David Herbert and Stefan Fisher-Høyrem (Eds.) **Social Media and Social Order**

David Herbert and Stefan Fisher-Høyrem (Eds.)

# **Social Media and Social Order**

Managing Editor: Katarzyna Inga Michalak

Associate Editor: Francesca Corazza

Language Editor: Michael Evans

ISBN 978-83-66675-60-5 e-ISBN (PDF) 978-83-66675-61-2 e-ISBN (EPUB) 978-3-11-076040-8

This work is licensed under the Creative Commons Attribution 4.0 International license (CC BY 4.0) For details go to http://creativecommons.org/licenses/by/4.0

#### **Library of Congress Cataloging-in-Publication Data**

A CIP catalog record for this book has been applied for at the Library of Congress.

Copyright © 2021 David Herbert, Stefan Fisher-Høyrem and chapters' contributors

Published by De Gruyter Poland Ltd, Warsaw/Berlin Part of Walter de Gruyter GmbH, Berlin/Boston The book is published with open access at www.degruyter.com.

Managing Editor: Katarzyna Inga Michalak Associate Editor: Francesca Corazza Language Editor: Michael Evans

Preparation and publication of this book was financially supported by the Research Council of Norway.

www.degruyter.com

Cover illustration: geralt/pixabay

## **Contents**

# **Acknowledgements —IX**

David Herbert and Stefan Fisher-Høyrem

**1 Introduction: How Do Social Media Change Social Order? The Deep Datafication of Society from Global to Local Scales (and Back Again) — 1** References **—9**

#### Nick Couldry


	- References  **16**

John D. Boy and Justus Uitermark

#### **3 The Dramaturgy of Social Media: Platform Ecology, Uneven Networks, and the Myth of the Self — 17**

	- References **—30**

Stefan Fisher-Høyrem and David Herbert

#### **4 Clusters of Prestige: Social Media and Social Order in the Norwegian Bible**

**Belt — 33**


4.2.4 Characteristics of Kristiansand's Social Order **—38** 4.3 Social Media in Kristiansand **—38** 4.3.1 Local Clusters **—40** 4.3.2 Local Top Users: Navigating Hierarchies **—41** 4.4 Reproducing the Social Order **—44** 4.4.1 Prestigious Consumption **—45** 4.4.2 City-Wide Symbols of Prestige **—47** 4.5 Concluding Remarks  **—49** References **—50**

#### Sagorika Singha


#### Liat Berdugo


Sarah Sobieraj and Shaan Merchant



Alicia Wanless and Michael Berk


# **Acknowledgements**

This volume has grown out of the conference *Social Media and Social Order* held in Oslo in Nov-Dec 2017 as part of the research project *Cultural Conflict 2.0: Religion, Media and Locality* (https://cc2.mediated.eu/; CC2). We are grateful to the Norwegian Research Council (Grant no. 231344) for funding the project, to the University of Agder for funding the open access electronic publication of the book, to Nick Couldry and Kim Knott our project advisors, and to our editor at De Gruyter, Beata Socha, for her patience, professionalism and help in bringing the project to fruition.

## **1 Introduction: How Do Social Media Change Social Order? The Deep Datafication of Society from Global to Local Scales (and Back Again)**

#### David Herbert and Stefan Fisher-Høyrem

**Introduction: How Do Social Media Change Social Order? The Deep Datafication of Society ...**

Researching the social/media relation today must mean more than merely describing how the latest platforms work, let alone celebrating their supposedly positive potential (democratic? expressive? socializing?). It must mean at least researching how social media platforms … have come to *propose* a certain version of "the social," and how users go on to *enact* it. It must also mean researching … and registering the fractured spaces from where alternative proposals of "the social" might be built (Couldry & van Dijck 2015, p. 2; emphases in the original).

The concept of "social order" is not an academic detail, still less a theoretical extravagance, but rather a highly practical term for registering how social worlds are, on the largest scale, being transformed through the interlinking of all we do on social media … and the systems of social governance, explicit and implicit, that are emerging to manage them (Couldry, this volume).

This book addresses the relationship between social media and social order at multiple scales, providing insights into how diverse social worlds are being reshaped by social media, as well as analysis and reflection on what this means and how critical publics might respond.

From the invention of the internet, successive waves of digitally networked technologies have been welcomed by social scientists and commentators as providing the means both to widen participation in discussion of matters of common concern (the public sphere model, following Habermas, 1989/1968) and for subordinate groups to self-organise and resist forms of oppression based on media and political elites' symbolic domination of society through the mass media (the emancipatory model, following Castells, 2012). Furthermore, social scientists have identified several mechanisms through which these forms of participatory and emancipatory agency might be exercised – including the propensity of networks to elude hierarchical domination (Benkler, 2006), their capacity to enable "contraflow" from peripheral to central media producers (Cottle, 2006), and the "virality" of user-generated content, which can on occasion generate networked publics capable of challenging major corporations and even governments (Castells, 2012).

However, it is our contention that, while networked technologies do indeed *sometimes* enable such dialogical and emancipatory outcomes, *normally* they enable rather different social processes – i.e. they tend to reinforce, elaborate and further embed existing forms of social order, rather than challenge them (Boy & Uitermark, 2019, pp. 2–3; Herbert, 2020, p. 10). Social order in this context should not be conceived as static, but rather as dynamic, constantly evolving and actively reproduced through continuous interaction. The tendency of networked technologies to re-enforce rather than challenge existing social order occurs partly because the networks formed through them are mostly composed on the basis of attraction, admiration and shared interest; for example, with social media networks being woven from the expressions of liking, following and retweeting etc., with those individuals who most exhibit the admired quality (and most actively promote it and often themselves), as the stars around which the network revolves, each forming a cluster with the network of clusters comprising the constellation of the platform. Platforms explicitly formed around the desire to engage with different others are comparatively rare, and often require vigilant forms of moderation to keep the discourse civil (see Herbert & Black, 2012). Furthermore, the formation of clusters is algorithmically shaped, as algorithms embedded in the platforms constantly suggest friends, content and products we might like in iterative response to our choices. The emergent process of tailoring content to users is thus one of continuous differentiation, which at an individual level ensures we receive suggestions we are likely to like, at a collective level connects us more intricately to those we already like and are likely to like (reinforcing our filter bubble, and likely existing power relations), and is continually commercially harvested, and may be used to enable differential (and hence potentially discriminatory) treatment by businesses and government.

In Chapter 2 project advisor for *CC2* and leading public intellectual on mediasociety entanglements Nick Couldry casts doubt on optimistic assessments of the impact of networked digital media for a related but different reason. Addressing the largest geographic scale possible, he asks: how does the corporate harvesting of data from networked digital devices impact on the ordering of social relations in general, at a global scale? In other words, what are the general features of a "datafied" social order? The question relates to the homophilic basis of social networks because the choices of individuals to link to networks of similar others forms a small part of the data on user activity collected by the corporations who run social media platforms and other networked digital services. On social media platforms it is used to suggest people we might know or like to connect with, and thus may shape our social networks directly; similar algorithms which drive the recommendation process are also deployed to recommend products and services based on our internet search and retail activity. Couldry's is therefore a large question, but an appropriate one within which to frame a collection which focuses on analysing specific instances of the development of how users enact social orders shaped by networked digital platforms – social media – and hence datafied – at city/site (Chapters 3 through 6), and national scales (Chapters 7 and 8).

But first, before considering Couldry's answer, why "social order" at all? Couldry argues that the concept is out of fashion, citing in support that the last major work came out in 1994 and is out of print, and giving as a possible reason the challenge to reified concepts of the social especially in the field of assessing the impacts of new technologies by proponents of actor network theory (Latour, 2005). In response, Couldry argues that "at a time when, through datafication, it appears that corporations and governments are intent on reconstructing social reality in ways that align with their interests, it becomes vital to pay close attention to processes of social construction" (Couldry, Ch. 2) – and that some concept of social order is essential in developing a critical account of such processes.

Furthermore, other evidence suggests social order may not be such a neglected or marginal topic in social sciences. In 2003 an American textbook on the topic appeared, with a second edition in 2009 and a well-maintained supporting website, suggesting at least modest currency of the term in US social science programs (Hechter & Horne, 2009; https://www.sup.org/socialorder/?ref=bookurl). This collection is useful for showing the influence of the concept of social order across a range of social science disciplines, for asserting its role as "a core theoretical issue in the social sciences" and for indicating the range of ways in which the "problem of social order" – understood by the authors as how people "coordinate their actions and … cooperate to attain common goals" (ibid. 1) is addressed through several traditions of conceptualisation and theorisation, including mobilising concepts of groups, hierarchies, markets and networks (ibid., vii–viii).

Notably absent from this collection, however, is any sustained attention to how media technologies might shape social order, which is the central question addressed in this volume, focusing on social media. The perspective also differs markedly in its Hobbesian framing – which constructs the "state of nature" as fundamentally conflictual and therefore legitimises the need for ordering mechanisms to reconcile the differing interests of individuals for the greater (individual and collective) good. In this way, the perspective aligns with what Zygmunt Bauman describes as "the gardening principle" of the modern state, as Schiel summarises the concept:

It refers to a state "managed" by its government like a garden. The gardener/ government applies rational methods based on scientific knowledge to create optimal conditions of growth for the "plants"/people (2005, p. 81).

The framing of an absence of social order as a problem that needs "solving" tends to legitimise the need for "solutions" of some kind; whether through markets or networks as "horizontal" or decentralised co-ordinating mechanisms, or through the state or some kind of hierarchy as "vertical" mechanisms of control. Bauman, in contrast, is suspicious of such mechanisms of social control, an approach exemplified by his characterisation of the modern state, which emphasises the disruptive and novel features of modern state power:

With the backbone of communal self-reproduction disintegrating or crushed, the modern state power was bound to engage in deliberate management of social processes on an unheard-of scale … it did not concentrate the previously dispersed powers. It presided over the formation of an entirely new kind of power, or unprecedented scope, depth of ambition, depth of penetration and ambition (Bauman, 1990, p. 157).

We think it is not exaggerating to say that Couldry sees the datafication of society as a project of similar scope, involving a parallel deepening of penetration of state and corporate mechanisms of social control. Couldry focuses on corporate data harvesting as the most ubiquitous global form of datafication, but locates this in relation to and stresses connections with state uses of digital data; for example, in China where there is less separation of state and corporate sectors and less emphasis on personal privacy to maintain that distinction than in the West. It is also where the most ambitious and overt attempt to harness big data to rank the social "trustworthiness" of citizens and distribute rewards and punishments accordingly – The Social Credit System – and hence directly harvest digital date to shape social order – is already being rolled out (Liang et al., 2018). Yet, Couldry also notes the interest of governments everywhere in harnessing digital data to police populations and shape behaviour, stressing the disproportionate impact of this on the lives of the poor and especially those relying on state welfare.

