ses de blessin' An' de co'n pone's hot. When de cabbage pot is steamin' An' de bacon good an' fat, When de chittlin**'s** is **a-sputter'n'** So's to show yo' **whah** dey's at; **Take** away yo' sody biscuit, **Take** away yo' cake an**'** pie, **Fu'** de glory time is comin', An' **it's** 'pr**o**ching **very** nigh, An' **yo'** want to jump an' hollah, **Do you** know **you'd** bettah not, When **you'** mammy ses de blessin' An' de co'n pone's hot. I **have heerd o'** lots **o'** sermons, An' I've **heerd o'** lots **o'** prayers; An' I've listened to some singin' Dat **has tuck** me up de stairs **Of** de **G**lory Lan**'** an' set me Jes' below de Mahst**er**'s **th'one** An' **have** lef' my **haht** a singin' In a happy aftah-tone. But dem **wu's** so **sweetly** murmured Seem to **tech** de softes' spot, When **my** mammy ses de blessin' An' de co'n pone's hot. — This version was published in the *Omaha*

*World Herald* (Omaha, Nebraska) on October 11, 1896[163] .

("An yo' feel jes' lak a racah / Dat is trainin' fu' to trot"), about being tired and sad ("When you set down at de table, / Kin' o' weary lak an' sad") and about having to contain emotions ("An' yo' want to jump and hollah, / Do you know you'd bettah not?"). But all these negative feelings go away "When yo' mammy ses de blessin' / An' de co'n pone's hot", two lines which are repeated at the end of each of the four stanzas. The smell of corn pone causes "joy" that "drives out

.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

de doubt", and in the last stanza, the mammy's blessing is even described as more touching and important than all sermons, prayers and singing: "But dem wu's so sweetly murmured / Seem to tech de softes' spot". Celebrating corn pone means celebrating one of the basic elements of southern foodways.<sup>16</sup> Edge (2009: 98) finds that

Corn – and particularly corn ground into cornmeal – makes up the second half of the southern food pantheon, with cornbread clearly standing as the region's staff of life. Known by various names – including spiderbread, pone, suppone, hot-water cornbread, dog bread, cracklin' cornbread, and hoecake – cornbread is the most elemental of southern foodstuffs. It serves as a totem of identity as well as a marker of class standing.

Even though corn pone is associated with poverty and the lower classes, it is celebrated in the poem precisely because of its important function as an essential element of nutrition and its connection to the fulfilment of other basic needs: of being cared for by a mother and the comfort provided by religious faith. The poem therefore illustrates the claim that corn pone "serves as a totem of identity" for poor southern (in this case Black) people.

The spelling of *corn pone*, in which the <r> is replaced by an apostrophe, prominently links the indexical values associated with the lexical item to nonrhoticity. As in other articles analyzed above, this puts particular emphasis on non-rhoticity as a form associated with Black American speakers. In one case, /r/ is also marked as absent in intervocalic position (*po'in* 'pouring'), like in *mah'ied* in the anecdote "Rich". What can also be observed here is that non-rhoticity is not represented consistently. However, the existence of different versions of the same poem makes it possible to compare the representations of non-rhoticity and in Table 4.4 I have listed those forms which have been represented differently in the versions (representations of rhotic forms are marked in grey). What can be observed is a major difference between the first version and the 1896 versions. Non-rhoticity is already represented in the 1893 version, but the degree of non-rhoticity has increased in 1896. This suggests that an effort has been made to represent non-rhoticity more consistently, and that it was probably Dunbar himself who made the changes when he included the poem in his collection in

<sup>16</sup>The term *foodways* is defined by Edge (2009: 97) as "the study of what we eat, as well as how and why and under what circumstances we eat it". Studying foodways thus encompasses several aspects, among them food events, food processes and "aesthetic realms that touch upon the world of food (country songs about food, quilts raffled at community fish fries, literary references to eating)".

### 4.1 Metadiscourses on phonological forms

1895. It is not possible, however, to determine with any certainty who decided on the exact form of the printed version (it is also possible that the editor of the newspaper changed the spelling), but it is not likely to be a coincidence that the degree of non-rhoticity increases after Dunbar became famous as a poet. It is striking that despite this increase of non-rhotic forms the two non-rhotic forms in the early version are represented as rhotic in one of the later versions (*fo'*/*fur* and *Mastah*/*Mahster*). This observation, combined with the presence of rhotic forms in all three versions, suggests that accuracy and consistency was not the goal in representing non-rhoticity. A further change observable in Table 4.4 that relates to the phoneme /r/ is the elision of pre-vocalic /r/ following a consonant in both 1896 versions (*throne* becomes *th'one*).


Table 4.4: Differences in the representation of non-rhoticity in Dunbar's "When de Co'n Pone's Hot".

With regard to other linguistic forms, interesting changes can be found as well. The most striking one is the complete absence of the stopping of voiced labiodental fricatives in the 1896 versions. For example, *oberflow*, *dribes* and *oben* become *overflow*, *drives* and *oven*. The consistency of this change suggests that for some reason Dunbar has chosen not to link this form to Black speakers anymore, when he published the poem in his collection *Majors and Minors* (the version which was most likely the basis for the two 1896 versions). Other changes are more sporadic, sometimes because they are restricted to a few lexical items in the first place. For example, *tek* becomes *take*, suggesting that a monophthongal face vowel is not represented anymore, but *tek* is also the only instance of such a monophthongal variant in the 1893 version. The word *little* is spelled both <lettle> and <leatle> in the 1893 version, but this represented change of vowel quality is restricted to this lexical item and also not represented in the later versions. Instances of eye dialect have been changed in both directions: *nacher* becomes *Nature* and *nature*

in the later versions, but *took* was also changed to *tuck*. On a grammatical level, the third-person singular form *hab* 'have' (*dat hab took*) becomes *has* (*dat has tuck*) in the later versions and the third-person plural -s is not represented in the 1896 versions (*But dem words so softly murmured* / *Seems to touch de softes' spot*). Another difference can be observed in the modal better construction: The modal is changed from *bettah* (as found in most articles above) to*'d bettah* (the contracted form of *had bettah*, the modal used in the cartoon with the two boys and their dogs). The past participle *heerd*, on the other hand, has been changed to *heard* in only one of the later versions.

Linguistic forms that are represented to (roughly) the same extent in all three versions are forms which have also been found in articles analyzed above: Voiced TH-stopping, final consonant cluster reduction, final consonant deletion, thirdperson plural -*s*. One interesting form is the price vowel because it is represented as containing a monophthong in *like* (*lak*) but not in any other words containing the vowel, not even in *my*, which has been spelled *ma* or *mah* in "Dat Deceptious Mule" and in "Cato's Soliloquy". So on the one hand, attention is drawn to the vowel through the parallelism in the fourth and sixth lines of the poem, but it is also restricted to this context. In general, it is noticeable, however, that the number of deviant forms is lower in the poem than in other articles, even more so in the later versions because no cases of labiodental fricative stopping are represented. This becomes especially obvious when comparing the poem to the short story "Dat Deceptious Mule" or to "Cato's Soliloquy". This creates the impression that while the voice of a Black speaker is created through the representation of linguistic forms, the number of forms is kept below a level where it would invite mockery and ridicule simply because of the frequency of forms. Taking into account that the poet is Black and that he celebrates a basic element of southern (including Black) foodways, it is rather the case that the linguistic repertoire of Blacks is positively evaluated. The corn pone as much as the linguistic forms might be associated with poverty and a lower social standing, but for Black people they are also indexically linked to the comfort of home and family and thus a symbol of their identity. Especially in the third version with the introduction praising Dunbar's qualities as a poet, a frame is established that leads the reader to expect a different set of values associated with Black speech than they normally find in anecdotes, reports, stories and cartoons. However, the fact that Dunbar labeled the section in *Majors and Minors* that contains the poem "Humor and Dialect" indicates the dilemma that as soon as a Black voice is evoked through the representation of language, the otherness created through this representation invites at least humor but potentially also contempt.

### 4.1 Metadiscourses on phonological forms

The advertisement that appeared four times in the *State* in December 1896 is an example which illustrates how the fame of the poem is exploited by a company to promote their product. Figure 4.17 shows the advertisement (the left part appeared on top of the right part). The advertisement is framed by the company's name, Lorick & Lowrance, which is printed in small letters on top and in big bold letters at the bottom. Below the company's name at the top of the advertisement, the title of the poem is printed in big bold letters as well, so that the readers' attention is drawn to the advertisement. Below the poem, it is contextualized by the comment "'De Co'n Pone's Hot' is very fine, no doubt, but nothing in comparison to our celebrated Cakes and Crackers". The phrase *Cakes and Crackers* is highlighted by being printed in big bold letters because they are the products advertised. The comment shows that the value of the products is established in relation to the corn pone celebrated in the poem. By first printing the poem, which constructs a high value of corn pone as a symbol of southern Black identity, and then reducing its value to "nothing" in comparison to their own product, the cakes and crackers seem even more valuable than without the comparison. The advertisement creates a stark contrast between corn pone as a home-made product, fulfilling the basic needs of poor Blacks (not only satisfying their hunger but also providing emotional comfort) and being at the heart of family and community gatherings, and cakes and crackers sold in large quantities to merchants by wholesale grocers, symbolizing thus the economic wealth of white people, both from the point of view of the consumers, who are able to afford buying fine cakes and crackers instead of having to bake corn pone themselves, and from the point of view of the grocers, who make money from selling their products to merchants, so that they reach a large number of people. Ultimately, the advertisement conveys that despite being celebrated by Black people, the corn pone is inferior to products like cakes and crackers and underlines Black people's lower social and economic status.

To conclude, all the articles analyzed above link non-rhoticity to Black speakers, usually in a southern context, although this is often not made explicit. Nonrhoticity as an index of Blackness predominates although it is noticeable that in the examples discussed above, direct representations of Black speech are only rarely contrasted with direct representation of white voices in the same article. In these exceptional cases, the white speakers are *not* depicted as non-rhotic (the lady in the anecdote "Rich" and the senior member of the firm in the anecdote "Too Hasty"). This is interesting because non-rhoticity is also found in representations of white southern speakers in other articles. It suggests that if ethnicity is foregrounded, Black speakers are depicted as non-rhotic, while white speakers remain rhotic, but when belonging to the south is foregrounded, non-rhoticity

4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Figure 4.17: Advertisement containing the poem "When de Co'n Pone's Hot", published in the *State* (Columbia, South Carolina) on December 30, 1896[164] , retrieved from America's Historical Newspapers

### 4.1 Metadiscourses on phonological forms

primarily functions as an index of place and not of ethnicity, and it is thus found in representations of white speakers as well. The following analysis will show how these links to southernness were established and which other social values and linguistic forms also played a role in the process.

The first article linking non-rhoticity to white southerners was published on September 14, 1875[32] , in the *Cincinnati Daily Gazette* (Ohio). It is a report on a shooting at a ball in Paris, Kentucky, where a young man of the Clay family opened fire on a Marshal, who had refused him access to the ball because of his "drunken state" and "disorderly temper". In the article, non-rhoticity is associated with the southern upper-class: It is represented twice in the word *sah* 'sir', which is in both cases combined with the mention of the Clay family and its description as "one of the finest families of Kentucky". The fact that these combinations are set apart from the main text through quotation marks indicates a change of voice from that of the neutral reporter to that of another person – the reader is invited to infer that it is the voice of a member of the mentioned upper-class family. In the last line, the quotation marks indicate a voice belonging to someone attending the ball, someone who is part of the upper-class "society". As non-rhoticity is the only phonological form represented in the quotation (in *bettah* and *heah*), it functions as a very salient index of southern upper-class speech in this article. It is also noticeable that *bettah* is used in a modal construction here (*bettah get out of heah*), as in most of the articles linking *bettah* to Black American speakers, and that the adverb *quick* is not morphologically marked. This shows that not only non-rhoticity but also the other forms represented here are not only associated with Black American speakers, but also with a group at the very opposite end of the social spectrum. What unites these articles is the negative attitude expressed towards these groups. In this article, the evaluation of the "sentiment of society" as "just and enlightened" is full of irony, because it is clear how unfairly the Marshal is treated. Even though he only defends himself against the attacker, he is the one who has to fear punishment – the implied reason is the high social position of the attacker. This young man is presented in a very negative light: Not only is he drunk and "in a disorderly temper", but he is also aggressive, reckless, without any consideration for other people's lives and, ultimately, also incapable of firing a gun accurately (which is not surprising, given his drunken state). The Marshal, on the other hand, is presented as very courageous because he opposes the young man despite his inferior social position and additionally, as very capable of doing his job because he manages to shoot his opponent with one or more bullets (without killing him). This contrast between the two figures only adds to the impression that the Marshal is treated very unfairly by the society people at

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

the ball. All in all, the article combines a link between non-rhoticity and white southern upper-class speech with a negative evaluation of that same class.

They had an interesting little society incident over at Paris, Ky., Friday night at the ball which wound up at the Bourbon County Fair. A young man, one of "the Clay family, sah," proposed to enter the ball-room in a drunken state and a disorderly temper. Being withstood by the City Marshal, he drew his revolver and opened a fusillade on the Marshal. Thereupon the Marshal drew his revolver and returned the fire. But the pistol of the scion of "one of the first families of Kentucky, sah," only snapped every time, while the Marshall's went off four times, and hit, too. The incident added much to the festivity of the ball, especially to the girls that went from the north side of the river. The condition of the scion of chivalry was thought critical. Of course, if he had shot the Marshal it would have been all right; but as it was the other way, the just and enlightened sentiment of society toward the Marshal found expression in such observations as: "Well, he'd bettah get out of heah d—m quick."

A negative evaluation of a southern upper-class figure is also expressed in an anecdote which was published in the *Cleveland Herald* (Ohio) on January 6, 1884[63] . It had originally been published in the *Chicago Inter-Ocean*. The main figure in the anecdote is Colonel Gutrippah, characterized in the sub-heading as a "Gentleman from Kentucky". The title of Colonel indicates that the gentleman carries the title Kentucky Colonel, which is an honorary title conferred by the governor of Kentucky to men who have done an exceptional deed or service to the state.<sup>17</sup> As stated in *The Kentucky Encyclopedia* (Kleber 1992: 493), the Kentucky Colonel "has come to represent the daring, glamour, dignity, wit, charm, and attraction of outstanding men who have claimed the title—the stereotype of a southern gentleman". Kentucky Colonels are therefore proud men, a characteristic which is used in the anecdote to create humor, as shown by the saying quoted in the sub-heading: "Pride Goeth Before a Fall". This saying summarizes and evaluates the events told in the anecdote: The Colonel is very proud of his skating skills, but then proves to be a complete failure when he loses control, collides with a woman (a "fat lady weighing something like 250 pounds"), falls down on his back and ends up with the woman sitting on top of him. Even though he is laughed at, the end of the anecdote shows that he tries to restore his pride by directing the attention away from himself and towards the weight of the woman, comparing her to a horse ("I nevah thought, sah, that a woman could weigh as much as a hoss"). This illustrates that not only his pride is not justified, but also that he is not a gentleman at all because a gentleman would be expected to admit to his shortcomings and be respectful and polite to women.

The positive stereotype of the southern gentleman therefore becomes the subject of ridicule in the anecdote and it is especially the excessive and unjustified

<sup>17</sup>The title "Kentucky Colonel" was originally a title given to members of the civilian state militia, but it was later granted to civilians as well (Kleber 1992: 493).

pride of the Kentucky Colonel that is criticized. His name, Gutrippah, already indicates the fictionality of the anecdote because it is a telling name which alludes to the characteristics of the figure (the colonel being a "good tripper on skates", implying that he is not steady on skates and likely to stumble and fall). This obvious fictionality ensures that the readers understand that the criticism is not directed against an actual person but at the figure of the southern gentleman in general. The linguistic forms represented in his speech (his voice is demarcated through direct quotations) are also linked to this rather negative image of the southern gentleman. Non-rhoticity is by far the most salient linguistic form. It not only occurs frequently, but it is also highlighted in several ways. First of all, it is very visible through its representation in the name of the Colonel, which is part of the heading (<Gutrippah> being a pronunciation respelling of *good tripper*). Secondly, it is emphasized through the repetitive use of the address term *sah* 'sir' and through its occurrence in parallel structures which occur in close proximity, as in "They ice [...] is hahdah, smoothah, and bettah in every condemned mannah". Thirdly, it is represented in the lexical item *hoss* 'horse', which occupies a prominent position in the anecdote because it occurs at the very end, and because it is central in characterizing the Colonel as disrespectful (by comparing a woman to a horse). There are very few other phonological forms: The repeated spelling of *the* as <they> and the spelling of *Kentucky* as <Kaintucky> indicate a use of a longer and diphthongized vowel in the lexical sets dress and comma, the spellings <sah> 'sir' and <whah> 'where' indicate not only non-rhoticity, but also a lower nurse vowel and a low monophthongized square vowel and there is an instance of unstressed syllable deletion in *'deed* 'indeed'. On the grammatical level, the speech of the Kentucky Colonel is marked by differences to the voice of the narrator and that of the reporter as well. The Colonel uses the firstand second-person plural pronouns *we all* and *you all* and there are differences in subject-verb agreement: *is* and *has* mark first- and second-person plural forms (e.g. in *you all is fooled*, *we all* [...] *has good ice skating*, *the* [...] *ice you all has up heah*) and third-person singular *don't* in *it don't last so long*.

The anecdote creates a contrast between the north and the south and given the context – the fact that it was published in northern newspapers (in Illinois and Ohio) – makes it not very surprising that the north is presented as superior to the south. The southern elite is characterized as proud and arrogant – exemplified by the Colonel who claims that he is better at ice-skating, an activity that is practiced much more in the north than in the south ("I drive everybody off they ice"), and therefore also at roller-skating. His claim that even the ice in Kentucky is better than in the north is so exaggerated that it creates ridicule of southern arrogance. That southern elitism and pride is unjustified becomes clear as the events in the anecdote unfold and the Colonel's incapability becomes the subject of laughter. Another aspect of southern pride is highlighted, when the Colonel refers to his home state as "old Kentucky" because this creates the impression that southern pride is based on the past rather than the present, inviting the implicit conclusion that northern pride is based on the present and the future. The grammatical forms deviating from the speech of the northern voices also implicitly create a picture of southern speech being 'incorrect' and 'inferior' to northern speech. With regard to the social values indexed by non-rhoticity, the article shows that the phonological form is linked to the figure of the southern gentleman, who, in a northern context, is criticized because of his excessive pride and arrogance, which is largely based on the past and lacks any justification in the present.

# **GUTRIPPAH ON SKATES.**

"PRIDE GOETH BEFORE A FALL."

# **How the Gentleman from Kentucky Went Boldly on the Ice and was Ingloriously Sat Upon.**

"I'll go with you all, sah," said Colonel Gutrippah. "I used to be a mighty fine skatah when I was a boy sah."

"I didn't know that they ever had skating in Kentucky, Colonel," said the reporter.

"That's whah you all is fooled, sah," replied the Colonel. "We all down ouh way has just as good skating as you all up heah, sah, only it don't last so long, sah. They ice in Kaintucky, when it does come, sah, is hahdah, smoothah, and bettah in every condemned mannah, foh all uses, including skating and mint juleps, sah, than they common soht of ice you all has up heah, sah. Yes, sah, 'deed I will go with you all and show they general assohtment of cussed mechanics that will doubtless be present how they business is done in old Kentucky, sah. Bet you all a hohn of Baeh Grass against a second-hand wooden tooth-pick that I drive everybody off they ice, sah."

"But, Colonel," said the reporter, "this is not ice skating. It's roller skating."

"Same thing, sah," replied the Colonel, putting on his fur overcoat. "Any man who can skate on ice can skate on rollers, sah."

[Immediately after stepping onto the skate rink, he falls down. He is helped up by two young men who skate with him for a while and then give him a shove that steadies him. However, the Colonel does not know how to stop and eventually collides with a "fat lady weighing something like 250 pounds". They fall down – the Colonel on his back and the lady on the Colonel's stomach – which causes the crowd to laugh and a trombone player to ruin his instrument because he falls off his chair laughing.]

The Colonel rose slowly and painfully to a sitting position, pulled down his vest and straightened his collar. Then deliberately, and with an air of determination, he took off the skates and flung them with great force across the rink. Taking the reporter's arm he limped painfully out of the building, and all in silence to the door of the hotel. Then he bent down and whispered in his companion's ear:

"I nevah thought, sah, that a woman could weigh as much as a hoss." —[Chicago Inter-Ocean].

The figure of the Kentucky Colonel also appears in a different text type in the collection of articles containing *bettah*. It is part of an advertisement which

### 4.1 Metadiscourses on phonological forms

was published four times in the *Fayetteville Observer* (North Carolina) on September 15, 1898[169] (see Figure 4.18). It advertises a clothing manufacturing company called The Royal Tailors, which was based in Chicago and New York and sold clothes nationwide. In the advertisement, readers are urged to buy clothes to dress well, based on the argument that only rich people can afford to dress poorly, presumably because they have already achieved the status and wealth that others hope to acquire. The main selling point is that the clothes tailored by The Royal Tailors are better than those of other tailors, and to emphasize this point, the voice of "a Kentucky Colonel" is quoted as saying "There aint no bad whiskey, sah, but I may say, sah, that some whisky is bettah than other whiskey, sah". The distinction between the voice of the company addressing the readers ("Better try The Royal Tailors") and that of the Colonel is reinforced when the company's

Figure 4.18: Advertisement using the figure of the Kentucky Colonel, published in the *Fayetteville Observer* (Fayetteville, North Carolina) on September 15, 1898[169] , retrieved from Nineteenth-Century U.S. Newspapers

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

voice draws on the whiskey analogy to advertise the quality of their clothes and, by doing so, imitates the Colonel's voice using quotation marks again ("while we are not prepared to say there "aint no bad" tailors"). Through this imitation and the explicit separation of voices through quotation marks, the company distances itself from the use of the form *aint* and the use of negative concord as well as from non-rhoticity: It uses *better* whereas the Kentucky Colonel uses the non-rhotic form *bettah*. As in all the articles linking *bettah* to southern speech above, the address term *sir*, whose non-rhotic pronunciation is also indicated by the spelling <sah>, is also linked to the Colonel's repertoire in the advertisement. Davies (2007: 175) calls the use of address terms like *sir* "a classic form of negative politeness in which social hierarchy is linguistically signaled". In the context of this advertisement, this has an important function: It directs the readers' attention to the existence of social hierarchies and shows them a way to signal a higher position in this hierarchy. The claim "There's a difference!", which is printed in bold and set apart from the rest of the text in the middle of the anecdote, underlines this focus on social hierarchy as well: Not only is there a hierarchical difference between good and bad whiskey and good and bad tailors, but there is also a hierarchical difference between rich and poor people and wearing the right clothes is supposed to get a person closer to the top. The name of the company, The Royal Tailors, also alludes to a hierarchical society by using the adjective *royal*, which not only establishes a link to English royalty, but also to the southern planter aristocracy with its close ties to England. The use of *aint* and negative concord indicates, however, that the Kentucky Colonel is not constructed as a linguistic (and also not as a social) model. These grammatical deviations from the 'neutral' company voice rather create the same hierarchical difference between the superior north and the inferior south as in the anecdote above. Given that The Royal Tailors is a company based in the north and the northeast, this is not surprising. However, using representations of language is a strategy to establish this superiority in a very subtle manner, and the fact that the advertisement was published in a southern newspaper suggests that the creators hoped that it would appeal to a southern audience. It is perhaps the slight ridicule of the Kentucky Colonel that is intended to motivate southern readers to advance in the social hierarchy and that, in order to achieve this aim, they do not have to carry the title of a Colonel nor be rich, but to simply buy clothes from the right tailor.

Another southern figure linked to non-rhoticity is the southern upper-class girl. She is portrayed in an anecdote which was originally published by the *New York Herald* and which was reprinted in *The Aitchison Daily Globe* (Kansas) on May 25, 1891[108] . The narrator describes two girls who ride on a street car in New York City – they are both young and pretty and wear good clothes, which is a

signal of a higher social status, but there is also the important difference that one of them is from New York and the other from the south. This difference is the main topic of the anecdote, which creates a representative incident, intended to characterize and differentiate New York and southern upper-class girls in general. The problem that the girls face on board the street car is that all seats are taken by businessmen on their way to work. While the northern girl accepts the situation as normal and tries to continue the conversation, the southern girl decides to change it and by talking loudly to her friend about how she expects the men to do everything they can to deny them the possibility to sit, she provokes two men to give up their seats to prove her wrong. In the end, she points out to her friend that she brought about the favorable change of situation because as a southern girl she knows how to "manage" men, or in other words, how to make them "behave bettah".

The anecdote addresses two interconnected issues: the characteristics of and the relationship between the male and the female gender and the difference between the north and the south of the United States. The northern men are portrayed as urban businessmen who are focused on work and success and who are not respectful or polite towards women (they do not offer to give up their seats on the street car). This characterization is summarized by the label "brutes" given to the men by the New York girl. This picture of the rude brute stands in opposition to the ideal of the southern gentleman to which the southern girl can be expected to be accustomed, which leads to her surprise when she realizes that none of the men is offering her and her friend a seat. These different male figures are implicitly linked to concepts of tradition and modernity, symbolized by the places that the street car passes in New York City: "Wall street running right up against old Trinity". Wall Street is a place associated with eastern urban values resulting from the new, modern business life, while the old Trinity church evokes older, traditional and more conservative values. Not only the male figures and their behavior towards women are linked to this opposition, but also the female figures. The northern girl is characterized as practical, strong and independent: She does not rely on the men to offer her a seat, but she can stand on her own feet and hold herself by holding onto the strap. She does not expect help, but she also does not need it because she can help herself. The southern girl, on the other hand, appears soft and weak at first, but it becomes clear that she uses this soft appearance to achieve her goals in an indirect way – in this case to get the men to give up their seats – so that she also appears strong, but in a different way than the northern girl. The author of the anecdote uses the voice of the southern girl to sum up this characterization by using an illustrative metaphor: "The iron hand in the velvet glove; that's a south'en woman's fo'te". She presents her

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

own character and behavior as a model to the New York girl by saying "If you nawthin gyuls understood managing yah men they'd behave bettah". This shows that the main aim of the anecdote is to create a representative incident based on which two ways of life are compared and evaluated: the southern, traditional one and the northeastern, modern one. The southern girl's feelings of superiority are constructed as typical of a southern (upper-class) attitude and it is this attitude and the conservative southern way of life that are humorously criticized in the anecdote.

The representation of language is important in conveying the criticism. Only the speech of the southern girl is marked as 'deviant' through pronunciation respellings – the speech of the northern girl therefore appears 'normal'. The linguistic form that stands out is non-rhoticity because apart from a palatal glide insertion (and a lower vowel) in *gyuls* ('girls') it is the only form that is represented.<sup>18</sup> By describing her as speaking "in soft, clear tones and unmistakable southern accents", non-rhoticity is explicitly linked to southern speech and to her soft and female character. This association with the south is particularly important considering that in the articles containing *deah* AND *fellah*, non-rhoticity is also linked to eastern urban speech, especially New York speech. In this anecdote, however, the New York girl is *not* portrayed as a non-rhotic speaker. This indicates that when differences between the south and the north are highlighted, non-rhoticity is an important marker of southern speech and northern speech is consequently *not* marked as non-rhotic (even in New York City). If, however, the focus is not on a regional difference but on social differences between speakers, non-rhoticity can also be used to mark a northern speaker as someone who wants to appear educated and 'cultured' and who wants to signal membership in upperclass circles through the imitation of British English speech, but also as someone who lacks authenticity. Nevertheless, these social differences are also linked to a regional difference, albeit not between north and south but between east and west. This suggests that non-rhoticity is restricted to a particular social group of speakers in the northeast, whose negative characteristics are highlighted by contrasting them with positive characteristics of western (not southern!) speakers. The association between southerners and non-rhoticity, in contrast, rather serves to emphasize the traditional and conservative ways of the south, where men are still gentlemen and where women are not independent per se but need to exert their power in rather indirect ways. Davies (2007: 175) describes indirectness in general as a "classic negative politeness strategy" in the south, and

<sup>18</sup>Note here that the insertion of a palatal glide after a velar consonant has also been used in representations of Black speech (*cyarry* in the article "Dat Deceptious Mule").

the anecdote reveals how this difference on the pragmatic level between indirect and direct forms of expressing one's wishes becomes indexically linked to the north-south difference in metadiscursive activity. The case of non-rhoticity consequently illustrates very impressively that one linguistic form can index very different and even opposite social values (southern and northeastern with regard to region, Black and white with regard to ethnicity) and that it is only in context with other linguistic forms and social values established in the text that specific indexical links are created and become interpretable by the reader.

All in all, the anecdote "She Got a Seat" ultimately presents the northern girl as tougher and more progressive because she can stand and does not need to sit like the southern girl, and because she does not have to use indirect politeness strategies to get men to help her. Even though the humor of the anecdote rests on the southern girl getting her way, it also leaves room for doubt whether it was really her 'management' that caused the men to give up the seat. One of the men says "No, no; we're not as bad as that", which creates the impression that even though northern men might not be like southern gentlemen, they can also not be characterized as uncivilized brutes without morals. They are rather used to modern women who are independent and, if they do need help, express their wishes directly.

# **She Got a Seat.**

It was about 11 o'clock, on the elevated road—and elsewhere.

At Twenty-third street two pretty young women in good clothes came aboard.

The morning rush was past, but still the car was full of comfortable business men studiously devoted to their newspapers.

Not a seat was vacant, and not one was offered. One of the young women hitched herself to a strap with an air of familiarity with the process; the other looked on and at the men with an expression of intellectual curiosity, not unmixed with scorn.

"**You'll see Wall street running right up against old Trinity**," said she of the strap; as if continuing a previous conversation, "**then I want to take you through one of the big office buildings, but we'll have to wait till papa**"—

"**Will we have to stand all the way down theah?**" asked her friend and evident guest irrelevantly, in soft, clear tones and unmistakable southern accents.

"No; not all the way," replied the New York girl, and then, chaperoning her sex instead of her section, she added sotto voce, "**some of these brutes will have to get out before Rector street—they'll have to give us a seat whether they want to or not**."

"**No, I don't think they will,**" said the southerner in the same soft, audible tone, and casting a meditative look about her; "**I think by their looks they'll stay aboard and lose money to keep us out of one.**"

A gentleman sitting on one of the cross seats with his back to them now rose with an amused expression, saying:

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

"**No, no; we're not as bad as that,**" and surrendered his seat, whereupon his vis-a-vis succumbed ruefully to moral suasion and gave up his. When the girls, with smiling thanks, were seated, the southerner winked merrily at her friend and said:

"**If you nawthin gyuls understood managing yah men they'd behave bettah. The iron hand in the velvet glove; that's a south'en woman's fo'te.**"—New York Herald

[emphasis mine]

Non-rhoticity, represented by the search term *bettah*, does not only occur as part of the representation of upper-class southern speech, but also of lower-class southern speech. An example is the article "A Relic of the Past", written by M. I. Dexter and published in *The Atchison Daily Glob*e (Kansas), on January 9, 1895[147] , and in the *Idaho Statesman* (Boise, Idaho) on January 18, 1895[148] . This article is a report about a sundial in St. Louis and the title "A Relic of the Past" indicates that the general topic is the transformation of St. Louis from an old town to a modern one. The first sentence explicitly describes how in this process the "old city is gradually ridding itself of many of the peculiarities that have heretofore stamped it as more of an old time southern than a modern western town". The sundial is constructed as a symbol of the past, an object that was once found in almost every southern town (especially in the southwest) but that is now so rare that its peculiarity attracts the interest of many visitors. The person linked to this relic from the past is a janitor who works at the courthouse where the sundial is located. He is interviewed and his description of the dial is rendered in direct speech, indicated through quotation marks (and highlighted in bold here). Phonologically, his speech is prominently marked by non-rhoticity, which is the only phonological form consistently represented. There are three more cases of pronunciation respellings, *they* 'the', *kin* 'can' and *jest* 'just'. The first one indicates a longer and diphthongized vowel in *the* and the forms *kin* and *jest* are most likely cases of eye dialect, indicating a weak form with a more central vowel (and not a non-standard pronunciation). On a lexical and pragmatic level, the high frequency of the address term *sir* (represented as *suh* here) is noticeable, as in the other articles analyzed above, which confirms its status as being consistently linked to a southern American linguistic repertoire and linking non-rhoticity to a social system marked by social hierarchy. On a grammatical level, the janitor exhibits a number of differences to the 'normal' voice of the reporter. Like the Kentucky Colonel in the advertisement, he uses *ain't* and *hain't* as negators as well as negative concord as in *they ain't no going back* and *that hain't got no bettah business*. And like Colonel Gutrippah in "Gutrippah on Skates", he exhibits a different pattern of subject-verb agreement, but in his case, it is the third-person plural form that is marked by an -*s* (*theah's not many*, *they comes*). However, unmarked forms occur as well (*they all know*). Furthermore, the janitor uses the

### 4.1 Metadiscourses on phonological forms

double demonstratives *this here* and *that there* (*this heah co'thouse*, *that theah old dial*), *what* as a relative pronoun (*theah's not many but what knows*), an unmarked past tense form in *They use to be two sundials* and existential *they* in the sentence just cited and also in the *they ain't no going back* cited above. Through the figure of the janitor, these linguistic forms are linked to the south and to an old and outdated way of life. By characterizing him as "quainter even than the old dials", the indexical link to old-fashioned behavior and attitudes is reinforced. However, the janitor is also described as "delightful" and his description of the old sundial also contains some elements of nostalgia: He makes a reference to the "fine" southern gentleman who is fairly rich (he is wearing a gold watch) and who uses the sundial to set his watch because it is very reliable. Even though the past is depicted in a positive light, it is made clear that this way of life is in the process of being transformed and that it is "entertaining" for 'modern' visitors precisely because it is not part of present-day life at the time. The difference between old-fashioned and modern is linked in the article to the regional difference between the south and the west ("an old time southern" vs. "a modern western town") and through the figure of the janitor non-rhoticity is linked to the former, and rhoticity, by implication, to the latter. The necessity for transformation can therefore be extended to language as well: Non-rhoticity and the other linguistic forms described above are portrayed as old and traditional and even though they are pleasingly peculiar they are not compatible with a modern way of life and therefore need to be changed. Non-rhoticity is thus portrayed as being as much a "relic of the past" as the sundial and the figure of the janitor.

A further aspect that is relevant in this context is the comparison made between the janitor and "our Uncle Samuel", a figure that became a symbol of the United States (Vile 2018: 321). On the one hand, the reporter finds that the janitor resembles Uncle Sam because of his facial features ("When I first saw his face, I was almost surprised because his trousers were not striped red and white and his vest was not blue and star spangled"). On the other hand, the shape of the body is the complete opposite to that of Uncle Sam ("instead of being thin and angular, he [the janitor] is round and stout and ruddy"). This description can be read as symbolizing the janitor's belonging to the American nation: While he is an American, he does not correspond to the ideal American, embodied in the Uncle Sam figure. He is a southerner who romanticizes the past (symbolized by his praise of the sundial, which is "running just as good as it did when it was first put theah") and who complains about the young men who destroy old things (like the dial part of another sundial) to accelerate the transformation to a modern way of life. He is protective of the past and fears, as he is quoted at a later point in the article, that "the county'll see fit to teah this grand old co'thouse down and

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

put up a great big steel frame modern building". All in all, non-rhoticity is therefore linked to a conservative, backward-oriented southern speaker who glorifies the past and fears the transformation to modernity.

# **A RELIC OF THE PAST**

# THE SUNDIAL BEFORE THE OLD COURTHOUSE AT ST. LOUIS **Views of the Janitor—A Man Who Is as Quaint as the Dial Itself—Fine Paintings That Cost a Right Smart Heap—The Past Linked to the Present.** [Special Correspondence.]

ST. LOUIS, Jan. 8. —Although this interesting old city is gradually ridding itself of many of the peculiarities that have heretofore stamped it as more of an old time southern than a modern western town, the transformation is still far from complete, and it is this very survival of an occasional feature of other days that renders the place entertaining to the curious visitor.

Standing within a 4 by 5 foot inclosure of iron fence, near the white painted old courthouse, for instance, is an iron post, topped with a broad, flat, circular plate. Rising from the center of the plate is a triangular projection, and there is an iron hinged cover attached, evidently intended to be closed when the weather is bad. Such a contrivance is a rare sight nowadays anywhere, but it would have been a poor sort of town indeed five or six decades ago, especially in the southwest, that was without one, and no one would have been at a loss regarding its use. But there is not a day in a twelvemonth those modern times, so the courthouse janitor told me today, that some stranger does not inquire curiously about the one in the St. Louis courtyard.

"**But of co'se,**" said he reflectively, pulling his chin whisker and turning his quid of shaved plug in his check, "**theah's not many but what knows it's a sundial when they comes to examine it. It's a fine dial, that. You see, it's running just as good as it did when it was first put theah. And many's the fine gentleman, with his kid gloves on, that steps up to that theah old dial, suh, at noon and sets his gold watch by it. They all know, suh, that, notwithstanding a sundial is old fashioned, they ain't no going back on sun time, suh.**

"**They use to be two sundials to this heah co'thouse. You kin see the post of the othah one out theah on the othah side of the yahd. But the dial paht is gone forevah, suh. You see, they's a pow'ful lot o' boys heah in St. Louis that hain't got no bettah business nights and Sundays than to pound co'thouse dials with rocks and things jest to destroy them.**

# [...]

This delightful janitor's name is Griffin, and he is quainter even than the old dials. In person, he resembles our Uncle Samuel, save that, instead of being thin and angular, he is round and stout and ruddy. When I first saw his face, I was almost surprised because his trousers were not striped red and white and his vest was not blue and star spangled. [...]

# [emphasis mine]

An article which creates a link between non-rhoticity and white southern as well as Black American speech and therefore relates both groups to each other was published in the *Northern Christian Advocate* (Syracuse, New York) on April 3, 1884[64] . It is an account of the missionary and social work of a religious group in the south (in Little Rock, Arkansas), written by a member of that group, Miss E. H. McIntosh, and directed at a northern and eastern audience (their "friends"). The group has installed a "Model Home" called the "Smith Industrial Home", which is connected to the Philander Smith College, an educational institution, and its chief object is to help Black Americans ("our work is with the colored people").

In the article, the author characterizes the people living in Little Rock through explicit comments and descriptions as well as implicitly by means of direct representations of their speech. She stresses that differences between southerners are mainly based on class and not on race: By saying that "both whites and blacks" are slow "except in the upper circles" and that "whites are little better than the blacks" in having a low morality, she groups both racial groups together as being "of the lower order". On the other hand, she points out that there are not only white upper-class people, but also "various grades" among Black people, a "colored gentleman" who is "intelligent, polite and interesting" being an example of a Black person with a higher social status. The statement that white lower-class people are not better than Black people functions as a disparagement of white people against the background that Black people are normally constructed as being socially inferior to white people. For the higher classes, white superiority remains unchallenged, as can be seen in the following remark: Even those "who are the most intelligent and most cultured among them [among the colored people] are sadly lacking in many things that they would be so much better for knowing". So even though there are degrees of ignorance, it remains a major problem and it serves to justify the installment of the "Model Home". Connected to ignorance is another social characteristic: backwardness. The author uses the voice of an imagined white southern speaker to characterize this particular group of people as backward because of their ignorance ("It is the ignorance of the people what make us so".). By using the pronouns *they* ("they said") and *us* ("make us so"), the author emphasizes that the voice is representative of the whole group of lower-class southern people. Using the voice of an insider of the group is an important strategy of characterization because it appears to provide support for the author's judgements and opinions from an insider's perspective. Similarly, she uses the representative voice to create the impression that their help is wanted by the people themselves: She lets the voice explicitly ask for it ("What we want is for you northerners to come down heah and teach us bettah"). This shows that the most important division created in the article is that between the south and the north, with most southerners being depicted as slower, less intelligent, less educated, backward and having a lower morality. They are not only presented as needing and wanting help, but also as deserving help because of their friendliness and good character, which is conveyed through their "good-natured" and "easy" laughs. A particular case to reinforce the impression of southern helplessness is presented through the description of the stereotypical "Topsy" figure: a fourteen-year-old girl who sucks her thumb like a much younger child, who is barefoot (she turns on her toe) and not properly dressed, and whose basic needs are not fulfilled (she is hungry and without a home, family, and friends).