Couldry argues that the genesis of the datafied order was largely unplanned (e.g. Google's accidental discovery that it could commercially exploit internet search data, Zuboff, 2019), and that this remains partly true of its ongoing development. For while "corporate intention becomes a major factor once the advantages of an emerging datafied social order become clearer to corporations," planned or coordinated action by key players is not required for what we suggest might be termed the *deep datafication* of society, a process through which "the principle of discriminationthrough-data (that is, automatically harvested data from online activities) has spread right across the social terrain" (Couldry, this volume). Thus, while corporations and governments have long sought, held and used data on citizens and consumers from multiple sources, networked digital platforms enable its gathering, use and integration on an unprecedented scale, reaching deep into the everyday lives of people across the planet – hence *deep* datafication.

Like Herbert and Fisher-Høyrem (Chapter 4), Couldry draws on Elias to conceptualise the emergence of a datafied social order. Elias understands a "figuration" as "a structure of mutually orientated and dependent people … the network of interdependencies formed by individuals" (Elias, 2000, p. 482). Elias argues that social order emerges through such figurations, "*through* the continuous interweaving of many interrelationships and connections, their progressive impact, as social actors try, successfully or otherwise, to live their lives *through* the web of interrelations in which they have largely no choice but to be entangled" (Couldry, this volume). Thus, what Couldry and Mejías (2019) call "data relations," referring to "relations that configure social life on a basis designed to optimise the generation and extraction of valuable data" become "a leading form of 'figuration' … for the era of datafication." (Couldry, this volume).

At the same time, Couldry argues that while "discrimination-through-data … has spread right across the social terrain," resistance is neither impossible nor futile; rather "to the extent that this order is resisted by social actors, it will not unfold exactly as I outline here. That is the point of analysing datafied societies from the point of view of social order, to alert readers to what is under way and help them imagine what resistance might feel like." (Couldry, this volume).

Elias' concept of figuration can be applied to many kinds of social relations, especially where there is a lack of strong central social control, or social norms leave space for negotiation. But the difference in a datafied social order is the way in which virtual forms of social interaction are mediated by the properties of digital systems, in particular by social media platforms. It is on this mediation, its effects on those enmeshed, their agency and resistance, that the substantive chapters that follow Couldry's essay focus. Each examines how users enact – using Couldry and van Dijck's terms – the versions of the social proposed by social media platforms across range of scales and sites.

Drawing on the dramaturgical perspective of Erving Goffman, in Chapter 3 Boy and Uitermark investigate the self-presentation and status displays characterising the online conduct and related careers of Instagram influencers. Instagram clusters tend to centre on very few user accounts, whose style of display sets the tone for the entire cluster, and whose influence enables them to access temporary job opportunities as party hosts or event promoters. The interviews reveal a vast and stress-inducing gap between their curated Instagram "frontstage" personas and the "backstage" precarity of their offline lives and continuous intense effort required to maintain their social media profiles. These interviewees are "central users" in their Instagram clusters and the platform has become central for their livelihoods, but one that has become a hard taskmaster, requiring constant updating to maintain the users' precarious position at the pinnacle of esteem.

Similarly, in Chapter 4 Fisher-Høyrem and Herbert draw on Elias' model of the early modern royal court, analysing how local top users of Instagram in a Scandinavian town compete for the attention of users of higher status, and how this quest for affirmation involves mutual (and self-) policing, shaming, and an anxious interweaving of online and offline lives. Combining computational and qualitative methods in their investigation of local online "clusters of prestige," they argue that what is celebrated in the local social order of the "Bible belt" town – specific locations, leisure activities, family structures, religious identities – is also celebrated as prestigious and worthy of esteem across local social media clusters, so that at this site Instagram tends to reinforce the social norms and hierarchies of the pre-existing social order.

In Chapter 5 Sagorika Singha also examines the interplay of small-town dynamics with digital media – in this case the media product being not an Instagram profile but the memes produced by local online "meme collectives" in the Indian region of Assam. Memes have revived and fuelled old sentiments, providing a platform from which users seek to influence regional and national politics through insider jokes and shaming practices playing on pop-cultural references. Singha explores the ambivalence of the effects of these meme collectives on the distribution of power in local social and political relations. On the one hand they demonstrate a decentralisation of "agenda-setting" media power away from educated urban elites in the major cities, empowering "small town" media users. As Singha comments "if designed and received in the right way, a meme has immense potentialities to provide required value to a cause."

But on the other hand, those so empowered tend to be amongst the more privileged within local communities. Furthermore, because "memes often appropriate dominant discourses" their use can (and does, in the examples discussed) reproduce and more widely disseminate gendered, stereotyping and populist discourses. The dynamic here seems to be similar to that observed by Rajagopal (2001) of the national televising religious epics on Indian politics two decades before: new publics are created and new participants become politically engaged, but at the cost of increased "confusion" of public discourse and polarisation of political life (ibid., p. 279; see also Herbert, 2003, pp. 112–3). Social media thus reconfigures the social order in ways that allow for the emergence of new political actors, but which tends to reinforce existing social hierarchies, prejudices and divisions.

In Chapter 6 Liat Berdugo also investigates the interaction between networked digital media and contested local spaces, in this case between Israeli security services and Palestinian and Israeli activists in contested territories on the West Bank, examining the liminal space that emerges between physical and mediatised conflict. Specifically, she investigates what happens when camera recordings of civic injustices in Palestine are met with camera recordings of the resistance – a "struggle for spectral power," and the implicit recognition from all involved parties of the force carried by the online spectacle, as activists' videos are posted on their website for a potentially global audience.

Berdugo's analysis highlights gendered aspects of the conflict, especially between the "vision of and visual documentation produced by Palestinian women" and that produced by "a special unit of exclusively female soldiers called *Tatzpitaniot* ('The Watchers')," whose "sole job (is) to watch live video streams of the IDF's network of 1,700 security cameras mounted along key sites in the West Bank and Gaza." Berdugo argues that the framing of the activities of the *Tatzpitaniot* in the Israeli Defence Forces' promotional material as defenders of Israel's borders draws on an essentialised construction of Israeli women as defenders of "the land," against which Palestinian counter-surveillance is positioned as a threat. In response, Berdugo invites readers to view the conflict through and "an ecofeminist lens, which demands a consideration of how the land and women – as feminized witnesses – have both been historically subjugated by a shared history of oppression" and for "a sightline that celebrates disobedient, insurgent ways of looking: ways that visibilise the very frame of an image as a means towards new kinds of resistance in conflict." Here, the use of networked digital technologies crystallizes around and, in some ways, amplifies existing divisions, with the IDF mobilising tropes of the female guardians of the land reaching back to pioneering days, and activists invoking the potential presence of a global audience for their videos in an attempt to redress the power balance. Yet Berdugo also points to the possibility of the invocation of a shared narrative and way of viewing which moves beyond binary constructs of "us" and "them."

Berdugo's analysis of the gender dimensions of the "competitive videography" may also be used to suggest another perspective on Amsterdam's Instagram influencers (Boy & Uitermark, Chapter 3). Both their leading interviewees are women, reflecting a tendency for women in their 20s and 30s to predominate among the most followed accounts on Instagram in Amsterdam (Boy & Uitermark, 2017, p. 617). It may be that the Instagram platform, with its emphasis on the curation and projection of stylish and glamorous self-images, is one that skilled female users are amongst those most adept at exploiting to attract followers. However, as Boy and Uitermark's analysis clearly brings out, this attention comes at the price of constant effort to maintain the image, articulating the precarity of a profile based on gendered criteria of glamour and attractiveness, and in a context in which the women's working lives are precarious, dependent on their capacity to keep drawing "the right crowd" to the venues they promote.

A different downside of gendered public attention to women on social media is exposed by Soberaj and Merchant's study of Twitter mentions of US legislators (Chapter 7), and their analysis takes on an intersectional dimension as it focuses particularly on the abuse directed at female legislators of colour in the US. Their analysis reveals that different users inhabit different digital worlds, in that online shaming and harassment affects female legislators and legislators of colour far more than their white (and/or) male colleagues. In a world where social media are crucial to the outcome of democratic elections, politicians are required to display themselves on several platforms, yet exposure can have huge personal costs for female candidates, especially women of colour. Here social media serves to amplify prejudices and power imbalances within the wider social order.

Our second American case study is provided by Wanless and Berk (Chapter 8), who show how social media users are "drawn into and participating in the creation and spread of persuasive messaging," something they call an "enhanced form of propaganda" that is increasingly invasive in nature. Using data from the 2016 US presidential campaign, they examine the "organised deployment" of techniques of "participatory propaganda" to greatly amplify the impact of political messaging by harnessing the trust people place in social networks, including giving credibility to messages of questionable provenance and veracity – and indeed whose provenance has been deliberately obscured. From the propogandists' viewpoint the method is particularly efficient, as rather than requiring the constant funding of political advertisements via mass media, "participatory propaganda campaigns run as long as the cause driving it matters to its members – or rather, those administrating such groups are able to produce content that engages and provokes followers." Social media here present a threat to social order rooted in democratic debate, because the social basis of trust in networks is used by participatory propagandists to mask the false basis of claims, resulting in systematically distorted public communication. How then to resist self-fuelling participatory propaganda? Wanless and Berk argue that "finding ways to identify and measure engagement within these networks to understand the driving rationale, as oppose to blocking them, should be a priority for those studying liberal democracies."

The overall construction of the collection thus follows an arc from outlining the general properties of a datafied social order through a series of actor-focused perspectives which examine how roles structured by social media are performed at various sites located in European cities, entangled in contested Middle Eastern borders, and embedded in provincial Indian small-town networks, then back to an increasing focus on the general properties of social media networks revealed through our American cases, while not forgetting the human costs for the recipients of abuse (legislators of colour) or the political costs of participatory propaganda for a deliberative understanding of democracy.

Our accounts will emphasise how the principle of differential treatment embedded in a datafied social order is becoming increasingly widespread across social fields (Chapter 2), and examine a series of cases in which social media is clearly implicated in the reshaping social order in ways which align with the principle: where social media creates new precarious hierarchies of esteem maintained at high cost by those at their pinnacle (Chapter 3), reinforces existing hierarchies (Chapter 4), includes a broader range of participants in political discourse but at the expense of reinforcing local hierarchies and dominant discourses (Chapter 5), reinforces gendered constructions of national identity (Chapter 6), amplifies the abuse received by women and people of colour in leadership positions (Chapter 7) and enmeshes users in the circulation of propaganda which resonates with their preconceptions, deepening societal polarization (Chapter 8).

This arc then, as Couldry writes, will alert readers "to what is under way," but what of the potential for resistance by social actors, to help readers "imagine what resistance might feel like"? Writing this introduction in late 2020, it is hard to miss the powerful global examples of #*MeToo* and #*BlackLivesMatter* as cases of digitally networked global movements which have been able to provide sustained foci of resistance to abuses of power, particularly to the gendered and racialised forms of abuse identified by Soberaj and Merchant (Chapter 7). The volume lacks case studies on this scale, but rather identifies scattered sites from which such resistance could be mounted, in the spirit of Zygmunt Bauman's contention that the purpose of sociology is to imagine alternative futures (Bordoni, 2016, p. 283).

These "fractured spaces" range from the personal awareness shown by the key witnesses in Boy and Uitermark's account of Amsterdam Instagrammers, to Berdugo's eco-feminist standpoint and Wanless and Berk's appeal for more research to understand the "driving rationale" for engagement with participatory propaganda networks. The first group articulate a discontent with their lifestyle and a desire to move to more secure employment, suggesting that even those most personally invested in the platform maintain a critical and external perspective which allows them to imagine how things might be otherwise. The second constructs a vantage point from which deeply divided Israeli and Palestinian women can be seen to share a common history of oppression, while the third articulates the desire to understand in order to empower people to choose to change; a good vantage point from which to launch into the collection.

#### **References**

Bauman, Z. (1990). Modernity and Ambivalence. *Theory, Culture & Society*, 7(2–3), 143–169.


Boy, J. D. & Uitermark, J. (2017). Reassembling the city through Instagram. *Transactions – Institute of British Geographers, 42*(4), 612–624.

Boy, J. D. & Uitermark, J. (2019). Theorizing Social Media with Elias: Status Displays, Mutual Monitoring, and the Genesis of New Sensibilities. *SocArXiv*. July 11. doi:10.31235/osf.io/phm5x.

Castells, M. (2012). *Networks of Outrage and Hope*. Cambridge: Polity.

Cottle, S. (2006). *Mediatized Conflicts*. Buckingham: Open University Press.

Couldry, N. & van Dijck, J. (2015). Researching Social Media as if the Social Mattered. *Social Media + Society*, *1*(2), 1–7.

Elias, N. (2000). *The civilizing process: Sociogenetic and psychogenetic investigations* (E. Jephcott, Trans. Revised ed). Oxford: Blackwell Publishers.

Habermas, J. (1989/1968). *The Structural Transformation of the Public Sphere*. Cambridge: Cambridge Univresity Press.

Hechter, M. & Horne, C. (2009). *Theories of Social Order.* Stanford, CA: Standford University Press.

Herbert, D. (2020). Social Media and Spatial Justice: Instagram, Status Competition and the Deepening of Urban Exclusion in Northern Europe. In Watson, S. (Ed.), *Spatial Justice in the City*  (pp. 7–25). London: Routledge.


## **2 The Social Construction of Reality – Really!**

#### Nick Couldry

Today, in countless diverse places, a new project is under way: the project of transforming the social fabric so it becomes amenable to a new kind of rule at a distance, rule through data. A new vision of social governance is emerging in many varied forms but based around a convergent principle; that the stuff of social life, its every element, will, indeed must, be reconfigured in ways that enable the extraction of data and so new regimes of governance and value extraction. Under these circumstances, we cannot think about media's role in society (let alone social media's role) without drawing on social theory. But which social theory exactly?

The societies we inhabit today are not societies *in the same way* as the societies of fifteen, perhaps even ten, years ago. They are different types of configuration that operate on different scales and through different flows from the societies of earlier eras. Indeed, because every point in space and time now, in principle, embeds a two-way computer connection (for influence and surveillance), the non-linear patterns of social relations operate in many more dimensions than our old models of social interaction can account for. The result is new forms of technological, institutional and social power that we have barely begun to characterise, although we already know one thing for sure: that they will depend on the continuous multi-scalar tracking of human life by technologies. We can also predict with confidence that those forms, taken together, will make possible *a new type of social order*, massively more aligned to corporate goals and economic ends than previous social forms.

We need to understand this emerging social order; indeed, it makes no sense to analyse social media without theorising the underlying "social order" of which their "socialness" is part. This affects the theorists we choose as allies.

The book that follows contains many finely detailed analyses of the uneven and often strange surfaces and conflicts that characterise this new social order: conflicts over identity and status, battles for political capital, personal attacks and strategies of self-defence, the shifting patterns of contention and self-advancement in urban space. They give the reader a vivid sense of the unsettling uncertainty of the datafied social world in locations as varied as Assam and Israel, Holland and the US.

In this opening essay, I want to strike a bass note that, I hope, will resonate through what follows. I would like to capture a few more general features of the type of social order that is emerging through "datafication" (Van Dijck, 2014), when every aspect of life is under pressure to reconfigure itself in ways that allow the extraction of data. It goes without saying that this is the core rationale behind the emergence of social media platforms, even if this is never their stated mission. But – and this is important to note – datafication as a process goes much wider than social media platforms. It is transforming other areas of business such as logistics, education and health, and it is changing business models of all sorts. That said, social media is a good place to focus analysis.

I will stop short of offering a specific theory of the new social order. There are, to be sure, rival theorisations here, and I have a stake in that debate (see Couldry & Mejías, 2019; Srnicek, 2017; Zuboff, 2019). But, whatever specific theory one prefers, what matters more here is to get a broader sense of key lines of change; lines that will help, I hope, bring into focus the details of the following chapters as you read them.

#### **2.1 The Question of Social Order: Returning to Elias**

The concept of "social order" has gone out of fashion. The last major book on the topic was written a quarter of a century ago and it is now out of print (Wrong, 1994): even that book complained of the topic's neglect. Meanwhile, and less directly, much has been said about technology's role in engineering the social world, in the wake particularly of Actor Network Theory. Indeed, a whole way of thinking about the social order through technology has developed that involves, in a sense, *not* thinking about it (Latour, 2005), and rejecting the whole tradition from Durkheim onwards of analysing the emergence of social facts from human beings' efforts to construct social reality.

There are, for sure, serious problems with versions of "social construction" that paid little attention to media technologies (Berger & Luckmann, 1966). But, as I have argued elsewhere (Couldry & Hepp, 2016), these can be addressed without abandoning entirely the question of what social world emerges from processes of social construction. Indeed, at a time when, through datafication, it appears corporations and governments are intent on reconstructing social reality in ways that align with their interests, it becomes vital to pay close attention to processes of social construction. What if those processes are not metaphorical – as the term "social construction" often seems to be – but literal: actual processes of *building a different material basis* on which humans' lives together can be configured?

Some remarkable intuitions of the future expansion of corporate power and rule emerged three decades ago in the wake of Foucault's extraordinary insights into neoliberalism. I am thinking of Deleuze's short essay on "the control society" (Deleuze, 1997), on which much has been written. That essay, however, remains vague about how its predictions will be actualized, and it cannot possibly have been based on a prediction of social media platforms, which were not predicted even by the engineers closest to the developments from which the internet developed and the world wide web was invented. So, we must look elsewhere for a theory of *how* today's social world is being transformed.

The place to look is not commentary predicting the emergence of social media themselves, for the issue is not about technology or software as such, but instead an earlier tradition of thinking about social order itself that remains extremely insightful to this day. I mean the work from the 1970s and 1980s and even before of the German sociologist Norbert Elias. Particularly useful is Elias's idea that complexity in social life emerges from interconnections between human beings, from the patterns of interaction that he calls "figurations" (see Couldry & Hepp, 2016).

Figurations for Elias are "processes of social interweaving" that have a "special kind of order" that "starts … from the connections, the relationships, and works … out from there to the elements involved in them" (Elias, 1978, p. 116). His most simple example is a game of cards or football or a dance in which everyone plays their part by being in relations with each other person playing. As he says, "the behaviour of many separate people intermeshes to form interwoven structures" (1978, p. 132). In Elias's approach to social order, by contrast with Durkheim's, two things are very important. First, he grasps the role that *material* infrastructures play: today, that means software, computer code, servers for storing data, the cloud. But second, Elias insists on thinking about the consequences of that material infrastructure, of technology, from the point of view of the human beings entangled within them and their *human*  goals. This was a point Elias made eloquently towards the end of his life: "People often seem deliberately to forget that social developments have to do with changes in human interdependence […]. If no consideration is given to *what happens to people* in the course of social change – changes in figurations composed of people – then any scientific effort might as well be spared" (Elias, 1978, p. 172, added emphasis).

There are many worries today about the role of social media platforms in politics, in government, in family life, in the lives of children, concerns that run through this volume. Those worries are important, but they do not get to the most important issues that datafication raises for social space and power: the problem of how social order is being put together. Right now, that debate is emerging in multiple places.

Think of the intense debate today in the US about the consequences of automatic data collection on the lives of the poor, which in the US disproportionately means the lives of black people. As the legal theorist Patricia Williams (2019) recently commented, "many of us imprison ourselves with ... technology by choice – the smart watches we wear on our wrists, the GPS tracking on our cell phones or car-location apps, the ... reassurances of Siri. They aren't perceived as disciplinary tools; instead they are marketed as ways to connect." Yet that, she suggests, is what they are: disciplinary tools of social order operating through processes of datafication. The most dramatic example of this new vision of social order through connection, through datafication, comes from China. In China there are the most socially integrated digital platforms: in effect "super-platforms" like Alibaba or Tencent which combine social media (something like Facebook and Twitter and WhatsApp), with sites for e-commerce (like Amazon) and with sites for personal finance. Unlike in the West, none of these platforms are securely encrypted and the government has a close relation with the owners of those platforms (it helped finance their building). Think of China's emerging "social credit system," whose general framework and key mechanisms were established by 2020 and which gives a score to every citizen depending on the data gathered about them online –their score for social responsibility.

In an important policy document outlining this new system, the Chinese government used an interesting phrase to describe its significance: "a market improvement of the social and economic order" (China Copyright and Media, 2014). So, we are brought back here to the question of social order, but this time not as a theoretical concept, but as vision of government, a practical plan for the management of society. A vision that, for the US, Patricia Collins (2019) goes so far as to call "the civic practice of nothing less than totalitarianism."

What these two rather dramatic perspectives on the forms of social order emerging in datafied societies – the US and China – bring out is that the concept of social order is not an academic detail, still less a theoretical extravagance, but rather a highly practical term for registering how social worlds are, on the largest scale, being transformed through the interlinking of all we do on social media (Elias's insight), and the systems of social governance, explicit and implicit, that are emerging to manage them.

We cannot, in other words, *do without* a concept of social order. In the next section, I want to address the question of what is distinctive about a datafied social order.

#### **2.2 Managing Social Order Through Data**

It might, initially, seem implausible to argue that a new type of social order is being made through the processing of data. This is, indeed, an extraordinary and epochal development. In explaining this a little further, let's remember Elias's key insight that new norms and what he called "social pressure" (Elias, 1987, p. 145) emerge without anyone exactly intending them, as a complex and, if you like, higher-dimensional side-effect of countless individual, group and institutional actors doing what they intend to do. The "special kind of order" Elias was interested in emerges *through*  the continuous interweaving of many interrelationships and connections, their progressive impact, as social actors try, successfully or otherwise, to live their lives *through* the web of interrelations in which they have largely no choice but to be entangled.

In Elias's view, there is absolutely no need to imagine a vast corporate conspiracy to build something like a social order. It would indeed be deeply implausible to claim that what is emerging today through data processes was all, from the outset, planned to occur in particular boardrooms (so, for example, Zuboff's (2019) account of the emergence of surveillance capitalism allows for plenty of contingency along the way, as for example when Google discovered its ability to predict human activities in great detail from crunching the vast datasets about people's online activities that it had, indeed, intended to amass to fuel its search algorithm). Acknowledging that, however, is very different from denying that corporate intention becomes a major factor once the advantages of an emerging datafied social order become clearer to corporations. We are now more than a decade into the era of social media and taken-for-granted fast internet connection in many parts of the world, and we are now in a very different phase in the evolution of datafied social orders.