The linguistic repertoires of the lower-class southern speakers are illustrated through the voice representing southern white speakers, a voice representing Black speakers in the beginning of the article and the voice of the "Topsy" figure (all the direct quotations are highlighted in bold in the quotation). The cooccurrence of representations of both white and Black southern speech makes this article ideal for exploring which linguistic forms these groups have in common and which forms mark them as different. Non-rhoticity is a phonological form associated with both groups and although it is quite frequent and noticeable through the spelling <ah>, it is also not represented consistently (it is not marked at all if /r/ precedes a consonant in the coda cluster). Apart from non-rhoticity, there is only one form represented in the spelling and linked to all three representations: the alveolar realization of -*ing* (e.g. in *tryin'*, *nothin'*). What differentiates the Black voices from white voices are forms that have been identified as typical of a repertoire of Black speech in articles analyzed above: final consonant cluster reduction (*tole* 'told'), unstressed syllable deletion (*bout* 'about') and TH-stopping (only represented once in *dis*) and the insertion of /j/ and dropping of /h/ in *yeah* 'here'. White southern speech is not additionally marked on the phonological level. However, on the lexical level white southern speech is characterized by the lexical item used to greet people (*How'dy*), the specific use of the greeting *good evening* (which becomes appropriate as soon as mid-day is past) and the phrase *I reckon* (which is neither part of the author's voice nor of the other voices which are represented by the author).

Considering grammatical differences, it is more difficult to identify forms that clearly mark a speaker as white or Black. What is noticeable is that the Topsy figure is distinguished on the grammatical level from the other speakers because she almost always uses reduced sentences in which the subject is deleted. The sentences are short and if two clauses are joined, they are not linked by a conjunction (*I'm just so hungry can't do nothin' 't all*). This syntactic simplicity is linked to the description of the simplicity of her life – she lacks the most basic elements needed for survival. The representative Black voice does not exhibit such a reduced syntactic complexity, however, so that this form seems to index a particular type of Black speaker rather than Black ethnicity in general. There is one case of copula deletion (*Our people slow*) in the representation of Black speech, but this is also found in the representation of white southern speech (*we too old*). As in two other articles representing Black speech, the double demonstrative *this here* (*dis yeah work*) is linked to the Black voice here as well, but as it is also used by the white janitor in "A Relic of the Past", it is doubtful that this form distinguishes Black from white speech. Noticeable elements of southern white speech are *am* as the third-person singular form of be (*the way it am done heah*), the

relative pronoun *what* and an invariant third-person singular verb form (*It is the ignorance of the people what make us so*), but all of these forms have been found to mark Black speech in articles above as well. A form linked to the white voice which has so far only been found in one article representing a white southern voice (Colonel Gutrippah) is the marking of the third-person plural (here combined with first-person plural *is* in *We makes lots of money, but we is always poah*). However, Colonel Gutrippah combines this with second-person plural -*s* marking which has also been linked to the Black preacher in "A Negro Revival", so it is again doubtful that this form indexes the ethnicity of the speaker. A form that has not been used in any of the articles above is the double marking of past tense in *Why didn't you tole us*, which is linked to the Black voice. There is also a regularized past participle (*knowed*), a form that has been found in the Black preacher's repertoire in "An Old Time Fo'th" as well.

To conclude, the article "First Impressions on Entering the Field" illustrates that the ethnicity of the southern speakers is primarily indexed on the phonological and the lexical level. On the grammatical level, a reduced, subject-less sentence structure does index a Black figure which is particularly simple-minded and poor, but this makes it a form that does not index Black ethnicity in general but only Black ethnicity co-occurring with other social characteristics. Before drawing any further generalizations based on all the articles analyzed so far, I will analyze articles linking *bettah* and thus non-rhoticity to the third major group of speakers: to mountaineers.

# **First Impressions on Entering the Field.**

by miss e. h. mcintosh.

Probably most of the readers of the Northern were interested in reading the account of the dedication of the "Model Home" at Little Rock, Ark., which occurred Feb. 25th: and perhaps they would like to know more of this Home which now bears the name "Smith Industrial Home."

[...]

The habits of the people who belong to this section of the country are slow. This may be said of both whites and blacks, except in the upper circles.

### [...]

The mind [of the colored people] is not more active than the body. When we told the people what we intended to do they said: "Why didn't you tole us befoah? Why didn't you come down heah and lecture three, foah months ago, so we'd knowed bout dis yeah work? Our people slow! They can't take this yeah thing so fast."

[...]

It is surprising to see how low is the morality of this people. I am told that the whites are little better than the blacks in this particular.

# [...]

When they meet here they say in greeting **"How'dy?"** and after twelve o'clock every one says **"good evening."** If you want anything done as it ought to be done they say: **"Oh! they don't do that way**

**here! I reckon you came from the North, you came from the East, no such thing done in the South—not in Arkansas anyway."** When you find them doing a thing backward and exclaim: **"Why! that is backward!"** they with a good-natured laugh will say: **"That's the way it am done heah,"** you will answer as you feel: "Yes! you do everything backward here!" And with another easy laugh they say: **"Yes, mam, we do everything backwards heah, but then you know ouah people doant know any bettah. It is the ignorance of the people what make us so. What we want is for you northerners to come down heah and teach us bettah. We makes lots of money, but we is always poah, for we doant know how to spend it. No use'n us old folks tryin' to learn; we too old; but you teach ouah girls so they know more than their mothers did."**

They are, however, not all of the lower order. We find various grades among them. We were shown through the post office of this city the other day by a colored gentleman, superintendent of the mails, as handsome a looking man as you would wish to see, intelligent, polite and interesting [...].

[...]

Then on the other hand we have seen the genuine "Topsy", who came in to ask for something to eat. When we would ask a question she would, with finger in her mouth, turn on her toe until you had the back of her head in full view, so I don't think she had on the original Joseph's coat but she had parts of it all over her. Her skirt hung to the floor on one side and showed the knees at the other. You judged her to be fourteen years of age. We asked her: "What do you want?"

# "**Want sothin to eat; I'm just so hungry can't do nothin' 't all.**"

"Where did you come from?"

# "**Come from no whah.**"

"Where do you stop?"

"**Stop no whah.**"

"Where do your friends live?"

# "**Don't live no whah. Have no friends. Just so hungry can't do nothin' 't all.**"

This is one of the kind that chews tobacco and 'dips snuff'."

[...]

We find that even those who are the most intelligent and most cultured among them are sadly lacking in many things that they would be so much better for knowing.

[...]

Friends, these are the people whom we have come here to help.

[...]

Then how necessary are these foundation stones! To these our toil shall be given and we hope also the prayers of many of God's dear children in the North and East.

For the qualitative analysis of articles constructing non-rhoticity as part of the linguistic repertoire of the figure of the mountaineer, I will give examples of three different text types: a report containing an anecdote, a humorous short dialogue and a long fictional story. The report and the fictional story contain longer stretches of direct representations of speech and were published ten years apart from each other (1885 and 1895). The humorous dialogue was published in the same year as the fictional story (1895), but the representation of speech is very

### 4.1 Metadiscourses on phonological forms

short and provides insights into which forms were focused on to differentiate the mountaineer's linguistic repertoire from that of other speakers.

The report names the object of interest in the heading of the article: "The Pool Tribe". It was published on September 20, 1885[69] in the *Galveston Daily News* (Houston, Texas), and it originally appeared in the *New York Sun*. The subheading provides the main information as to what the Pool tribe is and how it is evaluated: "A Queer Mountain Race that Live in Pennsylvania". The reference to the mountains is important – in the first paragraph this is specified further as "a spur of the Appalachian range" and this is where the people described by the author live. They are characterized as "non-descript" and "queer" and compared to prairie dogs, which evokes associations of animal-like, uncivilized behavior. By calling them a "mountain race", they are also presented as constituting their own race and thus distinguished from the white as well as the Black race. The report includes an anecdote about the reporter meeting a member of the Pool tribe. The man explicitly claims to belong to the group of "Pools" by saying "Wal, I'm one on 'em ". The anecdotal character of this part of the report invites the reader to regard this man as a typical exemplar of the mountain race and the description of the color of his skin as "a weather beaten, old-copper complexion" underlines his attribution to a different race. Furthermore, the adjectives "weather-beaten" and "bent" indicate a life outdoors and full of hardship. The reporter finds the man "half sitting, half lying on the flat surface of a big boulder, sunning himself", which suggests that he is not working anymore, and the man offers to help the reporter only if he gets whiskey in return, which, in combination with his repeated request for a drink, implies that he is addicted to alcohol. The reporter then tells the story of how they find a rattlesnake and how the old man catches it, kills it by crushing a stone on its head and then cuts off the rattles to get the money promised by the reporter. The mountain man therefore appears to be tough and unafraid. In the last part of the story he explains to the reporter why he is not afraid of rattlesnakes: His father has made him immune against the poison when he was a baby by cutting a hole in his arm and putting "some o' de pizon in it". This anecdote therefore draws several parallels between the man and the rattlesnake. Both the man and the animal were found by the reporter while they were sunning themselves and both have poison in them. When the mountain man says "Wal, I'm one on 'em", this could also be read as him identifying himself as a rattlesnake and not a Pool. The implicit message that can be constructed based on the anecdote is that mountain men are like rattlesnakes: They are dangerous, but they can be beaten.

This figure of the "Pool" belonging to the "queer" mountain race (the term *mountaineer* is not used in this report, but the figure is explicitly connected to

the Appalachian Mountains) is linked to several linguistic forms through representations of his speech in direct quotations (highlighted in bold). Non-rhoticity is quite prominent because it is frequently represented, for example in the address term *mistah*, which the mountain man uses several times. However, there are also cases in which the form is not marked as rhotic (*other* in the quotation below, but there are also several further forms in the remaining part of the story, e.g. *fadder* and *rattler*). Furthermore, voiced TH-stopping and alveolar -*ing* are frequently represented – especially the occurrence of TH-stopping is surprising because in the articles analyzed above it was only linked to Black speakers. Further forms shared by the mountain man and Black speakers in other articles are *little* spelled <leetle> to indicate a high front tense vowel, the choice vowel marked as having a lower onset in *pizon* 'poison' and the square vowel being represented as <ee> in *careful*, suggesting a high front monophthong, and as <ah> to indicate a lower and monopthongized vowel in *there* and *where*. This last form was also part of Colonel Gutrippah's linguistic repertoire. A phonological form not found in other articles is the devoicing of the final consonant in *have* (*haf* ). With regard to grammatical differences, the mountain man exhibits second-person singular -*s* marking and he uses the reflexive pronoun *deyselfs* – both forms have been shown to be linked to Black speakers as well. It is also noticeable that *bettah* occurs again in a modal better construction (*Bettah go*), as in seven articles representing Black speech and the first article representing southern upper-class speech. On a lexical level, the use of the phrase *I reckon* links the mountaineer's repertoire to white and Black southern speech. The phrase *I'll have to go you*, which seems to mean 'I'll help you', has not occurred in any article so far and therefore seems to be characteristic of the mountain man in this article.

# **THE POOL TRIBE. A Queer Mountain Race that Live in Pennsylvania.** [New York Sun.]

Towanda, Pa., September 1.—Two broad, low foothills, a spur of the Appalachian range, comprise almost the whole of Durell township, Bradford county, in the northern tier of Pennsylvania. To the surrounding country this double elevation is known as the Huckleberry mountain, or Pool Hill, the first named from the luxuriant growth of this popular fruit, which covers sides and summit of the mountain. The appellation Pool Hill it takes from the nondescript people who populate its broad sides like prairie dogs, better known as the "Pool tribe."

[...]

I camped two nights on the mountain a few weeks ago, and took a stroll up among the rocks, keeping an eye out for rattlesnakes. All at once I heard a call: "Say Mistah"

It was a bent old man, with white hair, and a weather beaten, old-copper complexion. He was half sitting, half lying on the flat surface of a big boulder, sunning himself. In his hand he held a heavy stick, with two short prongs at the end.

4.1 Metadiscourses on phonological forms

"Hello," said I; "what do you want?"

"**Whah's yo' goin'?**"

"Looking for a rattlesnake and a Pool," I answered.

"**Wal, I'm one on 'em, an' I reckon yo' won't have to go mo'n a mile to find th'other. Has yo' got anythin' to drink?**"

I had a pocket flask of snake-bite annihilator with me. Pulling it out, I said: "I'll give you a drink if you will show me a rattlesnake."

"**I'll haf to go yo', mistah,**" he said, as he climbed down from the rocks at a pace which left me breathless. Twice on the way he stopped, and, looking up blankly, said: "**Godlemity, but I'm dry, mistah.**" Both times I moistened the old man, and again we proceeded. Finally he said:

"**Bettah go a leetle kee'ful, mistah. Dey'll be up dah sunnin' deyselfs.**" Proceeding cautiously, we climbed up on a rock and looked over on the other side. There, stretched lazily out on a flat stone, lay a large snake.

"**Does yo' want him, mistah?**" said the Pool.

"I'll give you half a dollar for his rattles," said I.

"**I'll haf to go yo'**," was the reply, as he cautiously climbed down from the rock on which we rested and then crept along on all fours until within a few feet of the snake. Then he noiselessly straightened up and reached for the snake with his forked stick, planting it just back of his head. The reptile writhed and twisted and rattled his alarm, but the old man had him pinned fast. After watching him a minute or two, the Pool picked up a stone with his free hand and crushed the snake's head. [...]

# [emphasis mine]

The second article I have chosen for analysis is a short humorous dialogue with the heading "Safer, Too", published as part of the section "Multiple News Items" in the *Boston Daily Advertiser* in Boston, Massachusetts, on July 27, 1895[152] . It is taken from the *Chicago Record*. One of the conversational partners is a "Mountaineer", while the other is labeled "Visitor", which underlines that he serves as a 'neutral' point of comparison – he is not marked as belonging to a particular group or being from a particular region. The Visitor and the Mountaineer talk about a feud which has apparently been going on for about thirty years. The Visitor asks why the law does not settle it, to which the Mountaineer replies that it should go on until "only one gentleman is left, sah, and then we'll hang him". The dialogue marks the Mountaineer as having a belligerent and extremely violent character. A contrast is established between a civilized system, in which laws and institutions enforcing the law ensure peace, and an uncivilized system, in which people take the law into their own hands leading to lynchings and other forms of violence. The Mountaineer's aim to extinguish all gentlemen can be read as symbolizing his wish to end civilization and refined behavior in favor of uncontrolled violence. These social values are linked to several linguistic forms marking the Mountaineer's speech as different from that of the Visitor. Non-rhoticity is particularly salient because it is represented several times, especially in the address

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

term *sah* (but not in *thurty* 'thirty') and because it is the only deviant form apart from one instance of initial unstressed syllable deletion (*'Bout*) and one representation of a different quality of the nurse vowel (*thurty*). This illustrates that non-rhoticity has been chosen as a primary means of linking the Mountaineer's otherness to a different linguistic repertoire. The only form that stands out in the visitor's speech is third-person singular *don't*, which creates the impression that this form is widespread and not associated with particular places or social characteristics.

> **SAFER, TOO.** Chicago Record.

Visitor—How long has this feud continued?

Mountaineer—**'Bout thurty yeahs, sah.**

Visitor—Why don't the law step in and settle it?

Mountaineer—Well, **sah**, it's **bettah** to let it run on, **sah**, till only one gentleman is left, **sah**, and then we'll hang him.

The third example is a long fictional story written by Alfred R. Calhoun and entitled "A Mountain Missionary". The first chapter was published in the *Yenowine's News* (Milwaukee, Wisconsin) on March 3, 1895[149] . The complete story was published in the *Atchison Daily Globe* (Kansas) on March 15, 1895[150] . The main character in Calhoun's story is a white preacher, Father Peters, who is originally from Ohio but moved with his family into "the Cumberland mountains in southeastern Kentucky" to work there and convert the people living in the mountain region (labeled "mountaineers") to religion. He is the "mountain missionary" and he is characterized as a man of a very "intense religious spirit". In the beginning of the story, the mountain region is compared to his old home, northern Ohio, and the two places are presented as complete opposites: Whereas in northern Ohio "[w]ealth, or at least comfort, and intelligence were the rule", southeastern Kentucky is characterized by violence, poverty and ignorance. This contrast is underlined by contrasting the character of the preacher with a character from the mountain region: Bradley, the blacksmith, who is described as a "man of unusual physical strength", a "fighter" who had been "shockingly brutal and profane" and had "killed more than one man". However, as a result of the preacher's work, Bradley is described as having converted to religion and consequently as having undergone a "remarkable change". The story is set at the beginning of the Civil War and it describes how the mountaineers split up into secessionists and adherents to the Union, with Father Peters and Bradley belonging to the latter group. As the Union men are in a minority, it is dangerous to be open about their political views, and in the part quoted below Bradley warns Father Peters of the

danger posed by the secessionists and urges him to leave. Father Peters, however, refuses to flee, and in the course of the story he is attacked by his opponents (led by a mountaineer called Het Magoone), violence erupts, Bradley gives up his religious sentiments to become a fighter again (for the preacher and the cause of the Union) and ultimately even Father Peters takes up the rifle and shoots enemies before he is killed at the end of story.

The mountaineers are generally depicted in a very negative light and the narrator is condescending in his description of them. The main contrast established in the story is that between the north and the south, with the mountaineers representing the worst part of southern culture. The northern influence, symbolized by Father Peters, is the only positive element in the mountaineer region, but his success (visible in Bradley's conversion) is ruined when Bradley turns his back on religion again to fight back his opponents. The fact that even Father Peters takes up arms against his religious convictions conveys the strength of the negative southern influence – however, as both Bradley and Father Peters fight for the Northern cause, their fighting is justified, whereas the secessionist mountaineers are condemned because they have caused the fighting by their attack (especially by setting fire to the meeting house of the Union men). While Bradley's animalistic uncivilized fighting instincts cannot therefore be changed through religion, Father Peters has had enough influence on him that he at least uses them for a good cause. All in all, the culture of the north is portrayed as superior on all levels, and this superiority is constructed to a great extent in relation to religion. The fact that the mountaineer community views Bradley's conversion to religion very negatively (they find that "religion had took all the pluck and snap out of strong Dick Bradley") shows that they associate religion (and civilized behavior) with weakness, but the story ultimately presents the religious figure, Father Peters, as the hero: The mountain missionary becomes the martyr symbolizing the enormity of northern strength and efforts to bring some civilization to the southern mountains.

The speech of the mountaineers is represented as different from that of the narrator and that of the preacher and his wife and daughter. In comparison to the linguistic repertoire in the article "The Pool Tribe", published ten years earlier, non-rhoticity has remained a form marking the mountaineers' speech (*powaphul*, *no'th*, *fo'*, *ovah*, *heah*). This is not surprising given its salience in the humorous dialogue "Safer, Too", but it is noticeable that in this story not only do some forms remain rhotic (*yours*, *yer*, *thar*, *Lor'*, *sarmint* 'sermon') but that hyperrhotic forms occur as well (*kinder* and *tomorrer* 'tomorrow'). This combination of non-rhotic with hyper-rhotic forms is also present in eight articles representing

Black speech. Furthermore, there is one case of deletion of pre-vocalic /r/ following a fricative consonant (*f'om*). So far, this deletion has only been represented in the poem "When de Co'n Pone's Hot" (*th'one*), which makes it another case of an overlap between the linguistic repertoire of Black speakers and mountaineers. Further overlapping forms are final consonant cluster reduction (*mos'*, *Lor'*), a raised dress vowel in *get* (*git*), a lowered kit vowel in *if* (*ef* ), a choice vowel nearing price (*p'int* 'point' occurs in a part of the story not quoted here) and a lower square vowel (*thar*, *harabouts*) as well as alveolar -*ing* (*brewin*, *needin*). A lower square vowel in *where* is also found in Colonel Gutrippah's repertoire and thus linked to southern white speech as well. Alveolar -*ing* is also found in "The Pool Tribe". Two further forms shared with "The Pool Tribe" are representations of a lower nurse vowel (here *sarmint* 'sermon' and *larnt* 'learnt' in "The Pool Tribe") and metathesis of /r/ and a vowel (here *purtending* 'pretending' and *purty* 'pretty' in "The Pool Tribe"). In contrast to the linguistic repertoire represented in "The Pool Tribe" and to representations of Black voices, TH-stopping is not marked here. It is also not represented in the humorous dialogue "Safer, Too", which suggests that even though it is used by the mountaineer in "The Pool Tribe", it is not a salient form linked to mountaineer speech. It is a possibility that TH-stopping is used in "The Pool Tribe" to emphasize the "queer" race of the mountaineers (in contrast to white southerners) because since THstopping occurs in all representations of Black speech, it is constructed as an index of racial difference. In the story "A Mountain Missionary", no effort is made to present the mountaineers as racially different, but the emphasis is rather on cultural differences, especially in contrast to the north of the United States. It is also possible, however, that the linguistic repertoire associated with the mountaineer has changed over the ten years between the appearance of "The Pool Tribe" and "Safer, Too" and "The Mountain Missionary".

Grammatically, the mountaineers in this story use *ain't* and negative concord (*Thar hain't no one*) and demonstrative *them* or the double demonstrative *this*/*these here* (*these har hills*). These forms occur in representations of Black as well as of white southern speech. A frequent form that was also linked to Black speech is *a*-prefixing. An invariant participle in *some men so give ovah to sin* is reminiscent of invariant past tense forms found in "An Unwelcome Fifteenth" and "Dat Deceptious Mule". Regarding the pattern of subject-verb agreement, it can be observed that the second-person singular -*s* marking, which is salient in "The Pool Tribe", is not represented here (*Father Peters, yo've preached*). Firstperson plural marking seems inconsistent because the form *has* is used (*we'uns who's fought*) alongside unmarked forms (*we want*). Similarly, regarding secondperson plural forms, the form *is* occurs alongside unmarked forms, as in *yo' uns that don't want to heah Father Peters preach is free to leave*.

### 4.1 Metadiscourses on phonological forms

A very salient form that is only represented in "A Mountain Missionary" is the use of the first- and second-personal pronouns *we'uns* and *you'uns* (but *we* and *you* are used as well). They are highlighted in the following exchange between Bradley and Het Magoone: "'Who says so?' demanded Bradley. / 'We uns.' / 'And who's yo'uns?'" The pronouns are thus prominently linked to the mountaineers and serve as a contrast to white southern speech, indexed by the personal pronouns *we all* and *you all*. This contrast is not established in the story itself, but in relation to other stories, as for example the anecdote about Colonel Gutrippah. What is also noticeable is the frequent use of *as* as a relative pronoun (*the man ez built hit*, *the man ez kem down heah*), which contrasts with the use of the relative pronoun *what* in representations of Black or white southern speech in other articles (e.g. *de mule wot you bin zoonin' for* in "Dat Deceptious Mule").

Finally, the representation of the mountaineer's speech is also marked by the forms *hit* and *hain't* and the frequent use of eye dialect, e.g. *sespichis* 'suspicious', *ken* 'can', *kem* 'come', *dainjah* 'danger', *hull* 'whole', *ez* 'as', *religin* 'religion', which has so far also been observed in some articles representing Black speech. The use of eye dialect increases the perceived distance of the linguistic repertoire of the represented voice to that of the 'neutral' voice of the author or narrator and through the association of eye dialect with illiteracy the speaker is characterized as uneducated, which fits the depiction of the mountaineer as uncivilized and belligerent. An element of humor can be found in the representation of /f/ by <ph> in *powaphul* 'powerful', because the graphemes <ph> are normally only found in Greek loanwords, which are associated with a high degree of education. The contrast between this representation and the other forms linked to a lack of education (among others non-rhoticity in the same word) creates irony intended to entertain the educated readers. The impression of the uncivilized and uneducated mountaineer is strengthened further on the lexical level through the frequent use of *doggone*, as in *they've got so doggone sespichis* in the quotation below, because *doggone* is a swear word (according to the *OED* 2021, it is probably "a euphemistic alteration of *God damn*") and frequent swearing is associated with impolite and uncivilized behavior.

> **A MOUNTAIN MISSIONARY.** By Alfred. R. Calhoun [Copyright, 1895, by American Press Association.] CHAPTER 1. why he went to Kentucky

Mr. Robert Peters, or Father Peters, as he was called by the mountaineers, was born in Ohio. He was a Cambellite clergyman, and ten years before the war he, with his wife and daughter, moved into the Cumberland mountains in southeastern Kentucky. Only an intense religious spirit could have induced Father Peters to leave his home in the rich lands of the western reserve and to take

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

up his abode among the hills of the Cumberland range. It would be difficult in the United States to find a place and a people in more violent contrast with his old associates. Wealth, or at least comfort, and intelligence were the rule in northern Ohio. There was hardly an exception to poverty and ignorance in his new home.

### [...]

Bradley, the blacksmith, who was Mr. Peters' nearest neighbor, was a man of unusual physical strength, and before the coming of the preacher he had been noted as a fighter. Indeed he was credited with having killed more than one man. He had been shockingly brutal and profane and was known far and near as strong Dick Bradley to distinguish him from a cousin of the same name, who was not that strong. If the only good done by the clergyman had been the conversion of the blacksmith, his work in the mountains might well be considered a success. Bradley no longer swore nor drank, nor had he had a fight for years. This remarkable change, added to the fact that he was a man of few words, gave the impression to many that "religion had took all the pluck and snap out of strong Dick Bradley."

[...]

The blacksmith pulled his chair nearer, and with his big hands to the sides of his mouth to shield his voice he whispered:

"**I'm sorry to tell yo', Father Peters, that there's trouble a-brewin harabout, and lots of hit.**"

"Trouble to whom, Brother Bradley?"

"**Can't yo' guess?**"

"I cannot."

"**Waal, hit's to yo' and yours,**" said the blacksmith, with an emphatic shake of the head.

"But surely no one could wish to annoy me. I did not think I had an enemy in the world," said the alarmed clergyman.

"**That's jest hit,**" responded Bradley. "**Thar hain't no one ez doesn't allow yer the best preacher in the mountains, but the boys say they don't like yer politics, and so they'll make trouble.**"

Father Peters protested that he had not meddled in politics, and that since the breaking out of the war a few months before he had guarded his words that he might not give offense.

"**That's hit. That's why they've got so doggone sespichis. Now, Father Peters, yo' know I've allus been a good friend of yer's, ez I should be a blamed dog not to be, seeing that you took me by the hand and led me up to the light, bless the Lor! But thar's dainjah all about we uns, mos' powaphul dainjah, ez me and the wife allowed this night. And we said that ef so be yo' could go no'th fo' awhile till the trouble kinder blows ovah that hit'd be bettah fo' yo' and fo' yer friends.**"

[...]

[The secessionist Het Magoone addresses Bradley and Father Peters:]

"**We uns who's fo' the south ken wait to git even with yo' uns who's fo' the Union. Thar's plenty of time to settle them things, but what we want now is to git rid of a man f'om the no'th who comes down har ez a spy and purtending that hit's God Almighty's religin. Father Peters, yo've preached yo'r last sarmint in these har hills**."

"**Who says so?**" demanded Bradley.

"**We uns.**"

"**And who's yo'uns?**"

"**Mo and my friends.**"

"**Neither yo' nor yer friends owns a splinter of this house. From foundation log to clapboard hit's owned by the man ez built hit—the man ez kem down heah to lead us to do right—but thar's some men so give ovah to sin, Het Magoone among 'em, that they're bound to be damned, and I'm mighty glad of hit. Now, yo' uns that don't want to heah Feather Peters preach is free to leave. But, by G—, the next man ez tries to break up our worship will find himself needin a hull new top to his head!**"

# [emphasis mine]

To conclude the qualitative analysis of articles containing *bettah*, I have shown that the term and the phonological form it represents are linked to very different groups of speakers, including speakers contrasting on several social dimensions, like, for example, the white upper-class southern girl, and the Black lower-class man with fifteen children that he can barely feed. This underlines the importance of considering not only one linguistic form in isolation but the way that this form is combined with other forms to create contrasting voices linked to contrasting social characteristics and values. The analysis of the articles has revealed that non-rhoticity is not the only form that is linked to all three main speaker groups, to Black Americans, white southern Americans and mountaineers. Unstressed initial syllable deletion, a lower square vowel, particularly in the function words *here*, *there* and *where*, a lower nurse vowel, different patterns of subject-verb agreement, negation with *ain't* and negative concord, the double demonstrative *this here* (and related forms), the modal better construction, the forms *hit* and *hain't*, and the lexical item *I reckon* and the address term *sah* 'sir' all occur in the speech of at least one mountaineer, one white southern speaker and one Black American. How are differences between these groups indexed then?

One important element also regards the presence or absence of /r/: Many representations of Black American and of mountaineer speech exhibit several forms that are *not* marked as non-rhotic and, in addition, hyper-rhotic forms. In contrast to that, there is only one article in which several forms of white southern speech are not marked as non-rhotic and that is "First Impressions Entering the Field", which discusses lower-class white southern Americans and highlights their similarities to Black southern Americans. Hyper-rhotic forms do not occur in any of the representations of white southern speech. How can the co-occurrence of non-rhoticity and hyper-rhoticity in Black and mountaineer speech be explained then, especially if it is taken into account that the analysis of articles containing *deah* AND *fellah* revealed that these two forms constitute end points on a continuum representing the difference between nature (the hyperrhotic cowboy) and culture (the non-rhotic dude)? First of all, it is important to point out that there are also articles containing *bettah* in which such a continuum is visible, as the article "A Score" shows, which consists of a humorous dialogue

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

accompanied by an illustration of the speakers (Figure 4.19). It was published in the *Atchison Daily Globe* (Kansas) on September 11, 1888[89] , and represents the voices of two speakers, Mr. Delawney and a man labeled "Hard Up Gent". The Hard Up Gent is obviously poor and asks Mr. Delawney for a few cents, but the request is declined by Mr. Delawney, who arrogantly accuses the Hard Up Gent of not having brains enough to care for himself by saying "You had bettah ask for bwains instead of money". To this the Hard Up Gent replies "Well, boss, I asked

Figure 4.19: Humorous dialogue and illustration of a rich and a poor American, published in the *Atchison Daily Globe* (Atchison, Kansas) on September 11, 1888[89] , retrieved from Nineteenth-Century U.S. Newspapers

yer for what I thought yer had the most of", which not only shows that he is very smart but also critically exposes Mr. Delawney's unjustified arrogance and makes him appear more ignorant than the Hard Up Gent. The Hard Up Gent's speech is marked by hyper-rhotic forms (*yer*, *feller*), while Mr. Delawney's speech exhibits non-rhoticity in *bettah*, *faw* and *yaw*. As non-rhoticity is combined with a labiodental realization of /r/ (*bwains*), Mr. Delawney exhibits two typical forms of the repertoire of the dude or the swell. He therefore represents the end point of culture (pointing out the importance of education and intelligence to lead a successful life), while the Hard-Up Gent represents the end point of nature (having to struggle to fulfill his basic natural needs). The humor of the dialogue is created by showing how the poor man outwits the swell, who thinks so highly of himself, and it therefore serves to criticize upper-class arrogance towards the poor.

I suggest that by keeping rhotic forms and by adding hyper-rhotic forms to the voices of mountaineers and Black Americans, the indexical link to nature (as opposed to culture), which is also established in the dialogue in "A Score" to characterize the Hard Up Gent, is used and strengthened. While in the case of the Hard Up Gent, nature is presented positively by showing how his intelligence and wit make him superior to the over-cultured dude, it is portrayed in a mostly negative sense in the case of Black Americans and mountaineers, however. Nature stands for the absence of education and civilized behavior, which leads humans to rely on their physical strength and animal-like instincts instead of their intellect and morals. For example, the fifteen children of the Black man in "An Unwelcome Fifteenth" imply a lack of sexual restraint having negative consequences for the whole family, the Black man's use of religion to justify stealing a watermelon to satisfy his craving for the tasty ripe fruit in "Cato's Soliloquy" suggests a disregard of morals, and the mountaineer helping the reporter in exchange for whiskey in "The Pool Tribe" exposes him as as a person who is very fond of alcohol and whose greatest skill is killing rattlesnakes. In these cases, non-rhoticity does not index culture or refinement in any way, but the main indexical link created in these articles is that between non-rhoticity and lack of culture and civilization. In this line of argumentation, it makes sense that the articles representing white southern speech do not contain hyper-rhotic forms because southerners are not presented as uncivilized or animalistic. On the contrary, in several articles their system of cultural values is implicitly put in relation to that of the north, usually with the aim to criticize it as traditional, conservative, hierarchical and backward in contrast to the more modern and progressive northern one. It is therefore possible that non-rhoticity indexes culture and refinement in these articles as well, especially in those cases where upperclass southerners are portrayed, for example the southern girl in New York. As non-rhoticity also occurs in lower-class southern speech, however, I suggest that the main indexical link is that between non-rhoticity and southernness and this link is also present in the case of Black Americans and mountaineers because these groups are also predominantly associated with the American south in the nineteenth century. While non-rhoticity therefore indexes a variety of social categories and meanings, the addition of hyper-rhotic forms puts an emphasis on the indexical link to natural and uncivilized behavior and cancels potential links to culture and refinement.

Another form that is present in representations of Black American and mountaineer speech is voiced interdental fricative stopping. As this form is the only form next to non-rhoticity that is marked in all representations of Black American speech, it seems to index ethnicity most strongly. I have argued above that this link could also have been used create the impression of racial otherness of the mountaineers in "The Pool Tribe". However, considering the analysis of *deah* AND *fellah*, voiced interdental fricative stopping is also marked in the speech of the white Philadelphian "waifs", which suggests that while the form is an important index of ethnicity, it can also index a lower-class status, lack of education and uncivilized behavior more generally. In contrast to voiced interdental fricative stopping, voiced *labiodental* fricative stopping occurs in the speech of Black Americans only, which makes it a strong index of Black American speech. It is present in the majority of articles analyzed here, which makes it all the more striking that it was marked in the first version of Dunbar's poem "When de Co'n Pone's Hot", but not in the versions published after he had become famous. As Dunbar's poem is an exception in that it creates a positive image of Black American life and identity, this change could indicate that especially voiced labiodental fricative stopping might index very negative social values or perhaps that it became less frequent in actual use and thus less suitable as a marker of Black voices at the end of the nineteenth century. More analyses focusing on this form would be needed to support these suggestions, however.

Further forms which are frequently used in the articles to mark Black American speech are alveolar -*ing* and final consonant cluster reduction. Both forms appear in mountaineer speech as well, which shows that they do not only mark ethnicity, but that they can also mark lower-class status and lack of education. Less frequent forms which are also part of the repertoire of both Black Americans and mountaineers are the deletion of pre-vocalic /r/ in initial consonant clusters, the realization of the choice vowel nearing the price vowel, a raised dress vowel in *get* and a lowered one in *if*, demonstrative *them*, alternative reflexive pronouns and *a*-prefixing. In some articles, eye dialect is used to highlight

the illiteracy of the speakers. Forms that distinguish the two social groups are fewer, but I have shown in the analysis how they are highlighted: In the case of the mountaineers, it is particularly the personal pronouns *we'uns* and *you'uns* that contrast them with Black as well as with southern speech. Lexical items like the swear word *doggone*, the phrase *I'll haf to go you* and the address term *mistah* also occur only in representations of mountaineer speech. The relative pronoun *as* differentiates them from white southern and Black voices, which are sometimes marked by using the relative pronoun *what*. A very salient element used in some articles to distinguish Black American speech from mountaineers (and also from white southern Americans) is the use of malapropisms. They serve to highlight the failed attempt by some Black speakers (especially Black preachers) to appear educated and to mark a higher social position. That the mountaineers are not portrayed as using malapropisms shows that they are characterized as not even trying to appear civilized or educated – they are rather depicted as being content with their uncivilized life full of violence and hardship.

With regard to white southern speech, there is one form present in four of the five articles (next to non-rhoticity): the address term *sir* (spelled <sah>). The term is also used by the Black man in "An Unwelcome Fifteenth" and the mountaineer in "Safer, Too", which shows that it is used to index differences in social hierarchies between the speakers, but its frequent use to mark southern speech shows that it is used to emphasize the southerner's hierarchical social system. Forms restricted to southern speech are the lexical item *How'dy* and the different use of the greeting *good evening* and the personal pronouns *we all* and *you all*, but they all occur in one article only and are thus not as prominent as other forms here. (However, a detailed analysis of more articles could provide more insights on their prominence.) Two phonological forms unique to southern speech are the spelling of the article *the* as <they> and, in one instance, the spelling of the dress vowel as <ai>, which could be representations of diphthongizations typical of the so-called southern "drawl".

Considering all articles containing *bettah* analyzed here, a general observation that can be made is that representations of Black speech usually contain a much higher number of linguistic forms marked as different from other voices. Even very short stretches of speech, like Uncle Eben's "Philosophy" or the caption under the cartoon showing the two boys and their dogs, exhibit roughly the same amount of forms as the much longer quotations of Colonel Gutrippah's speech or the janitor's speech in "Relics of the Past". The highest number of deviating linguistic forms also occurs in two texts representing Black voices: the short story "Dat Deceptious Mule" and "Cato's soliloquy". Regarding the sheer frequency of deviating features, representations of mountaineers come second, but as the

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

example of the mountaineer in the humorous dialogue "Safer, Too" illustrates, a restriction to a smaller number of forms is possible, too. Those are non-rhoticity, unstressed initial syllable deletion, a lower nurse vowel and the address term *sah*, so phonological and lexical forms only, which illustrates the emphasis put on the phonological and the lexical level when only a reduced number of forms can be represented (due to the shortness of the dialogue). The voice that is least marked is that of the upper-class southern girl. While she exhibits several non-rhotic forms, the only other forms are yod-insertion after a velar consonant and a lower nurse vowel, which are both indicated in the spelling of only one lexical item (<gyuls>). This establishes a parallel to the representations of the swell and the dude, whose speech is also characterized by a reduced set of forms, which marks them as clearly different and links them to specific social characteristics, but which is at the same time not deviant enough to make them appear uncivilized or uneducated.

All in all, the analysis reveals the complexity in the creation of indexical links between linguistic forms and social characteristics and the high importance of context in the interpretation of these links. To illustrate this important conclusion, I will analyze one last example based on the findings generated in the analyses of the search terms *deah* AND *fellah* as well as *bettah*. It is a report about a new fashion item, shoes of yellow colored leather labeled "Yellow Shoes" in the heading, which was published in the *Evening News* in San Jose, California, on July 10, 1889[94] . The report quotes "a bootblack" as saying "You bettah take dem shoes off; you catch rumahtism suah!" Even without any further information about the bootblack, the representation of his speech indexes his ethnicity. The combination of non-rhoticity, demonstrative *them*, voiced TH-stopping and eye dialect could possibly also index a mountaineer, but it is highly unlikely that a mountaineer would work as a bootblack in a city. This particular combination of linguistic forms makes it clear to the reader that the figure of the dude or the swell and social characteristics associated with this figure cannot be evoked here. In the following part of the report, a "leading dealer in shoes" is quoted as saying that these shoes are particularly fashionable in the east and that they are sold in equal numbers to black shoes there, but that "in San Jose our customers are afraid of them, fearing they may 'swell'". This indicates that Americans living in the west are afraid of being negatively evaluated as eastern swells. The dealer is convinced, however, that this fear will "wear off however as many gentlemen are now wearing them, who have nothing, either in their manner or appearance to indicate the 'swell.'" This shows that there is an awareness of which perceivable signs index a swell, but language is not the focus here. Non-rhoticity in *bettah*, which could be interpreted as a form indexing a swell if one is aware of the fig-

### 4.1 Metadiscourses on phonological forms

ure and the signs indexically linked to it, is however not linked to the swell here at all. Although the article is concerned with both figures, the Black American and the white swell, and therefore offers a *potential* for recognizing the overlap in their linguistic repertoires, this overlap does not play any role at all, because the linguistic as well as the non-linguistic context clearly mark the bootblack as a Black speaker and because it does not invite the reader to create or draw on an indexical link between non-rhoticity and the swell that they might use or recognize in other contexts.