Why in particular, you might ask, should new ways of collecting data generate a new type of social order? Data, of course, has always been collected by governments, though on a massively smaller scale and intensity than today's everyday forms of corporate data collection. The link between data and social order derives not so much from the collection of data as from its use. As Oscar Gandy (1993, p. 15), a pioneer of research into corporate data collection back in the 1980s, noted, the *point* of gathering data is to make discriminations, to treat this entity or person differently from that entity or person. In the early states of datafication, it was discriminations between the customers of credit card companies and airline companies, to offer differential pricing for linked purchases. But now the principle of discrimination-through-data (that is, automatically harvested data from online activities) has spread right across the social terrain, including the actions of government.

Also spreading fast is the principle that institutional *rationality* now depends on the continuous gathering of data on human subjects so more discriminations can be made about them. Although this will vary depending on employee status (with higher status jobs being likely to be exposed *less* to continuous surveillance), many jobs today involve not just regular monitoring of key outputs, but continuous tracking of every dimension of an employment's activities and the generation from this of data-driven interventions to modulate and regulate the employee's behaviour. Similar principles apply to governments' regulation of those citizens who are dependent on state benefits or are in other ways subject to close management (for a useful recent survey, see Sánchez-Monedero & Dencik, 2019). Once again, the point of data gathering is not just to gather data, but to continuously manage behaviour through data uses.

While there might *seem* to be a huge distance between the explicit regulatory intent of state authorities and the business models of social media platforms that track their consumers, what is emerging across many different social and economic domains is a shared rationality of changing behaviour through data, based on continuous tracking and the constant modulation of signals and incentives.

The language of marketers is an instructive entry-point to this emerging rationale for the social order of datafied societies. Listen, for example, to Price Waterhouse Coopers (2014) speculating about a future where consumers will wear embedded chips that monitor their bodily and psychic mood continuously: "brands could even tap body cues to tailor messages ... sensor revealing that you're thirsty? Here's a coupon for smart water." Or listen to AT Kearney (2014), leading consultants in the insurance industry, commenting on the advantages of the so-called Internet of Things not for consumers, but for *insurers*: insurers, they say, could "use IoT*enriched* relationships to connect more holistically to customers and influence their behaviors." These are not random remarks, but early signs of a shift in business rationalities summed up recently in a report by Wharton Business school professors in the authoritative *Harvard Business Review,* which recommended the adoption of "four effective connected strategies, each of which moves beyond traditional modes of customer interaction and represents a fundamentally new business model. We call them respond to desire, curated offering, coach behavior, and automatic execution" (Siggelkow & Terwiesch, 2019).

Once again, we are looking less here at a conspiracy to dominate and influence and more at a rationality for ordering the world. Which leaves a final question: how do those tracked participate in this order? The short answer is that they are induced in countless different ways to enter into what Ulises Mejías and I call "data relations," relations that configure social life on a basis designed to optimise the generation and extraction of valuable data (Couldry & Mejías, 2019, Chapter 1). Data relations can be seen as a leading form of "figuration" in Elias's term for the era of datafication.

Such data relations can, however, have many different gradations: from the definitely voluntary choice of someone who wants to track their fitness or health and so uses a tracking app which sends data to a third party; the less voluntary "choice" of an employee encouraged by her employer to use a health app as a condition of obtaining work-related health insurance; the definitely *not* voluntary decision of welfare claimants or borrowers whose activities online are comprehensively tracked by state agencies or finance companies to generate a stream of information on their reliability and creditworthiness. And then there is the barely voluntary submission to tracking of users of social media platforms who may calculate that the seeming necessity of being on the platform such as Facebook – so as to connect with everyone else who is there for various practical purposes – makes it worth paying the price of being tracked by the platform. What is much less clear is how platform users calculate the bargain in relation to all the other uses of data relating to them gathered by and through Facebook and used by multiple parties unknown to them.

In these various ways, data relations work and the forms of power established through them come to stick. As we enter into ever more intermeshing data relations, something like a social order emerges.

#### **2.3 Conclusion**

In this short essay, I have only been able to give a bare outline of how a social order is being built through data processing, an order that is gathering a depth and intensity unrivalled by previous forms of governance and social power.

It is worth emphasising, however, that this order remains under construction; it is certainly not complete. Its eventual shape cannot at this stage be fully predicted. Indeed, to the extent that this order is resisted by social actors, it will not unfold exactly as I outline here. That is the point of analysing datafied societies from the point of view of social order; to alert readers to what is under way and help them imagine what resistance might feel like. This demonstrates again that the concept of social order is not a trivial addition to social theory, but a practical tool for analysing the world that corporations and governments are building for us.

In this essay, however, I have chosen to focus on general features of the emerging datafied social order, abstracting from the local complexities and tensions that are the subject of the essays which follow. A more complete picture can, of course, only be formed by reading them all together.

#### **References**


China Copyright and Media (2014). *Planning Outline for the Construction of a Social Credit System (2014-2020).* As last modified April 25, 2015. https://chinacopyrightandmedia.wordpress.

com/2014/06/14/planning-outline-for-the-construction-of-a-social-credit-system-2014-2020/ Couldry, N. & Hepp, A. (2016). *The Mediated Construction of Reality.* Cambridge: Polity.

Couldry, N. & Mejías, U. A. (2019). *The Costs of Connection.* Stanford: Stanford University Press.

Deleuze, G. (1997). Postscript on Control Societies. In *Negotiations* (pp. 177–182)*.* New York:

Columbia University Press.


## **3 The Dramaturgy of Social Media: Platform Ecology, Uneven Networks, and the Myth of the Self**

#### John D. Boy and Justus Uitermark

For good reason, critical observers of the contemporary digital media landscape have grown tired of Mark Zuckerberg's contention that there is something inherently good about "connection." Even when faced with overwhelming criticism, the Facebook founder and CEO has stuck to his claim, which seemingly counterbalances all evils the company might be accused of, that Facebook facilitates building "community." This is not surprising, since more connection means more users, means more revenue, means more profit for the company and its shareholders. But if we take Zuckerberg at his word, he sees connection as valuable because it shores up social ties, which have been shown to be beneficial to individuals in many ways; from lowering blood pressure to making it easier to find a job. In other words, Facebook and other platforms like it simply add more of a good thing to people's lives.

One reason this simple formula fails to convince is the growing recognition that social media are not simply *added* to what is already there: the institutions, interactions, and effects of everyday life. They have gradually come to permeate everyday life, and this interweaving has left the fabric itself changed. In other words, social media have become *ecological* (Postman, 1998).

For their part, sociologists have long asked whether online social ties develop at the expense of offline social ties. When it was found that this wasn't the case, they were reassured. Now we view the correlation between online and offline social ties somewhat differently. We no longer think of them as additive or interchangeable, but as *mutually transformative*. This changes the verdict. It means the difference they make is no longer linear. Instead, they effect a broad qualitative shift that isn't confined to any particular area but has reverberations throughout the social environment. This has consequences for how we study them. Although platforms have specific cultures, which is the stuff of media studies, there's also much to be said for understanding platforms as sites within a larger ecology.

Hoping to develop an understanding of how such ecological processes "scale down" (Breiger, 2015) to the level of daily life, we turn to the stories of two informants we met in the course of our research whose social position as "very online" women means that the platform ecology is particularly relevant to how their daily lives unfold. Their experiences give us some points of reference to evaluate how well the tools of the sociology of everyday life, particularly dramaturgical analysis, are suited to our task. Dramaturgical concepts, which study the self in interaction, are pervasive in the extant literature on digital media. It almost seems like Erving Goffman, the founding figure of dramaturgical analysis, has experienced a rebirth as an analyst of online life. Even so, our assessment of the dramaturgical approach is mixed. While on the one hand it helps us bring some of the cross-pressures that shape experiences within the platform ecology into view, it can also lead to an overly narrow perspective. That is especially the case when the self is merely taken as a strategic project subject to deception and manipulation by individuals. In such cases, the dramaturgical perspective obscures the "colonization of everyday life" (a term coined by Guy Debord) by platforms circumscribing what constitutes idiomatic identity displays and establishing a hierarchy among them. Insofar as dramaturgical perspectives account not just for individual strategies but also for normative pressures and systematic interdependencies, they provide a useful tool for making sense of platform ecologies.

#### **3.1 Very Online**

Throughout our discussion, we draw on some findings from five years of studying Instagram in Amsterdam using both computational and qualitative methods. Early on in our research on social media in urban life, we identified *uneven networks* as one of the defining features (Boy & Uitermark, 2017). For an impression, consider that among the roughly 80,000 Amsterdam Instagram users we included in a computational analysis, the top 50 users (0.06%) received more likes and comments between them than the bottom 30,000 (38.4%). The Gini coefficient, commonly used as a measure of income inequality, when applied to the distribution of attention comes to a stunning 0.76 – considerably higher than the Gini coefficient for South Africa's income distribution, the world's most unequal. We found this kind of unequal distribution fractally repeated at every level, from the city as a whole to individual neighborhoods and subgroups of users. Since we wanted to understand how personal appeal and self-exposure through platforms are experienced in everyday life, we figured that the most central users, those getting the most attention, offered a privileged window onto this world – kind of like how social scientists of another era studied the experiences of white-collar workers to understand the social alienation they saw as society's characteristic mentality. Or, to quote cyberpunk legend William Gibson, "The future is already here – it's just not evenly distributed" (quoted in Ratcliffe, 2016b).

Using a network metric, PageRank, that initially developed rank search results for the Google search engine, we could identify the hubs of the Instagram network. Not surprisingly given what we know about the "very online," a high number of these individuals were young women in their twenties and thirties. Many of them work in fashion, marketing, or entertainment – professions that give them facility in the visual idioms of micro-celebrity. The experiences of two women in particular shaped our thinking about the presence of Instagram in the practice of everyday life: Yvette and Sammy.1 We briefly introduce them here before returning to their experiences below in our discussion of dramaturgical perspectives on social media.

#### **3.1.1 The Insider: Yvette**

During the day, Loggerhead's is the kind of hamburger joint where you can get a burger with a truffle-glazed patty, but even the basic burger – Angus beef, lettuce, tomato, onions, slice of pickle – costs over ten euros. The restaurant, located on a busy market street, is small and unassuming and features the same unfinished wood paneling found in bars in gentrified neighborhoods the world over. At night, the back of the establishment is rumored to transform into a nightclub serving exquisite cocktails where you can dance till the early hours of the morning. Rumor also has it that access to the secret back room is only granted to those who know the password. Standing at the passageway between the burger joint and the extravagant speakeasy-style club is a door hostess. She's who you have to convince if you want to be part of what is happening in the back room. "Most people coming are wasted, that's why I call myself a 'door bitch.' I *try* to be polite," Yvette Legrand tells us, laughing.

In her early thirties, tall, with long blonde hair and wearing an elegant outfit, sometimes with a hint of butch, she cuts an imposing figure in the role of gatekeeper. But that's not her only role –not by far. On her Instagram profile, she calls herself an "Amsterdam insider," and her insider status derives not only from being able to decide who's in and who's out in the backroom at Loggerhead's. Yvette became an insider by entering the city's party scene and getting to know a lot of people not long after her thirtieth birthday. In a matter of just two years, she became highly connected among the movers and shakers of the city's nightlife. In her various other roles – as party organizer, event producer, and photographer – she has put them on a guest list, partied with them, and, of course, taken their picture. There's no question that she belongs.

What ties together these various activities in Yvette's life is the crowd of people in which she moves. In her understanding, this crowd, or "scene," is defined by a shared style above all else. "My photography is a bit more fabulous," she explains. "Either you like it or you don't." Those who like her photography are those she encounters in other parts of her life as well. Because she lives in De Pijp, walking distance from Loggerhead's and most other central venues where the city's party scene congregates, she moves in a relatively small radius and mostly interacts with likeminded people. "I have a certain kind of people that I see in my life wherever I go. They're all in the same crowd. So it's pretty easy for me, because I know a lot of them. When I go somewhere with a camera, they jump in front of my camera. I don't have to ask them, 'Can you

**<sup>1</sup>** We use pseudonyms for both people and places to hide our research participants' identities.

please…?'" Similarly, when she's the one hosting a party, she doesn't have to work particularly hard to get people to turn out. They come, because Yvette's involvement all but guarantees the party will have the right vibe. Yvette is connected, and she is also recognized as somebody who knows the right kind of people. Her recognition in the scene revolves around her persona – Yvette Legrand is her *nom de guerre*, and she projects a seductive and glamorous image through her appearances in the city's party venues and online, chiefly Instagram.

Other events she participates in, such as art openings, are a way for her to extend her social circles and to investigate what appeals to people in other scenes. Rather than just draw in the people she already knows, she tries to mix up the composition of her crowd by assembling members of different groups. "That's the most important thing to keep on going," she explains. "To be interesting in the scene, it has to be surprising." Filling a venue is no challenge when you're as connected as Yvette is; the challenge is to put the right people together. If four out of five attendants are gay men, the women and straight men will be alienated and won't want to come back. Instead, you have to engage in an activity that Yvette likens to scouting: searching for the right kind of people that will fit in. Another smaller event with an attendance of 80 to 100 that Yvette started serves as her scouting grounds. Each person she personally invites is able to bring two to three other people. "It's the best cocktail you could imagine," she says enthusiastically. The event brings together gay men, well-dressed women, and straight men who say to themselves, "It's a bit gay, but I actually like this!" Mixing this cocktail is a matter of mobilizing the right social connections, but also of having the right communications strategy.

#### **3.1.2 The Workaholic: Sammy**

In the course of our research we also spoke to Sammy, another young woman thriving on social media. Sammy has *a lot* going on in her life. She runs a program to match brands and social media influencers, organizes talent scouts, is responsible for the creative management for several well-known artists, designs marketing campaigns for various big brands, hosts evenings at one of Amsterdam's best-known clubs, and manages a co-working space for creative entrepreneurs. She's a self-diagnosed workaholic: "I never stop. I work 24/7. Go! Go! Go! There is no such thing as a holiday, and we don't do breaks."

Not that she minds. She knows no greater pleasure than creating content for marketing campaigns. She loves working with her best friend and long-time collaborator, Cori. Unlike Sammy, who is more into marketing, Cori is a "real fashion girl." She is also wilder, younger, and taller, so she appeals to a different subset of people in the scene. When they're at events, the two of them stand out. With her height and looks, Cori is naturally the center of attention. Sammy all too modestly states that she can also attract looks if she dresses up. As a self-professed workaholic, Sammy abhors the distraction of exhibitions or social events, but she found out that she and Cori are at their best when they're going out. She checks out how other entrepreneurs organize events or promote brands (and invariably comes to the conclusion that they can do better) and gets into contact with business partners or recruits for her influencers program. As long as she attracts attention, she knows opportunities, and new people, will present themselves.

Take her recent foray into deejaying. She had been messing around with a turntable for a couple of weeks when she got her first booking. At the time of the interview, her deejay career was accelerating. That was before she had actually purchased any deejay equipment. When she explains why she is so successful, she mentions looks and connections. Social media are important for both. Sammy consciously crafts a stream of images to maintain visibility, so she's on people's minds, allowing her to outcompete aspiring deejays who have spent years honing their skills. Having an Instagram profile helps her maintain connections and make new ones.

#### **3.2 Dramaturgy: The Self in Interaction**

The Canadian-American sociologist Erving Goffman famously conceived of social life as a kind of theater that binds people together as spectators and performers. Starting from face-to-face interactions between two people, he analyzed social interaction through the lens of dramaturgy –the performances and displays we put on in our daily lives. His best known statement of this perspective, *The Presentation of Self in Everyday Life* (1959), proposes to think of people in social situations as forming a team engaged in a performance together as performers and audience, the aim of which is to put on a show that accords with prevailing norms concerning conduct and decorum.

Branaman (1997) points out that, over the course of his career, Goffman used different metaphors to develop his dramaturgical perspective. Aside from drama, he frequently spoke both in terms of *games* and in terms of *rituals*. In general, the game metaphor directs our attention to the individual as a strategic actor, while the ritual metaphor directs our attention to the social order and actors' displays of fealty to it. Drama encompasses aspects of both games and rituals. This can be seen in his concept of "face": we have no choice but to preserve our "face" through ritual practices, but it also is subject to strategic manipulation, giving face-work a game-like character. Goffman didn't mean for any of these metaphors to be hypostatized. Instead, he regarded them as "scaffolding" to aid understanding through defamiliarization. All concepts in Goffman are thus provisional. They are not intended to reinscribe the definition of the situation, but to allow interpreters of social life to get a conceptual grip. Goffman never hesitated to tear down his scaffolding to construct a new framework, and neither should we today.

In this section, we review some of the applications of Goffmanian dramaturgy in the extant literature on social media. References to Goffman in the literature on social media are common. Since the 1990s, social psychologists, communications scholars, and scholars of human-computer interaction (HCI) have adopted his work to analyze what people do online, and publications discussing social media citing Goffman and using his central concepts now number in the hundreds. While these appropriations of Goffman indicate some of the ways in which dramaturgy helps us understand the social media–social order nexus, they are not in themselves without shortcomings. Our review seeks to assess the existing literature and clarify how the perspective we develop here relates to existing perspectives.

#### **3.2.1 Self**

Posts and profiles on social media are meaningful because they are the expressive products of users who put their selves on display. This was already true of personal websites, blogs and other pre-Web 2.0 formats, but social media heightened the emphasis on the self as the locus of significance because of the insistence, in the words of Facebook founder Mark Zuckerberg, that "[y]ou have one identity … Having two identities for yourself is an example of a lack of integrity" (quoted in Ratcliffe, 2016a). Foremost among the Goffmanian concepts to enter the literature is, therefore, the "self": its production, presentation, and evaluation.

The literature on social media using dramaturgical frameworks has two main emphases in its discussion of the self: the self as strategic project, and the self as myth.2

**<sup>2</sup>** There is also a third emphasis in research by behavioralist scholars, particularly (cyber-) psychologists, who frequently emphasize obsessive or pathological preoccupations with the self. Self-presentation on social media is of interest to these scholars to the extent that it appears to be associated with the so-called Dark Triad of personality traits: machiavellism, narcissism, and psychopathy. Other psychological phenomena connected to social media self-presentation include perfectionism, social comparison, self-esteem and social anxiety. In this literature, Goffman is selectively appropriated for emphasizing the presentation of self qua impression management, but this appropriation is suspect for a least two reasons. First, while Goffman himself took an interest in questions of mental health, he famously spoke out against pathologizing behaviors (Goffman, 1961a). What psychiatry classified as symptomatic behavior, he labeled "situational improprieties." Second, though it is easy to take Goffman's description of social actors continually occupied with the work of managing impressions and saving face to be an indictment of the superficiality and alienation bred in capitalist societies (see Gouldner, 1971 for a famous example), Goffman's dramaturgy does not, in fact, describe a character *type*. To reduce self-presentation to a form of obsessive or neurotic conduct is to misrecognize Goffman's entire intellectual project. Thus, while there are superficial similarities between this literature and other scholarship on the social life of social media, it is worth bearing in mind that this strand of work asks fundamentally different kinds of questions. As we argue elsewhere (Boy & Uitermark, 2019), narcissism and related labels seem ill-suited to making sense of the draws of self-display. We are better suited following Goffman's (Goffman, 1959) lead. As he said, we should not take the whole world to be a stage, but we shouldn't assume where the line can be drawn either. We

#### **3.2.1.1 Self as Strategic Project**

A significant strand of literature is concerned with the "balancing act" or "double bind" of projecting an ideal self-image in online spaces while remaining authentic (Davis, 2014). The "ideal" or "edited" self that is put on display has to strike the right balance between aspirational and acceptable (Ellison, Heino, & Gibbs 2006; Marwick, 2014). Overreach, and you get shot down; don't reach high enough, and you settle in place. Such balancing acts become especially acute in cases where self-presentation becomes bound up with self-branding and the incessant economic demand to enhance one's human capital (Hearn, 2008; Duffy, 2017). Livelihoods depend on successful self-presentation. Scholarship has especially focused on how cultural producers such as fashion bloggers, models and live-streamers are drawn into this balancing act.

A related recurring theme is the difficulty of distinguishing between frontstage and backstage regions in mediated communication. Social contexts are easily blurred and frequently collapse (Marwick & boyd, 2010; boyd, 2014). This happens, for instance, when parents friend their children on Facebook, only to discover a whole side of them that wasn't meant for them to see. Writing about Instagram influencers, Leaver, Highfield, and Abidin (2020) note that performers purposefully play with audience expectations of greater authenticity in backstage areas. Another example is when patrons supporting cultural producers financially through Patreon3 are promised access to unedited footage as a reward for their sponsorship. Users of social media are driven to be extra vigilant and strategic in crafting self-images in accordance with various imagined audiences (Dijck, 2013). This intensifies the game-like character of the online social drama.

These appropriations of Goffman are in line with his reliance on the game metaphor. Goffman had a lot to say about what he called "expression games" – social gambles to enhance one's status. Examples of these kind of strategic interaction abound in his work, though some of the most memorable ones deal with espionage and political intrigue (Goffman, 1969). Spies have to play an extreme version of the expression game as they strategically craft an identity allowing them to carry out their intelligence-gathering work. This requires them to practice a great amount of restraint, such as concealing or controlling emotions, even as they engage in fateful action.4 Overwhelmingly, dramaturgical perspectives on social media emphasize

especially should not assume it can be drawn around inherently pathological behaviors or traits (see also Warfield, Cambre, & Abidin, 2016).

**<sup>3</sup>** Patreon is a digital platform through which subscribers can access content produced by artists and performers.

**<sup>4</sup>** In Goffman's work, "fatefulness" implies that something may "spill over into the rest of life" (Goffman, 1967, p. 162). A gambler wants the situation at the gambling table to transform their life into one of affluence. Similarly, aspirational social media users may want their online personae to be a pathway to a better livelihood. An excellent literary depiction of such a fateful transformation can

the game-like, strategic nature of online sociability. Having and building an online presence is a gamble taken in hope of sprucing up one's social status.

The game-like cast of online interaction is a reflection of the competitive environment bred by social media which brings strategic considerations to the fore. Social media feature publicly visible personae, tools for reward and appraisal, and multiple clear indicators to gauge the impression made on others. Combined, these features amount to a perfect storm. They conspire to make impression management and self-presentation salient and ongoing concerns for those who maintain a public presence on social media.