# **YELLOW SHOES.**

They are Universally Condemned by the Bootblacks. **Said to Attract Rheumatism–The Sale of the Shoes Increasing–Extensively Worn in the East**

"**You bettah take dem shoes off; you catch rumahtism suah!**" said a bootblack to a young man who passed his stand this morning, wearing shoes of yellow colored leather.

As the young man passed on, the bootblack stated that before the wearer of the shoes had adopted the yellow leather for footwear he had patronized the shoe polishing establishment to the amount of \$2.50 to \$3 a month, and that if the custom of wearing canvas and light colored shoes continued, and it seemed to be growing in popularity, the trade of the shoe polisher would receive a severe blow.

A leading dealer in shoes, speaking of the sunset leather and canvas shoes, said "We do not sell a great many of these shoes but the demand seems to be increasing. The majority of calls for these shoes come from Eastern people and as we do not carry a great variety of styles we cannot always suit them. I am told that in the East the sale of them is equal to that of the ordinary black leather, but **in San Jose our customers are afraid of them, fearing they may 'swell'**. This will soon wear off however as many gentlemen are now wearing them, who have nothing, either in their manner or appearance to indicate the 'swell.' As a matter of economy the brown and yellow colored shoe will receive ready endorsement, as the amount of money spent with the bootblacks on one pair of black leather shoes would pay the original price of the shoes several times over."

[emphasis mine]

The last phonological form that I will analyze here is yod-dropping based on the collection of articles containing the search term *noospaper*. In the next section I will show how this phonological form connects to the linguistic repertoires and indexical links identified in the analyses above and how it also adds to them in ways that are important for defining an 'American' register.

# **4.1.2.4** *noospaper/s*

The first article in the databases that contains the search term *noospaper* was published on November 29, 1816[2] , in the *Western Monitor* in Lexington, Kentucky. It is a letter taken from the *National Pulse*, another Kentucky newspaper, and addressed to "Doctor Hun", that is to Dr. Anthony Hunn, who was the editor of the *National Pulse* at the time. The writer of the letter, who signs his name as

Fearnawt Burly, expresses his pleasure and agreement with a person who, in an article for a Frankfort newspaper, threatened someone ("that feller") who wanted to "skews" ('excuse') Gabriel Slaughter, who was the Governor of Kentucky at the time.<sup>19</sup> The threat is quoted and involves physical violence ("I will give it to you under the short ribs") and this violence is emphatically celebrated by the writer of the letter ("Yes, thats right! gouch hem! bite 'm! kick em, Hurrah for libberty!"). The political stance associated with this threat and celebration of violence is republicanism – the writer of the article for the Frankfort newspaper is said to be "like a troo republican". The writer of the letter in the *National Pulse* conveys his understanding of republicanism by emphasizing the high value placed on liberty ("Hurrah for libberty!") and the violent nature of their debates, which do not really allow for contradictions and arguments ("If one said tit to my tat—pang! He should have it behind his yeers"). The positive evaluation of this violent way of arguing and achieving liberty is underlined by contrasting Republicans with "book rats", a derogatory term for highly educated people, which suggests that the writer of the letter does not evaluate education and a civilized, well-informed argument very positively.

The telling name of the author (Fearnawt Burly) and the exaggerated praise of the Republicans already reveal that this letter has not really been written by a Republican writer, but that it is a piece of fiction full of irony used quite to the contrary to criticize Republican politics and manners. This is not surprising given that the *Western Monitor* was a Federalist newspaper. The linguistic form of the letter is an important part of conveying this criticism. Its most prominent feature is the highly frequent use of eye dialect to create the image of an uneducated and almost illiterate writer (e.g. <munstrusly> 'monstrously', <riter> 'writer', <Slawter> 'Slaughter', <wood> 'would', <manached> 'managed', <troo> 'true'). The style is more spoken than written, which is underlined for example by the use of the onomatopoeic word *pang*, and which creates the impression that the writer is not familiar with conventions of written texts at all. The frequent exclamatory sentences suggest that the speaker is emotionally highly involved, and they also serve to characterize the writer as a typical Republican – aggressive and impulsive rather than thoughtful and analytic. Spellings indicating an actual difference in pronunciation are very few. Next to yod-dropping in *noospaper*, there is also a case of initial yod-insertion before a vowel in *yeers* 'ears'. The spelling of *contradict* (<cunterdict>) indicates a metathesis of /r/ and the following vowel, and *feller* represents a case of hyper-rhoticity. Grammatical forms are

<sup>19</sup>Gabriel Slaughter, a member of the Democratic-Republican Party, became Governor in October 1816. Shortly after, he appointed John Pope as Secretary of State, which was criticized by Republicans because Pope was considered to be a Federalist.

### 4.1 Metadiscourses on phonological forms

first-person singular *is* and the regularized past participle *showed*. Against the background of the frequent use of eye dialect, these forms marked as constituting a differentiable linguistic repertoire are not very salient. The main social characteristics that they are indexically linked to are a Republican political stance, and connected to that, a way of argumentation emphasizing physical strength over civilized and educated behavior. Interestingly, the presence of the hyperrhotic form *feller* in this article indicates that the link between hyper-rhoticity and the nature end of the nature-culture continuum is already established early in the nineteenth century. The fact that the political debate takes place in Kentucky and in Kentucky newspapers could also link yod-dropping to the south, but there is no indication as to which place the south and southern speech forms should be contrasted to, which makes this not a very salient indexical link.

Doctor Hun!

### *From the National Pulse.*

I is munstrusly pleased with that riter in the Frankfurt noospaper, who like a troo republican cuts the matter short with that feller who wants to skews Slawter. He says, says he, "if you dont treet that [?] Mungomery with more respect, I will give it to you under the short ribs" Yes, thats right! gouch hem! bite 'm! kick em, Hurrah for libberty! Was I at Frankfert, I'd show them what it was to cunterdict mee! If one said tit to my tat—pang! He should have it behind his yeers. I wood not let them speechify matters in the sembly at all at all! If too of them did cunterdict each other I wood have a ring made round them on the spot, and then, hurrah bois, for libberty! Who hallers enough! Shall be in the rong—woodent that be a show and cheap way of carrying on the Government? That is the way Harry Clay fist his business—and if a man says Harry Clay is not a troo republican, he is a d—d lire! Let 'em come to mee till I put it to him under the short ribbs!

# […]

Docter, Docter! You dont know what we call heer Republicanism! If the war had lasted two or three yeers longer, we would have show'd you book rats how the rites of the people shood be manached!—Hurrah for liberty!

# FEARNAWT BURLY.

The second article containing *noospaper* is special because it was reprinted frequently in several newspapers over a time span of eight years (1826–1834) and across a geographical area including Louisiana and Florida in the deep south, North Carolina and Washington D.C. further north and five New England states (Maine, Vermont, New Hampshire, Massachusetts and Rhode Island) in the far northeast of the United States. The original article published in the *Louisiana Advertiser* is not contained in the databases, but it is indicated as the original source in all the other articles. The first article which appears in the databases was published in the *Carolina Observer* (Fayetteville, North Carolina) on June 21, 1826[6] . It is a long dramatic text consisting of stage directions and dialogues between three characters, Mr. Eagernoos, his son John and his neighbor Swallow, which constitute one scene which takes place early in the morning. Mr. Eagernoos sends

his son John to the neighbor to borrow a newspaper. He comes back without one and explains to his father how all of the neighbor's newspapers are already being read by other neighbors. This makes Mr. Eagernoos angry and impatient and he sends his son away to try again. After he exits, Swallow enters the scene and they talk about the problem of obtaining a newspaper to read in the morning, and when John returns several times empty-handed, Eagernoos finally decides to get a subscription to the newspaper to "not be so pestered with the trouble of borrowing from unaccommodating neighbors" anymore. The main message created by the scene is that it is beneficial to have a subscription to a newspaper instead of having to rely on neighbors to borrow them. The text therefore functions as an advertisement for the newspapers, which might be the main reason for the popularity of the article. Newspaper editors and printers were likely to have had an interest in illustrating the advantages of a subscription.

It is in the nature of a dramatic text that the voices of the characters are represented directly. A contrasting linguistic repertoire is created here by marking several forms in Swallow's speech as different from that of Eagernoos and his son. A very prominent form is yod-dropping because it is the only one that is particularly highlighted. Two strategies are used to achieve this: First, Swallow pronounces his neighbor's name as *Eagernoos* – the spelling <oo> indicates the yod-less pronunciation of the telling name *Eagernews*. Second, Eagernoos explicitly comments on Swallow's alternative pronunciation: "you are always inquiring after "Noos" as you call it". In addition to yod-dropping, Swallow also exhibits a metathesis of /s/ and /k/ in *ask* (*ax*), one instance of definite article reduction (*t'other*), an instance of invariant *be* (*be they reading them now?*) as well as an instance of demonstrative *them*, a relative pronoun *what* and a third-person plural -*s* marking in *them folks what brings the paper*. In general, it is thus Swallow's speech that is marked as deviant and it is indexically linked to his bad character, which becomes evident when he suggests that Eagernoos should only pay the five dollars asked for at the beginning of the subscription, but not the five dollars at the end because he can count on getting the paper despite not paying the bill. Swallow is thus constructed as a negative example of a subscriber, whose behavior is causing great damage to newspapers. With regard to yod-dropping, it is noticeable, however, that Eagernoos also uses the form *noospaper* once. However, as he explicitly distances himself from saying "noos", and as he is only shown to drop /j/ this one time, the link between his character and the linguistic form is rather weak. In general, the article indicates that yod-dropping is used in combination with other linguistic forms to underline negative character traits and behavior (Swallow's immoral attempt to get newspapers without fully paying for them) and to contrast them with the linguistically unmarked speech of a

character with positive character traits and behavior (Eagernoos's change from immorally attempting to read the neighbor's paper to subscribing for a paper, paying even more than is required and deciding not to lend it to anyone, is precisely the kind of behavior that is beneficial for the newspaper business). As the scene takes place in the south, an indexical link to southernness could also be formed, but, as in the first article, it is not very salient, especially considering that it is mostly Swallow whose speech is marked as deviant. The focus of the article is thus rather on social contrasts than on regional contrasts.

# From the Louisiana Advertiser.

Oh that my enemy would—Take a Newspaper.

"John! Oh John!—do you hear? run to neighbor Liberal's and ask him if he will oblige me by the loan of the morning's paper a few moments, just to look at the ship-news and the advertisements."

"That's just what I said yesterday morning, daddy, when I went to borrow the paper, and you know you kept it two hours and he was obliged to send for it."

"Well, then say something else to him, John, do you hear, John? and give my compliments, John, do you hear?

"Yes, daddy." (*Exit and returns*.)

"Well, John, have you got the paper!"

"No, daddy, neighbour Liberal is walking about the room waiting for Mr. Newsmonger to finish reading the Louisiana Advertiser, or Mr. Longwind to drop the Gazette, which he has got almost asleep over."

[…]

# (*Enter Swallow*.)

# "**good morning neighbour Eagernoos—any thing noo?**"

"New! fire and faggots, I have sent a dozen times to Liberal there, to request the loan of his paper, only for a moment, and he has the impertinence to refuse me."

### "**Refuse you?**"

"Not exactly refused me, but he permits such fellows as Longwind, Neitherside, Scribelerus, and Newsmonger, to pore over them for hours, not only (through a mistaken courtesy,) depriving himself, but his neighbors, from getting early intelligence of that is passing in the world."

# '**My goodness!—be they reading 'em now**?

"Yes" (*sighing*)

# "**Well, that's abominable! Why dont you take a Noospaper yourself?**"

"Why dont you take one? you are always inquiring after "Noos" as you call it."

# "**Why I did take one, but the printers dont leave it at my house any more, 'cause I hackeled about the price, and wood'nt pay him**."

"That's a good reason for the printer, if it is none for you. Well, John, did you get the paper."

"No, daddy, just as that Mr. Neitherside was done, in come Mr. Hookit and Mr. Knabit, and I come back."

"Confound my ill luck!—go back, do you hear? and ask Mr. Liberal if he will be kind enough—do you hear? kind enough to lend me any northern paper he may have, or if he has not one, ask him

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

to lend me yesterday's paper again, or the day before, or the day before that, or last Saturday's, or, do you hear? any of the last week's papers, do year?

"Yes daddy."

"I am determined on going right away and subscribe for a noospaper: I will not be so pestered with the trouble of borrowing from unaccommodating neighbors."

"**You are right, Mr. Eagernoos, the printers only ax five dollars right down, and then you have a whole year to pay t'other five dollars in, and then you can dispute the bill, and they will send the noospaper three months after that afore it is settled—them folks what brings the paper always throws it into a what had taken it, never thinking the subscriber is done over**."

"Here comes John—well John, have you got the paper? "No daddy, the neighbors borrowed all the old papers, and Miss Parrot sent to get the morning papers as soon as they were done with."

"The devil she did—then I may hang up my fiddle 'till sundown, for when she begins to read 'tis from alpha to omega. Give me my hat, John, do you hear? Never mind breakfast; neighbor Swallow, will you accompany me to the printing office? I will subscribe immediately; five dollars did you say? I will give twenty five before I would suffer such impertinence. If I lend *my paper* I wish I may be—."

[emphasis mine]

The articles above show that the search term *noospaper/s* appeared in newspaper articles already very early in the century and that in both cases the term is linked to a southern context, but it is not primarily used to mark region but rather to mark a negative character or political stance. Both these aspects, character and politics, are combined in a set of articles which have been called the "Nasby letters". The letters are largely responsible for the striking peak in the number of articles containing *noospaper/s* in the 1860s that I have described in §4.1.1. These letters were written by David Ross Locke, who was born in New York and became a newspaper reporter, editor, printer and owner in Ohio. He did not write the letters using his own name, however, but he constructed a character called Petroleum Vesuvius Nasby, a southern postmaster, who is described by Blair & McDavid (1983: 144–145) as "a bigot [and] an ignoramus, a hypocrite, a sluggard, an alcoholic, a coward, a bigamist, a thief, a corrupt politician, and a traitor". His name alludes to his explosive and unrestrained temperament, which is evaluated negatively. He embodies the political views and social behaviors which are in complete opposition to Locke's own views and norms and the letters are thus satirical pieces of writing intended to expose and criticize southern politics and culture during the Civil War years and afterwards. Locke began writing the letters in 1861 and continued until the 1880s and their popularity and wide circulation in the North made Locke one of the best-known humorists in America (Blair & McDavid 1983: 144).

The letter I have chosen for analysis here was published in the *St. Louis Globe-Democrat* (Missouri) on September 11, 1876[34] , and it was taken from the *Toledo*

*Blade*, an Ohioan newspaper owned by Locke. The sub-heading "Why the Nigger is a Trubble, a Worriment and an Irritashen—How a Hawty, Shivelrus People Hev Bin Obleeged to Succum to Force" already shows the satirical tone of the letter: Nasby's display of contempt for Black people does not fit his praise of southerners as haughty and chivalrous. These two topics are elaborated further in the letter. First of all, Nasby claims southerners to be superior to northerners: They are aristocrats and are thus superior to northern men who are just "mer labrin men, or mer men of biznis", who are not enlightened and not able to understand the situation in the south, which is why they need Nasby to enlighten them. Locke thus ridicules the southerners' pride in their aristocratic past and their emphasis on their feelings and traditional manners and behaviors, which, in their view, make them superior to hard-working northerners. Having a thoroughly negative character like Nasby express such a view has the effect that the southerners appear deluded and it emphasizes Locke's position that northern businessman do more for the progress of the nation than lazy southerners, who live off their plantations and other people's hard work. Similarly, it is clear that it is in fact Nasby who is ignorant and in need of enlightenment and not northern people. This concerns first and foremost his attitude towards Black people: Nasby's stupidity and lack of knowledge and intelligence disqualifies him and his negative views and rather convinces the readers to distance themselves from his positions. The letter thus rests on the depiction of Nasby as an uneducated and unintelligent brute, whose views are ridiculous and not to be shared by cultured and educated (northern) people, and this depiction is also achieved by means of language. The most notable element of the Nasby letters is the frequent use of eye dialect – examples in the short extract quoted below are *conclooshen*, *enliten*, *nacheral*, *biznis*, *hawty*, *shivelrus*, *succum*, *trubble*, *noboddy*, *bin*, *irritashen*, *absloot*. As pointed out several times already, eye dialect also functions here as an effective means to portray Nasby as so uneducated that he is unable to spell words correctly. It also creates humor and marks the character as inferior to the educated reader who detects the misspellings. Pronunciation respellings are much rarer. In the extract below, they mark alveolar -*ing* (*Reedin*, *bein*, *labrin*), which also extends to nouns (*feelins*), forms of connected speech (weakening in *hev*, *uv*, *kin*, *ez*, *wuz*, and elision in *em*) and the backing of strut (*onrestrained*, *oncontrolled*). While alveolar -*ing* and forms of connected speech are often used to mark the uneducated speaker, strut backing is a phonological form that marks Nasby's voice as different from that of voices represented in other articles. The representation of yod-dropping in the letter is interesting because by using the spelling <oo> to represent the vowel /uː/ in all words and not just in those in which yod-dropping occurs, Locke draws attention to the vowel, but he does not

highlight yod-dropping in particular.<sup>20</sup> This makes the Nasby letter similar to the first letter analyzed here because in both letters eye dialect predominates and <oo> is a case of eye dialect in some words (*troo*) and a marker of yod-dropping in other words (*noospaper*). This connection between eye dialect and yod-dropping reinforces the indexical link between the form and the uneducatedness of the speaker, which is also indicated in the Nasby letter on the grammatical level by the commonly used negation with *ain't* and negative concord (*ther ain't no question*), and on the lexical level by representing the pronunciation of *oblige* as /əˈbliːdʒ/.

So overall, even though the Nasby letters are largely responsible for the high frequency of articles containing *noospaper/s* in the 1860s, this does not mean that they contributed to the salience of yod-dropping in discourses on language because yod-dropping is not particularly highlighted here. The prevalence of eye dialect can also be found in other articles containing *noospaper/s* and published in the 1860s: They were written by the humorist writer Charles Farrar Browne under the pen name Artemus Ward and also reached a high popularity and a wide circulation. It is therefore rather eye dialect in combination with a few differential phonological, lexical and grammatical forms that become associated with humorist writings which pointedly ridicule uneducated speakers and their views – in contrast to other articles, the focus is thus more on their differential spelling than on their differential voices.

# **NASBY. Why the Nigger is a Trubble, a Worriment and an Irritashen—How a Hawty, Shivelrus People Hev Bin Obleeged to Succum to Force.** [From the Toledo Blade.]

Confedrit X Roads, wich is in the State of Kentucky, September 4, 1876.—

Reedin Northern noospapers for some weeks past, I hev come to the conclooshen that the people uv the North don't understand the troo status uv things down here, and I feel it my dooty to enliten em. It is not nacheral that a Northern man kin understand the feelins uv a Southerner. The Northerner never wuz a aristocrat like us—he never wuz a sooperior race. But bein mere labrin men, or mere men of biznis, or sich, they kin hardly be expected to comprehend how some things strike the minds of hawty, shivelrus people, which hev bin obleeged to succum to force.

That the nigger is a trubble to us ther ain't no question, and noboddy denies uv it. He is a worriment and an irritashen, and more than that an absloot noosence, and there never kin be peace so long ez he is onrestrained and oncontrolled.

[…]

<sup>20</sup>The question whether <oo> could not also represent /juː/ can be answered by looking at the spelling of *perpetually* as <perpetyooally> occurring later in the article. Here, the <y> seems to be used to represent /j/ before /uː/ (spelled <oo>), which shows that the absence of <y>, for example in <noospaper>, represents a pronunciation without /j/.

### 4.1 Metadiscourses on phonological forms

An article which also constructs links to uneducatedness in a southern context is a story which is not told by a reporter directly but indirectly using the voice of "E. H. Barclay, a New York traveler at the Lindell". It is headed "Out-of-Town People" and was published in the *St. Louis Republic* on August 18, 1894[143] . Barclay is quoted at the beginning as saying "I heard a good story once purporting to explain how the town of Rondo, Ark, got its name". This shows that the story has an anecdotal character because it is said to be based on facts rather than fiction. Barclay then tells the story of the steamboat captain James Crooks, "who sailed on the Red River" and got stuck because he was delayed and the water became too low to continue sailing. He and a passenger then used a steam sawmill and pool tables, which they had on board their boat, to start a business on the riverbank, and around this a town sprang up and because of the gambling resort it was named after the favorite game played there: Rondo. The "natives" of the town are characterized as uncivilized ("the natives were using cuss words, bowie knives and revolvers in settling disputes over the game of rondo") and they are described as spending much of their time gambling and drinking bad whisky. In the part of the story quoted below, the native inhabitants' first encounter with newspapers is described. They decided to get a newspaper to be "informed on the doings in the world" and because they were illiterate, they got a schoolteacher to read it to them in exchange for free whisky. The humor of the anecdote is created by contrasting the schoolteacher's conviction that he is educated and intelligent and thus superior to the other native inhabitants with his actual lack of education, which is exposed when he explains to the natives that the word *immigrant* designates "a little varmint about the size uv a gray squirrel".

The southern natives are therefore portrayed in a very negative light in the anecdote, the main characteristics highlighted being their ignorance, backwardness and lack of civilized behavior. The figure of the schoolteacher serves to reinforce these qualities because a teacher would be expected to possess a high degree of education and to function as a role model in the community, but he turns out to not be very educated either and his regular consumption of whisky also marks him as being just as uncivilized as the other natives. The heading "Out-of-Town People" also creates a contrast between rural and urban areas of the south, locating the town Rondo in the rural periphery. The voices of the southerners are constructed as different from the voice of the story-teller Barclay, who, as a New Yorker, is not only connected to the north but also to the urban sphere. Next to yod-dropping in *noospaper*, the native inhabitants of the area are also shown to use alveolar -*ing* (*expectin'*, *l'arnin'*), a lower square vowel in *thar* and *hyar*, yodinsertion after /h/ in *here* (*hyar*), a higher trap vowel (*thet* 'that', *dern* 'darn'), a lower nurse vowel (*l'arnin'*), a lower kit vowel in *if* (*ef* ) and a lower dress

vowel in *well* (*Waal*). Features of connected speech are marked as well (weakenings in *uv*, *wuz tuh* and elisions in *'em* and in *more'n*). There is also an instance of eye dialect (*wuddent*). Grammatically, the double demonstrative *this hyar* is a form used by the schoolteacher. All of these forms have also been identified in articles containing *bettah* as forms connected to Black, white southern and mountaineer speech, but it is noticeable that non-rhoticity is not marked here at all. This shows that while non-rhoticity *can* index southernness, it does not have to be present. This article also shows that yod-dropping is also not a very salient form, as it only occurs once and as it is not particularly highlighted. All in all, the representation of speech serves to underline the ignorance and uneducatedness of the (white) southern inhabitants in rural areas and they are depicted as using linguistic forms that evoke these values in other contexts as well, especially in contexts aiming at deriding mountaineers or Black Americans.

# OUT-OF-TOWN PEOPLE.

[…]

"Now, the natives at that time, with few exceptions, had never seen or heard of such thing as a newspaper. One of these ignorant natives asked what sort of a dern thing a newspaper was. The passenger explained how \$1 50 would pay for the subscription of a good weekly newspaper published in St. Louis, and thereby the citizens could keep themselves informed on the doings in the world.

"This was a startling innovation to the natives. A collection was taken up in a hat, and the money was sent off for a St. Louis newspaper. In due time the paper came, and then a new difficulty stared them in the face. Who could be found to read it? Upon somebody's suggestion the school teacher of the neighborhood was selected to do the reading. But the pedagogue had an eye single to business, and seeing that he had a monopoly on the intelligence of the community, he forced a bargain that he was to get his whisky all week free of charge for reading the paper regularly on the day of its arrival at the saloon.

"The first day the paper arrived the country pedagogue wet his whistle at the end of every other sentence, and occasionally when he struck a big word that had been used by a green reporter at a fire, he would stop at a comma, even, and dampen his throat with two glasses of whisky. He finally read an item stating that the corn crop of Texas was magnificent, and that the people of Texas were expecting a large immigration in consequence.

"'Hold on, thar! What's thet you read?' asked one of the natives.

"'Why, this hyar noospaper says that Texas has a big corn crop,' replied the school teacher, 'and thet they're expectin' a mighty big immigration on account uv it?'

"'Waal, what is immigration?' asked the illiterate native.

"'Why, you fool, immigration means immigrants coming to the State,' explained the school teacher.

"'Yes, but what sort uv a dern thing is an immigrant?'

"'Now, my friends,' replied the pedagogue, assuming a look of wisdom, 'but very few of you have got any book l'arnin' at all, an' ef I wuz tuh tell all the Latin name you wuddent know any more'n you do now. But an immigrant is a little varmint about the size uv a gray squirrel. I don't know much about 'em, but they're hell on corn.'"

### 4.1 Metadiscourses on phonological forms

An anecdote containing the search term *noospaper/s* that foregrounds differences between places or regions is headed "Didn't Know the Place" and it was published in the *Daily Arkansas Gazette* (Little Rock, Arkansas) on July 27, 1883[58] , but the *New York World* is given as its original source. It consists of a dialogue between "a man with Kentucky jeans on", which implies that he is from Kentucky, and a man whom he meets in the "Broadway corridor of the building", which implies a New York setting. The Kentucky man asks for the post office, and he is astonished when he finds that his assumptions about what a post office should look like, which people are found there and how the postmaster behaves turn out to be wrong. The Kentucky man's assumptions are used to convey to the reader the typical make-up of a post office in Kentucky: There are "fellers who sit around on the barrels and tell stories", there is "a fellow wots got the terbacker", there is "a minister [who] come in an' borry a postal card till he gits a whack at the plate" and there is "a postmaster [who] read all the papers an' postal cards before he sends 'em home". All these characteristics are negative and make the post office and the people in New York seem superior to those in Kentucky. This in turn is the source for the humor created by the Kentucky man's last statement: He finds the New York post office to be "the blastgamedest postoffice I ever seen" and decides not to trust them and to "send the letter home" himself, which is ironic because it is apparent to the reader that the New York post office is much more trustworthy than the Kentucky one. The anecdote therefore constructs New York as superior to Kentucky, with postmasters who are responsible men who are hard-working and not sitting around smoking and telling stories.

# **Didn't Know the Place.**

[New York World.]

"Where's the postoffice?" asked a man with Kentucky jeans on and wearing beard from ear to ear around the under part of his jaw that made him look as though he had only put it on for fun.

He was walking up and down the Broadway corridor of the building when he asked the question, and the man he asked told him that he was within the building.

"This is the postoffice?"

"Yes, sir."

"Where's the postmaster? I want to mail this letter."

"Oh, I suppose he's up stairs in his office."

"Well, that's good!" ejaculated the countryman. "Why ain't he here attending to his business?"

"He is, probably."

"That's good, again. I want to get a stamp of him and he'd ought to be here. And you call this the postoffice? Where's the fellers?"

"What fellows?"

"The fellers who sit around on the barrels and tell stories?"

"We don't have them here."

"Where's the noospapers wot you get out of the boxes an' read?"

"None here."

"Where's the fellow wots got the terbacker?"

"He ain't around."

"Don't the minister come in an' borry a postal card till he gits a whack at the plate?"

"Not here,"

"An' don't the postmaster read all the papers an' postal cards before he sends 'em home?" "No."

"Well, this is the blastgamedest postoffice I ever seen. They can't git any 8 cents from me. Guess I'll take the letter home myself," and he walked away toward Cortlandt street scratching under his hat.

With regard to the voices represented in the anecdote, it therefore not surprising that it is the Kentucky man's linguistic repertoire which deviates from that of the 'neutral' narrator and the New York man. The only form shared with the New York man is the use of *ain't*, which means that the form is associated mainly with spoken language here. Yod-dropping is only linked to the Kentucky man, which creates an association with southern (or more specifically Kentucky) speech. However, yod-dropping occurs again only once and it is not as salient as another contrast on the phonological level which is established based on the use of hyper-rhoticity: The Kentucky man uses the hyper-rhotic forms *fellers* and *terbacker* and the difference between him and the New York man is highlighted in the exchange "'Where's the fellers?'/'What fellows?'", which creates a parallel structure directing the readers' attention to the difference between *feller* and *fellow* and linking it to the difference between uncivilized Kentucky and civilized New York, which fits the values identified for hyper-rhoticity above. Further forms used by the Kentucky man are a raised dress vowel in *get*, relative *what* (*the noospapers wot you get out of the boxes an' read*, *the fellow wots got*) and third-person singular *don't* (*Don't the minister come in*). They mark him as southern (the last two forms have been linked to southern speech in articles above), but also as uncivilized and uneducated (the raised dress vowel has marked the speech of Blacks and mountaineers, and relative *what* has also been linked to Black speech in articles discussed above). On a lexical level, the Kentucky man's use of the colloquial figurative expression *get a whack at the plate* and of the swear word *blastgamedest* are noticeable, which reinforces the impression that his speech is not very elaborate and that his behavior is rather impolite. Overall, it is striking that apart from yod-dropping, there is no overlap between the repertoire of the southern speakers in "Out-of-Town People" and the Kentucky man in

### 4.1 Metadiscourses on phonological forms

this article. There are also only two forms shared between the Kentucky man and the two figures of the Kentucky colonel (Colonel Gutrippah and the colonel in the advertisement for The Royal Tailors): third-person singular *don't* and negation with *ain't*. From this follows that there is no linguistic repertoire that consistently indexes southern speech, but that there are several forms available which can mark southernness, but also other social values. That non-rhoticity is not marked in the speech of the southerners exhibiting yod-dropping shows that it is also not a prerequisite for indexing southernness: It can occur, but it does not have to.

This argument can be supported by the following humorous short dialogue between a "Small Kentuckian" and his "Pap", published in the *Morning Oregonian* (Portland, Oregon) on January 12, 1894[140] . The young boy asks about the "swearin' off that the noospapers air talkin' about", which his father disparagingly explains to him as being "just some Yankee custom" which southerners are not familiar with.

Small Kentuckian—Pap, what is this yah swearin' off that the noospapers air talkin' about nowadays? Pap—I don't know. It don't mattah, anyway. It's just some Yankee custom.—*Indianapolis Journal*.

As the article was published in the beginning of the new year, it is likely that newspapers contained articles about bad habits that people plan to swear off, that is about New Year's resolutions. The father's statement that New Year's resolutions do not matter in the south sheds a negative light on the region because it makes southerners seem like they are not interested in swearing off bad habits – habits which they are depicted as having in other articles, for example drinking whisky, chewing tobacco and swearing. This negative image is linked to the linguistic forms, which include yod-dropping *and* non-rhoticity here, and thus illustrate that southerners can exhibit both forms as well as only one of them. In addition, alveolar -*ing*, yod-insertion and dropping of /h/ before /j/ in *here* (*yah*) and third-person singular *don't* in *it don't mattah* are marked as southern here as well, but they also occur in the repertoires of Black speakers and mountaineers in articles above, illustrating again how many forms are shared by these groups and suggesting that several forms are more generally indexing a lack of eduaction than a specific social group. A form that has not been part of any of the other articles analyzed so far is a higher and fronter vowel in start (*air* 'are').

An article illustrating that yod-dropping is also linked to Black speakers is headed "Darkeygraphy". It was published first in the *Charleston Mercury* on April 13, 1858[20] , then in the *Daily Morning News* (Savannah, Georgia) on April 15, 1858[21] , and finally with slight changes in the *Columbus Tri-Weekly Enquirer* on

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

May 11, 1858[22] . Two Black speakers talk about the suicide of a Californian man at a hotel and the "colored gemman" (Lemuel) accuses "his colored crony, a waiter at a hotel" (Sam) of stealing money from the dead person – an accusation presented as justified because of Sam's unusually good clothes. Several linguistic forms are present here which, in combination, mark the speaker's ethnicity, for example voiced TH-stopping and voiceless TH-fronting (in medial position) as well as -stopping (in initial position), voiced labiodental fricative stopping (*neber*, *hab*), final consonant cluster reduction (*lass* 'last'), alveolar -*ing*, regularized past tense forms (*seed*), absence of do-support in wh-questions (*How you suppose I know?*, *What you mean to insenewate?*), malapropisms (*susancide*) and eye dialect (*nite*, *insenewate*). This shows again how an extremely negative stereotyping of Blacks as thieves, combined with the derogatory label "darkey" and the explicit ridicule of their speech as "amusing" is linked to the representation of a large number of differential linguistic forms. Yod-dropping is one of many of these forms and again not very salient, as it occurs only once.

Darkeygraphy.—The following sample of "darkey" talk is characteristic and amusing:

"So you had a bad susancide at your hous lass nite, Sam," said a colored gemman, on meeting his colored crony, a waiter at a hotel.

"Oh, yes, Lemuel, dat we had—it almost scart me into takin' a drink. He was jis from California, wid heeps of noospapers.

[…]

"Wus dere anything found in de pockets Sam?"

"How you suppose I know? Do you tink I'd put my hand in to feel? What you mean to insenewate?" "Oh, nuffin—only I neber seed you hab sich good close on afore, dat's all."

So far, I have shown that yod-dropping occurs in articles as part of the representation of southern voices and also of a Black voice. However, unlike nonrhoticity, yod-dropping is also used frequently to represent voices of speakers from other regions. Examples of such voices are Tom Blake, a New York newsboy who used to be a shoeblack in Brooklyn, a Wisconsin deacon and his wife Sarah Jane, Nort Kingsley, an old and grim hunter in Northern California, and Hank Borrows, a giant wagoner in Montana. I will briefly describe the articles that these figures appear in and discuss how the voices are constructed and how salient yod-dropping is as a characteristic of their speech.

Tom Blake is quoted as a witness of the situation of the shoe-blacking industry in New York City in an article which originally appeared in the *Brooklyn Eagle* and was then republished in the *St. Louis Globe-Democrat* (Missouri) on August 31, 1887[80] . He describes how the business has become difficult because of the

increasing competition from Italians and Black men, who offer big chairs, newspapers and a good polish which preserves the black color. This has caused him to become a newsboy at the Eagle – he earns more money there than by blacking boots. This contrast to Italians and Black people emphasizes his whiteness and his status as an American. His job marks him as belonging rather to the lower or lower-middle class. His speech is marked by one instance of yod-dropping and, in addition, one instance of hyper-rhoticity (*feller*). So even though he is a New Yorker, he is not marked as non-rhotic at all, which provides further evidence to the argument developed above based on the anecdote "She Got a Seat", that non-rhoticity is not primarily a maker of northeastern (or more particularly New York) speech, but rather a feature that marks social characteristics, like the imitation of English manners and speech to appear educated and cultured. The newsboy does not possess these characteristics; on the contrary, he is depicted as using alveolar -*ing*, which has been shown to rather mark the opposite. Lexically, the pronunciation of *Italian* as *Eye-talian* also underlines his lack of education and the shortening of *business* to *bis* marks his speech as colloquial. Grammatical forms include the negation with *ain't* and negative concord (*The bis ain't no good no more*), second-person singular -*s* and third-person plural -*s* marking and plural *is* in the existential construction *there's secrets* (compared to third-person plural *are* in the non-existential construction *They're just as careful*). These forms are all found in representations of southern, Black and mountaineer speech as well, so that, like yod-dropping in *newspaper*, they do not index a particular region here, but rather social characteristics like a lower degree of education and social standing.

# **EVOLUTION OF "SHINES." Rise and Progress of the Shoe-Blacking Industry.** [From the Brooklyn Eagle.]

[…]

Tom Blake, now an *Eagle* newsboy, who was a shoeblack under the old dispensation, delivers himself in the following terms:

"The bis ain't any good no more. Five years ago I could rake in a couple of dollars a day easy down by Fulton Ferry. Now, a box ain't worth 75c a day. The big chairs kills the bis. They're got up in fine style, an' you gets a read at a noospaper while the feller blacks yer boots, an' when he's through he brushes yer hat an' coat, an' al fur 5c. Beside that, there's secrets in the business. It's in the blackin'! I don't care how much you rub at a boot with bad blackin', the shine'll die off to black lead in an hour or two. The Eye-talians and neggers 'as got some wrinkle about blackin' as puts on a tip-top polish which stays there. They're just as careful about it as the Chinese is about their starch. But there's more money in selling the *Eagle* on a good route."

[…]

In the second example, the construction of voices is fairly complex: The article is a letter to the editor, which is supposedly written by a deacon in Wisconsin

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

to the *Milwaukee Daily Journal* on January 20, 1888[85] . There are several clues, however, that the letter is a fictional piece of writing, which has the purpose of expressing views about political topics more indirectly by presenting them as views from people living in Wisconsin. The main clue is the humor created through the contrast between the "subscriber", the figure of the deacon who is constructed as the writer of the letter, and his wife Sarah Jane. In the course of the letter, it becomes clear that it is in fact Sarah Jane who is smart and educated and who has developed a political view on tariffs based on several calculations that she did and that the deacon presents in the letter. At the very end of the letter, the editor adds a comment confirming that her calculations are correct and thus providing authoritative backing for her analysis. The deacon, in contrast, is portrayed as rather uneducated, which is revealed through several postscripts which the deacon added at the end of the letter. They show that his own writing is in fact quite different from that of his wife and that she has improved the quality of his letter to a large extent. He is depicted as a man who is not intelligent enough to realize that his wife is more educated than he is, although he does acknowledge that "shes a mitey good scholar". The humor rests on the irony that the deacon describes her as being ashamed of her writing, even though it is more correct than his own writing. The letter therefore implies that that there is a norm for correctness, which provides for example a clear answer to the question whether *rights* should be spelled <rites> or <rights> (again, Sarah Jane knows the answer better than her husband). This norm of correctness is violated sometimes in the letter, thus characterizing the subscriber as not educated enough to know these norms – the use of eye dialect is also prominent because it appears in the heading "Sarah Jane Figgers" and it serves to characterize the deacon as less educated than his wife because it occurs more often in the postscripts than in the letter itself (*bizness*, *rite*, *sez*, *mitey*, *littery*) and the deacon points out explicitly that his wife does not know that he added the postscripts, therefore implying that she did not have a chance to correct them. The only forms indicating a differential pronunciation are *noospaper* (the only instance representing yoddropping), *kivered*, indicating a fronted strut vowel, and words with final -*ing* spelled <in> to indicate an alveolar pronunciation. Grammatical forms comprise *a*-prefixing (*a studyin*), demonstrative *them* (*them letters*), negative concord (*she couldn't noways*) and subject-doubling (*my wife Sarah Jane she's a great reader*). Lexically, the adverb *noways* stands out (it is not contained in any of the articles analyzed so far) and the subscriber's reanalysis of *commas* as a singular form having the plural *commases* is also notable.

Compared to the letters analyzed so far (the letter addressed to Doctor Hun and the Nasby letter), this letter shares with them the characteristic of being

fictional, but it differs markedly in two respects. First of all, the subscriber and his wife are shown to be aware of the norms of correctness and to attempt to conform to them. Even though they do not achieve it completely, the number of eye dialect forms and forms indicating differences in phonology, grammar or lexicon is considerably smaller in this letter – they are rather the exception than the rule. This general difference can be linked to a contrast between north and south, as the fictional authors of the first two letters have a southern background, while the subscriber and his wife Sarah Jane live in Wisconsin. However, most of the forms found in the northern letter are also found in the southern letters, including yod-dropping. This indicates that these forms themselves index first and foremost a lack of education and not a place, but that the north, as a place, is indexed through the lower number of such deviating forms. Northern speakers are thus constructed as speaking more correctly than southern speakers. This relates in interesting ways to the difference between males and females – in the northern letter, the wife is constructed as writing more correctly than the subscriber and as I have shown in the analysis of *bettah*, it is the anecdote representing a female southern voice ("She Got a Seat") which contains the lowest number of deviating forms. The regional difference between north and south which is linked to differences in correctness and thus to the degree of education of its speakers seems thus more salient for male voices than for female voices.

# **SARAH JANE FIGGERS**

# **"SUBSCRIBER" AND HIS WIFE DIG UP SOME INTERESTING FACTS. The Profit the Wisconsin Farmer Derives from the Tariff Just \$2.27 1-2—The Need of Bounties for Badger Tillers of the Soil—Mr. Granger's Tobacco Bill.**

To the Editor of The Journal: I guess they have found out at Madison that I am writing for the noospapers for they have sent me by express (cost me 25 cents) a great big kivered book, full of figures about everything, all the cows and horses and pigs and sheep and hay and cranberries and old soldiers and everything you can think of in Wisconsin. And my wife Sarah Jane she's a great reader and bully on figgers and she sat up half the night a studyin of it and she said to me when she came to bed (I always go to bed at nine o'clock precisely) she said she believed we could figger out how much the farmers of Wisconsin made on their sheep by that tariff on wool they are makin' such a fuss about.

### […]

We think it would be best to throw this plagy tariff on wool overboard and the importers also, and give us farmers a bounty on wool as these great statesmen James Blame and John Sherman propose to do for sugar. Let them give us 10 cents a pound and wouldn't you see the wool crop in Wis. Go a kiting. We would soon raise 4 times or 10 times as much wool as we do now and old England could keep her wool at home and the cheating importers would have to shut up shop and the honest farmers would get their rights (should that be spelt rites or rights I say the first S. J. says the second.)

A Subscriber.

P.S.—Please send a copy of this to Senator Sawyer if any man can put this bounty bizness through old Philetus is the man.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

P.S.—I want Sarah Jane to rite them letters to you herself, but she sez she couldn't noways hav her riten seen by a Editor. Shes awful bashful. She changes the spelling in mine sometimes, and puts in comases and parenteses. Shes a mitey good scholar is S. J.

P.S.—I always like to put in the P. Ses, they look a kind of littery as Uncle Philetus sez. Sarah J. does not know that I have ritten these.

A Subscriber.

[We have looked up Sarah Jane's figures, and beg to say to A Subscriber that S. J. is quite correct. —Editor Journal.]

The two examples above show that *noospaper* is constructed as part of the repertoire of not only white and Black southern speakers but also of northeastern (New York City) and northern (Wisconsin) ones. The following article creates a link between *noospaper* and a speaker living in yet another place: in Northern California. It is a short story entitled "Off the Trail", written by Mariner J. Kent and published in the *Daily Inter Ocean* (Chicago, Illinois) on August 11, 1887[78] . The story is told from the perspective of the main character, a "reporter". He rides on a trail through the Sierra Nevada, even though he has been warned by "the old and grim Hunter, Nort Kingsley" that it is very dangerous and that another man who has taken the trail has not come back since. As predicted by the hunter, he encounters several difficulties: At one point his horse sinks into a patch of morass and has to struggle to get out, at another point they get into the middle of a landslide, with large rocks falling down around them and huge amounts of sand threatening to bury them, and finally, the reporter loses the trail and arrives at an "uncanny spot" where he finds the skeleton of the man that the hunter has told him about. In the end, however, he is lucky and finds the trail again and arrives at his destination – his horse, however, falls dead.

The extracts of the article quoted below contain the representations of direct speech in the beginning of the story and they show that the speech of the hunter is constructed as markedly deviant from the voice of the reporter. Yod-dropping, indicated through *noospaper*, appears once and it is thus less salient than other forms (as in the articles above). A more salient form is the lowered kit vowel which not only occurs in *ef* 'if' (as in articles representing Black or mountaineer speech), but also in *et* 'it' and *sence* 'since', suggesting that the raising is not lexically restricted. Furthermore, his speech is characterized by hyper-rhoticity (*ter* 'to', *yer* 'you', *er* 'a') and even though one non-rhotic form occurs as well (*fou'teen*), this seems to be an exception because no other forms are marked as non-rhotic. With respect to rhoticity, his repertoire is thus similar to the repertoire of the western cowboy and it reinforces the link between hyper-rhoticity and the west as a place full of hardships that have to be overcome in order to be successful and full of men who might not be civilized but who are tough and

### 4.1 Metadiscourses on phonological forms

capable of surviving there. In contrast to the southern mountaineer, who combines hyper-rhoticity with non-rhoticity and whose lack of education and civilized behavior is evaluated negatively, the hyper-rhotic western hunter is rather praised for his ability to live in "the unsought and untraveled fastnesses of the great stretching wilderness of mountains", which is full of dangerous natural phenomena. That the hunter is located at the nature end of the nature-culture continuum is also indicated by forms indicating aspects of connected speech like weakening (*thet* 'that', *hes* 'has') or elisions (*th'*, *an'*) and by the use of eye dialect (*thot* 'thought', *riting* 'writing'). It is noticeable that the reporter is located in the middle of the continuum as he is neither shown to be non-rhotic nor hyperrhotic: His use of *fellow* instead of *feller* puts particular emphasis on this absence of hyper-rhoticity and underlines the contrast to the western hunter. Finally, a lower nurse vowel is also indicated in the speech of the westerner (*war'nt*), as in that of Black and white southerners and that of mountaineers. Grammatically, a notable parallel to the southern mountaineer is the use of the relative pronoun *as* (*er chap ez thot thet he could tramp th' trail*). Furthermore, the hunter's speech is marked by the third-person singular past tense form *were* (*he war'nt*), a form that has not been found in articles above so far. A form that has also not been represented in any of the articles analyzed above is the use of *mought* as a lexical variant of *might*. So to conclude, although yod-dropping is linked to the western hunter, it is not a very salient form, whereas hyper-rhoticity is more prominent, indexing an uncivilized roughness and toughness, which is evaluated positively in this western context. This positive evaluation is also signaled by the old age of the hunter, which conveys experience and wisdom. A lowered kit vowel is as prominent as hyper-rhoticity and as it is consistently marked it could be an index of western speech or even California more particularly, but more evidence would be needed to support this tentative hypothesis.<sup>21</sup> Overall, the number of forms marking the westerner's speech as deviant is also not nearly as high as that of Black or white southern Americans or mountaineers and the use of eye dialect is also much more restricted.

> **OFF THE TRAIL. A Reporter's Wild Ride in the Sierra Nevada.** by mariner j. kent Author of "Shot Down a Flume," "Winning a Scoop," etc.

<sup>21</sup>This lowering of kit to dress marked by spellings like <sence> is described by Wolfram & Schilling-Estes (2016: 140) as a part of the Northern California Vowel Shift, which according to them is a "more recent vowel shift". The fact that it was already indicated at the end of the nineteenth century in the article "Off the Trail" raises the question whether the shift had already started earlier than these authors assume.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

[Copyrighted, 1887, by the author]

"Et's only fou'teen miles by th' trail an' et's nigh on ter thirty round by the road, which I allow ez safer by er chap ez thot thet he could tramp th' trail, an' he hes not turned up sence."

So urged the old and grim hunter, Nort Kingsley, as we stood in the shadows of Lassen Butte on a balmy August morning in 1877—one of those glorious mornings peculiar to the climate of Northern California. The twin peaks of Lassen towered heavenward in the midst of the mighty crests of the Sierra Nevada, where two mountain spurs seemed to have been hurled together.

"Thank you, Nort," I said, swinging myself into the saddle, but I'll take the trail and risk it. Perhaps I may run across the fellow you spoke of."

"Maybe yer mought," rejoined Kingsley, "but ef yer do et will be en er place unpleasant fur riting yer stuff for th' noospapers. He war'nt er pious chap yer see," added the old hunter, as he grasped my hand in a farewell clasp.

[…]

The fourth example of an article linking yod-dropping to voices from several different regions is a long fictional story written by W. Bert Foster and published in the *Salt Lake Semi-Weekly Tribune* (Salt Lake City, Utah) on November 15, 1898[171] . The character whose speech is marked by yod-dropping is Hank Borrows. As indicated by the title "Hank Borrows's Adventure", he is the protagonist of the story and the first sentence links him to the northwest of the United States because he is described as "a wagoner in the employ of the Government at a Montana post". He and his speech are contrasted with that of two other characters, Captain Langdon and Lieutenant Chester, two officers who are stationed in Montana but, in contrast to Hank, not native to the region. While the officers are young and inexperienced (the Lieutenant "was just out from West Point", a military academy in New York), Hank is old and has accumulated a lot of knowledge and experience over the years. Furthermore, Hank calls the officers "gentlemen", which indicates their higher social status, but at the same time it is clear that they depend on Hank in this context – not only on his experience, but also on his strength, which is conveyed by the narrator describing him as "a giant in stature—a mighty man of bone and muscle, who could 'pack' a mule-load if necessary, and who knew all the trails and mountain passes within 200 miles of the fort". In the dialogue quoted here, Hank tells the officers that he has discovered signs of another hunter and that it will be dangerous for them as long as the hunter is in the same valley. The officers assume that Hank refers to another human and while Hank lets them believe it at first, he later tells them that he is actually referring to a bear. Furthermore, Hank sees signs of an upcoming blizzard. Even though Hank is depicted as calm and not worried, it is established that the three characters are in a dangerous environment with forces of nature (like snowstorms and wild animals) threatening their lives. The story thus revolves around the contrast between nature and culture again, with the officers representing the

### 4.1 Metadiscourses on phonological forms

side of culture (as men affiliated to an elite academy in the northeast) and Hank representing the side of nature (as a physically strong man from the west who is not afraid of fighting the forces of nature). It is not surprising that Hank's speech is marked as deviant from that of the officers: Next to the one instance of yoddropping in *noospapers*, it is noticeable that he exhibits hyper-rhoticity (but not non-rhoticity) like the Northern Californian hunter and also like the Montana cowboy losing a fight to a dude in the anecdote "A Muscular Dude".<sup>22</sup> This provides further support for the strong indexical link between hyper-rhoticity and a roughness and toughness needed in natural environments as well as a lack of civilized behavior which can be evaluated negatively (as in "The Muscular Dude", where the cowboy appears as threatening, wild, and drinking), but also positively as in "Off the Trail" and the story here, where the speakers' experience and their tough and calm nature are depicted as helpful because they provide support for easterners unused to such difficult environments.

In addition to yod-dropping and hyper-rhoticity, Hank's speech also exhibits forms which have also been found in many articles analyzed above: alveolar *-ing* (*huntin'*), a raised dress vowel in *git*, yod-insertion before /h/ followed by a dropping of /h/ in *yere* 'here' and a lowered square vowel in *thar* 'there'. There is one instance of final consonant cluster reduction in *tol'*, but given this low frequency it could also be interpreted as a marker of connected speech, like the weakening of the vowel in *kin* 'can'. A phonological form that has not been represented so far is yod-coalescence, indicated by spelling *Indians* <Injuns>. As on the phonological level, grammatical forms which distinguish Hank's speech from that of the officers and that of the narrator have been commonly found in other articles: the double demonstrative with *this here*, negation with *ain't* and the past tense form *seen*. Lexically, the use of *worritin'* instead of *worrying* stands out because it has also not been part of any of the other articles; the phrase *I reckon* and the alternative pronunciation of *well* as *wa-al* have been found in other articles above. The latter form is explicitly connected to a "drawl" here ("'Wa-al,' drawled Hank again"), a description which probably also applies to *ya-as* 'yes'.

# **HANK BORROWS'S ADVENTURE By W. Bert Foster.**

Hank Borrows was a wagoner in the employ of the Government at a Montana post. In winter when teaming was impossible, he hunted in the foothills some distance from the fort. Occasionally some

<sup>22</sup>There is one instance of a representation of a non-rhotic form: *worse* spelled <wuss>. This form belongs to a set of words in which loss of non-prevocalic /r/ occurred very early (see §3.3.5 for a detailed discussion) and it is likely that these non-rhotic forms (e.g. *hoss*, *cuss* and *fust*, where /r/ is deleted before /s/) are linked to a different set of social meanings than non-rhotic forms where /r/ occurs in a different environment (e.g. in final unstressed -*er*, as in *bettah*).

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

of the younger officers accompanied him, and it was considered something of an honor to go with Hank upon one of his hunting trips. The venture was sure to yield some good returns in the way of game and pelts.

Hank was a giant in stature—a mighty man of bone and muscle, who could "pack" a mule-load if necessary, and who knew all the trails and mountain passes within 200 miles of the fort.

About Christmas time one winter Hank started out with two of the fort officers—a Capt. Langdon and Lieut. Chester¾bound [sic] for a certain valley, some sixty miles from the fort, which the old guide kne wwell [sic].

## [...]

There was enough snow on the ground for good tracking and Langdon and Chester looked forward to some excellent sport. But when Hank returned to camp the first night, after spending the day setting beaver traps along the river, he looked grave.

"I tell ye how it is," he said, after supper; "I kinder wish't I hadn't brought you gentlemen into this yere valley—that I do!"

"What's the matter?" queried Langdon.

"Indian signs?" demanded the Lieutenant, eagerly. He was just out from West Point, and had yet to experience his first Ute and Bannock campaign.

"No, no," returned Hank. "Injuns don't monkey around in the hills this weather. I've seen signs wuss'n that."

"What was it? Think there'll be a blizzard?"

"We kin weather a blizzard," said Hank, calmly, "an' we may git one. The signs is propeetious, as the weather prophets say in the Denver noospapers. But that ain't what's worritin' me. I seen tracks today that tol' me thar was somebody in this yere valley that can't stay here if I'm goin' ter occupy it, too."

"Another hunter, Hank?" asked Capt. Langdon. Isn't there room for two parties?"

"Don't be hoggish, Hank," added Chester. "This is a free country."

"Tain't free enough for him an' me," replied the old wagoner, with a curious smile on his rugged face. "There ain't room for both of us in this yere valley, an' I wish I hadn't brought you gentlemen into it."

"There must be no shooting scrape, Hank," said the Captain sternly. "I thought you were a decent, quiet sort of a man—"

"I am—mostly," said Hank, grinning behind his hand. "But I tell ye there ain't none of us safe while this chap's erlive in this yere valley."

"What d'ye think he's doing here—hunting?" questioned the Captain.

"Ya-as, I reckon he's huntin'," said Hank.

"Who is he? What's his name?" demanded Lieut. Chester.

"Wa-al," drawled Hank again, his eyes twinkling, "I call him 'Ole Ephr-am.'"

[...]

So far, the analysis of articles containing *noospaper/s* has focused on those articles which contain direct representations of speech. There are, however, also some articles in the collection of articles containing *noospaper/s* which discuss yod-dropping explicitly and use *noospaper* as an example to illustrate the phenomenon. This search term thus provides an excellent opportunity to compare

### 4.1 Metadiscourses on phonological forms

the indexical links created indirectly in the representations of speech in letters, anecdotes, short humorous dialogues and fictional stories with those created in explicit metadiscursive discussions of the form.

The first article in the databases which discusses yod-dropping explicitly was published originally in the *Washington Star*, a newspaper not contained in the databases but indicated as the source in the article cited below, which appeared in the *Daily Arkansas Gazette* (Little Rock) on September 24, 1879[37] . The article was reprinted on August 29, 1880[38] , almost a year later, in the same newspaper before it appeared in two northeastern newspapers (in the *New Hampshire Sentinel* and in the *Northern Christian Advocate*) on December 30, 1880[39] , and in the *Omaha Daily Herald* (Nebraska) on December 6, 1881[46] . It needs to be noted that the later reprints have shortened the original article to some extent and that they also refer to different sources (like the *Southern Letter*, the *New York Weekly Review* or *Hall's Journal of Health*). This indicates that the article was very popular and attracted the interest of many newspapers and journals over a time period of at least two years.

The article is a contribution by a reader – his status as a person not working for the paper is indicated by his addition of "if you will allow me the space". His intention in writing the article is that he wants to draw attention to "a fault in English pronunciation", which, as indicated by the heading, concerns "The Pronunciation of 'U'". He describes this pronunciation as "giving the long 'u,' which is in so many [of] our common words, the sound of 'oo'". This shows that the "fault" is conceptualized as a differential pronunciation of the vowel and not as an elision of a consonant. This provides further insights into the question of salience: Since /j/ is not represented in the spelling, it is harder to represent its absence graphically and by choosing the vowel grapheme <oo> to mark yod-dropping, the difference can easily be perceived as a vocalic one. This makes yod-dropping different from /h/-dropping, which is unambiguously marked in the spelling by omitting <h>. Nevertheless, since yod-dropping is the only linguistic form discussed in this article, it is of course very salient here. The examples given by the author show that the elision of /j/ occurs in words where /j/ follows an alveolar consonant (*institute*, *duty*, *student*, *Tuesday*, *avenue* and *dupe*), but not in words where it occurs after other consonants (*beauty*, *pew* and *cupid*). His evaluation of the form as a fault and a vulgarism which is not authorized by any dictionary of the English language creates an indexical link between the form and incorrectness, that is a failure to conform to established norms, and thus also to a lack of education. He also links the use of the form to regional difference, however, by claiming that yod-dropping is "exceeding common in the north, rarely heard in the south, or in England, but which seems to be spreading here". The extent

to which the pronunciation is used in the north is quantified by his claim that "ninety-five out of every hundred northerners will say institoot, instead of institute". Only the most highly educated men in the north (the names he gives as examples are "Wendell Phillips, Charles Sumner, George William Curtis, Emerson, Holmes") do not use the 'wrong' variant, which strengthens the link between education and correct pronunciation and which serves to underline the urgency of the author's appeal that especially teachers and students need to be made aware of the error in order to be able to correct it. That the author finds yod-dropping to be a purely northern form, which is not found in the south at all ("It is a fault that a southerner also never falls into"), is remarkable given that it has been linked to direct representations of southern speakers in articles above. Yod-dropping is also constructed as an exceptional case by the author himself who states that the southern speaker "has slips enough of another kind" – among them the "vulgarism to call a door a doah", which refers to non-rhoticity. Non-rhoticity is thus constructed as a southern form and yod-dropping as a northern form, the difference being that non-rhoticity and its evaluation as incorrect and vulgar are already well known and shared by a large number of people in 1879 ("as we all admit"), while yod-dropping is judged to be much less salient, which justifies the writing of the article ("As many of our teachers have never had their attention called to this, I hope they will excuse this notice"). This confirms an observation made based on the analysis of the articles above, namely that yod-dropping is not a very salient form, especially in comparison to non-rhoticity and hyperrhoticity.

# **The Pronunciation of "U."** [Washington Star.]

As the schools have just opened, and as everybody reads your paper, if you will allow me the space, I wish to call the attention of the teachers and pupils to a fault in English pronunciation, **exceeding common in the north, rarely heard in the south, or in England, but which seems to be spreading here**. (We have faults enough in the south without grafting some northern ones upon them.) I refer to the **vulgarism**—if I may so term it—of giving the long "u," which is in so many ot [sic] our common words, the sound of "oo."

For instance, **ninety-five out of every hundred northerners** will say institoot, instead of institute, dooty instead of duty—a perfect rhyme to the word beauty. They will call new and news noo and noos—a perfect rhyme to pew and pews—and so on through the dozens and hundreds of similar words. **Not a dictionary in the English language authorizes this**. In student and stupid the "u" has the same sound as in cupid, and they should not be pronounced stoodent and stooped, as so many teachers are in the habit of sounding them.

If it is a **vulgarism** to call a door a **doah**—as we all admit—isn't it as much of a vulgarism to call a newspaper a **noospaper**? **One is northern and the other southern**—that's the only difference. When the London Punch wishes to burlesque the pronunciation of servants it makes them call the duke the dook, the tutor the tooter, and a tube a toob. **You never find the best northern speakers**, such as Wendell Phillips, Charles Sumner, George William Curtis, Emerson, Holmes, and **men of that class**, saying noo for new, or Toosday for Tuesday, avenoo for avenue, or calling

### 4.1 Metadiscourses on phonological forms

a dupe a doop. It is a fault that a southerner also never falls into. He has slips enough of another kind, but he doesn't slip on the long "u." As many of our teachers have never had their attention called to this, I hope they will excuse this notice.

# [emphasis mine]

The second article linking *noospaper* to an explicit discussion of yod-dropping appeared eight years later than "The Pronunciation of 'U'", on August 13, 1887[79] , in the *Grand Forks Daily Herald* (Grand Forks, North Dakota). Its heading, "Teaching in the West", indicates again a relation to teaching and education. It also draws attention to a particular region: the "West". In the article, the place is specified further as "a school room in lower Arizona", and the main aim of the article is to report the prescriptive linguistic rules given and enforced by a teacher there to the readers of the newspaper in North Dakota. One of these rules is cited as "They must soften the u in such words as duty and opportunity, and not pronounce them dooty and opportoonity". As in the article "The Pronunciation of 'U'", the difference in pronunciation is conceptualized as a vocalic difference, with the 'correct' vowel being 'softer' than the 'incorrect' one. However, in contrast to that article, yod-dropping is not as salient here because it is only one of several rules and not highlighted in any way. In general, the article shows that yod-dropping was found in western regions because it was apparently deemed necessary to teach the alternative variant in school. The fact that it was written by a western correspondent for a northern newspaper indicates that it is judged to be relevant in the north as well. Moreover, the correspondent explicitly points out in the article that the rules "might find a conspicuous place in many families in the east, where the educated ear is so frequently offended by the mispronunciation of the most common words", which shows overall that the forms listed in the article are relevant for all American regions. The south-north difference is not emphasized here – it is even possible that the correspondent conceptualizes "the east" as comprising the northeast as well as the southeast. The forms are thus explicitly evaluated as incorrect, and in addition, they are implicitly constructed as non-regional. Taking into consideration other forms judged to be incorrect, like for example alveolar -*ing* ("drop final g's"), negation with *ain't*, or elisions which occur in connected speech ("gray deal for great deal"), it becomes clear that they are forms which I have identified in direct representations of several different speakers as well and that they thus serve to signal primarily a lower degree of education or cultural refinement and not the belonging to a specific region.

# **TEACHING IN THE WEST.**

A western correspondent of one of our daily papers gives some **rules** that are posted conspicuously in a **school room in lower Arizona**. The teacher is an enthusiast in the use of good English, and

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

insists on its use in the school room. The rules that this teacher has made imperative in his school room **might find a conspicuous place in many families in the east**, where the educated ear is so frequently offended by the mispronunciation of the most common words. His rules are:

"My scholars must not pronounce dreadful, dretful; or catch, ketch; or **newspaper, noospaper**; or society sassiety; or February, Febuary; or Massachusetts, Masschusetts; or eleven, leven; or height, hithe; or drought, drowth; they must not say fur for for, or git red of for get rid of. They must not say anywheres or nowheres, or anyways, or a long ways, or those sort of things, or those kind, for that sort and that kind. They must not say he don't for he doesn't and they must never use the word ain't. **They must soften the u in such words as duty and opportunity, and not pronounce them dooty and opportoonity**. They must not drop final g's, or leave out of words their h's. They must not half pronounce, must not say gray deal for great deal. Every word demands the full, authorized, verbal mention of all its letters."

The correspondent says this teacher's method of teaching is very original, and his success, considering the environment of his pupils, marvellous. […]

> —Christian Union. [emphasis mine]

In contrast to the article "Teaching in the West", the difference between the north and the south found in "The Pronunciation of 'U'" is also emphasized in the explicit discussion of yod-dropping in the article "The New Woman", published in the *Daily Picayune* (New Orleans, Louisiana) on September 15, 1895[154] , and which contains a sub-section about "The Speech of Southern Women", which is quoted below. The main part of the sub-section consists of a quotation of a passage of Theodore Mead's book "Our Mother Tongue"23, in which he generally criticizes "the speech of American women" but at the same time praises the speech of "the women of the southern states" as "pleasant and correct". As a reason he gives "the pure and more sonorous use of the vowels by southern women than by their northern sisters, and [...] the less harsh employment of the consonants" and the two linguistic forms illustrating this point are the *absence* of yoddropping, described as "giving its full significance, for instance, to 'u,' instead of giving that vowel the sound 'oo'", and non-rhoticity because even though he acknowledges that "the ignoring of the 'r's" could be seen as an offense, he also finds that the realization of the /r/ is often too harsh and rolling, which makes it sound menacing and unfriendly. Mead thus creates an association between the linguistic forms yod-dropping and rhoticity and the geographical region of the north as well as the social values of masculinity, harshness, unpleasantness and unfriendliness. In contrast, the correctness of yod-retention is highlighted and the incorrectness of non-rhoticity is downplayed by marking it as pleasant and connecting it to "sweet" female voices.

Mead furthermore attempts to explain the difference between the north and the south by suggesting that the southern forms are a result of "the influence

<sup>23</sup>The author's last name is spelled incorrectly in the article; it is Mead and not Meade.

that the negroes have had in the speech of the southern people", thus linking non-rhoticity and yod-retention to Black American speech as well. He suggests that "the negro, by nature, by instinct, discards what is harsh, discordant and unmusical to the ear", thus creating unusually positive associations between linguistic forms and Black speech. However, this theory is strongly contested by the author of the article, who, after quoting Mead, evaluates it as "very peculiar" and suggests that it is in fact the other way round: "whatever the softness or refinement has come into the negro character and speech has been acquired from contact with the whites of our section". He emphasizes the connection to refinement by postulating a difference between the "ordinary corn field negro" and "the traditional black mammy or ladies' maid", the former's voice being harsh and unpleasant and the latter's voice being soft and pleasant as a result of the contact with whites. This link to culture and refinement connects this article to the articles containing *deah* AND *fellah*, in which non-rhoticity is also shown to be linked to fashionable and refined behavior in the (urban) northeast. However, the view that non-rhoticity is a sign of a particularly refined character is criticized and ridiculed in those articles, while it is at the basis of the argument developed by the author of the present article. Considering the results of the analysis of articles containing *bettah*, it is thus remarkable that non-rhoticity is evaluated so positively in this article; however, it is not surprising that the positive evaluation is connected to female speech because it is also the southern girl in "She Got a Seat" who, even though depicted as non-rhotic, is not characterized as uneducated or unrefined.

# **The Speech of Southern Women.**

Mr. Theodore Meade has published a little book on "Our Mother Tongue," in which he takes occasion to criticise the speech of American women, saying, among other things, that it is really a notable thing, a something that is instantly remarked, when an American woman in speaking has a pleasant voice and uses it with good modulation.

"But," he continues, "there are women in America who, as a rule, even in conversation, have quite sweet voices and a method of speaking which could be made at once pleasant and correct. I allude to the women of the southern states. This is due, no doubt, to **the pure and more sonorous use of the vowels by southern women than by their northern sisters**, and to **the less harsh employment of the consonants**. The useful but dangerous 'r,' which plays such a sad havoc in the speech of the majority of Americans, is all but ignored by southern women, and many of them are as innocent of 'r's' as cockneys are of 'h's' Probably, however, the musical effect comes to a greater extent from a proper use of the vowels—from giving **its full significance, for instance, to 'u,' instead of giving that vowel the sound 'oo.'** A southern woman would never speak of reading a 'Noo York noospaper;' she would not eat 'stoo,' nor would she go out in the 'doo,' She would, however, open the 'doe' and walk over the 'floe.' She knows her 'u's' very well, but has little acquaintance with any 'r's' **But even the ignoring of the 'r's' entirely has not in it nearly of the offense of giving them more than their due significance**. A 'dore' and a 'flore,' where the 'r' in each word is long and somewhat rolling, seem something else than what we should be accustomed to, while the pleasant greeting from a friend, 'Good-morning,' with a roll in the 'r,' has in it something of the sound of menace rather than friendliness.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

"I have often wondered why **southern women** should use these vowels and consonants **more pleasantly** than other Americans, and I have reached a conclusion which many southern women my sisters and other kinsfolk among the rest—will no doubt fail to acquiesce in. They certainly did not achieve these pleasant results through training, because, if anything, they are not as carefully trained at school as girls in the northern states. Nor could it have been entirely by inheritance, for there is not a great difference between the ancestry of the north and south, though in the latter section the people may be a trifle more homogeneously English. The difference, I am persuaded, is due in a very great measure, if not entirely, to the **influence that the negroes have had in the speech of the southern people**. Children, who start out in life with voices as sweet as the chirping of birds, and tones as pure as the notes of a flute, learn their first words from the nurses in whose charge they are, and southern children are universally reared by negro nurses. […] Now **the negro, by nature, by instinct, discards what is harsh, discordant and unmusical to the ear**, and for that reason adopts that which is pleasant and musical to the ear. […]"

This is certainly a very peculiar theory to advance. The southern woman will hardly appreciate the conclusion that her naturally sweet and harmonious tone of voice has been acquired from negro association. As one who knows whereof she speaks, it is a noteworthy fact that **whatever the softness or refinement has come into the negro character and speech has been acquired from contact with the whites of our section. Take the ordinary corn field negro** and contrast the voice with that of the **traditional black mammy or ladies' maid**, and the **roughness of the one's voice and the softness of the other** is at once apparent. No, Mr. Meade, your theory as to the softness of the voices of our southern women is not tenable, as you would find out by short residence.

### [emphasis mine]

Right at the end of the century, on October 20, 1899[177] , an article was published in the *Macon Telegraph* (Georgia) which also argued against the nationwide distribution of yod-dropping. This is already indicated by the heading "By No Means Universal" and the text itself is not only part of the discourse on yod-dropping in America but it also explicitly describes its interaction with the discourse in the English press. The author reacts to a statement made by "Mr. Archer" in the English *Pall Mall Magazine* that yod-dropping is an error which "is distinctly and universally American". This is denied by the author of "By No Means Universal", who first argues against the distinctiveness of the error by citing an English newspaper article, published by the *Manchester Guardian*, which points out that yod-dropping can be found in England as well. In this article, Dean Alford's *Plea for the Queen's English* is quoted as stating that yod-dropping is "a very offensive vulgarism", which is "most common in the Midland countries, but found more or less almost anywhere" and which "arises from defective education, or from gross carelessness". After having established that yod-dropping is not distinctly American, but also an English form, the author of the article in the *Macon Telegraph* continues to argue against the supposed universality of the form by pointing out that it is not universal in any "section" of America except in the northwest. The south is highlighted as a region where the 'error' is "rarely if ever made even by the uneducated". This implies that, according to the author's judgement, yod-dropping occurs to some extent in other regions, the northeast

and the southwest, but that it is a variable form. With regard to the south, the article therefore agrees with the view presented in "The Pronunciation of 'U'" and "The New Woman", but with regard to the north, it differs from these articles by making an explicit distinction between the northeast and the northwest. However, the author does not deny that yod-dropping is also found in the northeast; he just claims that it is not universal (not as universal as the 95% suggested by the author of "The Pronunciation of 'U'"). The evaluation of the form is the same as in the other articles: It is regarded as incorrect and as signaling a lack of education. Considering that the article takes a perspective which compares England and America, yod-dropping therefore becomes a matter of (linguistic) superiority and inferiority in this article and it is thus clear why the author emphasizes that yod-dropping is not a distinctive and universal form in America: It would make the English in America appear less correct than the English in England. Choosing yod-dropping also allows the author to create a positive evaluation of the south with regard to linguistic correctness, a view which he or she could expect to be liked by the readers of the southern newspaper.

# **By No Means Universal.**

The **English press**, following the lead of Mr. Archer who recently discussed "The American Language" in the Pall Mall Magazine, finds much to say about our pronunciation, particularly the substitution of the "oo" for the "u" sound, as in "dooty" and "Dooey," claiming **that this error is distinctly and universally American**. One writer says that an American novelist, "one of the most vigilant of writers," mimics the speech of an Englishwoman by printing her pronunciation of the word stupid "styoupid" ("stewpid would have been better, being simpler and giving the correct sound of the word when rightly pronounced.) But the Manchester Guardian declares that **the error is found in England as well as in America**. It says:

As a matter of fact, this pronunciation is quite English, as Lowell showed in the case of so many habits of speech that we call distinctly American.

"There is a very offensive vulgarism," wrote Dean Alford in his 'Plea for the Queen's English,' "most common in the Midland countries, but found more or less almost everywhere, giving what should be the sound of the 'u' in certain words as if it were 'oo;' calling 'duty' 'dooty,' Tuesday 'Toosday;' reading to us that the clouds drop down in the 'doo; exhorting us' "Dooly" to do the "dooties" that are "doo' from us,' asking to be allowed to see the 'noospapers" and this is not from incapacity to utter the sound, for though many of these people call new 'noo,' no one ever yet called few 'foo,' but it arises from defective education, or from gross carelessness."

As The Telegraph pointed out when reviewing Mr. Archer's article, **this error of pronunciation is by no means universal in America**. In no section is it universal, **unless be it in the Northwest**. As for the **South**, the improper substitution of the "oo" for the "u" sound is **rarely if ever made even by the uneducated**, granted that the speaker be a native.

[emphasis mine]

However, there is also an article which suggests that the evaluation of yoddropping is slowly changing in America at the end of the nineteenth century. The article does not contain *noospaper/s* because I found it in a different exploratory

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

search (using the search terms *dooty* AND *pronunciation*), but it offers an important insight into the changing attitude towards the form in question, which is why I add it to the analysis here. It is entitled "Lesson in Pronunciation: Words That Are Frequently Mispronounced Even by Educated People" and it was published in the *Idaho Avalanche* (Silver City, Idaho) on October 18, 1890[104] . It quotes an article by Alfred Ayres, which had been published in the *New York Times*, and which consists of a list of frequently mispronounced words. Yod-dropping is described and evaluated in connection with the word *adduce* and it is first of all noticeable that in contrast to the articles above, yod-dropping is conceptualized as a consonantal difference here (and not as a vocalic one) because the author speaks of the introduction or omission of the "sound of y". This allows him to construct two extremes: the complete omission on the one hand and the "clear and perfect" realization of /j/ on the other hand. Interestingly, Ayres suggests avoiding both extremes – the omission because it is incorrect and careless and the clear realization because it is "over-nice and consequently smacks of pedantry". Although this presents the readers with the difficulty of finding out how exactly they should pronounce words like *adduce*, it illustrates a change away from a negative evaluation of yod-dropping. Ayres even points out that "it is not easy to introduce the sound of y" and thus provides the readers with an excuse for a yod-less pronunciation. That yod-retention comes to be associated with pedantry can be interpreted as a sign that yod-retention has become a form used by an educated elite, a small group of people whose insistence on 'correct' pronunciation is not generally appreciated, but rather viewed as an ostentatious display of their education and social status.

# LESSON IN PRONUNCIATION. **Words That Are Frequently Mispronounced Even by Educated People.**

There are in our English at the least three or four thousand words that are frequently mispronounced. Some of these, writes Alfred Ayres in the New York Times, are the following:

[…]

Adduce. When, in the same syllable, long *u* is preceded by one of the consonants *d*, *t*, *l*, *n*, *s*, or *th*, it is not easy to introduce the sound of y; hence careless speakers omit it, pronouncing duty dooty; tune, toon; lute, loot; nuisance, noosance, etc. And yet to make the *u* in these words as clear and perfect as in mute, cute, etc., is over-nice and consequently smacks of pedantry. The two extremes should be avoided with equal care.

To conclude the qualitative analysis of selected articles containing *noospaper/s*, I have shown that the main indexical link created is that between yod-dropping and a lack of education. I have analyzed articles containing direct representations of speech, in which the form was linked to voices of different characters, as well as articles containing explicit discussions of the social and regional distribution

of the forms. It is striking that while there is some overlap, there are also differences between these types of articles, underlining the importance of taking different text types into consideration. In the fictional letters, anecdotes, reports, stories and short humorous paragraphs, speakers exhibiting yod-dropping are usually portrayed as uneducated, but there are differences with regard to other social values which, interestingly, correlate with region. Southern characters are usually portrayed in a negative way, with the important difference that in earlier articles, yod-dropping is used to differentiate negative from positive southern figures (the Republican whose political views and behaviors are satirized in the letter to "Doctor Hun" in 1816 and the neighbor in the dramatic text published in 1826 in the *Louisiana Advertiser* who does not pay for his newspaper subscription), whereas in the later articles, the negative portrayal of southerners is rather contrasted with a positive image of northerners (the man in Kentucky jeans who does not recognize the post office in New York in 1883, the father of the "Small Kentuckian" ridiculing the "Yankee" custom of New Year's resolutions in 1894 and the "Out-of-Town People" impersonating southern rural ignorance which is not found to the same extent in the north in 1894). Black speakers are also represented as dropping the yod and they are associated with very negative social values as well, but only in very few articles like for example in "Darkeygraphy" (1858). In contrast to the negative portrayals of southerners and Black people, figures from the north or the west are characterized more positively, despite their lack of education or lower social position: the New York newsboy, a former bootblack, is depicted as someone who knows about business and who is socially upwardly mobile (1887), the illiteracy of the deacon in rural Wisconsin is mainly represented to emphasize the level of education and intelligence of his wife (1888) and the two tough hunters in mountain regions in the southwest and in the northwest are depicted as strong and tough men who help easterners to survive in an environment dominated by natural forces and not cultured civilization (1887 and 1898).

The explicit discussions of yod-dropping confirm the link between the phonological form and a lack of education by calling it a vulgarism and an error and by classifying it thus as incorrect and deviating from a norm. However, in contrast to the representations of direct speech in fictional texts, three articles point out explicitly that yod-dropping is *not* found in the south: "The Pronunciation of 'U'" (1879), "The New Woman" (1899) and "By No Means Universal" (1899). Furthermore, in "The New Woman" yod-retention by southern women is attributed to the influence of Black speech, a view which is in contrast to the representations of yod-dropping as part of Black speakers' voices. This shows that explicit and implicit metadiscursive strategies do not necessarily create the same indexical

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

links and can also contradict each other. It is noticeable that despite the emphasis on the south as a yod-retaining region in the explicit discussions of the form, there is at the same time no common agreement on where yod-dropping is actually found in America. In the three articles which consider it a non-southern form, emphasis is placed on "the north" as a region where it does occur; in "By No Means Universal", it is more specifically the northwest which is highlighted as a region (the only region) where it is universal. "Teaching in the West" (1887), however, places the form in the east and implicitly also in the west (because it needs to be taught there), which creates the impression that it is fairly universal in America. That the universality of yod-dropping is a perception that plays a role in linguistic metadiscourses is indicated by the article "By No Means Universal" because the very need to counter the claim of universality means that this claim is present in the first place.

Taking both types of strategies to create indexical links into account (implicit and explicit ones) therefore reinforces the impression that yod-dropping is not a salient marker of region, but rather of social characteristics like a lack of education or civilization. However, as the article "Lesson in Pronunciation" shows, even this link seems to become weaker because the presence of /j/ is associated with pedantry here, thus marking a speaker negatively as paying too much attention to correctness. In general, the analysis showed that yod-dropping is not particularly highlighted and very often occurs just once, whereas for example non-rhoticity or hyper-rhoticity occur much more frequently when they are represented in articles. This is of course not surprising as /j/ following an alveolar consonant in an onset cluster is of course much less likely to occur in speech than post-vocalic /r/ or a final vowel to which an /r/ can be added. Nevertheless, the authors of the anecdotes could have developed strategies to draw attention to the difference between forms with yod and without yod, but in most cases, they did not employ such strategies. The articles which focus on discussing only yod-dropping (or yod-dropping and non-rhoticity) in more detail evidently draw much attention to the form; they are, however, very small in number (at least in this collection of texts based on the search term *noospaper/s*). Overall, yoddropping is thus not an index of a specific region or specific social group, but rather of more general social characteristics, and it is also not a very salient one.