Metrics can be an immaterial mark of distinction, and they can also translate into economic opportunities. But in absence of metrics, there are other ways of waging competition on social media. On Instagram (which has stopped displaying the number of likes on posts), appearing alongside celebrity users and being tagged by them is a coveted trophy (Abidin, 2018). On Twitter, the practice of "one-upping" or "dunking on" other users, of concocting "owns" and "burns" in response to others' status updates, is an almost unescapable part of the experience. All of this competitiveness is no doubt intensified by the reigning neoliberal meritocracy which "places a strong need to strive, perform, and achieve at the center of modern life" (quoted in Day, 2018).

Yvette and Sammy's experiences both speak to the importance of expression games. Looking through their carefully curated feeds, it was undeniable that they gave a lot of strategic thought to their appearance. Yvette likened this to being an artist on a stage, and she insisted that aside from that, she is a "normal person" who attends small-town family gatherings as well. She finds it important to not come over as too icy and unapproachable, so she tries to smile and not take herself too seriously. In discussing this with us, Yvette revealed that she was aware of a tension between her desire, on the one hand, not to get too sucked into the frequently competitive game of self-presentation on social media and, on the other hand, managing the impression others have of her. Resolving this tension is a matter of careful reflection, which means her social media presence becomes a larger preoccupation than she'd like it to be. The game never ends.

Sammy saw this, too, and in fact she thrived on the excitement of always being on and always having to remain vigilant. Moving through different scenes at a dizzying pace, Sammy emphasizes she has to act fast, very fast, to keep up with developments and seize new opportunities. But you couldn't tell just from looking at her posts. There she seems to be kicking it. She is having a drink, doing a dance. She's acutely aware that the images are contrived. "Everything I put up there is tongue in cheek.

be found in Stagg (2016). Almost overnight, the protagonist Colleen rises to influencer status, and the resulting transformation of her life is whiplash-inducing.

You know *exactly* how you present yourself. You're constantly feeding. There's a sort of marketing narrative. It's ridiculous, but it's there."

#### **3.2.1.2 Self as Myth**

Strategic considerations are a big theme both in Goffman's oeuvre and the lives of the "very online," but they are not the end of the story. For his part, Goffman sharply rejected Georg Simmel's "embarrassing effort to treat sociability as a type of 'mere' play, sharply cut off from the entanglements of serious life" (Goffman, 1961b, p. 21). Understanding social life requires more than a study of individual, strategic action. It calls for understanding something that Goffman elsewhere called "interpersonal ritual" that revolves around the self as the last remaining sacred myth.

Scholars interested in gender display and other parts of Goffman's oeuvre concerned with the ritual dimension of social life understand that strategic action isn't everything. Baker and Walsh (2018) are recent examples, drawing on Goffman's *Gender Advertisements* (1979) to unpack the "ritual idiom" that underlies gender expression in "clean eating" (a sort of minimalist, healthy cuisine) posts on Instagram. They note that proper adherence to the ritual idiom requires observance of a gendered social hierarchy. In his study of advertising, Goffman observed that, almost without fail, women appear physically smaller in visual displays, and their lower positions accentuate the subordinate roles they play. Posed pictures are thus an indicator of the ritual idiom within a society. In their case study of "clean eating" posts, Baker and Walsh (2018) find that it is the idiom of hegemonic masculinity rather than "emphasized femininity" that structures the top displays. Muscle presentation and other displays of virility and strength serve to affirm membership in a community of lifestyle icons. Displays that fall back on a ritual idiom merely draw from a stock of existing display conventions. They are rituals that seek to make behavior legible *as ritual*; they are hyper-ritualized (see also Hancock & Garner, 2014).

This aspect of Goffman's work highlights that self-presentation is not merely strategic, but also a vehicle for establishing social bonds. These bonds revolve around "sacred" objects, namely selves, which participants in the interaction have no choice but to perform. As sacred objects, our selves are the myth we are collectively invested in, turning images of "epidermally bounded containers" (Goffman quoted in Wissinger, 2015) into meaningful social encounters. To the extent that social media are a vehicle of social integration (see Boy & Uitermark, 2020), the integration depends on the myth of the self.

To Goffman, the self is not a utility function, but a sacred object. The reasons we show it consideration are not simply "strategic" in the narrow sense of manipulative. Through our performances, we express deference to the social order through demeanor, propriety, and the display of our connections." Through these expressions, we accord proper levels of veneration to selves – our own and others' – with regard to their position in hierarchies of status. Importantly, showing proper veneration often means *avoiding* expression rather than engaging in strategic expression games. It's about what we hold back, not just what we put out there.

Despite the evident joy both derive from expression games, the strains of paying homage to the myth of the self are apparent in Yvette and Sammy's experiences. When we met Yvette, she was dressed down in yoga pants and looking forward to getting a good night's rest. For the next four days, she was booked, either as the "door bitch" at Loggerhead's, or photographing events. This meant multiple late nights coming up, and yoga and resting were Yvette's way of anticipating this exhausting stretch and maintaining a degree of balance in her life. Yvette expressed ambivalence about her ability to maintain balance. "I can do a lot," she told us. "But when it's my own party and I need to be there till five o'clock in the morning, sometimes I really don't feel like it." She has to be the last one standing, not just because she bears responsibility as the organizer, but also because she has to keep everyone's spirits up till the end.

Yvette is feeling the strains of playing her role. "I'm always watching and they're always watching me. … You can see me in social media, being in front of everything. I want to be more in the background. It's pretty exhausting." Her way out is not an exit, but a pivot. She aspires to become an entrepreneur whose involvement in urban scenes is not in the foreground, as an artist, but more behind the scenes. She hopes to achieve this by building a web-based guide to gay Amsterdam, advertising events and locations and providing booking and ticketing services, all from her unique insider perspective. She's built a prototype, but it hasn't really gotten beyond the stage of a passion project. Even so, Yvette hopes this venture will become a "second step" in her career as an insider, allowing her to capitalize on her connections and her persona without having to perform all the frontstage work the scene currently demands of her.

Her planned pivot holds the promise of an eventual reprieve from all this exhaustion. She thinks perhaps she'll even be able to hire somebody to share the load with her. Meanwhile, she has to achieve her hoped-for balance in other ways. Reading about spirituality and practicing yoga are part of her practice of self-care and have been since before she became a scenester. "I'm really into growth. Once in a while I really need to go inside myself for real, like working, healing," she explains when we ask about her interest in spirituality. "That's something I missed in the last two years, when I got really into this world, which is a lot of, of course, like, fake and tough and competition and whatever."

The sense that her world is permeated by fakeness and competition is another source of exhaustion, if not anxiety. Yvette feels a desire to steer against it, to be a beacon of authenticity in the fake world in which she moves. That's the reason she doesn't really like social media, and why she tries to restrict her use of it to building a portfolio of her various activities. Aside from Instagram, where she feels the need to be active, she also maintains a site on Tumblr, a Vimeo profile, and a Facebook Page promoting her work as a photographer. At the same time, because she is aware that everything on social media tends to "look so much more beautiful than it really is" – in fact, that's the whole point of social media as she understands it – she also wants to show another side of herself. She mentioned making an effort to share things from other parts of her life, such as taking an elderly acquaintance out for a walk or visiting her family in a small town in the south of the Netherlands.

Sammy also told us that she'd recently been looking to make a change. She was hunting for a job. "Not a nine-to-five job," she emphasized, but still something more "corporate." Her attempts so far had been unsuccessful. For one position, she spent an inordinate amount of time on the application. She put together a video clip with help from some friends. She took four days to write an application letter to get it absolutely right. And still, in the end, she didn't get the job. It's not that she can't do the work, because she's already doing it, and at a high level. It's just that, in this industry, short-term or one-time contracts are the norm. While it is exceedingly difficult to get a steady job, it is – at least for Sammy – remarkably easy to move from one gig to the next. Under these structural conditions of intense competition and short-term engagements, people like Sammy are successful as long as they maintain their reputations and keep up with their excruciating work schedules.

Sammy's carefully curated Instagram profile only captures the glamorous aspects of her brutal life-work regime. "I was working five days full-time on a day job. But you can't tell someone what they need to do and then not show up at night to see how they do it. It's not officially part of your job but you have to do it." At 32, she feels she's getting old. She works out in the gym and has a policy of not drinking alcohol or using drugs. "I look to London and New York for examples. Those people get up at six in the morning and get home around one. Sure, they're paid different kinds of amounts, but you have to persist if you want to get somewhere." Working from a canal house in Amsterdam's historic central district and shaping the image of big clubs and large corporations, she has achieved much more than most people in her industry can hope for, but she remains in a precarious position. Her devotion to her career is taking a toll. She suffers from spells of exhaustion. She shares much of the glamour and the suffering with her friends but she can't really relate to her family anymore. When we asked her whether it was difficult to explain what she's doing, she replied that they have stopped asking. In turn, she stopped going to family celebrations.

Bound up as it is with myriad strains and anxieties, self-presentation is not merely strategic, but also ritualized. That means that, while they can tailor their performances to different audiences and use all manner of strategies to improve their status, neither Yvette nor Sammy can avoid self-presentation or turn it off when the costs outrun the payoff.

#### **3.2.2 Interaction**

Dramaturgy is an interactionist theory. Following his Chicago School forebears in the symbolic-interactionist tradition, Goffman (1959) states that performances create a shared "definition of the situation" between performers and audiences. In other words, we inhabit a shared social world in which our roles and behaviors are meaningful to each other because of the displays we put on for one another. Every moment of our social existence is an accomplishment rather than a natural outgrowth of our human nature or the functional product of the social order. Both performers and audiences constantly have to affirm, confirm, or correct the impressions fostered in their shared drama. Understanding social life requires close attention to these ceaseless maneuvers. As a "sociology of occasions" (Goffman, 1967, p. 2), Goffman's dramaturgical approach is interested in the *syntax* of momentary relations rather than their specific content or longer-term structures. It's all about process.

Goffman thought that face-to-face interaction was the site *par excellence* to study this process. This wasn't because he deemed other interactions less meaningful; he mainly wanted to redeem the most basic forms of interaction from the condescension of mainstream social scientists dismissing them as trivial. Given this emphasis, it is no surprise that the literature takes differences between face-to-face and mediated interaction as occasion to go beyond Goffman (Cetina, 2009). The *context* of interaction becomes more complicated. "The presentation of self in everyday internet still corresponds to Goffman's playacting metaphor," writes Jia Tolentino (2019, p.14). "But the internet adds a host of other, nightmarish metaphorical structures: the mirror, the echo, the panopticon." While there is broad-based consensus in the literature that we need to study more than the face-to-face, there is no consensus on what that should look like.