# **4.1.3 Indexical links to social personae: a quantitative overview**

To complete the picture resulting from the detailed qualitative analyses carried out in §4.1.2, I will now add a quantitative analysis that aims at providing an overview of the more general patterns that can be found when all articles are

### 4.1 Metadiscourses on phonological forms

taken into account and not just selected ones. This analysis will focus on the central social personae and social groups identified in §4.1.2 by asking first in how many articles these groups become linked to particular search terms and thus to the respective phonological forms and secondly, how the frequency of articles linking the social personae to the forms changed over time. Before presenting the result of the analysis, a methodological issue needs to be addressed: It concerns the identification and classification of the social personae in the articles. One problem is that the social persona that the form is linked to is not always explicitly mentioned. Sometimes, just a name is given; sometimes no indication is given at all. In these cases, a decision needed to be made to either classify the social persona as 'unclear' or to use available clues to identify the social persona. Dudes and Black Americans could often be identified based on their names – typical names of the dude were for example Cholly, Chappie, Algie as well as Mr. Delawney or Young Mr. Sissy, while typical Black American names were for example Uncle Rastus and Aunt Sheba (the address terms Uncle and Aunt being an indicator of Blackness), but also Sambo and Ebenezer. I largely avoided identifying social personae solely based on their linguistic repertoire because this would have led to a circularity in my argument. Since my question is who the forms are linked to, I could not assign the voice represented or described to a social group or persona based on the forms. This was especially the case with Black Americans because there is some overlap in their linguistic repertoire with southern white speakers and mountaineers, so if in doubt, I classified the social persona as 'unclear'. However, in most of these unclear cases, the linguistic repertoire points to a Black person. In the case of the dude, there were a few cases where the articles consisted only of a dialogue without any indication of the names of the speaker. If the content *and* the linguistic repertoire pointed to the dude figure, I classified the social personae as dudes. This classification is less likely to be wrong because the linguistic repertoire of the dude is clearly distinct from other repertoires; especially the realization of /r/ with a labiodental approximant is almost always associated with the dude. In most cases, however, the persona linked to the form could be identified clearly, either through explicit labels like "dude", "darkey" or "mountaineer" or through descriptions like the following which is taken from a fictional story written by Laura B. Marsh and published in the Daily Inter Ocean on April 2, 1893[125] : "One of them was a white boy—that was Jimmy Price. The other, the hired boy, was perhaps a few years older, and black as ever African could be. For this reason he was called 'Snow'".

The second problem was that sometimes several social personae are indexed in one article. This was especially the case in articles which discuss how a particular form is used by fashionable English people and imitated by Americans for this reason. In most cases, I have decided to only consider one social persona – the more important one, which in these cases is the American figure because it is the main motivation for authors to write the article in the first place. Both links were equally prominent in only two articles containing *dawnce*. I therefore counted the respective article twice, once as linking the form to English people and once as linking the form to white Americans. This has the disadvantage of making the overall number of articles containing *dawnce* seem larger, but as there are only two of these cases, this is negligible.

The main social personae or social groups that I have used for the quantitative analyses here are the following ones: First, the white American, which subsumes white Americans of all genders, ages and regions. Within this rather large group, I distinguished one sub-group of white Americans, the American dude. I counted all articles as linking the form to the dude which appeared in the 1880s and 1890s (because this is when the figure appeared in the newspaper discourse) and in which the form is linked to a male figure with the typical characteristics of the dude. This means that I included articles describing male American swells and "Anglomaniacs", even if they were not labeled as dudes, but I did not include articles featuring female personae like the society girls or salesladies. Further subdivisions did not make sense for the analysis of all search terms because for some search terms, the number of articles linking the forms to white Americans (not dudes) was so small that no sub-groups could be determined. However, the analysis of articles containing the search term *bettah* revealed the sub-groups of white southerners and mountaineers, which is why I will take a closer look at the social personae linked to *bettah* in an additional analysis. The second social persona I used for the quantitative analysis is the English speaker. In the overwhelming majority of cases, these English speakers were male, but I did not restrict the group to males only, so I just labeled it "English". Thirdly, I distinguished the group of Black people. This group comprised all people whose skin color was described as black or "colored" – in most cases those people were Black Americans, but there were also a few articles in which the form was linked to Black people in other, non-American, contexts, which is why I chose ethnicity as the main criterion for classification. If the article contained links to a social persona that could clearly be identified (it was thus not "unclear"), but that could not be assigned to any of the groups described above, I classified it as "other". If no characterization of the persona linked to the form was provided at all, I classified it as white American because I assume that this is how the reader of an American newspaper would categorize the persona as well. This means that the characterization of the personae in the "other" group always involved some form of othering which highlighted characteristics that mark the person as un-

### 4.1 Metadiscourses on phonological forms

American – often by emphasizing their status as immigrants and their (former) nationality (e.g. Dutch, German, Irish, Indian).

Figure 4.20 and Figure 4.21 show the number of articles linking the social personae to the search terms (and thus to the phonological forms they represent) and the percentage of articles, respectively. A striking observation that can be made based on these figures is that for almost all search terms one social persona predominates to a great extent. Almost all articles containing *hinglish* link /h/-dropping and -insertion to English speakers – only 7 out of 107 articles link it to white Americans (1) or other social personae (4); in two articles the social persona is unclear. *Noospaper/s* is primarily associated with white Americans – only 6 articles contain links to English speakers and another 6 articles contain links to Black speakers. The number of "other" social personae is highest for this search term (14 articles) – a reason for this is the high number of Dutch and German people portrayed as speakers dropping the yod. Twousers is almost exclusively associated with the dude's linguistic repertoire (92 out of 103 articles), which is interesting because even though the labiodental realization of /r/ is constructed as an English form imitated by American dudes, it is only in one article that an English speaker is actually portrayed as using the form (the article with the young Oxford swell discussed in §4.1.2). This indicates that the indexical link to Englishness has become weak and that it is only indirectly present in the indexical link to the dude, the link that has come to predominate. In contrast to twousers, the search term *deah* AND *fellah* is still linked to English speakers in a comparably large proportion of the articles (25 out of 157 articles), which shows that the indexical link to the characteristic of being English is still stronger for non-rhoticity than for the realization of /r/. Nevertheless, the dude figure is linked to non-rhoticity in the overwhelming majority of articles containing *deah* AND *fellah* (116 articles). The search term *bettah* shows that non-rhoticity is also associated with white Americans who are not dudes and, most notably,Black speakers, who are linked to *bettah* in two thirds of the articles (246 out of 374). This shows again how the lexical level interacts with the phonological level in *deah* AND *fellah* and it underlines the importance of using several search terms to confirm whether the results can be attributed mainly to the phonological form or whether the lexical item(s) play a role as well. With regard to the dominance of one social persona, the search term *dawnce* is an exception because even though it is linked to white Americans in almost half of the articles (34 out of 75 articles), the proportion of articles associating the form with other social personae is also fairly large (dudes in 16 articles and English speakers in 12 articles).

Comparing the results for *dawnce* and *bettah*, it is noticeable that even though the number of articles linking *bettah* to white Americans and dudes is relatively

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Figure 4.20: Social personae linked to phonological forms (number of articles per search term)

Figure 4.21: Social personae linked to phonological forms (percentage of articles per search term)

### 4.1 Metadiscourses on phonological forms

low in comparison to those linking *bettah* to Black speakers (81 vs. 246 speakers), it is higher than the number of articles linking *dawnce* to white speakers and dudes (50). In terms of absolute numbers, the group of white Americans linked to *bettah* is thus prominent enough to deserve closer analysis. Such an analysis makes it possible to assess the relevance of the groups of white southerners and mountaineers in quantitative terms. Figure 4.22 shows the number of articles linking *bettah* to these two groups as well as to dudes and reveals that all three groups are almost equally large, which shows that none of the groups are marginal even though they are not as salient as Black Americans (if salience is based on frequency).

Figure 4.22: Number of articles linking white Americans to *bettah*

In addition to the absolute numbers of articles, it is also important to look at the development of the links between phonological forms and social personae over time. Figure 4.23 focuses on the social personae and shows that from the 1830s on, there is a fairly stable number of articles linking phonological forms to English speakers (mostly between 2.7 and 4.3 articles per million). Links to white Americans also occur throughout the century, with a particularly large proportion in the early second half of the century, due to the high number of articles containing yod-dropping (particularly in the 1860s). Links to Black speakers occur later (in the 1850s), but they increase towards the end of the century and

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

constitute a large proportion of indexical links in the 1880s and 1890s. The social persona that appeared latest is unsurprisingly the dude and it can be seen that this figure quickly assumes a prominent position among all social personae. The largest number of articles containing the search terms appeared in the last two decades and the diagram shows that Black speakers and the dude are the most salient social personae who are linked to the phonological forms investigated in the present study in this important time period.

Figure 4.23: Social personae linked to phonological forms in the course of the nineteenth century

Concerning indexical links to southerners and mountaineers, a separate analysis shows the temporal development of the appearance of social personae in articles containing the search term *bettah*. Figure 4.24 shows that the links to southerners increase in the last three decades of the nineteenth century, especially towards the end. Mountaineers appear in the 1880s, but only in a low proportion of articles, which then increases greatly in the 1890s, while the articles containing links to the dude decrease slightly. This shows that while in the 1880s the main white figures linked to non-rhoticity (represented by *bettah*) are dudes and, to a much lesser extent, southerners, the proportion of links to southerners and mountaineers becomes much greater in the 1890s and surpasses the number of articles linking it to the dude.

Finally, some important observations can be made by looking at each social persona separately and analyzing how often they were linked to the different

Figure 4.24: Number of articles (per million) linking white Americans (including dudes) to *bettah* (temporal development)

search terms in each decade. For English speakers, it is not surprising that they are linked to *hinglish* most frequently in all decades, but the 1860s and, most strikingly, the 1880s show that other search terms are also linked to English speakers (see Figure 4.25). Especially with regard to non-rhoticity, the figure shows that *deah* AND *fellah* is much more often linked to English speakers than *bettah*. As *deah* AND *fellah* is most prominently linked to the dude, who is portrayed as imitating the English, it is not surprising that the link between this particular search term and English people is stronger than that of *bettah*, which is not as strongly connected to the dude, but also to several other social personae.

With regard to white American speakers, Figure 4.26 shows that until the last two decades, they are indexically linked almost exclusively to *noospaper/s*. In the 1880s, they are as often linked to *noospaper/s* as to *dawnce*, and, to a lesser but still significant extent, to *bettah*. Towards the end of the century, the links to *bettah* increase and *bettah* becomes the form most frequently linked to white Americans because the links to *noospaper/s* decrease slightly and the links to *dawnce* decrease considerably. Figure 4.27 shows the development of the search terms linked to the dude and it can be seen that there are no great changes, except the increase of articles containing twousers, which is responsible for the overall increase of articles creating indexical links to the dude figure in the 1890s. Concerning Black speakers, it is not surprising that they are almost exclusively

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

linked to *bettah*. Figure 4.28 shows additionally that while Black speakers occur already in articles in the middle of the nineteenth century, it is particularly in the 1870s and 1880s that their number increases to a very large extent.

Figure 4.25: Phonological forms linked to English speakers (temporal development)

Figure 4.26: Phonological forms linked to white Americans (temporal development)

Figure 4.27: Phonological forms linked to dudes (temporal development)

Figure 4.28: Phonological forms linked to Black speakers (temporal development)

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

To conclude, these quantitative overviews show very well that while English speakers and white American speakers are indexically linked to phonological forms for most of the nineteenth century, the other social personae appear in larger numbers of articles in the last three decades of the century. In fact, Black speakers and dudes are the social personae indexed most often in articles in the 1880s and 1890s and while Black speakers are closely tied to non-rhoticity represented in *bettah*, dudes are portrayed as non-rhotic speakers using *deah* AND *fellah* as well as users of twousers and to a lesser extent of the back vowel in bath (represented by *dawnce*). With regard to non-rhoticity in *bettah*, I have also shown that white southerners and mountaineers become increasingly important in the last two decades as well, but in absolute numbers, they do not appear as frequently as Black speakers. Before discussing how the results of the qualitative and quantitative analyses of the indexical links between social personae and values and phonological forms can be interpreted within the framework of enregisterment, I will present the results of the analysis of indexical values linked to lexical forms in the next section.

# **4.2 Metadiscourses on lexical forms**

In this section, I will present the analysis of articles containing two variants of a variable: firstly *luggage* and *baggage* and secondly *pants* and *trousers*. The first pair of forms will be analyzed quantitatively in §4.2.1 based on a collection of articles containing both forms appearing within 10 words of each other. The collection is restricted to the database NCNP because it yielded enough articles for a quantitative analysis and since considerable differences to the database AHN are not expected, there are no disadvantages of restricting the analysis to one database. Both pairs of forms will then be analyzed qualitatively in §4.2.2 and §4.2.3 to show which indexical values and social personae become associated with each lexical variant.

# **4.2.1** *luggage* **vs.** *baggage***: frequency and temporal and regional distribution of articles**

Using the search term *luggage* n10 *baggage*, I collected all articles containing the two lexical forms within a range of 10 words in the database NCNP. This included lexical items in which the forms function as part of a compound, e.g. *baggage car* or *luggage van*. The search yielded 185 articles, and the first analysis revealed that they could be divided into two groups: Either they contained indexical links

### 4.2 Metadiscourses on lexical forms

between *luggage* and *baggage* and British English or American English respectively or no such links could be found. I therefore used two categories for the following quantitative analyses, "AE/BE yes" for the former group (72 articles) and "AE/BE no" for the latter group (113 articles). In one case, an indexical link to Canadian English was created, and I included the article in the "yes" category because it still distinguishes a North American and a European context. In another case, it was not an English speaker that was indexed but a white American who can be regarded as belonging to the groups of Americans imitating English speakers. Due to the implicit connection to English speech, I also included the article in the category "AE/BE yes".

Figure 4.29 shows the development of the number of articles containing *luggage* and *baggage* over the course of the nineteenth century. The black area shows the proportion of articles containing indexical links to British English and American English and the grey area shows the proportion of articles containing no such links. It can be seen that the overall number of articles increases already towards the mid-century (1830-1850), then decreases slightly until the 1870s, followed by a sharp increase in the 1880s and a smaller decrease in the 1890s. In contrast to the development of all articles containing the two lexical items in close proximity, the number of articles in which *luggage* is indexically linked to

Figure 4.29: Number of articles containing *luggage* and *baggage* (within 10 words of each other) and indexical links to American English or British English

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

British English speech and *baggage* to American English speech increases only in the second half the century, especially in the 1880s. It is only in this decade that the proportion of articles containing indexical links is larger than that without such links.

It is striking that the temporal development of articles containing indexical links between the lexical items and American and British English is similar to that of several sets of articles containing phonological search terms. This can be seen in Figure 4.30, which shows that *dawnce*, twousers, *deah* AND *fellah* and *bettah* also increase significantly in the 1880s. The decrease in the 1890s also mirrors closely that of articles containing *dawnce* (and to a lesser extent those containing *deah* AND *fellah*). This observation thus highlights the importance of the last two decades of the nineteenth century for the creation and circulation of indexical links.

Figure 4.30: Number of articles containing the search terms in NCNP (per million): comparing phonological search terms to the *luggage* n10 *baggage* (AE/BE yes)

With regard to the regional distribution of articles, Figure 4.31 shows the number of articles per state of those 72 articles which indexically link *luggage* and *baggage* to British English and American English respectively. What can be observed is that Wisconsin and Kansas are the states with the highest number of articles (12 articles each), two states in which also many articles containing twousers, *deah* AND *fellah* and *bettah* were published. At the same time, however, several

Figure 4.31: Regional distribution of articles containing *luggage* n10 *baggage* which link the lexical items to British English and American English respectively

articles were also published in California in the southwest (8 articles) and in Massachusetts (6) and Maine (4) in the northeast, which shows that the circulation of articles differentiating *luggage* and *baggage* as English and American forms is not restricted to a particular region. The absence of articles in Pennsylvania, in the north and in the southeast is as noticeable as the presence of articles in Utah because that differs from the distribution of articles containing other search terms (except for Utah, where also a fairly large number of articles containing *hinglish* was published, which is interesting because in the case of *luggage* and *baggage* it is also the contrast between England and America that is the focus of the creation of indexical links). In general, articles do not appear primarily in the northeast (like those containing *hinglish*), but they also do not primarily appear in the (mid)-west and Pennsylvania (like those containing the non-rhotic search terms and twousers). Due to the absence in the north and in the southwest, their distribution does not seem to be as wide as that of the other search terms. However, as the number of articles is fairly low, it is only possible to identify some tendencies with regard to circulation.

To conclude, the quantitative analysis shows that while articles in which both lexical forms are used within 10 words of each other appear already in the first half and the middle of the century in relatively large numbers, the number of

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

articles containing indexical links between the forms and American and British English only increases notably towards the end of the century. This follows the same pattern as that of those phonological search terms associated with the dude and with Black speakers, and it underlines the importance of the last two decades in the creation of indexical links. The regional distribution shows that articles evaluating *luggage* and *baggage* as British and American forms are not restricted to any particular region, although the absence of articles in the north (except of course Wisconsin) and the southeast is noticeable. The low number of articles needs to be taken into consideration as well when comparing this set of articles to that containing other search terms. The next section will thus focus on a qualitative analysis of selected articles to show how the indexical links to American and British English were created, how the use of the American form (*baggage*) was evaluated and how the context (including the region) possibly influenced this evaluation.

# **4.2.2** *luggage* **vs.** *baggage***: indexical values and social personae**

When analyzing indexical links between *luggage* and *baggage* and social values as well as social personae, it should not be forgotten that the majority of articles do not contain such links at all because both terms are used without any apparent difference in either semantic or social meaning. An example of such an article is a travelogue published in the *Daily Rocky Mountain News* on May 3, 1874[28] . The author describes how "baggagemen" transfer "luggage" to "baggage vans", which shows that the meaning of *luggage* and *baggage* must be the same because the latter contributes the same meaning as the first to compounds (the collection of property that people take with them when they travel is transported by men and by vans).

# **First impressions of Denver.** […]

Polite and attentive **baggagemen** attend to the transfer of the passenger's **luggage** from the cars to elegant **baggage vans**, and equally courteous conductors act as escort to omnibusses and carriages, as fine as can be found in the streets of New York, Boston or Chicago, drawn by magnificent matched teams.

[emphasis mine]

A second example of an article without indexical links is a Supreme Court decision published in the *Daily Evening Bulletin* in San Francisco on August 09, 1890[103] . In the decision, *baggage* and *luggage* are used once in a coordinated noun phrase with *and*, which suggests a possible difference in semantic meaning, but they are also used in a coordinated noun phrase with *or*, which suggests that

### 4.2 Metadiscourses on lexical forms

there is no semantic difference, and that they are alternative terms for the same thing. As linguistic precision is particularly important in legal contexts, it is noticeable here that even though both terms are used, they are not distinguished in meaning, which indicates that the Supreme Court acknowledges that both terms are in use, but that there is no semantic difference, at least none that matters legally. There is also no indication of any different social meanings connected to the different lexical forms.<sup>24</sup>

# **Supreme Court Decisions.**

[…]

The Court finds that at Kansas City, Mo., in January 1888, the plaintiff engaged passage on the defendant's railroad to Coitin, in this State, bought a ticket, paid his fare and checked his trunk, containing, among other things, one lady's gold watch and chain of the value of \$150; [...], which the Court found to be "proper articles of **luggage and baggage** for the plaintiff to carry as such." [...] And [the appellant's] contention is that said articles or none of them constituted what in law is defined to be **luggage or baggage**.

[emphasis mine]

The first article which creates indexical links to the forms and thus distinguishes them not semantically but with regard to their social meaning was published in the *Weekly Ohio Statesman* (Columbus, Ohio) on September 24, 1845[13] . The correspondent writes a letter to the newspaper from Liverpool and after using the term *baggage*, he adds the information "always called luggage in England" and thus explicitly links the term *baggage* to American English and *luggage* to British English.

# **Foreign Correspondence of the Statesman.**

Liverpool, *Clayton Arms Hotel*, August 18th, 1845

Travelers frequently speak of the annoyances of the Custom House. Mr. Willis, in one of his recent letters, invokes the retaliation of American Custom House officers upon English travelers, on account of the trouble he received in the passage of his baggage here. He was probably in a bad humor.

To make you acquainted with all the particulars of passing **baggage (always called luggage in England,)** I will relate exactly what took place with mine. […]

[emphasis mine]

<sup>24</sup>The author of the article, Associate Justice of the Supreme Court of California John Sharpstein, quotes the definition of *luggage* in the Civil Code: "Luggage may consist of any article intended for the use of a passenger while traveling, or for his personal equipment". This suggests that the lexical item used in legal texts is*luggage*. Nevertheless, Sharpstein uses both terms, *luggage* and *baggage*, to express the same meaning in his explanation of the Supreme Court decision in the article. It therefore seems as if *baggage* is so commonly used that it appears in legal contexts as well, despite the fact that *luggage* is the term used in the legal text itself.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Even though the article above explicitly creates a difference in social meaning between *baggage* and *luggage*, neither of the forms is evaluated positively or negatively. The author states the difference neutrally as a difference between the two countries. This neutral perspective is adopted in almost half of the articles which link the forms to American and British English respectively (34 out of 72). In the rest of the articles, an evaluation of the terms is expressed – sometimes explicitly, but sometimes also rather implicitly. In the vast majority of cases (35 articles), the American form, *baggage*, is evaluated positively. Only in 3 cases can a decidedly negative evaluation of *baggage* be found, and in one exceptional case, the author states that he is undecided about how to evaluate the form.

As a first step, I will analyze some examples of articles which illustrate how *baggage* becomes evaluated positively as an American form. In the collection of articles obtained for this analysis, the first article conveying such a positive evaluation is headed "The English Language" and it was published in the *New Hampshire Statesman* (Concord, New Hampshire) on September 21, 1866[24] . It is a letter from a correspondent of the *Cincinnati Gazette*, in which he writes about his experiences in England, including differences between English and American speech. The fact that he labels the English expressions "peculiarities" indicates that he regards the American variants positively as a norm from which the English variants deviate. Even though he recommends the readers to study the English forms, this is merely for practical reasons – in his view, knowing about the differences will help the American traveler to communicate in England. The article thus attests to an endonormative orientation visible already in the 1860s. Differences to British usage are observed, but they do not call into question the use of the American variant in any way.

# **The English Language**

A correspondent of the Cincinnati Gazette, now in England, writes the following:

Before leaving America I thought it desirable to pay some fresh attention to the French language, in order to travel successfully in France. **It never occurred to me that I ought to study the English language, with a special view to facilitating my travels in England**. Nevertheless, I have sometimes been put to inconvenience from my ignorance of **some peculiarities in the English mode of speaking**.

# [...]

From time to time I have noticed the peculiarities of speech which have struck me ; and if your readers will not accuse me of pedantry, I will give them some extracts from my observations, promising not to deluge them with a general jail delivery of the contents of my notebook.

**Many of these peculiarities occur in connection with one's travels.** Thus *railroad* is always here a "railway." Nobody speaks of *cars*. The train is made up of "carriages." They do not talk of selling *tickets*, and they have no *ticket office*, so-called. But when they give you a ticket and take your money, they are said to "book" you, and the office is a "booking office." *Baggage* **is "luggage," and a baggage car is a "luggage van."** […]

[emphasis mine]

### 4.2 Metadiscourses on lexical forms

The second article in the collection with a positive evaluation of *baggage* was published in 1873 and the third one in 1882. The latter is particularly interesting because it combines questions of semantic difference with questions of technical progress and superiority. The title of the article is "English Recklessness" and it was originally published by the *New York Tribune* and then reprinted as a part of a collection of several articles about railroads, first in the *St. Louis Globe-Democrat* (Missouri) on December 20, 1882[55] , and then in the *Cleveland Herald* (Ohio) on December 22, 1882[56] . The author comments on differences between English and American "railroad management", which manifest themselves in the use of different terms but also in the use of "methods of operating railroads" and "devices for promoting the comfort of passengers". The author of the article aims to show that while Englishmen regard American innovations as reckless, they are actually reckless themselves because due to a lack of safety measures an accident occurred on a train in England that, according to the author, could not have occurred in America. There is thus a clear ironic undertone, when the author speaks of "the hazards of American recklessness". In the same vein, it becomes clear that he does not share the British view that "having different names for the same things" is "an unnecessary way of asserting their [American] individuality", but that, quite to the contrary, Americans have a right to be self-confident due to their superior position regarding technology and operating systems and thus, implicitly, a right to their own names. The use of *baggage* instead of *luggage* is therefore not only linked to American usage, but it also indexes autonomy, individuality and superiority based on technological innovation and progress.

# **English Recklessness.**

[From the New York Tribune]

Englishmen have been harassed these many years by American methods of railroad management. The divergence of views dates from the introduction of the system in both countries. The English were annoyed at the outset because the Americans persisted in calling a shunt a switch, a driver an engineer, a stoker a fireman and luggage baggage. This seemed a very unnecessary way of asserting their individuality. Not satisfied with having different names for the same things, the Americans devised characteristic methods of operating railroads, and introduced many devices for promoting the comfort of passengers.

# […]

Englishmen have had for thirty years a characteristic word which they have applied to American railroad methods. That word is "reckless." They have adhered with dogged persistence to many inconveniences of their own system rather than expose themselves to the hazards of American recklessness.

# […]

American railroad officials will read the details of this accident [in England] with a grim smile. They have been censured so often by English critics of the recklessness of their methods that they will derive consolation from this exhibition of incapacity and carelessness. Passengers in American sleeping-cars are not allowed to smoke in their berths, whether they are drunk or sober, nor can

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

they take their reading-lamp to bed with them; and if an engineer receives warning that something is wrong and that the train must be stopped instantaneously, he is not prevented by the printed regulations of the company from saving the lives of the passengers.

Another important value indexically linked to *baggage* is that of authenticity. This value is established, on the one hand, in relation to Englishmen and, on the other hand, in relation to the group of Americans who hold English linguistic forms in high regards. The latter set of articles connects the discourse on *luggage* and *baggage* to the discourse on phonological forms linked to the American dude and other white American speakers imitating English speech. However, I will discuss an article first that establishes authenticity based on national identity. It is also headed "The English Language" and it was published in the *Galveston Daily News* (Houston, Texas) on October 31, 1886[72] , and in the *Morning Oregonian* (Portland, Oregon) on December 07, 1886[73] . It describes and lists a large number of linguistic differences between England and America, among others lexical items belonging to the "lingo of railways". The respective paragraph illustrates that *baggage* and *luggage* are one pair among many others, which makes the form not particularly salient in the context of this article. Nevertheless, the attitude expressed towards the American forms is also valid for *baggage*: The author constructs it as a sign of national identity and a marker of "a true and sincere American" and thus creates a link to American nationality as well as to authenticity. Authenticity is, in fact, at the basis of his argument because he finds that if a high value is placed on authenticity, it cannot be detrimental for an American to signal his nationality. On the contrary, hiding one's nationality would imply that the speaker is untrue and insincere. Another important aspect that becomes visible in this article is the connection to discourses on language beyond newspaper articles: The author explicitly refers to "letters from both sides of the Atlantic", "books on travel" and "conversation[s] among returned travelers", which all contribute to the discussion on linguistic differences between England and America. This also shows that discourses on language in England and America are not separate from each other, but that they interact in many ways. Also within newspaper discourse, the article shows the connection between England and America as well as between different regions in America: The author cites an article by a London correspondent of the *Argonaut*, a political journal based in San Francisco, California, thus emphasizing the connection between America and England, and the fact that this article was published in Texas and Oregon shows that discourses on language in America spread to regions far apart in the United States.

### 4.2 Metadiscourses on lexical forms

# THE ENGLISH LANGUAGE

The subject of English pronunciations of different words, names, and places totally out of accord with the spelling, together with the **difference that exists in England and America in respect to the expression of the same thing**, has been frequently touched upon in letters from both sides of the Atlantic, referred to incidentally in books of travel, and often made the topic of conversation among returned travelers, whether from Liverpool or New York, says a London correspondent of the Argonaut. But I don't think the subject has ever been thoroughly gone into. Yet a knowledge of these pronunciations and expressions would be extremely useful to the American traveler in England.

# […]

**The lingo of railways differs wonderfully.** Railroad is railway; the track is the line, and the rails the metals; the cars are the train; to switch is to shunt; a turnout is a siding; a locomotive is an engine; an engineer a driver, and a fireman a stoker. The conductor is the guard, a car is a carriage, **baggage luggage, a baggage car a luggage van**, and a freight train a goods train. A depot is a terminus or a station, and a switch-tender a pointsman or signal man. […]

[…]

Now, I don't mean to contend, and I hope I shall not be misunderstood as contending, that it is necessary, if, indeed, in all senses desirable, for any American visiting England to in the least sink his **national identity**. Far from it. There is nothing that a true and sincere Englishman likes and admires more than **a true and sincere American**. It will be no detriment to a man among true Englishmen that he is an American. But while it is not desirable to hide one's nationality, and is a sign of bad form to do so, it is equally lacking in good form to parade one's birthplace and make a blowing-horn of it. […]

# [emphasis mine]

The article thus illustrates very importantly that nationality is connected to authenticity and that this connection forms the basis for evaluating the use of *baggage* positively. The same argument can also be found in other articles, culminating for example in the simple conclusion at the end of a short paragraph published in the *Atchison Daily Globe* on June 16, 1888[88] : "The American language for Americans and the English for the English".

The value of authenticity not only played a role when the focus was on the contact between English and American people, for example through traveling, but also when an inner-American context was foregrounded because, as with several phonological forms, the use of *luggage* is not only indexically linked to English speakers, but also to a group of Americans imitating English speech. There are two articles which are particularly illustrative examples of how this group is characterized and how this characterization is embedded in the article to express a political view. The first example was published in the *Rocky Mountain News* on August 29, 1883[59] , and its main aim is to criticize a "technical and mischievous decision of Judge Nelson in the Circuit court of the United States, at its sitting in Boston, August 22, upon the admission of Chinese laborers to this country". The decision interprets a prior act of Congress against such an admission by stating that it applies only to Chinese laborers who come directly from

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

China, which means that Chinese immigrants may enter the country if they arrive from the ports of other countries. The author of the article views this very negatively because he finds that the Chinese pose a threat to the American middle and working classes and he constructs these classes as being in opposition to the wealthy classes, who have an interest in cheap Chinese laborers to promote their own wealth, and whose interests are represented by Judge Nelson in the Boston Court. The following paragraph shows how he constructs this class difference and links it to language:

[…] because the **middle and laboring classes** of this country, the **intelligent artisans** and the **inventive mechanics**, are the **genuine American citizens**, with no aspirations toward British manners or British strong government, the **mania** which, in common with **dudeism**, is at present so prominent among the **wealthier classes, whose baggage is suddenly changed to "luggage,"** while the Yankee twang of a few years ago, with its nasal independence of harmony and men, has given place to **a badly affected English drawl** and an **assumed indifference** to the everyday affairs of our everyday world. [...]

[emphasis mine]

The adjectives used to describe the middle and working classes are "intelligent" and "inventive" and by describing them as artisans and mechanics he creates an image that rests on physical skills and force combined with intelligence and inventiveness and it is implicitly conveyed that he regards these traits as crucial in nation-building (because the physical aspects connect to the image of "building" a nation, while intelligence and inventiveness are associated with progress). These Americans are marked as genuine and thus as authentic and their interest in nation-building and progress differentiates them from the wealthier classes, which are depicted as oriented towards England and indifferent to American affairs. This is connected to language because the "badly affected English drawl" indicates their lack of authenticity as Americans. The link to the social personae of the "Anglomaniac" and the dude (discussed in detail in 4.1.2) is alluded to by calling their "aspirations toward British manners or British strong government" a "mania" and by using the term "dudeism", which implies that by the middle of 1883, the characterological figure of the dude has already been firmly established. The lexical items *baggage* and *luggage* are very salient here because they are the only items linked to the repertoire of the British-oriented wealthier classes. Furthermore, the author underlines the lack of authenticity in their language use by claiming that "baggage is suddenly changed to 'luggage'", which shows that *luggage* is not evaluated as a natural, long-established form intrinsically connected to American people, but as something unnatural that is put on quickly by some Americans in an opportunistic attempt to signal social characteristics deemed desirable by the wealthier classes but viewed negatively by the author of the article.

### 4.2 Metadiscourses on lexical forms

A second example of an article characterizing the group of American speakers indexically linked to *luggage* was published ten years later, on September 23, 1893[132] , in the *Milwaukee Sentinel* (Wisconsin) and on September 29, 1893[133] , in the *Morning Oregonian* (Portland, Oregon). It is a comment on an article criticizing the appointment of James J. Van Alen as minister to Italy, published in the *New York World*. The author claims to disagree with the criticism, but his text is so full of irony that it rather reinforces the criticism to a great extent. The main reason put forward against the appointment is that Van Alen did not earn it based on his hard political work and accomplishments but based on a large sum of money that he contributed to the Democratic campaign fund. The statement that "His friends affirm that he 'bought the office like a gentleman'" extends the criticism to an entire group of people: the social and political elite in the northeast of the United States. The place linked to this elite is Newport (Rhode Island) because this is where they spend the summer in their "million-dollar cottages", drive around in their "swell carriages", ride "dock-tailed horses" and play tennis in "stunning tennis suits". However, the most important signs indexically linked to these people, in the author's view, are the monocle and the linguistic repertoire. The monocle (or eyeglass) also plays a prominent role in articles characterizing the dude and, not surprisingly, the language is described as being "as near the English as any man not born in England can hope to acquire". As in the article criticizing the court decision, *baggage* and *luggage* are very salient because the only other forms mentioned are the variants *brasses* and *checks* and *patron* and *president*. Expletives are mentioned generally, as well as a distinct English accent, but no concrete forms are listed, which shows that the readers are expected to have a general idea of the linguistic repertoire linked to rich Newport people. It is important that this group is distinguished from "mere Americans" who speak "vulgar American" English. The adjectives represent the view of the wealthy classes – their "fine contempt" for other social classes – and serve to underline the ironic stance of the author who aims to characterize the Newport elite as snobbish, condescending and not authentically American despite the fact they are born in America. Their orientation towards England is also depicted as detrimental to America because they follow English political ideas, which are in England's interest rather than in America's, and because they "draw revenues from America to spend in Europe" and thus harm the economy. In short, Mr. Van Alen is constructed as a prime example of an elitist northeastern American who is not authentic, whose political position is not earned through hard work but bought with money, and whose political ideas and economic and social behavior are harmful to the nation. The elite's bad influence on 'normal' Americans is represented by the trainmen who adapt linguistically to Newport speech, which

reveals the prestige that this linguistic repertoire has acquired, but at the same time, the Newport elite and their attempt to be as English as possible is portrayed so negatively that the readers are urged *not* to follow the example of the trainmen and align with their language use. That they only come near to sounding English but do not fully achieve it (the argument being that only people born in England can speak British English) further underlines the lack of authenticity of Newport speakers and their uselessness as a model for other Americans.

# MR. VAN ALEN'S APPOINTMENT.

In making war on President Cleveland for **the appointment of Mr. Van Alen as minister to Italy**, The New York World exhibits a narrowness such as we might expect in a provincial newspaper. As long as the offices are to be used for the reward of party workers, there is no reason why the rewards should be limited to the vulgar bosses and those employed by the vulgar bosses. While it is proper that the Poles should be recognized, it is equally proper that the real gentlemen who honor America by being born in this country, should share in the spoils. At **Newport in summer** we see the heights to which **the most favored American** may reach. We see it in the **million-dollar cottages**, in the **swell carriages** and **dock-tailed horses** and in the **stunning tennis suits**, but we see it **most** in the **monocle** and the **expletives** and in **the language** generally, which is **as near the English as any man not born in England can hope to acquire**. Even the trainmen who run into Newport have learned to speak of **"the brasses for the luggage"** instead of the **vulgar American "checks for the baggage."**

**Mr. Van Alen is the most English** of all the people at Newport. He **lives most of the time in dear old London**, he **has a fine contempt for mere Americans**, and **his ideas of American politics are acquired through The London Times**. Last year he spent a few months at Newport, in response to the suggestion of The London Times that the British in America should organize to advance the cause of tariff reform, he drew on his London bankers for money to fit out the Newport Cleveland club, of which he became the patron, as they say in England, or the president as the vulgar Americans say. Mr. Van Alen did more. He agreed with the Hon. William H. Whitney to give \$50,000 to the Democratic campaign fund if Mr. Cleveland, in event of Democratic success, would appoint him to some office abroad where he could maintain his social life and where his relations would be with Europeans rather than Americans. This sum was much greater than that raised by The World for missionary work in the West.

We see no reason to complain of Mr. Van Alen's appointment. He has cultivated **a close resemblance to His Royal Highness**, the prince of Wales, **he speaks with a distinct English accent**, he **wears white spats**, and to the army of Americans who **draw revenues from America to spend in Europe** he will be a gladness and a joy. His friends affirm that **he "bought the office like a gentleman,"** paying the large price with a draft on London without grumbling. He has a clear title to the office, and since his appointment is eminently satisfactory to the gentlemen and ladies at Newport, who are condescending enough to tolerate America for a few months each year, we do not see why The New York World should complain.

[emphasis mine]

The articles analyzed so far thus indicate that there was a process of discursive negotiation, in which the association of *luggage* with English usage is a reason for some Americans to adopt the form and for other Americans to reject it. This negotiation is a reflection of the struggle between an exonormative and an endonormative orientation and it is illustrated very well by a short paragraph which pointedly comments on the use of *luggage* and *baggage* as well as some

### 4.2 Metadiscourses on lexical forms

other lexical items (*gown* vs. *dress*, *ticket office* vs. *booking office*, *car* vs. *van*, *engineer* vs. *driver*, *fireman* vs. *stoker*). It was published in the *Atchison Daily Globe* on September 15, 1887[81] , and it shows how the positive indexical values of 'correct' and 'proper' are re-evaluated as 'un-American' and 'not authentic': By stating that *gown* "is English, and may be more proper than dress", the author acknowledges the prestige accorded to the form, which is based on an exonormative view in which English forms are regarded as a model. The continuation of the sentence with "but it is English, and that ought to settle the matter as far as every true American is concerned" implies that this exonormative view is rejected in favor of an American norm. As the value 'proper' is associated with the English form, the American form must be evaluated positively in a different way, and this is done by presenting authenticity as a goal and linking it to nationality. This link is signaled by the noun phrase *every true American*, which implies that those who do not follow the American norm are not really Americans. Another important opposition is that between "anglo-maniacs" and "sensible Americans", which relates the use of the different forms to the opposition between sickness and health. Overall, it is striking that the writer of the article protests against a new development and urges its readers to contain it by warning against the boldness of its proponents. It is thus not the case that the American forms are characterized as 'new' and promoted to replace 'old' English forms, but, quite to the contrary, American forms are presented as already established (as 'normal' forms used by 'sensible' Americans) and are now under threat by the new forms (used by "anglo-maniacs"), so that they need to be defended.

We protest against the word "gown" as applied to a "dress." "Gown" is English, and may be **more proper** than "dress," **but it is English**, and that ought to settle the matter as far as every **true American** is concerned. In this connection we would suggest that the word **"luggage"** as applied to "baggage" is beginning to be used, and **should be frowned down**. The first thing we know the **anglo-maniacs** will become so bold as to call a ticket office a booking office, a car a van, an engineer a driver, and a fireman a stoker, and then life to **sensible Americans** will become simply unbearable.

# [emphasis mine]

The view that an established American usage is to be preferred as a norm over an English norm is also presented in an article headed "Good Form in England" and published in the *Atchison Daily Champion* on February 07, 1891[106] , and in the *Rocky Mountain News* on March 01, 1891[107] . The article was originally published in a New York newspaper, the *New York Ledger*, and it is notable that its argumentation for not adopting 'English' forms like *luggage* relies more on practical issues than on emotions connected to values like nationality and authenticity. According to the author, the differences between English and American usage

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

are so numerous ("we could enumerate a thousand other peculiarities") and so full of detail that it is simply too time-consuming and too difficult ("Americans who ape English usages almost always blunder") to acquire the English forms. In addition, the acquisition of English forms is also portrayed as unnecessary and not worth the effort (the time "could be put to a better purpose").

# **GOOD FORM IN ENGLAND Terms Considered Proper Among the English Gentility**

To adequately indicate the divergences between the ways of English society and our own would require a volume, but some striking examples may be given in a few paragraphs.

[…]

**Americans who ape English usages almost always blunder** in the use of crests. [What follows is a list of differences, among others *baggage* and *luggage*.]

[…] We might enumerate **a thousand other peculiarities**, but we have cited enough to show that **an American citizen should not easily acquire what in England is called "good form" without an expenditure of time that could be put to a better purpose**."—N. Y. Ledger.

[emphasis mine]

Taking into consideration all articles containing the search term *luggage* and *baggage* in the database NCNP, it is noticeable that the number of articles containing *explicit* comments on the lexical items are far more numerous than in the set of articles containing the search terms representing phonological forms. This is not surprising given that pronunciation respellings invite humor and are thus likely to be used in humorous articles which convey evaluations of forms in rather implicit ways. However, there are also some humorous articles focusing on the difference between *luggage* and *baggage*. The article "Terms" is a good example in this regard. It was originally published by the *Detroit Tribune* (Ohio) and reprinted in the *Milwaukee Sentinel* (Wisconsin) on July 01, 1894[142] , and it creates salient indexical links between *baggage* and *luggage* and different American regions. It consists of a very short dialogic exchange between an "Omaha Coxeyite" and a "Boston Coxeyite", in which the former asks "Any baggage?" and the latter replies "If you mean luggage, no". The shortness of the dialogue draws a maximum of attention to the lexical items *luggage* and *baggage* and to the different figures using them. Coxeyites were a group of people who were part of a labor movement led by Jacob Coxey and called "Coxey's Army" in newspapers (Barber 2007: 327). They demanded that the federal government create jobs to counter the massive unemployment caused by an economic crisis, and in order to draw attention to their demands they organized a protest march from Ohio to Washington in 1894. Around 300 supporters left from Ohio, but many others traveled to Washington from different places, so that there were 1,000 people marching down Pennsylvania Avenue on May 1, 1894 (Barber 2007: 327). That

### 4.2 Metadiscourses on lexical forms

Coxeyites were rather poor people is emphasized in the dialogue by pointing out that they did not carry any baggage. Their similarity with regard to social characteristics reinforces the focus on the different regions that become indexically linked to *baggage* and *luggage*. Omaha, the biggest city in Nebraska, is representative of the western parts of the United States, while Boston is representative of the northeastern parts. The humor of the dialogue builds on this contrast between the Bostonians' social position (being a poor Coxeyite) and his use of *luggage*, a linguistic form associated with the wealthy northeastern elite. The article thus humorously criticizes the prestige of the variant *luggage* in the northeast that extends even to poor Coxeyites. Doing so, the article mainly serves to reinforce the contrast between the northeast and the west that has been found in articles containing *dawnce*, *deah* AND *fellah* and twousers as well.

# **Terms.**

Detroit Tribune: Omaha Coxeyite—Any baggage? Boston Coxeyite—If you mean luggage, no.

Against the overwhelming majority of positive evaluations of *baggage* there are only three articles in which the form is evaluated negatively. One of these article presents a Supreme Court Decision, which was published by the *Daily Evening Bulletin* (San Francisco, California) on July 20, 1886[71] . It states that "The terms baggage and luggage signify one and the same thing. The former is the form in general use in the United States, while in England the latter prevails". This statement suggests a neutral perspective, which links the two forms to the two different nations. However, the next statement is: "Our Code has adopted the English expression". Choosing the British form over the American form signals to the reader that the British form is more appropriate, at least in such an official and institutional context. The American form is in turn implicitly evaluated as inappropriate and consequently as inferior to the British form.

A more explicitly negative evaluation is expressed in the article "'Luggage,' not baggage", published in the *Los Angeles Daily Times* (California) on September 19, 1882[53] . As in the first example, it is acknowledged that *luggage* is the English term and *baggage* the American term, but the latter is judged to be "inadequate, incomprehensive, unsatisfying and limited" and to be used with "undiscriminating and imbecile inappropriateness". The words used to evaluate *baggage* show that the term is evaluated negatively mainly on linguistic (and not on social) grounds. It is seen as not adequate to designate that which it should designate and thus as limited. However, the adjective *imbecile* also links the linguistic deficiency constructed in the article to a human characteristic, based on which users of the form can be linked to stupidity. The English form is thus linked to being

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

superior because it is "objectively" more appropriate and used by speakers who are smart and educated enough to know that.

# **"LUGGAGE," NOT BAGGAGE. Graphic Account of a Desperate Struggle for it on the San Francisco Wharf.**

[Following is an extract from a private letter received in the TIMES office from a personal friend recently arrived at San Francisco from the Hawaiian Islands. It will be read with deep feeling by the average sea traveler:]

[…]

I found, for the first time, the vast and comprehensive appropriateness of the English word **"luggage,"** as compared with our **inadequate**, **incomprehensive**, unsatisfying and **limited** term **"baggage,"** which we apply with **undiscriminating** and **imbecile inappropriateness** to things that never could, would or should be "bagged," "corraled," or otherwise gathered together, in any other way than by infinite lugging." […]

[emphasis mine]

In a similar vein, lexical differences between England and America are discussed "objectively" in the article "Springs of English", published in the *Daily Inter Ocean* (Chicago, Illinois) on March 20, 1892[116] . The author calls for a separation of the judgement of lexical forms and nationality. He rejects any focus on correctness or incorrectness but suggests concentrating on the "pure value of the word". Linguistic criteria are thus also regarded more highly than social criteria and the "pure value" can be determined by finding out the way which "represents the idea" best. This in turn is determined based on linguistic economy: "the least friction of letters [which] convey the greatest amount of thought". However, in the case of *baggage* and *luggage*, the proposition is to "sacrifice economy to truth" and to "give to it the comprehensive Latin name 'impediments'", so that not only *baggage* but also *luggage* is evaluated negatively as inferior to a third term.

> **SPRINGS OF ENGLISH.** One in Old England, the Other in America. Anglicisms, Americanisms. Curious Differences in Words and Expressions. Differentiation of a Divided People—Habits of Speaking and Eating Contrasted. Old Customs Preserved.

London, March 12.—*Special Correspondence*—

TRADITION and habit are such linguistic tyrants, that it is not easy to place oneself in a perfectly unprejudiced frame of mind to judge between words of like significance in the same language. From the nature of the case it is impossible to find for the English tongue an unbiased umpire, yet both in England and America there must exist minds of a sufficiently judicial temper **to separate** themselves for the moment **from their nationality**, and, presuming either to be equally correct, decide upon the **pure value of the word**; which the better represents the idea and which will, with the least friction of letters, convey the greatest amount of thought; for in the **economy of language** is its greatest strength.

### 4.2 Metadiscourses on lexical forms

The "baggage car" is a "luggage van," and, **of course, "baggage" is "luggage:"** but why—as it is neither universally "bagged" nor "lugged"—not **sacrifice economy to truth** and give to it the comprehensive Latin name "impediments?"

# [emphasis mine]

To conclude the qualitative analysis of *luggage* and *baggage*, it is striking how positively *baggage* is evaluated as the American form that indexes not only American nationality but also authenticity. It is strongly connected to discourses on non-rhoticity, labiodental /r/ and the back bath vowel because like *luggage* those forms index English speakers or American speakers who are not genuine and sincere and who often fail in their ridiculous attempt to imitate English speech. The oppositions between different classes and regions are found for *luggage* as well: The wealthy upper classes in the northeast of the country use *luggage* because they suffer from 'anglomania', while the sensible middle and working classes in the north, midwest and west of the country use *baggage*. Whether the positive evaluation of *baggage* can also be found for *pants*, the American variant of *trousers*, will be the central question guiding the qualitative analysis in the next section.

# **4.2.3** *pants* **vs.** *trousers***: indexical values and social personae**

The search for articles containing *pants* within ten words of *trousers* revealed three broad categories of articles, according to which the following qualitative analysis will be structured. They are largely equivalent to those categories established for *baggage* and *luggage:* First, there are articles in which the two forms are used synonymously, which means that no indexical links between the lexical items and different social values are created. Secondly, there are articles in which those links are created and in which *trousers* is evaluated positively as a form that should be preferred to the form *pants*. The third category comprises articles in which the links are also created but in which *pants* is the term that is determined as preferable. The last part of the analysis will focus on one particular text type, namely advertisements, because they provide interesting insights into how the different social meanings linked to the lexical items have an effect on the use of the terms in a context in which language is used to persuade people to buy particular items from particular stores and in which language use has thus direct economic consequences.

The first article in the collection of articles obtained by searching for *pants* and *trousers* occurring within a range of ten words was published in the *Fayetteville Observer* (North Carolina) on May 16, 1861[23] , a month after the outbreak of the Civil War. The article is addressed to volunteer soldiers and provides a set

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

of instructions on proper dress, hygiene, and other aspects deemed important to ensure the "success and efficiency of any army". By referring to the leg garments as "Pants or Trousers", the author marks the terms as alternatives, and he does not give any indication as to what motivates the use of the different forms. The fact that *pants* is named first, before "or trousers" is added, and the fact that at the end of the sentence only *pants* is used ("outside of pants") creates the impression that *pants* is less marked (and possibly more common) than *trousers*. Nevertheless, the absence of any social differences linked to the alternative lexical items makes the article a clear exemplar of the first category.

# FOR VOLUNTEERS.

The Duke of Wellington occupied himself a good deal with details of very much the same character as those which we propose to speak of in a few brief papers. He knew how much they had to do with the success and efficiency of any army; and, while in command in Portugal and Spain, found time to discuss in his correspondence the size of "*camp kettles*"

1. […] **Pants or Trousers** should be of same color and material with Blouse, cut *full* around hips and knees; and when practicable, leather gaiters should be provided *a la Zouave*, to be worn outside of pants. […]

[emphasis mine]

The second example of an article belonging to the first category was published in the *Atchison Daily Champion* (Kansas) on May 26, 1888[87] , and therefore at a time when discourses on language in newspaper articles were highly prominent in America. Despite the increasing interest in language use that is apparent in numerous articles of the 1880s analyzed above, in this article several variants are used without any reason being provided for the use of the different forms. The topic of the article is the problem of pants bagging at the knees and it thus requires a repeated reference to the garment in question. The author uses *pantaloons* first ("how to prevent pantaloons from bagging at the knees"), followed by *trousers* ("Your trousers will bag") and later *pants* ("A great many men pull their pants up on their knees"). *Trousers* is the only variant used a second time ("The smallest part of the trousers is that around the calves".). The article thus creates the impression that the author, who is identified as a tailor at the end of the article and who is thus professionally involved in the subject he discusses, is aware of the existence of different terms but that he deliberately avoids choosing one over the others. That he uses *trousers* twice could be interpreted as a slight preference for *trousers*, but this preference is not at all salient.

# **Bagging at the Knees.**

A great many inquiries are made as to how to prevent **pantaloons** from bagging at the knees. There is only one answer to these, it can't be done. Your **trousers** will bag, and you can't help it. The bagging can be lessened by frequent pressings and taking good care of them, but as long

### 4.2 Metadiscourses on lexical forms

as men bend their knees in walking their pants will bag. The skin would also, if it didn't settle back. A great many men pull their **pants** up on their knees when seated to prevent their bulging. This is very foolish. The smallest part of the **trousers** is that around the calves of the legs, and, of course, by pulling them up and bending the leg a greater strain is brought to bear on the cloth so it would not stretch so much, but so far no tailor has succeeded in hardly lessening the cause of the complaint. The tailor who does make the discovery will at the same time make a fortune.—Tailor in Globe-Democrat

# [emphasis mine]

That tailors are regarded as experts not only on the clothing they create but also on the linguistic terms that are used to designate such items of clothing is shown in the following article, which, in contrast to the article above, belongs to the second category because the tailor explicitly argues in favor of using the word *trousers*. The topic of the article is the same as the one above, the problem of bagging pants, and it was published only one year later, on February 23, 1889[91] , in *The Wisconsin State Register* (Portage, Wisconsin). The article had originally been published in the *New York Mail and Express*, which is important because it indicates a different regional context in comparison to the article above: The tailor's opinion in the 1888 article was first expressed in the *Globe-Democrat*, which is a St. Louis newspaper, while in this article a reporter interviews a "fashionable tailor up town", which indicates that the tailor is from New York. This New York tailor is quoted as saying that "There is but one way to prevent bagging, and that is to pull the trousers up a few inches over the knees whenever you sit down". Both the adjective *fashionable* used to describe the tailor and the self-confident manner in which he provides his expertise establish the tailor as a person of authority. When asked about his preference for the word *trousers*, he argues that it is "the proper word", while *pantaloons* is not proper because it "referred originally to the clown in the pantomime" and *pants* is not proper because it is "a very inelegant and inadequate term". The evaluation is thus mainly based on associations to clownishness on the one hand and elegance on the other hand. The association with clownishness is motivated by drawing on the etymology of the word *pantaloons*, but the elegance of the term *trousers* is not motivated. However, by describing the "bifurcated garment" as "noble", he connects the elegance of the items to the (suggested) elegance of the word *trousers* and thus creates a reason for using his preferred variant. In general, his evaluation of the terms is rather presented as a matter of fact, which does not require much argumentation. Again, this is possible because he portrays himself as a professional whose opinion can be trusted – a position that is underlined by his comment "You see, I have studied up so as to be posted in every branch of my profession". His last statement "A tailor may use a goose and not emulate one" plays on the

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

two meanings of *goose* – 'a tailor's iron' and 'a simpleton' – and underlines that he wants to be taken seriously as a professional.

> WHY TROUSERS BAG. A Fashionable Tailor Tells How to Prevent it Without Trouble. [...]

"Why do I use the word **trousers**? I think it the **proper word**. **Pantaloons** referred originally to the **clown in the pantomime**, and **pants** is a **very inelegant and inadequate** term with which to designate the noble bifurcated garment. As to breeches, it is worse than either pants or pantaloons. You see, I have studied up so as to be posted in every branch of my profession. A tailor may use a goose and not emulate one." —New York Mail and Express

[emphasis mine]

The preference of New York tailors for *trousers* also becomes visible in the following statement, published in the *Yenowine's News* (Milwaukee, Wisconsin) on February 23, 1890[102] , as part of a collection of entertaining short paragraphs in a section labeled "By-the-bye": "A New York tailor advertises 'Pants for gents, and trousers for gentlemen.'" The advertisement thus links the difference between *pants* and *trousers* to the difference between *gents* and *gentlemen*. For *pants* and *gents* this is done on a linguistic level: The link between the forms rests on the one hand on their phonological similarity and on the other hand on them being the result of the same morphological process involving shortening and suffixation with a plural -*s*. But social aspects played a role as well: The clipping of *gentlemen* to *gents* was often negatively viewed as slang and it was mentioned together with the clipping of *pantaloons* to *pants* as the following article shows, which was published on January 23, 1856[16] , in the *Charleston Mercury* (South Carolina).<sup>25</sup> In this article, which is entitled "Avoid slang words", the "author of the behavior book" is cited as stating that "There is no wit [...] in a lady [...] in calling pantaloons 'pants' or gentlemen 'gents' [...]". The shortenings are rather labeled "slang words" and explicitly evaluated as "detestable", which is why they have to be avoided if a lady would like to belong to the "best society".<sup>26</sup> This evaluation provides support for interpreting the New York tailor's advertisement as highlighting the prestige of the word *trousers*. The creation of indexical links works on multiple levels here as the quality of the linguistic terms is linked to

<sup>25</sup>The article was quite popular: The search in the database NCNP yields five tokens, of which two articles were also published in the south (in the *Daily South Carolinian* in South Carolina, and in the *Fayetteville Observer* in North Carolina), and two articles were published in the east (in the *Daily National Intelligencer* in Washington, D.C., and in the *Bangor Daily Whig & Courier* in Maine).

<sup>26</sup>It is notable, however, that the lady is *not* advised to use *trousers* instead of *pantaloons*. This suggests that the latter term is viewed as normal and acceptable in 1856, at least in South Carolina, where the article was published.

### 4.2 Metadiscourses on lexical forms

the quality of the items of clothing and to the quality of the people wearing the items of clothing. However, the present context of the advertisement also needs to be taken into consideration: Its appearance in a section with short humorous paragraphs and the introduction of the slogan by "a New York tailor advertises" (emphasizing the place New York) implicitly conveys that the New Yorker's evaluation of *pants* and *trousers* is funny and different from evaluations found in other places, such as Wisconsin, where this article was published. The indexical links created by the advertisement are thus revalorized by ridiculing the New Yorkers' concern with being gentlemen and using specific terms to signal this status. This shows that the line between the three categories is blurry – an article can contain links to negative social values and present a case of a positive evaluation at the same time simply because of the way it recontextualizes the information in the article and because of the context it is published in.

A similar case of revalorization can be found in an article in which the positive evaluation of *trousers* in New York is also described explicitly. It was published in the *Daily Picayune* (New Orleans, Louisiana) on May 16, 1896[161] . The title, "No Pants", draws attention to the form in a humorous way because at this point it is not clear that the subject of the article is the lexical form *pants* and not the object it designates. This creates the potential of surprise to readers who might expect that the article is about pants not being worn by some people. The author of the article then describes the social values linked to the forms not only in New York, but also in Boston: As in the articles above, elegance and correctness are contrasted with vulgarity, and a link between the quality of the item of clothing and the term used to designate it is established (*pants* being "hand-me-down articles in a ready-made clothing store", *trousers* being "the creation of swell tailors"). The homophony of the noun *pants* and the third-person singular form of the verb *pant* is used to link the use of the term *pants* to dogs and thus implicitly to animal-like behavior, in contrast to *trousers*, which is linked to men. In addition, the author emphasizes the importance of social elites and authoritative figures in the determination of the social status of the term. The "elegant and correct Bostonian" is representative of the Boston social elite, while in New York a specific person is named: Theodore "Teddy" Roosevelt, who was the New York City Police Commissioner (the head of the New York City Police Department) at the time. Mentioning Theodore Roosevelt also serves to underline the mechanisms behind the spread of the usage of a lexical item. The political power exercised by him is seen as having direct consequences on language use – not just by people under his direct command (policemen) but also on the general public ("What Teddy Roosevelt says goes in New York"). The fact that this description was published in a New Orleans newspaper has an important effect on the interpretation

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

of the indexical links described: By focusing on the evaluation of *trousers* and *pants* in Boston and New York without relating it to evaluations found in other places, the two cities are implicitly marked as special and contrasted with the rest of the country. Furthermore, the last line in particular reveals an ironic and humorous undertone ("from this on 'pants' will be the exclusive property of dogs, who can have them creased or not as the weather permits"). The author here alludes to the importance attached to fashion by upper-class men in Boston and New York and ridicules it by ironically suggesting that in these two cities even dogs wear pants with creases. This ridicule creates intertextual links to articles about the dude, for example to the one discussed above in which Cholly tells his friend that he cannot drown himself because this would take the creases out of his "twousers". It therefore becomes clear that the author views the situation in Boston and New York rather negatively. As in the article quoting the advertisement by the New York tailor, the links between *pants* and negative social values (inelegant, incorrect, animal-like and thus uncivilized) are revalorized as being specific to New York and Boston, which marks these places negatively as being dominated by a social and political elite that exerts its power even on the language that people use.

# No Pants.

The social status of trousers was settled in Boston a long time ago. The **elegant and correct Bostonian** pointed out that **a dog pants**, but **a man wears trousers**. Others held that these articles of masculine attire were **"pants"** when they were the hand-me-down articles in **a readymade clothing store**; but **the creation of a swell tailor** were **"trousers."** In **New York**, Mr. Roosevelt has just made an important decision on this subject. He has ordered that the **vulgar word "pants,"** referring to bifurcated garments, shall not occur in any report made to the police department. The police are to officially speak of nether clothes as "trousers," or else be silent on the subject. What Teddy Roosevelt says goes in New York, and from this on "pants" will be the exclusive property of dogs, who can have them creased or not as the weather permits.

[emphasis mine]

A different way of conveying negative values indexed by *pants* is used in the following article, which was published in the *Raleigh Register* (North Carolina) on July 29, 1885[68] . It is a short paragraph containing a brief narrative telling the story of "the maiden" who decides against marrying Fitznoodle because his use of *pants* reveals his uneducatedness and ignorance (""I'll wed no man so ignorant," she said; 'he uses "pants" for "trousers"!'"). The last comment by the narrator, "And she was right", shows that the negative evaluation is emphasized and that, in contrast to the two articles above, no revalorization can be observed. It is interesting that in this case no indication as to the place and context of the conversation is given, which implies that the negative evaluation of *pants* is not restricted to a particular region, but rather is a universal phenomenon. This

4.2 Metadiscourses on lexical forms

shows that while indexical links are revalorized in some articles, they are not in others, indicating that competing evaluations circulate in newspaper discourses on language.

> **ITEMS ABOUT WOMEN** Grave and Gay—Lively and Severe […]

"His pants alarm me so," the maiden said (referring to her poodle, which an unruly cow a chase had led) as she walked with Fitznoodle. "Are they too tight?" the untaught Fitz replied, his indignation rising. (He thought the maiden's mind he occupied, and her soliloquizing). Sharply she turned, and in her pretty head her eyes glowed like a mouser's; "I'll wed no man so ignorant," she said; "he uses 'pants' for 'trousers'!" And she was right.—*Chicago Tribune*.

It is striking that in none of the above articles evaluating *pants* negatively and *trousers* positively was a reference to British English usage made. In the case of *luggage*, the argument could be found that *luggage* is more proper because it is the English term – but this or a similar point was not made to argue for the use of *trousers*, despite the fact that the association between *trousers* and British English speech can be found in several articles. The article "Pants" (September 30, 1875[33] ), which I already analyzed in §4.1.2, begins with the statement "Pants or, as they call them in England, trousers—are not a subject which the average citizen cares, except in the presence of his tailor, to discuss". However, instead of evaluating *trousers* positively as 'proper English', the characterization of English people in the article is very negative, so that I assign this article to the third category of articles analyzed in this section: The use of *trousers* is constructed as *not* desirable because it is English and because English people are no role models. I have already described the ridiculous figure of the "young Oxford swell" and the many different kind of "twousers" in his possession. This shows a connection between the lexical and the phonological level because it is not only the lexical item *trousers* but also the realization of /r/ as a labiodental approximant that is linked to this figure. In addition, there are more negative characteristics of English swells mentioned in the article:

But touching pants, we read with interest, and congratulate the swells thereon, that among the most pronounced authorities in London checks are once more the rage. The check is so large that it takes two men to carry it, or two pairs of trousers to show off the pattern. The swell puts on one pair in the morning, and another in the afternoon. He might with advantage pin upon the first, "to be continued in our next." Who pays the cheques we do not presume to say. As the boy is father to the man, perhaps he pays the tailor himself. But then, as every swell knows, it is such a deucedly low thing to pay a tailor at all, perhaps he doesn't. May he not hand in his checks, and leave them a legacy to his creditors. But that is a subject beyond our province. We started in on pants; the more we think of their career and their cheques, the more we are inexpressibly reminded of Adam and original sin.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

The author thus not only ridicules the fashion of large check patterns, but he also criticizes English swells for often not paying their tailors. This habit is also commented on in a later article, in which it is probably not only seen as applying to English swells but also to American dudes. It was published on October 1, 1892[122] , and contains just a short comment on the difference between *pants* and *trousers*:

THE PANTS QUESTION. The difference 'twixt pants and trousers is (I think no one has said it) That pants are always sold for cash And trousers bought on credit. —Indianapolis Journal.

It is not made explicit who buys the trousers on credit and the pants with cash, but as the 1875 article indicates, the characteristic of the swells to live on credit and not pay for the fashionable clothes and other items important for their lifestyle is already circulating in newspaper discourses. Using this background knowledge, the readers can infer the respective social personae. Buying something for cash, on the other hand, implies not only having actual financial resources, in contrast to the swells' possibly just pretended wealth, but also honesty and integrity that is associated with hard work in contrast to the swells' careless attitude and leisurely lifestyle.

An article in which the author expresses an explicitly positive attitude towards the word *pants* was so popular that it was published in several newspapers across the country. The original was published in the *Milwaukee Sentinel* on October 9, 1889[97] , and four reprints can be found in the NCNP alone, in the *Wisconsin State Register* on November 16, 1889[98] , and on November 23, 1889[100] , in the *Atchison Daily Champion* (Kansas) on November 19, 1889[99] , and in the *Bismarck Daily Tribune* (North Dakota) on January 4, 1890[101] . All of these newspapers are (mid)western or northern ones and the first sentence shows the opposition between them and the New York newspaper *The Sun*. The author reacts to the "regular quarterly attack on the word pants" in the New York newspaper by insisting that the "average American" has adopted *pants* and "regards the word trousers as an English affectation". The article thus connects to the discourses on *luggage* and *baggage*, in which the same argument can be found. In this article, however, the author extends this argument by linking the linguistic usage question to the American political system and its underlying ideology: By finding that *pants* is "a good democratic term" because it has been adopted by a majority and is not only used by an elitist minority, the author appeals to the pride Americans take in their democracy. Nevertheless, the author also counters the linguistic arguments

### 4.2 Metadiscourses on lexical forms

put forward by the New York newspapers. First of all, using the etymology of the words to argue that *pantaloons* is the best term to designate the object because it originally referred to items of clothing covering entire legs and feet, whereas *trousers* used to designate only those clothing items covering the hip and the thigh. Secondly, the author advances an argument for the shortening of the word *pantaloons* which is based on an iconic relationship between the word and the object: "As the modern leg coverings are pantaloons cut short, why shouldn't we cut the word short and call it pants?" Overall, this article thus suggests that discourses favoring *trousers* are primarily advanced by New York newspapers and that they are countered by midwestern and northern newspapers by establishing the superiority of the "average American" over the northeastern British-oriented elite, a superiority mainly based on sheer numbers, which is a decisive factor in a democratic society, but also on good linguistic arguments which can be read and interpreted as establishing the position that the 'average' midwestern/northern Americans are no less educated than northeastern Americans.

# **"Pants" It Must Be.**

The **regular quarterly attack on the word pants** appears on time in **The New York Sun. The war is useless.** The American people have adopted it, and protests, ridicule and arguments are all wasted. Whether we like it or not, pants is here to stay. The **average American** regards the word **trousers** as an **English affectation**, and is no more disposed to adopt it than the word waistcoat for vest or topcoat for overcoat. Since the word pants will stick in the face of all opposition, it is sensible to make the best of it. **And there is nothing very bad about it.**. Both The Sun and The Herald declare that pants are not pants, but trousers; but it is also true that trousers are not trousers, but breeches; and that breeches are not anything worn off the stage. Originally trousers were applied to breeches worn by pages—a hip and thigh covering. Pantaloons resemble the leg coverings of today more than trousers or breeches—for pantaloons cover the entire legs and feet. **As the modern leg coverings are pantaloons cut short, why shouldn't we cut the word short and call it pants?** Besides, we have some justification in this in the word pantalet, derived from the word pantaloon. The pantalet, as may be seen in old prints, was a leg-covering for women and children which reached to the shoe-top and resembled the modern made leg-covering more than trousers as originally known. The word trousers comes from the French trousse, a bundle—or a bunch about the hips.

Let us accept **pants** as **a good democratic term**, since there is no way to get rid of it.—Milwaukee Sentinel

[emphasis mine]

The prestige accorded to the word *trousers* is also ridiculed in several short humorous paragraphs. An early example was published on October 17, 1873[26] , in the *Bangor Daily Whig & Courier* (Maine):

A Vermont paper says, "Here we are in our new trousers." It "pants" for more subscribers, probably.

As in some of the articles discussed above, the pun is created based on the homophony of the noun and the conjugated verb *pants*. The implicit criticism

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

conveyed by the statement is that the Vermont newspaper does not use the word *trousers* because they regard it as a better term but solely because they hope to benefit financially by appealing to rich people who prefer *trousers* over *pants*. The use of *trousers* is thus linked to a lack of integrity and authenticity.

Ten years later, on December 11, 1883[62] , a short paragraph consisting of a dialogue between a "youthful Bostonian" and his "Mamma" was published in *The Wisconsin State Register* (Portage, Wisconsin). It ridicules the influence on language use exerted by the social and intellectual elite in Boston: The boy asks his mother whether he may call his trousers *pants* while Mr. Holmes and Mr. Lowell are absent from the city, a question which presents *pants* as the normal and preferred term and *trousers* as the term which is only used because influential men like Oliver Wendell Holmes and James Russell Lowell advocate its use. By employing the figure of a young boy, this contrast between authentic language use and the insincere language use resulting from discourses dominated by an elite is highlighted because it can be expected that children are less susceptible to social pressure – what matters most to them is their mother's permission. On the other hand, it also illustrates how far-reaching the elite's influence is because it causes mothers to forbid their children to use the word *pants*. The relationship between the boy and his mother can also be read as mirroring the relationship between 'average' Bostonians and the Boston social elite, the first group being infantile and looking to the second group for instructions on how to speak and behave. The article can thus be interpreted as an implicit call to the reader to act as an emancipated adult and choose *pants* over *trousers*.

Youthful Bostonian—Mamma, aren't Mr. Holmes and Mr. Lowell both absent from the city? Mamma—I believe they are, dear. Y. B.—Well, can't I call my trousers "pants," just while they're away?—*Life*

As can be expected, the discourses on *pants* and *trousers* also involve articles making reference to the "Anglomaniac" and the dude figure. A good example is the article "As an Englishman", published in 1887, which I have already discussed in §4.1.2. The description of the young American trying to appear as English as possible also includes his avoidance of the word *pants* because he considers it a "vulgar blunder". Furthermore, the article also illustrates the connection between language use and other performable signs by not only describing the "young American Englishman" as using the word *trousers* but also as following the English fashion of wearing trousers turned up. The "Anglicization" process is reflected in his style of wearing his trousers. If he wears them turned up in the rain, there are still traces of Americanness left in him, but if he wears them turned up in the sunshine, he has become a "thorough Englishman".

### 4.2 Metadiscourses on lexical forms

He would almost die for shame should he make such a vulgar blunder as to say "pants." The word he uses is "trousers," "breeches," or "bags." He will tell you confidentially, "I pwefew to say bags; it's awfully English; the best fellows all say it, you know." In this way does the young citizen proceed to Anglicize himself.

# [...]

### With trousers turned up.

I know a young American Englishman who runs to the window every morning on rising to see if he will have an opportunity of turning his trousers legs up. If the day looks fine he comes from the window with a disappointed air and says "Too bad, by Jove. It isn't going to wain afteh all." Once he has become a thorough Englishman, however, he will walk through Broadway the sunniest day in the year with his trousers turned up.

# [...]

The fashion of wearing trousers turned up is also illustrated in cartoons, as shown in the two following examples.<sup>27</sup> In the cartoon "How Cholly Got Left" (see Figure 4.32), Cholly is labeled a "Masher", which, according to the *OED* (2021), is a slang term first attested in 1872 that could not only be used in the United States for a "a fashionable young man" but also for "a womanizer; a man who makes indecent sexual advances towards women, esp. in public places." The reader is thus invited to interpret the dude's rather harmless offer to the young lady as an indecent advance and to laugh about his failure when the lady takes the umbrella but leaves him standing in the rain. This contributes to the picture of the dude as being unattractive and unsuccessful in his attempts to impress women. His turned-up pants are clearly visible because the patterned outside is contrasted with the plain inside material. The name Cholly, his behavior, his pants and the eyeglass thus clearly signal his belonging to the group of dudes.

The second example depicts Cholly wearing his pants turned up even in the sunshine (see Figure 4.33). In this case, attention is drawn to the bottom part of his legs by coloring his socks black, while his shoes and pants remain white. His fancy outer appearance is generally ridiculed in the cartoon: The caption contains a dialogue between him and "Miss Soldier Girl" in which they discuss his uselessness as a soldier, which culminates in the girl's suggestion that he could function as a wigwag signal – a suggestion which alludes to his conspicuous clothing that is likely to be noticed as well as to his incapability of carrying out any task that is more complicated than going back and forth repeatedly and automatically.

The link between the dude's use of the lexical item *trousers*, in combination with the labiodental realization of /r/ (indicated by *twousers*), and his huge concern for fashion is also established in the humorous dialogue between Percy Pad-

<sup>27</sup>The cartoons discussed in this section have been obtained by searching the NCNP for articles containing the name Cholly and illustrations.

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Figure 4.32: A cartoon illustrating the dude wearing his trousers turned up in the rain, published in the *Atchison Daily Champion* (Atchison, Kansas) on December 28, 1890[105] , retrieved from Nineteenth-Century U.S. Newspapers

4.2 Metadiscourses on lexical forms

Figure 4.33: A cartoon illustrating the dude wearing his trousers turned up in the sunshine, published in the *Milwaukee Journal* (Milwaukee, Wisconsin) on September 22, 1898[170] , retrieved from Nineteenth-Century U.S. Newspapers

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

dleford and Daniel Maginnis (published on April 20, 1895[151] ), which I have already discussed in §4.1.2 because it is part of the collection of articles containing the search term twousers. They talk about Hawold Montmowenci who is, according to Percy, a martyr "of the modern type" because he gave up drinking and eating (and smoking) in order to be able to keep up with changes in fashion and who ultimately died of starvation. One of these changes in fashion related to trousers, as the following extract shows:

"Hewoic cweature!" said the pwopper young man. "But his twials and twibulations were not yet ovah. **Just then the fashion changed fwom tight twousers to loose twousers**."

"Put him in a hole again?" asked the horse doctor?

"**Yaas. But he was made of the twue stuff. He had a gweat and hewoic soul, and he gave up his foah cwackers a day and bought him a pair of loose twousers.**"

[emphasis mine]

This shows that the dude not only had to take the bottom part of his pants and the right pattern into consideration, but also the fit of the pants. That his concern for fashion poses a threat to his life is of course a great exaggeration, but it can also be found in other articles, as in the following cartoon, published in the *St. Louis Globe-Democrat* on February 5, 1887[77] , and taken from the magazine *Life* (see Figure 4.34). The cartoonist expresses a warning that the dude "could be 'carried away' by the fashion" by depicting how his cane gradually becomes so big that it literally carries him away. The heading "Verbum Sap" is a phrase which is used, according to the *OED* (2021), "in place of making or concluding a full statement or explanation of something". The drawings are thus presented as providing a clear conclusion to the matter of the dude's over-emphasis on fashion, which emphasizes its negative consequences. Even though 'fashion' comprises mainly items of clothing and accessories here, linguistic forms like *trousers* are connected to the fashion as well, as the analyses above have shown.

The overall conclusion that can be drawn from the analysis of these articles is thus that there are discourses favoring the word *trousers*, which are particularly widespread in northeastern cities, in particular Boston and New York, and highly influenced by the social and political elite. The word is connected to educatedness and indexes elegance as well as the belonging to higher social classes. The prestige of *trousers* is also indirectly visible through its adoption by the dude, who regards it as fashionable and hopes to impress others by using it. On the other hand, there are discourses ridiculing the high prestige attached to *trousers* in the northeast – the dude figure is at the heart of this ridicule, and it is in this connection that the Englishness and thus the un-Americanness of *trousers* is emphasized. But other articles also focus on the lack of authenticity of the use of

### 4.2 Metadiscourses on lexical forms

Figure 4.34: A cartoon depicting the dude being carried away by his own handle, published in the *St. Louis Globe-Democrat* (St. Louis, Missouri) on February 5, 1887[77] , retrieved from Nineteenth-Century U.S. Newspapers

*trousers* instead of *pants* by comparing people using *trousers* to children who are told which form to use by their parents, for example. *Pants* is presented against *trousers* as a genuine and democratic form because it used by the majority of average and authentic Americans, and even though the shortened form is still considered as 'slang' by some people who prefer the long form *pantaloons*, it is also motivated in one of the articles analyzed above as being in an iconic relationship with the object, which has also been shortened.

The qualitative analysis thus reveals the existence of competing discourses which are indicative of a negotiation process on the basis of the social meaning of *trousers* and *pants*. Given this competition, it is particularly interesting to focus on the text type of advertisements because their main aim is to appeal to the

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

readers and convince them to buy the company's product. The choice of the linguistic item is thus of particularly high importance for the company and it reveals information about how the creators of the advertisements assess the prominence of the different indexical values in discourse and how they use linguistic means to position their company and their product socially.

The first example of an advertisement was published in the *North American* at the very end of the century, on October 5, 1899[176] (see Figure 4.35). The *North American* was published in Philadelphia, Pennsylvania, which makes it likely that the company, Miller's, is also based in Philadelphia. The advertisement is interesting because it not only prefers the term *trousers*, which appears in big capital letters at the top of the advertisement, but it also explicitly distances itself from the word *pants*, as can be seen in the line below trousers, which contains the line "Not Pants, Mind You" written in small capital letters. The advertisement thus makes use of the value of 'elegance' being indexically linked to the word *trousers* – by stating that they sell trousers and not pants they implicitly call attention to the high quality of their product and the social status that potential buyers can convey by wearing the product. This shows that the discourses favoring *trousers* are not restricted to New York and Boston but extend to Philadelphia as well and that the indexical meanings of the lexical items are so well known by the end of the century that they do not require explicit mention in the advertisement.

The second and third example are advertisements published in the *St. Louis Globe-Democrat* in Missouri by the company F.W. Humphrey & Co. In contrast

Figure 4.35: Advertisement making use of positive indexical values linked to *trousers*, published in the *North American* (Philadelphia, Pennsylvania) on October 5, 1899[176] , retrieved from Nineteenth-Century U.S. Newspapers

### 4.2 Metadiscourses on lexical forms

to the example above, they avoid choosing a particular term, but opt for making the choice of a name for the product the subject of explicit discussion in the advertisements. The title of the advertisement published on January 28, 1887[76] , is thus "What's in a Name?", followed by a subtitle in smaller print "That which we call a rose by any other name would smell as sweet" (see Figure 4.36). This reference to Shakespeare's play *Romeo and Juliet* not only serves to convey the position of the company in the discourses on language, namely that it is not the name of the product that matters but the product itself, but it also serves to support this position – especially against the members of the intellectual (Britishoriented) elite who are expected to prefer the term *trousers* but who are also likely to accept an argument based on Shakespeare. In the text following the title and the subtitle, an anecdote is told about an "uncultured Chicagoan" who was so

Figure 4.36: An advertisement containing an explicit discussion about the use of the terms *pants* and *trousers*, published in the *St. Louis Globe-Democrat* (St. Louis, Missouri) on January 28, 1887[76] , retrieved from Nineteenth-Century U.S. Newspapers

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

"narrow-minded in the nomenclature of 'leg-dressings'" that he would not buy pants in a store where they call them *trousers*. The exclusive preference for the term *pants* is thus linked to a lack of culture and to the city of Chicago and presented negatively in contrast to the city of St. Louis and the company's view of allowing "the individual to exercise his own sweet will and call them pants, trousers or breeches". Individuality and freedom are thus presented as the most important values, which are superior to questions of culture, elegance, or social position.

The other advertisement by the same company was published on January 12, 1887[75] , and even though they chose the term *trousers* here as a title, they emphasize in the text that the choice of terms is "a matter of individual fancy" and not important (see Figure 4.37). The mention of the possible label of *dude protectors*

Figure 4.37: An advertisement using *trousers* in the title but giving several terms as alternatives in the text, published in the *St. Louis Globe-Democrat* (St. Louis, Missouri) on January 12, 1887[75] , retrieved from Nineteenth-Century U.S. Newspapers

### 4.2 Metadiscourses on lexical forms

in addition to the variants *pants*, *pantaloons*, *trousers* and *leg coverings* shows that the creators of the advertisement were aware of the negative indexical meanings linked to *trousers*, but that they chose to position themselves above such evaluations. Nevertheless, their preference for *trousers*, which also appears in big letters in the lower part of the advertisement ("All-Wool Trousers!") suggests that even a company in the midwest does not opt for *pants* (nor for *pantaloons*) in choosing a title for their advertisement, which implies that the indexical link between elegance, high quality and the form *trousers* must be regarded as quite strong. However, the explicit discussion of their position in the debate on the choice of lexical item shows that they have to defend using *trousers*, which also indicates that they expect several people in the midwest to have a preference for *pants*.

A different strategy of social positioning in discourse is adopted by the company Polack's Clothing House in the following advertisement (see Figure 4.38), published in the same newspaper as the two examples above (but five years earlier, on May 16, 1882[50] ). They use the title "Pantaloons!", which suggests that they expect many people in St. Louis to find this term appealing. In addition, both *pants* and *trousers* are used in the text to designate the products sold without any indication of different linguistic or social meanings. This shows that the strategy employed here is similar to that used in Humphrey's advertisement, namely the decision to avoid choosing just one term, but in contrast to Humphrey's, Polack's Clothing House chooses *pantaloons* as a title and avoids any implicit or explicit references to discourses on language.