#### **3.2.2.1 Interactive Affordances**

Overwhelmingly, the literature tries to grasp the other structures Tolentino (2019) refers to with the concept of *affordances*. These are the "possibilities for action" offered by the platform ecology (Norman, 1988). The study of affordances directs attention to features of social media as socio-technical systems shaping how they are used by both the producers and the consumers of these systems. While the work of building an online profile – profile-work – bears some comparison to the embodied rituals associated with face-work (Silfverberg, Liikkanen, & Lampinen, 2011), it is distinct in that it affords different levels of control over one's public image and requires different kinds of work.

boyd (2010) discusses four key affordances of social media: persistence, replicability, scalability, and searchability. These affordances crucially allow mediated performances to extend through time and space because they can be saved, spread, queried, and replayed. While users can and have shaped the affordances of social platforms – the hashtag being a well-known case in point –the literature stresses that product design is typically dictated by the twin demands for profitability and growth. Because business models are invariably advertisement-based, and because venture capitalists backing major tech ventures demand steep growth curves, social media affordances are optimized for interactions that bind user attention to screens for as long as possible, and that scale up indefinitely. They also try to squeeze as much data out of users as possible to target advertisements or strengthen their positions as data brokers. For that reason, Facebook users are encouraged to share their likes and dislikes in ever more detailed online profiles. These profiles are thus no longer (primarily) symbolic representations of selves, but inputs into larger algorithmic systems.

Despite this strong emphasis in the literature, our conversations with Yvette, Sammy and other participants do not suggest that varying affordances play a big role in their daily lives. Even though they navigate multiple platforms with different affordances and have seen platforms change their affordances, their practices and anxieties had little relation to the shifting socio-technical landscape.

#### **3.2.2.2 Network of Mediations**

Another way to broaden our perspective on the context of interaction is to regard it as a level within a larger social whole. Henri Lefebvre ([1961] 2002), an influential sociologist studying everyday life from a critical perspective, observes that there is no such thing as "immediate" interaction, because even face-to-face interaction is mediated by language. Thus, to understand the context of interaction, we must have a model of the larger "network of mediations" (Lefebvre [1961] 2002, 2:141) within which interaction plays out. These mediations, broadly understood as interdependencies linking the "micro" and the "macro," shape the strategies of self-presentation and shape the normative order that makes self-display compulsory.

Our research with Yvette, Sammy and other participants suggests that the endemic insecurity induced by flexible work conditions and precarious livelihoods plays an important mediating role. Neither the imperative to engage in expression games nor the normative demand to remain visible to others would be as acute if they had the option of "dropping out" for a while. This dependency thrusts the weight of the world on individuals, giving them no choice but to maintain the myth of the self through its ongoing reproduction in commodity form. This demand is contradictory because it gives the semblance of unfettered self-expression, which further dissimulates the normative demands of the situation. The result is that individuals accuse each other of being "fake" or engaging in insincere signaling behaviors rather than seeing the larger demands and interdependencies that are at play.

#### **3.3 Conclusion**

Our review of dramaturgical perspectives in the literature on social media leaves us with a mixed assessment. Particularly when Goffman's work is filtered through social psychology (a common reference is Leary & Kowalski, 1990), we find a version of dramaturgy that focuses chiefly on individuals and their strategic action. This is often associated with Goffman's game metaphor. Undoubtedly, the competitive environment of social media brings strategic considerations to the fore. Our research supports that view, which is widely shared by commentators. But our research also suggests that such an individualistic and rationalistic version of dramaturgy is illsuited to our task of making sense of the social media–social order nexus, since it works with a very thin conception of the social. The social only enters the picture to the extent that it requires individuals to engage in balancing acts. It does not account for the normative demands impinging on users.

Yvette and Sammy maintain a presence on social media as a way of claiming and maintaining membership in their respective worlds. Each has an idiom of displays that members must enact to be eligible as members. Yvette has to be seen partying, with a smile on her face, till the party is officially over. Sammy knows about the "marketing narrative" she has to adhere to in her scene. Though she finds it ridiculous, she knows better than to not follow along. Both have learned the pervasive idiom of microcelebrity, though both desire a slower, more muted life.

While we could see their presentations of self as expression games helping them to succeed, their stories make it clear that this doesn't do justice to what they are doing. In each case, the price paid seems out of proportion to the benefits reaped. Sammy in particular remains precarious. Both engage in coping strategies like yoga to mitigate the stress induced by the requirement to maintain decorum. They also envision more structural solutions that would allow them to rely less on ephemeral collaborations and shallow images. Yvette wants to do more backstage work; Sammy is looking for a corporate job. As they seek an out, they run into formidable obstacles. In their line of business, short-term engagements are the norm, which means they have to ceaselessly pay homage to the prevailing norms of decorum to remain in view of potential clients or collaborators.

Studying social media as a site of order requires more than an account of individual strategic action. Crucially, we need to see social media as a source of obligation, which requires us to study contradictions, emotions, and strains. We also need to scale up the analysis from face-to-face encounters. This necessitates not just accounting for the affordances of social media as socio-technical systems, as is common in the extant literature, but understanding the "network of mediations" that shapes the economy of display individuals are drawn into.

#### **References**

Abidin, C. (2018). *Internet Celebrity: Understanding Fame Online*. Bingley: Emerald.

Baker, S. A. & Walsh, M. J. (2018). 'Good Morning Fitfam': Top Posts, Hashtags and Gender Display on Instagram. *New Media & Society*, *20*(12), 4553–70. https://doi. org/10.1177/1461444818777514.

Boy, J. D. & Uitermark, J. (2017). Reassembling the City Through Instagram. *Transactions of the Institute of British Geographers*, *42*(2), 612–24. https://doi.org/10.1111/tran.12185.


Hancock, B. H. & Garner, R. (2014). Erving Goffman: Theorizing the Self in the Age of Advanced Consumer Capitalism. *Journal for the Theory of Social Behaviour*, *45*(2), 163–87. https://doi. org/10.1111/jtsb.12062.


## **4 Clusters of Prestige: Social Media and Social Order in the Norwegian Bible Belt**

Stefan Fisher-Høyrem and David Herbert

#### **4.1 Introduction**

This chapter uses data from the southern Norwegian city of Kristiansand (population 90,000), asking whether social media constitute a vector for challenging the local social order, or rather reproduce and reinforce existing established-outsider relations. We combine computational and qualitative methods, while drawing on analytical concepts from Norbert Elias' figurational sociology, whose primary concern is how social figurations are created and maintained through notions of prestige and shame, as well as models of behavior (Elias, 1978). Specifically, we utilise Elias' conceptualization of "established-outsider relations," which he proposed as "a universal human theme" of creating and maintaining social distinctions within the same community (Elias, 1994, p. xv).

For Elias, the fundamental characteristic of social figurations is the distinction between a dominant "established" group and other "outsider" groups; indeed, categories such as gender, race, or class he sees as secondary to this most fundamental social figuration (Elias, 1994, p. xviii).5 The superiority of the established group stems not primarily from numerical size or use of force – the group is often a numerical minority – but from a comparatively high degree of internal cohesion and extent of communal control (Elias, 1994, p. 86). Internal cohesion is created and maintained by certain collective norms "which are apt to induce the gratifying euphoria that goes with the consciousness of belonging to a group of higher value and with the complementary contempt for other groups" (Elias, 1994, p. xviii). Put another way, the two key elements in established-outsider figurations are *power* and *morality*, or rather, power reinforced by a sense of moral superiority (Elias, 1994, p. xv).

Elias's framework lends itself particularly well to analyses of social media, since they are constructed precisely to facilitate mutual recognition (or ignoring) between users competing for signs of prestige and seeking to reinforce their social status within particular networks and groups.

**<sup>5</sup>** For a discussion of Elias' theory of established-outsider relations and race, see e.g. (Stanley, 2016).

#### **4.2 Social Order in Kristiansand**

In the following, we propose that Kristiansand's social order includes a powerful 'established' network of businesses, politicians, and other influencers explicitly or implicitly affirming hetero-normativity, nuclear family values, sports and outdoor life, and a loose mix of 'conservative' Evangelical Christianity and mercantile interests. While this group might be numerically inferior, its norms are reproduced and made manifest across a range of social fields, such as the promotion of the city as a whole as well as new neighborhoods, the field of cultural production, and public rituals.

#### **4.2.1 Evangelical Church Culture**

Named the buckle of the Norwegian bible belt, Kristiansand is the administrative and economic center of the country's southernmost region – a region which for the past century has been known as a stronghold of conservative Evangelical Christianity. The close association between regional Christian mission ministries, free churches, and the mercantile middle and upper classes has molded the political, religious, and cultural landscape of the region and its coastal 'capital' for at least the past 120 years, and is well-known and accounted for by historians. (Abrahamsen, 2014; Seland, 2012, 2014). Even today, the region – but in particular Kristiansand – remains something of a national political anomaly in its comparatively strong support for the relatively conservative Christian Democratic Party (KrF), together with weaker support for the Labour Party (AP)(Røed, 2010). Many of the Christian families who, with the help of willing and generous brothers in the Christian associations, became wealthy during the late nineteenth and early twentieth centuries, still remain so today.6

Through the 1990s and onwards, the region and its municipal center has maintained a high level of religious activity according to almost all common indicators (Botvar et al., 2010; Botvar & Aagedal, 2002; Magnussen et al., 2013). 24% are religiously active at least once a month, compared to 7% nationwide. The region has a high proportion (15% in 2008) of people who feel some association with one of its many free Evangelical churches, compared to the rest of the country (3% in 2008) (Magnussen et al., 2013). The region has high levels of part-time work among women, and studies suggest an ideal of "the good mother" – working less in order to "spend time with her children" – remains a strong norm, to the point where nonworking mothers might sanction other mothers who "work too much," which again leads to full-time-working mothers reporting feeling like outsiders (Magnussen et al.,

**<sup>6</sup>** As Elias also points out, it is not uncommon to find a certain allusion to family bonds in studies of high- and upper middle-class communities, where networks of families often may operate as a powerful factor in the social stratification and structure of the local community (Elias, 1994, p. 3).

2005).There are more married couples with children than in other regions, and more married couples with many children (Ryen, 2002).

Other research has shown how in this particular region the most important factor explaining the lingering skepticism towards gender equality – even more than social background, age, education, or income levels – is the conservative Christianity found in the free churches and associations (Magnussen et al., 2005, 2013). So-called 'traditional men' in this region are more conservative than their peers in other parts of Norway: they believe in God; they tend to see religion as relevant and important on a daily basis; they support religious organizations, and they are particularly oriented around so-called 'traditional family values'. While on a national scale this group tends to score low on experience of contentment/happiness, in this region it scores the highest, enjoying higher social status, higher education, and a stronger financial situation (Ellingsen & Lilleaas, 2010).

Even religious people in the region who are "sporadically active" are more conservative than the same group nationwide, which some researchers have suggested "supports the hypothesis of a Christian conservative hegemony" (Halvorsen, 2004). Some have suggested that the effects of conservative Christianity extend beyond the "core congregation," so that conservative values (with regard to family structures, sexual norms, and so on) somehow spill out of the specific church contexts. While one might contest this by saying it ascribes too much power to what is arguably still a numerical minority, Elias' theory of established-outsider relations explains how the fact that key members of the "established" group might hold key positions in the community means that their norms come to influence people beyond their relatively close-knit circle, and despite their lacking numbers. Indeed, it has been shown that the moral ideals associated with this group remain strong even for people who have to some extent left the former religious faith; that is, the ideals continue to influence their perspective on what it might mean to be a "good person" and lead a "good life" (Henriksen & Repstad, 2005).