Figure 4.38: An advertisement using *pantaloons*, *pants* and *trousers*, published in the *St. Louis Globe-Democrat* (St. Louis, Missouri) on May 16, 1882[50] , retrieved from Nineteenth-Century U.S. Newspapers

A similar avoidance of choosing one term over the other can also be observed in the following advertisement (see Figure 4.39), published by Kohn, the Clothier and Hatter in the *Morning Oregonian* on May 2, 1893[126] . Even though they use the term *trouser* in the title "Special Trouser Sale!" and in the first line of the text ("We will give you the choice of any pair of trousers in the house [...]"), they then use *pants* when listing the prices ("\$9 Pants, \$8 Pants, \$7 Pants"). In the text at the 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

Figure 4.39: An advertisement using *trousers* and *pants*, published in the *Morning Oregonian* (Portland, Oregon) on May 2, 1893[126] , retrieved from Nineteenth-Century U.S. Newspapers

bottom, *trousers* is used again so that overall, *trousers* is the dominating form, but the fact that *pants* is used as well suggests that the company must have decided that it does not want to repel potential customers who prefer *pants* to *trousers*.

The last example of an advertisement (see Figure 4.40) shows that pants can also appear in a title, which is even more notable as the advertisement appeared in the *North American*, so in the same paper that published the advertisement by Miller's with its exclusive preference for *trousers*. However, *pants* is not the only term in the title, but all three variants are used: "Pants, Pantaloons, or Trousers". The first sentence of the text can be interpreted in two ways: "Take your choice" can either be understood as taking the choice between the different products sold by the company as well as taking the choice between the different terms in the title. That the company labels their products "Thompson's Patent Cut Trousers" implies a preference for *trousers*, but this preference is not foregrounded in the advertisement.

4.2 Metadiscourses on lexical forms

Figure 4.40: An advertisement using the word *pants* in the title (in addition to *pantaloons* and *trousers*), published in the *North American* (Philadelphia, Pennsylvania) on May 26, 1891[109] , retrieved from Nineteenth-Century U.S. Newspapers

### 4 Results: metadiscursive activity in nineteenth-century U.S. newspapers

To conclude, the advertisements in particular show that none of the variants (*pantaloons*, *pants*, *trousers*) was evaluated so negatively that it was generally avoided. While the analysis of *luggage* and *baggage* suggested that *baggage* had become the form that was overwhelmingly regarded as positive by the end of the nineteenth century because it was linked to American nationality and authenticity, the analysis of *trousers* and *pants* revealed that while *pants* was associated with similar positive values, the links between *trousers* and positive values like elegance, high quality and a high social position were also still strong. The process of negotiating indexical values linked to the terms was thus still ongoing at the end of the nineteenth century.

# **5 Interpretation: key values and phases in the enregisterment of American English**

# **5.1 Indexical values and the enregisterment of American English**

The overarching goal of the present study is to identify enregisterment processes of American English by systematically investigating and describing the cultural construction of American English as a discursive variety. The analysis carried out in §4 focused on newspaper articles and a specific set of five phonological and two lexical variables to answer several research questions relating to the creation of indexical links between these linguistic forms and social values and characterological figures. So far, the focus has thus been on the forms and their indexical values, which can be conceptualized as several indexical fields (in the sense of Eckert 2008, see §2.2.2). The next logical step is to interpret these findings with regard to the role played by these indexical values in the construction of a register whose forms have 'American' as their main indexical value and which can thus be labeled "American English". In this chapter, I argue that this construction process revolved around three central values, the nationality value, the authenticity value and the non-specificity value, and that those values affected the process in different phases and in connection with different linguistic forms.

# **5.2 The nationality value: delimiting American English against British English**

As the central aim of the present study is to answer the question of when and how an American register was constructed or continued to be constructed in the nineteenth century, it is obvious that nationality must be a central value in the process because the adjective *American* already indicates a relation to the American nation. This relation is also one of the reasons for focusing on the nineteenth

### 5 Interpretation

century because at this time the United States of America had been formed as a political union – a nation that was politically independent from other nations. However, the question that remained open and that guided the analysis is if there were linguistic forms which came to index the quality of being 'American', and if so, how these indexical links were constructed. The analysis of metadiscursive activity in newspaper articles has revealed that the construction of American linguistic forms revolved centrally around their differentiation from British English forms – and that one crucial value in this process was accordingly the nationality value. The linguistic form that was most important in these metadiscourses was /h/-dropping and -insertion. While the importance of delimitation from British English has of course always played a central part in prior studies on the historical development of American English (see §2.4), the role played by /h/-dropping and -insertion in this process has not been identified so far. A reason for this is that /h/-dropping and -insertion is not a form that was in use in the United States. Research on its distribution and evaluation in metadiscursive activity has focused on England because it was in this geographical area that not only variability was found on the level of structure but where the form also played an important role in delimiting a repertoire of 'standard' forms from a repertoire of 'non-standard' forms and thus in the discursive construction of a standard British English register. However, the analysis conducted in this study shows that Americans drew precisely on these metadiscourses in England to construct their American register. Two central values in these American metadiscourses were uniformity and superiority – the complete absence of the form in America was interpreted as a marker of American uniformity, and this uniformity in turn was part of the argumentation that established linguistic superiority and connected this to a general cultural superiority.

To shed more light on the construction of American English as uniform based on the absence of /h/-dropping and -insertion, it is important to note that first of all uniformity has been a central value in standardization processes and the construction of national identities throughout Europe. In the case of French, for example, Lodge (2013: 6) emphasizes:

the central role which language has played over the past two centuries in the definition of French national identity – the standardized variety of French is more than an efficient vehicle for communication across the vast length and breadth of France; it serves as a powerful symbol fostering among French people a sense of national solidarity (internal cohesion) and a feeling of their uniqueness in comparison with other nations (external distinction).

### 5.2 The nationality value

Consequently, since the French Revolution, non-standard dialects have been perceived as a threat to uniformity and this even resulted in political action and the persecution of these dialects "with great ruthlessness" (Lodge 2013: 6). This example shows that an ideology emphasizing inner linguistic homogeneity was important for emerging nation states to justify and legitimize their status as a political unit. It is within this ideological framework that /h/-dropping and -insertion became an ideal form for emphasizing the linguistic unity of the United States, especially in contrast to England, which was constructed as linguistically heterogeneous because of the variable use of this form. Secondly, the value of uniformity also tied in with the specific American ideology of being a nation marked by democracy and equality. In England, discourses on /h/-dropping and -insertion served to emphasize social differences between speakers: It became an index of a lower social class and a lack of education and cultivation (see §3.3.2). Americans used this link to construct the absence of the form as a marker of their social uniformity, in the sense that it marked their lack of emphasis on class differences and the greater level of education of *all* Americans. We can thus observe the identity relations that Bucholtz & Hall (2005) have called adequation and distinction (see §2.2.2): The linguistic and social differences between different groups within the United States were downplayed by emphasizing their similarity (indicated by /h/-retention), while their social and linguistic dissimilarity to Britain was highlighted. Despite this effort on constructing distinctiveness, discourses surrounding /h/-dropping and -insertion also show that Americans still exhibited an exonormative orientation: Their patterns of argumentation rested on the negative indexical values that the form had previously acquired in England. They could not use the form to construct an American English standard against a British English standard because in England the form was also evaluated as non-standard, but this is precisely the reason for why the analysis of enregisterment has to go beyond the distinction between non-standard and standard and look more closely at how standardness is constructed and how this interacts with the specific linguistic forms that are foregrounded. Americans used /h/-dropping and -insertion to argue that their standard is superior because the 'correct' retention of /h/ was not a form used by only a small group of upper-class speakers but by all speakers who were deemed relevant for the imagination of the American nation. This social uniformity, connected to a strong belief in democracy and equality, made the American standard appear better than the English one, despite the fact that /h/-retention was in fact part of both standard registers. This reveals a clear picture of how the shift from an exonormative to an endonormative orientation proceeded in discourses on language: The linguistic norms constructed in England were taken as a basis for establishing American

### 5 Interpretation

linguistic superiority and thus to create the self-confidence necessary for the subsequent establishment of norms that were in contrast to British norms.

The analysis of articles containing *hinglish* in §4.1.2 has shown in detail how this linguistic self-confidence was created and transmitted. Characterological figures surfacing in anecdotes and humorous paragraphs played an important role because they embodied the struggle over superiority and because they contributed to the construction of a register of British English that stood in contrast to an American one, thus emphasizing not only the linguistic but also the cultural differences between the two nations. The Englishmen appearing in these anecdotes were usually characterized as arrogant and as feeling superior to Americans, but their dropping and insertion of /h/ was used to depict them as being "in fact" vulgar and ignorant. The enregisterment of British English in America thus proceeded by linking /h/-dropping and -insertion to British people. The indexical meanings 'vulgar' and 'incorrect' had already been present in metadiscourses in England, but they were re-interpreted in America as applying to the majority of British people, thus differentiating between two British registers: on the one hand "standard British English", which was linked to a small number of upperclass British speakers, and on the other hand a general "British English", which was linked to the rest of the people living in Britain. This general "British English" was constructed as inferior to "American English", and the anecdotes describing encounters between arrogant Cockney speakers complaining about American speech and ways of life and competent and relaxed American speakers who put them in their place created representative incidents and invited Americans to identify with the American figures. Interestingly, in two of the anecdotes discussed in §4.1.2 the Americans were judges, so figures who have the power to decide on what is right or wrong, which gave the incorrectness of /h/-dropping -insertion even more weight. The construction of the vulgar British speaker in contrast to the educated and sensible American speaker has striking parallels to the enregisterment of American English in England in the nineteenth century, as shown by Hodson's (2017b) analysis of British fiction from 1800 to 1836. She finds that the stereotype of the "vulgar American" was already present in the earliest novel in her study, Susan Ferrier's *The Inheritance* (1824), but that this stereotype was not linked to specific linguistic forms yet. However, in an 1853 revision of the same novel, Hodson identifies several changes that Ferrier made with regard to the direct speech of the American figure, Lewinston. She interprets this as evidence that by the middle of the nineteenth century a repertoire of linguistic forms associated with American speech had emerged. As Ferrier did not make any changes to the characterization nor added metalinguistic comments, Hodson concludes that she probably expected her readers to recognize these forms

### 5.2 The nationality value

as American, which suggests the existence of an American register. The vulgarity of the American figure remained in place. Like the English figures in American newspaper articles, the American figure in Ferrier's novel is not only depicted as vulgar and lower class, but also as arrogant and blunt, with "a firm belief in the superiority of American customs" (Hodson 2017b: 38). Hodson's findings illustrate the importance of interpreting enregisterment processes with respect to the sociohistorical population that is involved in these processes: In England, American English was enregistered differently than in America. They also show that the mechanisms are similar: By depicting members of the other nation as vulgar and ridiculing their arrogant belief in their superiority, the cultural and linguistic superiority of one's own nation is constructed. It also suggests that the metadiscourses are linked: It is likely that Americans were aware of how their speech and behavior were evaluated in England and that they reacted to it by applying the same argumentation but twisting it in their favor. This view can be supported by the anecdote of the American traveler in England who is complimented on his good English. The anecdote was quite popular in America and it shows how knowledge of English perceptions of American speech came to circulate in the newly formed American nation.

The quantitative analysis of the frequency of articles containing *hinglish* in §4.1.1 showed that they appeared relatively early and they kept appearing at a relatively low but stable frequency throughout the nineteenth century, which suggests that the nationality value and the delimitation of American English as different from and, more importantly, superior to British English did not lose its importance. The presence of /h/s "in their proper places" continued to be important for the enregisterment of American English; with regard to region, it seems to have been more important in the north and northeast than in other parts of the country. In general, /h/-dropping and -insertion was a highly salient form and in several cases it was the only one that marked the difference between British and American speakers. Even though upper-class speakers were not associated with the use of the form, the analysis has shown that the negative indexical meanings linked to the form had become so well-known that they could be used to characterize even this speaker group negatively. So while it can be argued that a general British English register was contrasted with a standard British register, the former being inferior to American English and the latter being equal and pointing to shared norms, /h/-dropping and -insertion was also used to shed a negative light on British people in general and to assert American superiority.

With regard to lexical forms, the nationality value also played a role – more so in the case of *baggage* and *luggage* than in the case of *pants* and *trousers*. The analysis in §4.2 has shown that if indexical links were created, *baggage* was favored

### 5 Interpretation

over *luggage* because it was considered the American form. Therefore, *baggage* became enregistered as American English in contrast to *luggage* becoming enregistered as British English. In this case no differentiation between a standard and a general British English was made, which provides clear evidence of an endonormative orientation: The American form was valued highly even though it contrasted with a form that was considered standard in Britain and that was not evaluated as vulgar and incorrect like /h/-dropping and -insertion was. It needs to be noted that in contrast to /h/-dropping and -insertion, these metadiscourses occurred primarily in the last two decades of the nineteenth century.

However, the situation was more complex for *pants* and *trousers*. Although indexical links between *trousers* and British speakers could be identified, *trousers* is also constructed as the 'proper' American form in several articles. These are signs of conflict with regard to which form should be favored. Although the nationality value features importantly in some articles arguing for *pants* being the democratic American term since it was used by the majority of speakers in the American nation, it is not at all present in others, where *trousers* is associated with correctness, elegance and cultivation and not with a British nationality. I argue that *pants* and *trousers* present a case where another value comes into play, adding to and interacting with the nationality value: the authenticity value.

# **5.3 The authenticity value: delimiting an authentic American English against an inauthentic American English**

Authenticity was a crucial value in the enregisterment of American English because it helped speakers to make sense of linguistic differences that occurred *within* America, that is, differences between *American* speakers, which could thus not be interpreted as linking speakers to different nationalities (British vs. American). As pointed out in §2.2.2, in recent research in sociolinguistics and linguistic anthropology, authenticity is not viewed as a quality that is inherent in language itself, but as a claim that speakers make – a claim that sets a process of authentication in motion and which is often accompanied by a process of denaturalization. In order to construct speech forms or other practices as authentic, differing speech forms or practices are constructed as unnatural and not genuine. I argue here that authenticity played a central role in enregistering a set of linguistic forms, among them a back bath vowel, non-rhoticity, a labiodental realization of /r/ and the use of the lexical item *trousers*, as *non*-American – not because the users of the forms were not Americans, but because they were constructed as not

### 5.3 The authenticity value

being *authentic* Americans. In the same process, the alternative variants were thus enregistered as genuine American English forms, used by 'true' and 'real' American speakers. The analysis shows that characterological figures played a key role in this process. The negotiation of what is authentically American revolved around the opposition between the American dude on the one hand and the American cowboy, hunter and farmer on the other hand. These figures were essential in linking authenticity (or lack thereof) in language with authenticity in other practices, a connection which is also regarded by Johnstone (2014: 109) as being always present in authentication processes. The linguistic forms used by the dude were constructed as a register which was indirectly also linked to the evolving British standard register. The forms used by the dude were evaluated as being 'fashionable' in England among educated upper-class speakers, which constitutes evidence of an enregisterment of these forms as "standard British English" in England and a recognition of this process in America. Some articles describe how the indexical values that the forms had acquired in England, i.e. signaling education, cultivation and membership in upper classes, also became relevant in America. These observations tie in with linguistic theories which assume that non-rhoticity was adopted in eastern parts of the United States during the nineteenth century because Americans imitated British speech. The newspaper articles show that this view had already been put forward and transmitted by contemporary observers. They also confirm Schneider's (2007: 288–289) point that even during the endonormative phase exonormative tendencies were still present. However, the dude figure is used to exploit precisely this link to British English speech in order to construct the use of these forms as inauthentic. By parodying the way that men in eastern urban centers of the United States imitated not only fashionable British English forms of speech but also other fashionable practices, like dressing according to English fashion (especially by wearing specific styles of trousers and accessories like the eyeglass and the cane), attention was drawn to the unnatural and affected nature of this imitation – a denaturalizing process that aimed at denying speakers who use this form their 'true' American identity.

The dude thus constitutes an embodied metapragmatic stereotype of a speaker whose orientation towards British English norms resulted in a lack of authenticity as an American. The contrasting characterological figures embodied precisely this authenticity as Americans that the dude lacked. Even though the cowboy, the hunter and the farmer were portrayed as uncivilized and uneducated, they also possessed a great amount of practical knowledge based on experience, they had adopted a pragmatic stance towards life, and their strength, toughness and self-reliance made them independent of external (British) influences and fash-

### 5 Interpretation

ions. The speech of these 'genuine' Americans was marked by several linguistic forms which established a contrast to the speech of the narrators and the linguistic forms present in other newspaper articles, for example negation with *ain't*, alveolar -*ing*, a lower kit vowel in *if* and hyper-rhoticity. Even though not directly visible in the spelling, the absence of representations of non-rhoticity, back bath vowels, and labiodental realizations of /r/ signaled to the reader that their speech was rhotic, that their bath vowel was front and that their /r/ had an alveolar or retroflex realization. In terms of enregisterment, this means that these figures contributed to the enregisterment of a rural northern and (mid)western American English which was constructed through negative values like lack of civilization and education, but also through positive values like self-reliance, pragmatism, strength and, most importantly, authenticity.

In §2.2.2, I pointed out that studies on enregisterment have often focused on the enregisterment of regional or social accents and that authenticity was an important value in this process. In Pittsburgh, for example, the revalorization of linguistic forms as positive indexes of place was linked to people expressing their pride in being authentic Pittsburghers. The present study has shown that social and regional aspects connected to authenticity in different ways in nineteenthcentury enregisterment processes of American English. The dude was a figure that was linked to northeastern cities, in particular New York City, while the figures of the American cowboy, hunter and farmer were associated with rural areas in the north and the west. The analysis of newspaper articles suggests that metadiscourses which were circulating mostly in the northeast (but also spread to urban centers in the west) and which evaluated non-rhoticity, back bath and labiodental /r/ and *trousers* as signs of cultivation, education, elegance and a belonging to higher social circles were countered by metadiscourses that relied chiefly on authenticity and indirectly also on the nationality value in order to link the alternative variants to a real and genuine American identity.

For phonological forms, the discourses which constructed a positive evaluation of the variants analyzed here can be accessed only indirectly because these variants were criticized and ridiculed in the articles, but for the lexical form *trousers* these competing metadiscourses could be traced in more detail. For example, in the advertisement by the company Humphrey's, published in the *St. Louis Globe-Democrat* January 28, 1887[76] , the "uncultured Chicagoan" was described as rejecting the use of *trousers*, which provides evidence of a link between the form *pants* and midwestern cities. However, the adjective *uncultured* also reveals the existence of a link between *pants* and a lack of cultivation. Conceptually, the midwestern cities were midway on the continuum between urban northeastern centers and rural midwestern areas; they were geographically located in the

### 5.3 The authenticity value

midwest, but they also constituted urban areas. The advertisement solved the dilemma of choosing a lexical variant by leaving the choice up to the individual. The metadiscourses analyzed in this study reveal the cultural models available for motivating such a choice: one emphasizing the value of cultivation and the other emphasizing the value of authenticity. That the latter was endonormative by establishing a model of authentic American linguistic behavior probably contributed to its later success. Articles constructing the American forms as authentic often pointed out the exonormative orientation of the competing model and the social pressure that contributed to its success in the northeast, for example by contrasting the young Bostonian's wish to use *pants* (because this is the form he would normally use) with the elite's prescription to use *trousers* (which motivates his mother to forbid the boy to use *pants*). This indicates that Bucholtz & Hall's (2005) processes of authorization and illegitimization were at play as well – linguistic authority exerted by a social elite was challenged.

The regional distribution of articles containing *dawnce*, *deah* AND *fellah*, *bettah* and twousers described in §4.1.1 has shown that metadiscourses surrounding these forms circulated most prominently in the (mid)west and the north. It was thus particularly in these regions that the stereotypes of northeastern men, embodied in the dude figure, circulated: men who are unintelligent, lazy, living off inherited wealth and spending their time buying and showing off fashionable clothes and accessories and trying to impress women. However, several articles, especially those aiming at entertaining the readers, were taken from magazines, which were often published in the northeast and had a nationwide readership, which suggests that such discourses were also present the northeast. A further indicator of this presence is the ambivalence that Bonfiglio (2002: 48) identifies in Grandgent's comments on non-rhoticity in 1899 (see §2.4). Grandgent associated non-rhoticity with cultivation but also with decadence and lack of vitality, so the positively evaluated energy and strength of the rhotic hard-working American farmer or cowboy was also found in expert discourses on language, which were still dominated by the northeastern elite.

The frequency at which articles containing *dawnce*, *deah* AND *fellah*, *bettah* and twousers appeared in newspapers in the course of the nineteenth century clearly reveals that these metadiscourses rose to prominence in the 1880s. Until the 1870s only very few articles containing these features could be found, and none were published before 1840. So claims to authenticity, which were in opposition to constructions of a standard American register based on British 'fashionable' forms, were developed particularly towards the end of the nineteenth century, and, as Grandgent's comments in the last year of the nineteenth century suggest, continued well into the twentieth century. This confirms and adds

### 5 Interpretation

to Schneider's (2007: 288) observation that the purist movement started after the Civil War and rose to prominence in the 1870s and 1880s. The present study shows, however, that these purist discourses favoring British variants, which were in the process of becoming constructed as 'standard', 'cultivated' and 'upper class' in England, were countered by constructing the alternative variants as being part of an *authentic* American register.

Finally, it is important to note that social personae and characterological figures which played a prominent role in the metadiscourses described in Chapter 4 were mostly male. However, the few cases where female figures were used are particularly interesting. The anecdote about salesladies working in the department store in Denver, Colorado, not only shows that metadiscourses indexically linking non-rhoticity, back bath and labiodental /r/ to upper-class membership, education and cultivation had reached western urban centers, but it also indicates that the effect of these metadiscourses were criticized and ridiculed and that this criticism was not directed at the members of the upper classes themselves, but primarily at the members of the lower middle classes who attempted to imitate upper-class speech to advance socially. The salesladies lacked authenticity and were thus in stark contrast to the proud, pragmatic, intelligent and down-to-earth Montana girl, who was presented as superior to the dim-witted and affected New York dude. The encounter between the Montana girl and the New York dude is also a good example to illustrate the role attributed to masculinity and femininity in these discourses: Particularly in the case of non-rhoticity, the weakening or loss of /r/ was linked to femininity and physical weakness, so that even male figures like the dude were portrayed as effeminate, whereas the rhoticity of the cowboy was associated with masculine characteristics and gave figures like the Montana girl a masculine touch. This impression of masculinity was often increased by adding hyper-rhoticity to the register of the (mid)western and northern authentic American. The enregisterment of hyper-rhoticity in combination with other forms like for example *ain't* also indicates, however, that the cowboy and his speech were not constructed as models for *all* Americans despite their authentication as genuine Americans. This reveals a third value that came into play and was indispensable for creating a register that came to be recognized by all as 'American': non-specificity.

5.4 The non-specificity value

# **5.4 The non-specificity value: delimiting American English against more specific regional and social American Englishes**

What I regard as the non-specificity value can be illustrated well by using the following quote by Krapp (1919: ix), who explains his understanding of the term *standard speech* at the beginning of the twentieth century:

The term standard speech, it will thus be seen, has been used by the author without a very exact definition. Everybody knows that there is no type of speech uniform and accepted in practice by all persons in America. What the author has called standard may perhaps be best defined negatively, as **the speech which is least likely to attract attention to itself as being peculiar to any class or locality**. As a matter of fact, speech does not often attract notice to itself unless it is markedly peculiar. For the most part, when one is listening to the speech of others, one is intent upon getting the meaning, not upon observing the form. In consequence there is likely to be, even in what we may justly call standard speech, a considerable area of negligible variation, negligible, that is, from the point of view of the practical use of language. To the conscientious and critical listener, many of these variations may seem reprehensible, but only so by the test of some theoretical or ideal standard. [emphasis mine]

First of all, Krapp distinguishes between a standard "in practice" and a "theoretical or ideal standard" in order to justify the presence of variation in speech that would be evaluated as 'standard' by many Americans despite the experts' demand for uniformity. Secondly, he finds that those linguistic forms are evaluated as 'standard' which are not "peculiar to any class or locality", so in other words, linguistic forms which are non-specific. I argue that Krapp's distinction between an "ideal" and a "real" standard can be captured within the theoretical framework of enregisterment by separating expert metadiscourses from non-expert metadiscourses. While the uniformity value plays a great role for experts who write dictionaries, grammars and other prescriptive texts aiming at codifying a standard variety comprising a large number of forms, I have shown above that in newspaper discourses the uniformity value mainly played a role in relation to *one* form: /h/-dropping and -insertion. Non-specificity, by contrast, was an important value in newspaper discourses with regard to almost all linguistic forms analyzed in this study. In order to answer the question of which forms became

### 5 Interpretation

enregistered as 'American' for *all* Americans, one must address the question of which forms were constructed as belonging to 'specific' American registers and thus as being too specific to become part of a cultural model that all Americans could potentially orientate towards.

Non-rhoticity constitutes the prime example to illustrate the role of the nonspecificity value. The absence of post-vocalic /r/ became part of several specific registers: not only of the register of the urban northeastern higher social classes, embodied most prominently by the dude figure, which is why I will refer to it as the "dude register" in the following discussion, but also of a "southern American register", of "a mountaineer register" and of a "Black American register". The analysis of the frequency of indexical links between *bettah* and social personae has revealed that the association between non-rhoticity and Black speech dominated in quantitative terms, which indicates that people frequently came into contact with this link. However, as non-rhoticity occurred in combination with a large number of other linguistic forms to index Black speakers, not only phonological, but also grammatical, lexical and pragmatic ones, it was less salient as a shibboleth of a Black American English register than as a shibboleth of a dude register, which consisted of a very restricted set of forms, causing nonrhoticity to stand out more. The southern American English register and the mountaineer register occupied the middle ground with regard to the salience of non-rhoticity: They also comprised a larger number of forms, including grammatical ones, which were completely absent in the dude register, but usually not as many as the Black American register. This has important implications for the non-specificity value: The Black American register is most specific in the sense that it is constructed as comprising the largest number of forms which are specifically (though in many cases not exclusively) linked to it in newspaper articles, whereas the dude register is constructed as least specific because it contains only a small number of forms specific to the register. This in turn had the consequence that the Black American register was constructed as most deviant and least 'normal', whereas the dude register, despite its few salient differences, appeared least deviant. In the case of phonological forms, specificity is marked through spelling, whereas in the case of grammatical forms, specificity is marked through the contrast to other speakers and especially to the majority of other newspaper articles. I argue that the register that was constructed as generally American and thus non-specific consisted of forms that were *not* constructed as shibboleths of specific social or regional registers. Non-rhoticity was part of several specific registers, with different degrees of salience. In contrast to that, hyper-rhoticity became enregistered as part of the rural (mid)western register, embodied for example by the cowboy. Against this, rhoticity was constructed as non-specific –

### 5.4 The non-specificity value

as the form used by authentic Americans who are in no way "peculiar", to use Krapp's wording again.

Yod-dropping presents an interesting case because even though the yod-less pronunciation is marked by a specific spelling, thus creating the basis for linking this specific pronunciation to a specific group or locale, the analysis revealed that the form has come to index rather general social characteristics like uneducatedness and lack of cultivation. Even though it was described in articles containing explicit metadiscursive comments as a form occurring in the north and west (and not in the south), it also appeared in representations of direct speech of southerners throughout the nineteenth century, which underlines the non-specificity of the form with regard to region. Nevertheless, in those articles where it was presented as a "fault" of the north and west, one which was not found in the south, this again constructs the south as a specific region that is indexically linked to linguistic forms, in this case to yod-*retention*. In fact, it could be argued that the continuing enregisterment of yod-retention as southern, which is visible in some articles, makes yod-retention too specific to become part of a non-specific American English register. I analyzed one article which provides evidence that yod-retention was increasingly evaluated negatively by linking it to pedantry towards the end of the nineteenth century, which indicates that the indexical links to negative social characteristics weakened and thus the non-specificity of yoddropping increased. In general, the occurrence of articles containing *noospaper/s* throughout the nineteenth century, without great changes in frequency except for the 1860s, when they appeared frequently in the Nasby letters, shows that yod-dropping had been part of metadiscursive activity in newspaper articles for a long time, but it was not very salient and it was not involved in any claims to nationality or authenticity. However, based on its (increasing) lack of specificity, it became enregistered as part of a non-specific general American English register, while yod-retention was enregistered mainly as (specifically) southern.

The importance of specific registers for the discursive construction of a standard register as the register of an imagined national public is also highlighted by Frekko (2009) who shows convincingly that in the case of Catalonia today, the national public is imagined as fragile precisely because such specific registers do not exist. She finds that an imagined uniformity is as important as an imagined diversity, at different recursive levels:

At one taxonomic level, registers are erased in order for one register imagined as standard and homogeneous to count as the named language in contrast with other named national languages. At a lower recursive level, these

### 5 Interpretation

registers must be imagined to exist in order for the language and its corresponding national public to be able to account for "everyone" in the projected national public. (Frekko 2009: 71)

This can be transferred to the case of nineteenth-century American English and it confirms claims and observations made by Cooley (1992) and Minnick (2010) (see §2.4). In the process of being contrasted with British English, internal differences were erased in metadiscursive activity and particularly the retention of /h/ served as a marker of this uniformity. Yet, linguistic differences were imagined at the same time because the existence of sets of forms to index specific (groups of) speakers was indispensable to the existence of a set of forms that could be imagined as indexing all Americans.

# **5.5 Conclusion: the emerging American English register**

To conclude the interpretation of the detailed analyses carried out in Chapter 4, I argue that the enregisterment of American English was based on three main values: nationality, authenticity and non-specificity. Already fairly early in the nineteenth century, nationality was connected to discourses on linguistic uniformity and superiority, which focused on /h/-dropping and -insertion to enregister a national American English in contrast to British English and establish the superiority of the former against British metadiscourses enregistering American English as vulgar and incorrect in the first half of the nineteenth century. By contrast, authenticity came to play a major role towards the end of the century, particularly in the last two decades, and was used to enregister forms like non-rhoticity, a back bath vowel, a labiodental realization of /r/ and the lexical items *luggage* and *trousers* as a northeastern social register through evaluating them as affected and unnatural. Even though the focus thus shifted from delimiting British from American English to delimiting authentic from non-authentic American English, British English still played an indirect role in the latter process because what made forms used by northeastern Americans non-authentic was the fact that these forms were fashionable in England and thus indexical of a (changing) standard British English. However, these metadiscourses also indirectly provide evidence of a positive evaluation of these forms at least for some speakers, which are based on positive evaluations in England, where they are seen as indexing cultivation, education, elegance and a belonging to higher social circles. In this study, these competing evaluations in metadiscursive activity could be illustrated most clearly through the analysis of *pants* and *trousers*.

### 5.5 Conclusion: the emerging American English register

They show that the negotiation process of which forms became enregistered as American for all Americans did not end in the nineteenth century but continued in the twentieth century. Finally, the construction of further specific American registers was important for the enregisterment of a general, non-specific American English. An important register was the southern American English register because it was not only linked to a region but also to social characteristics, specifically to white southerners. However, with regard to linguistic forms, it was marked by a considerable overlap to the Black American register, a social register based on ethnicity. Social and regional values also overlapped with regard to the mountaineer register. Speech forms, among them non-rhoticity, were linked to uncultivated, uneducated, non-religious, immoral and violent speakers living in the southern mountain regions. The non-authentic dude register was of course also a specific register that served to delimit non-specific and authentic American forms against specific and non-authentic ones, which shows that all three values, nationality, authenticity and non-specificity, interacted in the process of enregistering American English. As in the case of the dude register, the most important period for the construction of the other specific registers in newspaper discourses were the last two decades of the nineteenth century. Only yod-dropping, represented by *noospaper/s*, appeared throughout the nineteenth century, with particularly high frequencies in the 1860s, and these articles provide evidence of an earlier delimitation of a register which, although not very specific, differentiated educated, hard-working and well-mannered Americans from uneducated, uncivilized and lazy Americans, exemplified by David Ross Locke's Nasby figure.

All in all, the results of the analysis of metadiscursive practices in nineteenthcentury American newspapers show that the identity relations described by Bucholtz & Hall (2005) play an important role in the construction of a register that had 'American' as its main indexical value. Adequation and distinction figured prominently in the emergence of a model of *American* linguistic and social action, not only through the emphasis on national and linguistic uniformity in relation to British heterogeneity, but also through the distinction of a general, nonspecific American register from several more specific registers within America. These processes of adequation and distinction were tied to processes of authentication and denaturalization because 'true' and 'genuine' Americans were constructed as avoiding linguistic forms indexing British speech. Finally, processes of authorization and illegitimization also became visible through the competing discourses on which forms should be used in an American context and on who had the power to exert linguistic authority (either directly through the prescription of forms or indirectly through the use of the forms): the northeastern so-

### 5 Interpretation

cial elite or the western masses of 'average' Americans. As both of these groups were conceptualized as consisting of white and predominantly male speakers, it becomes clear that other groups were excluded in such negotiations of authority: Blacks and mountaineers, both constructed as ethnically other, were completely absent, and white southerners and women in general only rarely played a role, which shows that they were illegitimized and that the variants that they used and that were in contrast with variants used by white, non-southern males, were not even considered to become part of a national *American* register. The implications of these results on theories of the emergence of new varieties of English and other theoretical discussions will be discussed in the following chapter, in which I will also address the limitations of this study.

# **6 Implications and limitations of this study**

# **6.1 Limitations**

Every study on enregisterment faces the difficulty that there is a large variety of ways to typify language use, so that a systematic study of such typifications that ultimately lead to registers requires a selection of typifications that are the focus of the study. In the present study, the focus is on typifications which occurred in newspaper discourse in the United States during a time period of 100 years, which means that all observations and conclusions based on the analysis of these articles are necessarily restricted to this particular discourse – even though it is possible to identify connections to discourses in other regions (Great Britain in this case) and discourses in other types of media, for example in books (which can be subdivided into different genres typifying language in different ways, for example non-fictional prescriptivist texts and fictional literary works). Although newspapers were chosen because of their wide regional and social circulation, other typifications of language need to be studied as well to complement the picture.

A second limitation results from the chosen search methodology. The focus on pronunciation respellings is likely to lead to an overrepresentation of articles that contain representations of direct speech and an underrepresentation of articles that contain explicit metalinguistic comments. However, I have argued that representations of direct speech are necessarily linked to speakers using these forms and that repeated associations between speakers, their social characteristics and the linguistic forms they use lead to the creation of metapragmatic stereotypes, often embodied in characterological figures, which offer more concrete models available for social alignment than abstract descriptions of language, so that these rather implicit ways of typifying linguistic forms serve an important function and are thus worth focusing on. Nevertheless, it would be a worthwhile task to search for articles by choosing search terms that are likely to yield more articles containing explicit comments, for example by searching for descriptions like *drop their aitches*. Furthermore, pronunciation respellings

### 6 Implications and limitations of this study

already signal deviation and otherness through their alternative spelling, so that the articles are more likely to contain negative evaluations of the phonological forms represented by the search terms. Positive evaluations of these forms are thus likely to be missed or at least underrepresented. However, in several articles, positive evaluations are indirectly visible because they are the reason for people's negative responses, thus rendering revalorizations of linguistic forms visible to the observer. The complementation of the analysis of phonological forms by an analysis of lexical forms also revealed metadiscursive activity which did not rely on pronunciation respellings and showed on the one hand that the lexical forms fit into the pattern found for phonological forms (including the temporal development), but that in the case of *pants* and *trousers* competing evaluations become more directly observable.

In general, all the results of the study are only based on an investigation of a small set of linguistic forms. The advantage of this approach is that these forms can be analyzed in a very systematic and comprehensive way, but this approach of course has the disadvantage that metadiscourses which are potentially relevant for the enregisterment of American English are not taken into consideration. For example, Hodson's (2017b) study on the enregisterment of American English in British fiction revealed that discourse markers like *I guess* or *I calculate* were linked to vulgar American speech and these forms have also occurred in my first exploratory search for articles containing the phrase *American language*, which suggests that an analysis of metadiscourses surrounding these forms in America would yield interesting results to complete the picture. Furthermore, I found early non-rhotic forms like *fust*, *hoss* and *cuss* in the newspaper articles (early because post-vocalic /r/'s were elided first in contexts where they preceded /s/, see §3.3.5), and they seem to have been enregistered in different ways than forms like *deah* or *bettah*, where post-vocalic /r/ is not in pre-consonantal position. For example, Hank Borrows, the Montana wagoner in the article published in 1898 (discussed in §4.1.2), represents the western rural American in contrast to the eastern city dude, but he is portrayed as using early non-rhotic forms like *wuss* 'worse'. Exploring these forms would thus yield insights into how phonological context also matters in metadiscursive activities.

Not only the number of forms is restricted, but also the number of articles found for each form. If compared to the overall number of articles contained in the databases, which is close to 78 million, the number of articles collected for each analysis seems very small. However, as the articles needed to be checked manually as to whether the search term was identified correctly by the recognition software and as the main part of the analysis rested on the qualitative analysis of *all* articles containing the respective search terms, it would not have

been possible to investigate a considerably higher number of articles using the methodological framework developed for this study. Despite the small numbers of articles, however, I could identify clear patterns, which are indicative of the fact that the articles analyzed in the study constitute only the metaphorical tip of the iceberg. For example, if it is considered that *bettah* alone yielded 374 articles and that the articles containing *bettah* also contain several other lexical items which are marked as non-rhotic, for example high frequency words like *heah* 'here' and *theah* 'there', it can be assumed that the number of articles containing similar instances of metadiscursive activity is in fact much higher.

Finally, it is a problem that in the majority of cases, the author of the article is not mentioned and/or the origin of the article is not entirely clear. This makes it difficult to identify the actors in the discursive processes analyzed above. However, it is at the same time also important to note that it was precisely the medium of newspapers that allowed the people creating and publishing the articles to recede into the background in favor of a seemingly impersonal point of view. People were usually not told who had "experienced" the anecdotes or created the cartoons and this had the effect that they were more likely to generalize and not see the content as being the result of a personal subjective view only. However, sometimes a general political stance of the editor and therefore of the newspaper was known and this might not only have had an influence on what was published but also on how people interpreted the content. A closer analysis of who those authors and editors were and how they influenced metadiscursive practices in newspapers would nevertheless enrich the analysis.

To conclude, the present study cannot give a complete picture of how American English was constructed as a discursive variety through enregisterment, but it offers several important insights into enregisterment processes in the nineteenth century that are based on a systematic investigation of a large body of evidence and that have several implications for modeling the emergence of new varieties of English, for a theoretically informed description of the historical development of American English and for modeling of language change in general. I will discuss these theoretical implications in detail in the following sections. It remains a task for further studies to complement the picture drawn here by using further search terms (including search terms that are not pronunciation respellings), by extending the time frame or by focusing on typifications of language occurring in other media.

### 6 Implications and limitations of this study

# **6.2 Implications for modeling the emergence of new varieties of English**

The theoretical arguments and the results of the analysis provide support for Schneider's (2007) claim that identity constructions are central for the emergence of new varieties of English. The theoretical framework I have developed in Chapter 2, which relies heavily on Agha's (2007) theory of enregisterment and other theoretical positions developed by linguistic anthropologists (especially Silverstein 2003, 2016 and Bucholtz & Hall 2005) as well as sociolinguists (especially Eckert 2008, 2014) and discourse-linguists (especially Spitzmüller & Warnke 2011 and Spitzmüller 2013), is suitable for shedding light on precisely this interaction between linguistic forms and identity constructions that Schneider postulates and that was subject to heated debate because it contrasts with Trudgill's (2004) deterministic theory of new-dialect formation (see §2.1.3).<sup>1</sup> I argued that a crucial step in this debate is to carefully distinguish between different types of varieties, structural varieties, perceptual varieties and discursive varieties, because only a careful distinction makes it possible to systematically investigate how exactly these levels influence each other. Conceptualizing registers (in Agha's sense) as discursive varieties has the advantage of emphasizing that they are not independent of language use (which can be thought of as a complex system, following Kretzschmar 2015b), but interact with it. In contrast to structural varieties, the definition of discursive varieties puts speakers' ideas about and evaluations of linguistic forms in the center: In the process of enregisterment, speakers typify linguistic forms and link them to typified social personae and practices. The emerging registers in turn influence the production of linguistic forms: According to Spitzmüller's model of social positioning (see Figure 2.4), registers fulfill their function as cultural models of action in interaction because actors position themselves socially (or get positioned by others) not only through alignment with the actor(s) they communicate with directly but also through alignment with these cultural models. These processes of social positioning and alignment constitute identity in interaction and emphasize the relational nature of identity, as pointed out (among others) by Bucholtz & Hall (2005). The theory of enregisterment and the model of social positioning thus provide a detailed picture of the central element of Schneider's Dynamic Model: the interaction between identity and language use.

In this study, I have identified several registers in nineteenth-century America based on empirically observable metadiscursive activities. I could show for

<sup>1</sup> It also provides a model for the concept of "positive feedback" that plays a role in Kretzschmar's (2014) view on how new varieties emerge.

### 6.2 Implications for modeling the emergence of new varieties of English

example that since the 1880s non-rhoticity had been discursively constructed as being part of an urban northeastern American register, a southern American register, a mountaineer register and a Black American register, which suggests that for speakers who did not want to position themselves as being part of such groups or as having characteristics associated with these groups, the use of rhotic forms must have seemed more attractive than that of non-rhotic forms. However, the analysis of indexical values associated with linguistic forms revealed the complexity of evaluations, which could contrast with each other. To use non-rhoticity as an example again, the newspaper articles also show that the form must have been evaluated very positively as a sign of cultivation, education, upper-class membership and a close relation to European/British ancestry, especially in northeastern cities, but probably also in the coastal south and perhaps even in (rather urban) regions further west (as indicated by the salesladies' conversation in Denver, Colorado). Furthermore, I also found glimpses of a positive evaluation of non-rhoticity within the group of Black American speakers because the form was used by the Black poet Paul Dunbar in his celebration of "de co'n pone", which suggests that for this group of speakers, the use of non-rhoticity was attractive as a signal of African American identity.

A model of linguistic and social action indexically linked to the value 'American', which could then serve as a point of reference for speakers to position themselves as Americans, was thus constructed in a complex relation to other models. The fact that these models were linked to stereotypical personae meant that speakers did not exhibit positive or negative alignment with any sort of abstract national identity, but with more concrete models of types of speakers and behaviors. Agha's (2007: 176) concept of negative alignment, that is, a process of self-differentiation and not of convergence, is important in this context because signaling an American identity could also be done by *not* aligning with social personae and practices which were regarded either as inauthentic or too specific to be enregistered as generally American. To put it simply: According to the model of social positioning, Americans living at the end of the nineteenth century who shared the evaluations conveyed in articles ridiculing the dude figure were likely to realize the post-vocalic /r/ and use the word *pants* in order to distance themselves from this social persona and signal their identity as true Americans. The existence of such social personae and characterological figures make the models more concrete than the type of abstract colonial identity that led Trudgill (2008a) and Mufwene (2008) to argue against identity being a major force in the emergence of new varieties. This confirms Eckert's (2016: 82) hypothesis that "[a]ccommodation in colonial situations may have more to do with emerging local social types or stances in the colonial situation than with some ab-

### 6 Implications and limitations of this study

stract colonial identification". Trudgill's view that identity is only relevant *after* a stable set of forms constituting the new (structural) variety has emerged is also not convincing given the results of this study: While in the case of /h/-dropping and -insertion it could indeed be argued that the retention of /h/ has already been at least the majority variant (if not the only variant) in the United States before it became the focus of discursive activity, in all the other cases the articles suggest that there was considerable variability in the use of the forms *at the same time* that they were subject to implicit and explicit metadiscourses in newspapers. Furthermore, while Trudgill (2004: 88) sees "focusing", so the stage at which identity supposedly plays a role, as "the process by which the new variety acquires norms and stability", I have shown in the present study that metadiscourses did not just stabilize a particular set of forms as American, but that they were the site of a negotiation process in which several forms competed for the value 'American'. It is thus more convincing to regard language use and metadiscourses as interacting forces, revolving around identity: Expressing and ascribing identity through the creation and recognition of indexical values in every instance of communication leads to the creation and recognition of cultural models of actions, which then become points of reference for expressing and ascribing identity in interaction again.

Another point that has been put forward in arguments against identity is that acting based on identity is somehow intentional and that it is unconvincing that the formation of a new variety is a planned and goal-oriented process (see Mufwene 2008, as discussed in §2.1.3). However, Schneider's (2007: 95) view of speakers' social identity alignments that are at the heart of the Dynamic Model is different:

Note that there is no implication made here that these developments have anything to do with language consciousness: accommodation works irrespective of whether the feature selected and strengthened to signal one's alignments is a salient marker of which a speaker is explicitly aware or an indicator which operates indirectly and subconsciously.

Bucholtz & Hall (2005: 606) support this position in their discussion of the concept of "agency":

From the perspective of an interactional approach to identity, the role of agency becomes problematic only when it is conceptualized as located within an individual rational subject who consciously authors his identity without structural constraints. [...] [A]gency is more productively viewed as the accomplishment of social action.

### 6.2 Implications for modeling the emergence of new varieties of English

While it is clear that the newspaper articles that were of interest for the present study are the result of conscious acts aiming to draw attention to the linguistic forms in question and thus contributing to the enregisterment of these forms, this does not imply that speakers' alignment with the social personae and practices associated with these forms is in any way conscious. Neither does it imply that the alignment is restricted to only those forms targeted in the articles. If speakers are exposed to speech by actual speakers that they link to such typified social personae, for example in direct interaction, they might (consciously or subconsciously) pick up on other forms and link them to the register as well.

To conclude this discussion, the complexity of understanding and studying identity has already been noted by Schneider (2000: 361):

It is clear that identity as the determining factor of speech performance is a concept more complex and scientifically more difficult to grasp than region and class, encompassing psychological, sociological, and pragmatic components which are fuzzy in themselves; but it is something we need to understand in an increasingly complex and multifarious postmodern world, an appropriate challenge for language variation study in the new millennium.

The present study has not only shown that this challenge has been successfully addressed by linguistic anthropologists and sociolinguists, but it has also provided a methodological approach to study register formations systematically and thus to provide insights into which cultural models, and, most importantly from a linguistic perspective, which linguistic forms became available for speakers to constitute identity in interaction.

While this study therefore supports the essential role of identity in the emergence of new varieties, it also challenges Schneider's (2007: 30) claim that "to a considerable extent the emergence of PCEs is an identity-driven process of linguistic convergence [...] [which] is followed by renewed divergence only in the end, once a certain level of homogeneity and stability has been reached". As argued in Chapter 5, I rather support the view that the existence of different registers is an essential prerequisite for the construction of a unifying national register. I thus agree not only with Kretzschmar's (2014) point that linguistic diversity has always existed, during all stages postulated by the Dynamic Model, but that this diversity has also always been socially indicative. While Schneider (2007: 296) writes that in the last phase of the emergence of a new variety,

diversification happened because the various regional, social, and ethnic groups recognized the importance of carving out and signaling their own

### 6 Implications and limitations of this study

distinct identities against other groups and also against an overarching nation which, while it is good to be part of, is too big and too distant to be comforting and to offer the proximity and solidarity which humans require.

I postulate that at least on the discursive level, the construction of regional and social registers was essential for creating models of the negative and deviant "other" against which a neutral national standard register could be constructed. What could mark the starting point of a new phase after the fourth phase of endonormative orientation is the increasingly positive revalorization of social and regional registers. In the newspaper articles analyzed in this study, all specific registers were constructed largely through linking the speech forms to *negative* social personae and values: the uncivilized, uneducated and immoral Black American and mountaineer, the old-fashioned, quaint and also rather uneducated southerner and the affected, ignorant, lazy and unsuccessful city dude. As shown in studies on the enregisterment of regional dialects in the twentieth century (see §2.2.2), this negative evaluation was replaced by a pride in local practices, including the use of linguistic forms indexically linked to the region. Pittsburgh is a prime example for such a changing revalorization and it is interesting to see that the authenticity value which played such an important role in delimiting an authentic national variety at the end of the nineteenth century then became important in the re-enregisterment of a local variety in the twentieth century.

# **6.3 Implications for a theoretically informed history of American English**

It is one of the important contributions of Schneider's Dynamic Model that it provides a framework for a theoretically informed history of new varieties of English, including American English. Schneider's own account of the developmental phases of American English constitutes an excellent starting point for such a description. The analysis conducted in the present study adds to it and suggests that some aspects deserve closer attention.

An often-debated issue in research on the history of American English is the role of British influence on linguistic developments in America. This debate revolves especially around the development of non-rhoticity and rhoticity in America (see §2.4 and §3.3.5). Some linguists (e.g. Fisher 2001) argue that non-rhoticity became common in the northeast and the coastal south because it was associated with prestigious British speech and that it also declined in frequency because

### 6.3 Implications for a theoretically informed history of American English

British English lost its prestige (see for example Labov's (2006: 296) claim that rhoticity becoming the prestige form in New York City "reflected the abandonment of the earlier prestige form of Anglophile English"). This view is contested by other linguists, most prominently Bonfiglio (2002), who argues for a greater role of an inner-American struggle centering on race and ethnicity. Furthermore, I have also pointed out that there were different views with regard to the temporal development of the changes in prestige. Fisher (2001), for example, argues for a loss of prestige of non-rhoticity after the Civil War and he points out that this shift in prestige was particularly visible in New York, where the colonial elite had lost its influence. By contrast, Bonfiglio (2002) argues for a change in norms in the first half of the twentieth century, while Labov (2006) suggests the years of World War II as the crucial point in time for this change.

The present study contributes to this debate by showing how the prestige of non-rhoticity (and rhoticity respectively) can be studied more systematically within a discourse-linguistic framework. It provides evidence for an increasing amount of attention paid to non-rhoticity in metadiscursive activities in newspaper articles in the second half of the nineteenth century, especially in the last two decades. The occurrence of articles containing evaluations of non-rhoticity was not regionally restricted, but they were especially frequent in the (mid)west and north of the United States, that is, in areas which according to the available evidence have always been rhotic. It can also be assumed that the articles were read by many people: Newspapers in general were widely read, several articles were reprinted in one or more newspapers and many of the articles contained humorous texts (sometimes accompanied by visual elements) and thus provided entertainment for their readers. The metadiscursive activities surrounding postvocalic /r/ therefore support Fisher's postulation that non-rhoticity started to lose prestige already in the late nineteenth century. Furthermore, the qualitative analyses of these activities also demonstrate that positive and negative evaluations of /r/ must have co-existed and that British influence did indeed play a role. The writers of the articles explicitly or implicitly convey that English fashions and speech were regarded as desirable in northeastern cities and consequently imitated by Americans – first by upper-class speakers and then by middle-class speakers who wanted to be part of higher circles of society. The negative evaluation of this process of imitation provides evidence of a revalorization of the form: Users of non-rhotic forms were depicted as Americans who lack authenticity by pretending to be English. Metadiscursive activities of this kind are therefore a clear indicator of the shift from an exonormative orientation to an endonormative orientation postulated by the Dynamic Model. Most importantly, they suggest that this shift was noted by a large number of speakers and not

### 6 Implications and limitations of this study

just an educated elite who was interested in language. By means of newspaper articles, indexical links between phonological forms and social values could circulate widely even before the advent of radio broadcasting. This study therefore provides support for those accounts that attribute the change in prestige to British influence. It gives a detailed picture of how the link between non-rhoticity and Englishness led to the construction of the social persona of the inauthentic American, embodied by the characterological figure of the dude, which becomes available as a model for negative social alignment – the dude's character traits are represented so negatively that they do not invite imitation but rather distancing. This in turn makes contrasting 'genuinely' American figures appear attractive: the (mid)western farmer, cowboy or hunter. This is in line with Bonfiglio's (2002: 231) finding that constructions of the "western hero as an instantiation of the proper American male" played an important role in the standardization of American English. However, as pointed out by Minnick (2010), the newspaper articles demonstrate that these constructions were already prevalent in the nineteenth-century, especially in the last two decades. Furthermore, these western American heroic figures were set in opposition to the eastern elite (via the dude figure) and not to Black Americans, so that the claim that the (mid)western parts of the country were constructed as 'racially pure' in contrast to 'ethnically contaminated' areas in the northeast and the south could not be confirmed for the nineteenth century.<sup>2</sup>

However, this study also provides support for the important role of race in the enregisterment of American English. The quantitative analysis of social personae linked to *bettah* showed that in the vast majority of cases these personae were Black Americans. The qualitative analysis of these articles demonstrated that racial othering was achieved in several ways – they all encompassed a negative portrayal of physical traits and social characteristics and, as in the case of the dude, relied to a great extent on humor and ridicule. The enregisterment of Black speech, which included non-rhoticity as well as a large number of other linguistic forms, therefore also created a cultural model that invited negative social alignment. This demonstrates the extent to which metadiscourses were shaped by actors who were white and who had the power to mark Black speech as the deviant "other" and its forms as indexing negative characteristics only. A central argument of Bonfiglio's (2002) study is thus confirmed, although the analysis

<sup>2</sup>As newspaper articles representing Black social personae usually located these personae in the south and in northeastern cities (that is, in those areas where most African Americans lived), indexical links between region and race were of course created. However, in the articles analyzed here, these links were not salient because region was not foregrounded in relation to the value of ethnicity but in relation to the value of authenticity.

### 6.3 Implications for a theoretically informed history of American English

conducted in the present study suggests that the factor race was already of more importance in the late nineteenth-century than postulated by Bonfiglio. Overall, the discourse-linguistic analysis conducted in this study therefore does justice to the complexity of "prestige". Prestige depends crucially on who evaluates a linguistic form in which context – a point which is illustrated well by the author of a newspaper article by stating that "[a] hardy backwoodsman may not appear so well in a drawing-room, but he is more attractive on his own ground" (September 8, 1889[96] ).

A further contribution of the present study is that it sheds further light on the relation between endonormative orientation and diversification in the history of American English (and thus between phase 4 and 5 of the Dynamic Model). As pointed out above, early endonormative tendencies could already be found in relation to discourses surrounding /h/-dropping and -insertion, which established a sense of linguistic superiority over British English, although the evaluation of this form as incorrect was still shared between Great Britain and America. Diverging evaluations of forms could be observed with striking frequency in the 1880s and 1890s – American forms were "defended" against fashionable English forms, thus providing evidence of the existence of an American norm that was judged superior to a British norm. The term *Hinglishism*, used in an article published in 1881, is an excellent indicator of this development: It not only marked British English as inferior based on /h/-insertion, but also as deviant from an American norm by playing on the negatively connotated term *Americanism*. At the same time, the increasing representations of inner-American linguistic diversity, that is, of specific social and regional forms of speech, also attest to the stability of the American norm at this point. Diversity does not appear threatening anymore when there is a stable agreement on the existence of forms which are "universally" American; in that case, diversity rather contributes to the stabilization of a new American norm (see the argumentation in §6.2). This suggests that the last two decades can be classified as a transition phase in which discourses emphasizing the presence of a positively evaluated American linguistic norm overlapped with discourses showing an interest in inner diversity. The developments observed in newspaper articles in the present study tie in with the "cult of the vernacular" in American literature described by Jones (1999) (see §2.4), which demonstrates that discourses in different genres were likely to have influenced each other.

A final point concerns the characterization of the different phases of the Dynamic Model. Busse (2015: 92–93) for example notes that Richard Grant White (1870) "explicitly disclaim[s] any right of the Americans to set up their own linguistic standard, independent from Britain", which seems to be a position that

### 6 Implications and limitations of this study

does not fit into a phase of endonormative orientation. However, he cites Schneider's observation of a pro-English movement in the 1870s and 1880s as well as Kretzschmar & Meyer's (2012) position that exonormative tendencies did not cease to exist until the end of World War II. This illustrates a point that Schneider (2007: 277) makes himself, namely that there has always been an ambivalence with regard to language attitudes in America. Pride in American forms coexisted with a "complaint tradition" upholding British forms as correct. The phases of the Dynamic Model are thus not to be understood in absolute terms, but as describing *tendencies*. Accordingly, Schneider (2007: 277) suggests at one point that "[i]n *quantitative* terms [...] nativization made the balance tip toward the former position" [emphasis mine]. The present study shows how such quantitative claims can be empirically tested: The case of *baggage* and *luggage* shows for example that there were only three articles in the collected sample that valued the English form *luggage* more highly than *baggage*. It remains a task for further research to identify the quantitative evaluative patterns of *pants* and *trousers*, which, according to the qualitative analysis, are not likely to show such a clear preference for *pants*.

To conclude, the present study suggests a way of approaching the description of American English based on the theoretical framework of the Dynamic Model, but in which the level of structure and the level of discourse are described and investigated in their own right in order to shed light on their interaction. The enregisterment processes identified here show how discursive patterns can be identified – this study thus presents a first step towards a "feature-by-feature social-reasons account" (Trudgill 2008b: 279) which shows convincingly how social factors and identity influence the shape of a new variety.

# **6.4 Implications for theories of language change**

A final implication of the present study relates to general theories of language change. I would like to draw attention here to Baxter et al. (2009), who use a mathematical model of language change and computer simulation to test Trudgill's (2004) theory of the emergence of New Zealand English. Their model thus relates directly to the discussion of models of the emergence of new varieties of English and can thus also be related to my study. It is based on Croft's usagebased evolutionary framework of language change (Croft 2000, 2006), which proposes that language change is an evolutionary process that can also be found in other areas, particularly biology. The underlying theory is Hull's (2001, 1988) General Analysis of Selection (GAS). It consists of the following central elements:

### 6.4 Implications for theories of language change

first, the *replicator*, which in the case of language is the lingueme, an element of linguistic structure that corresponds to a sociolinguistic variable, secondly a process of *replication*, which is language use in face-to-face interaction, and thirdly the *interactor*, which is an entity that interacts with its environment to cause replication, which in the case of language is the speaker interacting with other speakers.

Within the model, language change is characterized as a differential replication of variants in the process of interaction between speakers. This basic process can be modeled in different ways, depending on which factors are assumed to affect replication. On the one hand, the process can be marked by selection, which means that one variant is selected over other variants. This selection can be influenced by the differential social value assigned to the linguistic variant (*replicator selection*) or to the speaker (*weighted interactor selection*). It can also be neutral, which means that the frequency of interaction between speakers is the only factor influencing the process (*neutral interactor selection*). On the other hand, no selection can be involved, which makes "the successful propagation of a variant [...] a function of the frequency of the variant" and models language change as a purely probabilistic process (*neutral evolution*).

Baxter et al. (2009: 270) argue that Trudgill's (2004) deterministic model of new-dialect formation corresponds to a model of neutral evolution because "his theory is invoking the same usage-based processes as we are, namely that speakers alter their behavior in response to the language they hear around them, and those usage-based processes are probabilistic". As Trudgill's model is also based on accommodation between speakers in interaction and thus relies on the frequency of interaction, Baxter et al. (2009: 271) find that "neutral interactor selection" is also a possible mechanism that Trudgill allows for. It is this latter model that they use to test Trudgill's claims – they basically take the empirical data obtained in the analyses conducted by Trudgill (2004) and Gordon et al. (2004) and apply the model of neutral interactor selection to find out whether the changes in frequency can be explained by this particular type of mechanism (see Baxter et al. 2009 for the mathematical details). The result of the application is "compelling evidence that neutral interactor selection is unlikely to be solely responsible for the fast convergence of the New Zealand dialect to a homogeneous, stable variety" (Baxter et al. 2009: 284), which means that although they do not deny that the frequency by which language users interact with each other and are thus exposed to linguistic variants does play a role in language change, they strongly argue that "weighted interactor selection and/or replicator selection must be added into the model" (Baxter et al. 2009: 291).

### 6 Implications and limitations of this study

Baxter et al.'s (2009) results therefore not only support the important role attributed to social factors in Schneider's Dynamic Model, but they also suggest a way of making an analysis of enregisterment processes, such as the one conducted in my study, fruitful for providing support for a model of language change that includes mechanisms of weighted interactor selection and/or replicator selection. By means of such an analysis detailed information can be obtained about which social values are attributed to linguistic forms and also to social groups linked to the use of the forms. Baxter et al. (2009: 269–270) point out that factors like the prestige of a variant are not easy to measure – this study, however, suggests a way of approaching such a measurement. By using a well-defined corpus like the databases of newspaper articles, it is possible to quantify metadiscursive activities, as I have shown for example with regard to the numbers of articles associating linguistic forms with different social groups or the numbers of articles evaluating baggage positively, negatively or neutrally. How exactly such a quantification could be carried out and integrated into such a model of language change remains a task for further research, but it would definitely be a valuable addition to the discussion of the role of social factors in language change.

# **7 Conclusion**

The present study started out with the question of whether statements about American English, like the one made by the editor in a newspaper article in 1882 saying that the American language "is the English language with the "H's" in their proper places" should be of interest to linguists studying the emergence of American English as a new variety of English. The theoretical argumentation and the empirical analysis conducted in the present study strongly suggest that this question must be answered in the affirmative. The theory of enregisterment as well as recent research in sociolinguistics and linguistic anthropology provide a detailed account of how language variation on a structural level and the recognition of such variation on a discursive level interact: Registers, defined as cultural models of (linguistic) action, are constructed through metapragmatic and metadiscursive activities that create indexical links between linguistic forms and social values, as in the statement above, which links /h/-retention to the value 'American'. Registers in turn influence speakers' selections of linguistic forms in the process of social positioning in interaction, which is essentially an act of identity. These theories thus provide a detailed account of the mechanisms underlying the central tenet of Schneider's (2003, 2007) Dynamic Model, namely that speakers' attitudes towards and evaluations of linguistic forms are highly relevant because they influence speakers' linguistic choices in a process of (conscious or unconscious) social identity alignment. The relevance of registers, which I see as equivalent to discursive varieties and as different from structural and perceptual varieties, motivates the application of the theory of enregisterment, which I have visualized in a model in §2.3, in order to investigate the discursive construction of American English in the nineteenth century.

One of the central achievements of the study is the development of a methodological approach to investigating historical enregisterment processes that is based on a discourse-linguistic framework: DIMLAN. The systematic collection of newspapers articles in two large databases by means of specific search terms allowed for quantitative as well as for qualitative analyses of metadiscourses surrounding five phonological and two lexical variants. I analyzed close to 1,200 newspaper articles by focusing on the intratextual layer in order to determine which indexical links were created in these articles and how they were created. Based on these analyses, larger patterns could be discovered and quantified to some extent in order to do justice to the transtextual layer. Particularly the temporal development and the regional distribution of metadiscursive activities in newspaper articles could be investigated by means of quantitative analyses.

The findings of the study provide important insights into the enregisterment of American English, even though the picture is, of course, far from complete since the focus is restricted only to seven linguistic forms. The analysis of the newspaper articles showed that different variants played different roles in the process of enregisterment at different times: /h/-dropping and -insertion as well as yod-dropping were present in newspaper metadiscourses throughout the nineteenth century, but while the former was a very salient form that was essential in delimiting an American English register from a British English register, the latter was much less salient and became more and more non-specific so that it could ultimately index the value 'American' and contribute to the delimitation of a nonspecific American English register from several specific American English registers. For /h/-dropping and -insertion, the nationality value was consequently most important: Based on nationalist ideologies placing a high value on uniformity, the absence of /h/-dropping and -insertion was constructed as a marker of American English and as a sign of its superiority over a heterogeneous and thus inferior British English. The ""H"s in their proper places" therefore functioned as a basis for American's linguistic self-confidence. By contrast, yod-dropping was characterized mainly by its non-specificity. Even though it was marked as deviant through the spelling <oo>, it became linked to an unmarked standard American English because the alternative variant, yod-retention, became increasingly associated with southern speech on the one hand and pedantic behavior and attitudes on the other hand. In comparison to these two forms, all the other forms occurred in metadiscourses much later, and they all exhibited a dramatic rise in the last two decades of the nineteenth century.

Non-rhoticity, the back bath vowel and the labiodental realization of prevocalic /r/ were all linked to the characterological figure of the dude, which became central in the delimitation of an authentic American English register from an inauthentic American English register. The dude was portrayed as being an American, but not a true and genuine one, because he held English fashions, including linguistic forms, in high regard. His attempts at imitating Englishmen were subject to humor and ridicule in the newspaper articles, which conveyed to the reader that such behaviors, and manners of dressing and speaking were not acceptable for 'real' Americans. More peripheral, female figures like the salesladies or the society girls confirm that the authenticity value was important in countering the positive evaluations of non-rhoticity in the urban

northeastern centers. Articles which contained both *luggage* and *baggage* in close proximity and which linked these variants to British or American English respectively also rose in frequency in the 1880s. They show that the nationality value remained important because the overwhelmingly positive evaluation of *baggage* as the American variant is based on the strong indexical link between *luggage* and Englishness that made the form an inadequate alternative in America. By contrast, the qualitative analysis of articles containing *pants* and *trousers* demonstrated that in this case the evaluation of *pants* was more contested and rather revolved around the authenticity value. In articles favoring *trousers*, positive associations with educatedness, high social standing and elegance could be identified, whereas articles favoring *pants* constructed the use of the word as a sign of being a true American. That *trousers* was also linked to the dude and occasionally combined not only with a labiodental /r/, but also with non-rhoticity (e.g. in *twousahs*), strengthened associations between *trousers* and a lack of authenticity. This made *trousers* a form that illustrates that enregisterment includes several types of perceivable signs: The word not only linked the level of the lexicon to that of phonology but it also connected linguistic to non-linguistic signs because the newspaper articles also drew attention to the style of trousers worn by the dude. In contrast to the inauthentic dude figure, the authentic American was embodied by figures like the strong and tough American cowboy, hunter or farmer. Through the contrast with the dude, it became implicitly clear that these figures used rhotic forms, front bath vowels and non-labiodental (probably retroflex) realizations of /r/. In some cases, these American figures were represented as using hyper-rhotic forms – I have argued that this created a continuum with rural, uncultivated, but nevertheless true, strong and hard-working Americans using hyper-rhotic forms on the one end, and urban, affected, ignorant, effeminate and lazy Americans using non-rhotic forms on the other end.

Next to the nationality and the authenticity value, the non-specificity value also figured prominently in the final decades of the nineteenth century. Nonrhoticity was constructed as specific not only through its association with the dude but also through indexical links between the form and southern Americans, mountaineers, and, most frequently, Black Americans. Non-rhoticity was not only salient, but the groups it was linked to were portrayed in a negative light in these articles, which made rhoticity, the non-specific variant, appear positive by implication.

The enregisterment of American English thus proceeded by way of delimitation against other registers: British English on the one hand, but also inauthentic as well as more specific regional and social American Englishes on the other hand. These findings have several implications for further research. First, they provide

### 7 Conclusion

support for the Dynamic Model because the forms that have been enregistered as American in nineteenth-century newspapers are also forms that are now regarded as Standard American English and that have also been shown to have increased in frequency since then. The analyses conducted in this study show in detail how cultural models of (linguistic) action were constructed and how they circulated and thus became potentially relevant to speakers' social positioning in interaction. The concept of prestige, which was often used in previous research in rather simplistic terms, was elaborated in more detail by identifying the social values based on which forms were constructed as prestigious and how conflicting evaluations of forms are negotiated. This study thus provides a first step towards addressing the need for studies that investigate the connection between the discursive and the structural level. It needs to be complemented by more studies on enregisterment processes that use other search terms, investigate other metadiscursive genres or focus on other linguistic forms. Moreover, it needs to be combined with detailed studies of actual language use in nineteenth century America in order to find out how the emerging registers actually affected speakers' linguistic choices. Focusing on a specific context (a region and/or social group) will make this task easier.

Secondly, this study supports the claim that a theoretically informed description of American English needs to distinguish carefully between the structural and the discursive level, while at the same time considering their interaction. This means that on the discursive level attention needs to be paid to which linguistic forms are subject of reflexive activities and how these forms are enregistered in the process. I have shown how this study provides a more detailed perspective on Schneider's account of the emergence of American English, particularly on the shift from an exonormative to an endonormative orientation, by showing how these orientations become visible in metadiscourses relating to specific linguistic forms. A description that includes enregisterment thus comes closer to the "feature-by-feature social-reasons account" demanded by Trudgill (2008b: 279).

Finally, I have argued that the present study can inform models of language change like the mathematical model by Baxter et al.'s (2009) which is based on Croft's usage-based evolutionary framework. Baxter et al.'s test of Trudgill's model of new-dialect formation was based on a model of neutral interactor selection to do justice to Trudgill's claim that social factors do not play a role. That this test failed suggests that in order to account for the emergence of new varieties (at least for the test case of New Zealand English) mechanisms of weighted interactor selection and/or replicator selection need to be included in the model. This means that the value placed on the interactor (the speaker) and/or the replicator (the linguistic form) need to be taken into account. The present study has

suggested an approach to determining such values in quantitative terms, but it has also emphasized that this can only be achieved in combination with detailed qualitative analyses.

The editor's comment on the American language "with the "H"s in their proper places" not only constitutes one quantifiable data point indicating a positive evaluation of /h/-retention and an indexical link to the value 'American'. It is also part of a larger discourse that extends to other places and other genres, which proves one point: that metadiscursive activities are as complex as language use itself.

# **Appendix**

Regional distribution of newspapers, issues and articles published in 51 states and territories of the United States in the nineteenth century and contained in the databases AHN and NCNP (ranked according to the number of articles).


### Appendix


# **List of sources**







Hickey, Raymond (ed.). 2010. *Varieties of English in writing: The written word as linguistic evidence*. Amsterdam & Philadelphia: John Benjamins.

Hill, Archivald A. 1940. Early loss of [r] before dentals. *PMLA* 55(2). 308–359.


*cies of colonial English: Studies in transported dialects*, 121–157. Cambridge & New York: Cambridge University Press.






Adams, James Eli, 194 Agha, Asif, 2, 3, 35, 38–43, 48–51, 64, 65, 68, 81, 84, 87, 112, 117, 124, 128, 136, 190, 398, 399 Algeo, John, 3, 89–91 Altendorf, Ulrike, 157 Amana, Harry, 115 Angermüller, Johannes, 75, 76 Anonymous, 151 Babel, Anna M., 63 Bailey, Guy, 16, 101, 102 Bailey, Richard W., 3, 89, 90, 127, 135, 136, 146, 147 Barber, Lucy G., 354 Barker, Hannah, 113 Bartlett, John Russell, 91 Bauer, Kathryn M., 63, 64, 68, 74 Bauer, Laurie, 29 Bauman, Richard, 67, 177, 178 Baxter, Gareth, 406–408, 412 Beal, Joan C., 57–60, 63, 66–68 Becker, Kara, 148 Becker, Richard A., 120 Bergs, Alexander, 88 Berruto, Gaetano, 25 Blair, Walter, 304 Blommaert, Jan, 26, 75, 77, 83 Bogle, Donald, 253 Bonfiglio, Thomas Paul, 96–100, 106, 147, 148, 154, 155, 387, 403, 404

Bradley, David, 134, 135 Briggs, Charles L., 67 Brinton, Laurel J., 88 Britton, Derek, 142, 143, 156 Bucholtz, Mary, 49, 63, 69, 381, 387, 393, 398, 400 Burrows, Simon, 113 Busse, Ulrich, 405 But, Roxanne, 65 Campbell-Kibler, Kathryn, 63, 64, 68, 74 Chambers, Jack K., 21 Chaudron, Adelaide de Vengel, 154 Clark, Urszula, 63 Cloud, Barbara, 113 Cole, Debbie, 63 Cooley, Marianne, 94, 106, 109, 392 Cooper, Christopher, 133, 135 Cooper, Paul S., 63, 65, 66 Copeland, David, 112, 114 Coupland, Nikolas, 29, 30, 33, 47 Cox, Karen L., 259 Cramer, Jennifer, 63, 72 Crisma, Paola, 126 Croft, William, 406 Cukor-Avila, Patricia, 244 D'Onofrio, Annette, 55, 65, 68 Davies, Catherine Evans, 274, 276 Davis, Daniel R., 90, 92 DeSante, Christopher, 120

Dicken-Garcia, Hazel, 116 Diessel, Holger, 257 Dillard, Joey L., 19, 147 Docherty, Gerard J., 156–158 Dong, Jie, 63 Drager, Katie, 55 Du Bois, John W., 61 Eberhardt, Maeve, 65 Eckert, Penelope, 2, 45–47, 55, 63, 67, 69, 71, 379, 398, 399 Edge, John T., 264 Eggert, Elmar, 63 Ellis, Alexander J., 130, 144, 151 Elmentaler, Michael, 63 Elphinston, J., 143 Elwyn, Alfred Langdon, 91 Everett, George, 116 Fairclough, Norman, 66 Fallows, Samuel, 91 Feagin, Crawford, 130, 131, 149 Ferris, William, 255 Fischer, David Hackett, 19 Fisher, John Hurt, 145, 148, 402, 403 Foulkes, Paul, 156–158 Fowler, Joy, 228 Frekko, Susan E., 63, 391, 392 Gal, Susan, 49, 50 Gale, 119 Garrod, Simon, 32 Giles, Howard, 10 Goebel, Zane, 63 Gordon, Elizabeth, 407 Grandgent, Charles H., 137, 155 Gumperz, John J., 45 Hall, Kira, 49, 63, 69, 381, 387, 393, 398, 400

Harris, Skylar, 237 Hay, Jennifer, 55, 144 Henry, Eric S., 65 Hickey, Raymond, 123 Hill, Archivald A., 142 Hodson, Jane, 66, 382, 383, 396 Hofer, Matthew, 196 Holmes, Janet, 29–31 Honeybone, Patrick, 123 Hughes, Arthur, 129, 130 Hunt, Tamara L., 181 Huntzicker, William E., 112 Hutchison, Robin Brandt, 196 Irvine, Judith T, 49, 50 Irwin, Patricia, 148 Jaffe, Alexandra, 124, 125 Jespersen, Otto, 136, 151, 154, 156, 157 Johnstone, Barbara, 2, 3, 51–54, 57– 59, 63, 64, 66–69, 75, 76, 78, 111, 385 Jones, Charles, 133–135, 143, 144, 150, 151 Jones, Gavin, 93, 405 Jones, Stephen, 138 Kelsey, Mavis Parrott, 196 Kerswill, Paul, 29–31 Kiesling, Scott F., 38 Kleber, John E., 270 Knights, Peter R., 147 Kökeritz, Helge, 133 König, Werner, 25 Krapp, George Philip, 127, 130, 132, 136–139, 145, 147, 154, 389 Kretzschmar, William A., 1, 8, 17–20, 23, 24, 26, 27, 31–33, 35, 398, 401, 406

Kurath, Hans, 130, 137, 139, 145, 146 Kytö, Merja, 101 Labov, William, 16, 45, 51, 96, 148, 228, 403 Laferriere, Martha, 138, 140 Lanehart, Sonja, 103, 105 Lass, Roger, 126, 129, 130, 133–135, 141 Le Page, Robert B., 9, 18, 31, 32 Leigh, Percival, 135 Leimgruber, Jakob, 24, 26, 28, 66 Lodge, R. Anthony, 380, 381 Longmuir, John, 135 MacMahon, Michael K. C., 130, 133, 134, 144 Madsen, Lian Malai, 63 Managan, Kate, 63 Mather, Patrick-André, 228 McDavid, Raven, 130, 137, 139, 145, 146, 304 Mencken, H. L., 91 Meyer, Charles F., 24, 27, 406 Meyerhoff, Miriam, 20 Milroy, James, 99, 100, 126 Milroy, Lesley, 57, 58, 99, 100 Minkova, Donka, 126, 141, 142 Minnick, Lisa C., 95, 98, 109, 392, 404 Mitchell, Melanie, 17 Molee, Elias, 91 Moll, Andrea, 65 Montgomery, Michael, 100–103, 139, 145, 154 Mufwene, Salikoko S., 8, 10, 29, 30, 32, 104, 105, 399, 400 Mugglestone, Lynda, 127, 128, 135, 136, 150, 151, 179 Nagy, Naomi, 138, 148

Nares, Robert, 133 Nettels, Elsa, 262 Newell, Sasha, 63 Newman, Michael, 145 Nicolaisen, Wilhelm F., 178, 190 Niebuhr, Oliver, 63 Niedzielski, Nancy, 72 Nord, David P., 112–115, 117 Norton, Charles Ledyard, 91 Orton, Harold, 142 Park, Jonghyun, 63 Paulsen, Ingrid, 86, 87 Pawley, Christine, 115 Peter, Benjamin, 63 Phillips, Betty S., 130, 131 Pickering, Martin, 32 Pickl, Simon, 25–27, 102 Picone, Michael D., 123 Pitts, Ann, 130 Plichta, Bartłomiej, 73 Pratt, Teresa, 65, 68 Preston, Dennis R., 68, 72–74 Ramisch, Heinrich, 126 Readex, 119, 120 Remlinger, Kathryn A., 63, 66, 68, 75, 111 Riley, Sam G., 116 Roberts, Julie, 138 Romero, Sergio, 63 Ruano-García, Javier, 63 Savage, W. H., 135, 143, 144 Scharnhorst, Gary, 196 Schele de Vere, Maximilian, 91 Schilling-Estes, Natalie, 317 Schintu Martínez, Paula, 65 Schleef, Erik, 66

Schlüter, Julia, 126 Schneider, Edgar W., 1–3, 7–13, 15– 17, 19, 20, 22–25, 28–31, 33, 34, 103, 130, 131, 148, 156, 239, 251, 385, 388, 398, 400, 401, 406 Sheridan, Thomas, 127 Silverstein, Michael, 2, 35–37, 43, 44, 55–59, 398 Simpson, David, 3, 92, 93 Slotta, James, 63 Smith, Charles William, 136, 152, 153 Snodgrass, Mary Ellen, 231 Sóskuthy, Martin, 143 Sparks, David, 120 Spitzmüller, Jürgen, 2, 4, 35, 61, 62, 70, 71, 74, 78–84, 98, 108, 111, 398 Sudbury, Andrea, 144 Tabouret-Keller, Andrée, 9, 18, 31, 32 Thomas, Erik R., 138, 148, 149 Tillery, Jan, 16 Trudgill, Peter, 1, 8–10, 16, 17, 20–22, 28–30, 32–34, 59, 398–400, 406, 407, 412 Tuten, Donald N., 29, 30, 32, 33 van der Auwera, Johan, 241, 257 Vile, John R., 279 Walker, John, 130, 135, 143 Wallace-Sanders, Kimberly, 235 Walton, Shana, 124, 125 Warnke, Ingo H., 2, 4, 35, 70, 71, 78– 84, 111, 398 Watson, Kevin, 123 Watt, Dominic, 157 Webster, Noah, 13, 154

Wells, John C., 130, 133, 157 White, Richard Grant, 405 Wickham, Hadley, 120 Wilce, James M., 63 Williams, Quentin E., 65 Williams, Ralph Olmstedt, 91 Wolfram, Walt, 317

Young, Elizabeth, 252

# The emergence of American English as a discursive variety

Do speakers' identity constructions influence the emergence of new varieties of a language? This question is at the heart of a debate about how the process of the emergence of postcolonial varieties of English can best be modeled. This volume contributes to the debate by linking it to models and theories proposed by anthropological linguists, sociolinguists and discourse linguists who view identity as a social and cultural phenomenon that is produced through linguistic and other social practices. Language is seen as essential for identity constructions because speakers use linguistic forms that index social *personae* as well as specific social practices and values to convey an image of self to other speakers. Based on the theory of enregisterment that models the cultural and discursive process of the creation of indexical links between linguistic forms and social values, the argument is made that any model of the emergence of new varieties needs to differentiate carefully between a structural level and a discursive level. What emerges on the discursive level as a result of processes of enregisterment is a *discursive variety*. The volume illustrates how the emergence of a discursive variety can be systematically studied in a historical context by focusing on the enregisterment of American English as it can be observed in nineteenth-century U.S. newspapers. Using a discourse-linguistic methodological framework and two large databases containing close to 78 million newspaper articles, the study reveals a complex pattern of indexical links between the phonological forms /h/-dropping and -insertion, yod-dropping, a lengthened and backened bath vowel, non-rhoticity, a realization of prevocalic /r/ as a labiodental approximant as well as the lexical items *baggage* and *pants* on the one hand and social values centering around nationality, authenticity and non-specificity on the other hand. Qualitative analyses uncover the social personae associated with the linguistic forms (e.g. the American cowboy, the African American mammy and the "Anglo-maniac" American dude), while quantitative analyses trace the development over time and show that the enregisterment processes were widespread and not restricted to a particular region.