#### **4.2.2 Family Friendly**

From the 1930s, tourism gradually became one of the region's primary economic engines, with Kristiansand as a natural center point. During the first decades of the twentieth century, as artists began to explore the variations in the southern landscape, these traits gradually turned into "positive" stereotypes in the increasing promotion of *Sørlandet* (*the South land*) and Kristiansand as a tourist attraction associated with summer, seaside, and hospitable locals (Johnsen, 2010). The accusations that had been levelled at southerners during the decades of decline in the sail shipping industries – that they were slow, lacked initiative, and were generally politically incompetent – transformed into characteristics with more positive connotations: the stereotypical southerner (and Kristiansander) was now portrayed as relaxed, calm, and unassuming (Hundstad, 2008).

From the early 2000s, an image of positive homeliness took a particularly strong hold of the promotion of the city as an object of tourist consumption, mainly following the lead of the regional tourism and business cluster USUS's project *The Children's City*. Tourist brochures began showcasing sunny archipelagos, children laughing and swimming, and happy families visiting the local zoo and amusement park *Dyreparken*. In an interview, one former project leader at USUS described how this project gradually incentivized businesses not usually catering to families with children to find ways of doing so. She recounted the story of a local bar and bistro owner who, when hearing of the project, exclaimed "Oh God, if we don't accept children in here, we'll lose all of our summer business!"

Kristiansand's self-promotion can also be understood in relation to urban policy development in the field of cultural production, in particular the foundation *Cultiva* and its surrounding debates. In 2002, the municipality sold portions of its share in the local water power plant Agder Energi to a state-owned company, and set up the public foundation *Cultiva* with a NOK1.4 billion trust fund of which the returns are to be used to support the city's cultural economy. As the first and only of its kind in Norway, *Cultiva* emerged as "the principal cultural entrepreneur in the city of Kristiansand, seeking to play a pro-active role also in the financial and industrial sector[s]" (Lysgård & Tveiten, 2005, p. 498). As such, it became subject to extensive public debate, in particular regarding its stated purpose and underpinning ideology (Bille, 2013; Lysgård, 2013). In 2005, researchers Lysgård and Tveiten put it thus:

Kristiansand is a small town with a traditional bourgeoisie citizenship marked by a Nordic brand of Christian-democratic political values. Experimental art, avant-gardism and political critique have not been in the forefront of this culture. The moral-based and traditionalist concept of culture accepted as legitimate in the bourgeois and religious parts of the city's cultural life is, therefore, challenged by Cultiva, and has led to a discussion about how the money should be spent and by whom. This is actually a question about who should control the power of the cultural discourse in the city, the discourse about what is acceptable according to ethical and moral values. The broad, liberal and innovation-oriented concept of culture adapted by Cultiva where culture is meant to 'put economy to work' is challenging the more or less hegemonic traditionalist discourse of culture in the city of Kristiansand (Lysgård & Tveiten, 2005, p. 496).

At the time of writing, one long-time member of the Cultiva steering board is Kristiansand's vice mayor from the conservative Christian Democratic party. In 2009, another well-known conservative Christian politician from the Conservative Party became new chair of the board. Soon after, in April 2011 Cultiva surprisingly changed their official strategy to the following: "Kristiansand will become the leading city of culture for children in the Nordic countries. With this the city will become the best place for growing up and a city that children and their families alike must visit" (*About Cultiva*, 2017). Instead of having artists apply for funding (a process that had been criticized for being overly complex and oriented towards questions of economic profitability), the board would seek to initiate projects themselves, and then look for potential partners. The abrupt change spurred strong negative reactions from the cultural milieus in the city – like when more than 50 artists, critics, actors, and directors issued a joint statement protesting the change and demanding the board members' resignation – but at the time, these protests had little effect.

#### **4.2.3 Physical Activity**

Kristiansand's self-promotion and image-building is equally evident in the marketing of new neighborhoods. Perhaps the largest and most widely advertised new development since 2015 is the area of *Lauvåsen*. Primary targets are families with parents aged 35–45 with an average of 1.7 children. According to the project leader, a competing project run by another developer had already taken the slogan "The Children's Neighborhood," so they decided to make theirs "the Sportiest Neighborhood in Norway." Houses include extra storage space for sports equipment, and a new 5 km running track (parts of which will be fitted for parents walking with prams) is being prepared in the surrounding woods. Everyone moving to the neighborhood will receive free membership at the local gym.

In general, the population in Kristiansand appears to value physical activity and exercise. The city has seen impressive growth since the 1970s in the area of sports and outdoor life. While this has been a general development on a national level, historians have pointed out that in the southern region one important factor was the free churches' change in attitude, from viewing sports as "worldly" and potentially damaging to the faith, to supporting physical exercise and the importance of visible Christian 'role models' in various social fields. In the following decade many sports clubs saw such a dramatic increase in Christian members that some of them – like the main football club IK Start – even came to be seen as "Christian clubs" (Justvik, 2012).7

At the time of writing, Kristiansand boasts over 110 local sports clubs, and near 60,000 memberships (many are members of two or more clubs) after a 28% increase over the past decade, primarily in football, handball, swimming and gymnastics. The same decade similarly saw a strong increase in physical exercise across generations. Of the ten largest cities in Norway, Kristiansand has the highest coverage of sports centers and arenas per capita (*Kommunedelplan Idrett- Og Friluftsliv, 2015-2018*, 2014). Local businesses and even the city council sponsor the national population health project *Cycle to Work*, enabling Kristiansand's citizens to participate for free. In 2014,

**<sup>7</sup>** IK Start still enjoys an annual Church service in their honor at the local church.

participation had reached a national record high of 10%, five times higher than other comparable cities.

#### **4.2.4 Characteristics of Kristiansand's Social Order**

From the above we might glean some general characteristics of Kristiansand's social order.

One characteristic is a particular church culture. During interviews with project leaders and key people found through "snowballing" methods, remarkably many informants referred to "church culture" or "church power" as a key factor in the city's social order. These were less precise attributions than expressions of a general sense that in Kristiansand it matters whom you know, and that decisions can often appear to have been made after church services as much as around official negotiation tables.

Another characteristic is a strong focus on nuclear families. We see it in how the *Children's City* project not only seeks to cater to this demographic for tourism purposes, but also in the effect it has on local businesses, who even outside of the tourist season have to target families with children, even if this would not be their 'natural' customer base. A further indication is the changes made to *Cultiva'*s statement of purpose, which at the time when this research was conducted primarily targeted nuclear families with children.

A third characteristic we see across various arenas in Kristiansand is a general celebration of physical exercise and experiencing nature. This is evident in the high number of memberships in sports clubs, and in the mediatization of and official subsidizing of fitness events such as *Cycle to Work* or *Terrengkarusellen*. Similarly, the promotion of a large new neighborhood as "the sportiest neighborhood in Norway" speaks to the importance of physical exercise as a symbol of prestige and high status.

In isolation, none of these are peculiar to Kristiansand; many cities promote themselves in similar ways. However, we contend that the particular combination of these gives Kristiansand's social order a distinct local "flavor" which is recognizable as "typical" of the place. In the next section, we turn to the question of whether social media in Kristiansand tend to challenge these norms, or rather serve to reinforce and reproduce them.

#### **4.3 Social Media in Kristiansand**

Social media are frequently imagined as vectors of transformation and disruption, and as a result very little existing research considers the continuities and conservative schemas that are reproduced by these platforms (Benkler, 2007; Castells, 2008). It is often overlooked how algorithms reward engagement in the form of likes, comments, mentions, referrals, and so on, which means that online visibility depends on affirmation, and that different forms of affirmation provide different levels of visibility. In contrast to such accounts, Alice E. Marwick convincingly argues that just like the tech-entrepreneurial scene in San Francisco where they were first designed, "social media applications encourage people to compete for social benefits by gaining visibility and attention [...] adopt[ing] self-consciously constructed personas and market themselves [...] to an audience or fan base" (Marwick, 2013, p. 5). One point being made here is that the technology itself is structured in order to create and facilitate ever-more conforming clusters of users with similar sensitivities and aspirations. In these clusters, most signs of affirmation go to a few very active and comparatively central users with a lot of followers. Receiving affirmation from a cluster of users, or indeed from one central user within such a cluster, ensures visibility within that cluster.

Again, Elias' figurational sociology lends itself particularly well to this reading of social media logics. In his book *The Court Society* (1983), he describes the social life at the seventeenth-century royal court, which has several striking parallels to the social dynamics on social media platforms. For instance, in order to rise in the social hierarchy of the court, the important thing is not so much to be rich and powerful, as to be *perceived* as such.8 People at the center of this universe display their status with ever-new attire and performances, which people in their orbit seek to emulate and mimic. In the same way, social media users tend to orbit central celebrity users – so-called influencers – within specific fields of interest that one simply *has to* follow in order to belong.

Furthermore, like at the royal court, social media users tend to display themselves in situations signaling high status, while at the same time trying to stand out from the crowd as authentic and original. Success in this endeavor is relative to feedback in the form of likes, comments, tags, shares and mentions. This "traffic" makes users visible to others, just as their own affirmative gestures determine what and who will (over time) become visible and available to them. Negative feedback most often takes the shape of scrolling past content, unfollowing, or quietly ignoring – though outright shaming also occurs.

Drawing on these theories, the *Cultural Conflict 2.0* project has sought to investigate how social media is reshaping social relations in particular cities, looking especially at how social order is generated, reinforced, and challenged both on social media platforms and at their intersection with life in the physical city. What then can be said about the social order and social media in Kristiansand? We began our investigation by focusing on Instagram, because this platform's strong emphasis on visual communication and affirmation in the form of following, likes and affirmative comments, together with a comparatively limited possibility for verbal messaging,

**<sup>8</sup>** I am grateful to Justus Uitermark and John Boy for highlighting this point.

means it lends itself particularly well to this kind of analysis. With Instagram we were able to combine interviews with computational methods, as we had formerly done in Amsterdam (Boy and Uitermark, 2016). In addition, and on a more practical note, this data was freely available before the platform was acquired by Facebook in June 2016.

#### **4.3.1 Local Clusters**

In order to locate user networks and clusters, we identified geotagged posts from Kristiansand posted between June 2015 and May 2016, from users who post at least every 14 days (to eliminate potential tourists). We assigned a relational tie every time users expressed mutual appreciation in the form of likes, comments, mentions, tags, and so on. In this way, we identified 8 Instagram user clusters in Kristiansand, each centering on about 10–15 top users. The pattern of distribution of social status typical of Instagram also holds on this level: most likes and comments go to the top users, who could be seen as modelling the successful symbolic entrepreneur in their respective clusters.

We manually viewed all the posts posted in this period by the ten top users in each cluster and compared each of their feeds against ten randomly selected users in the same cluster. Most of these users had an open account, so we were able to see all their posts. The few who had private accounts generally seemed (from the look of profile pictures, or from emojis or hashtags used in their bio) to affirm the general sense of their respective clusters' circulating symbols of prestige (so, for instance, a private account near the top of a cluster focusing on CrossFit might typically have a profile picture of someone performing CrossFit exercises).

Table 4.1 below shows the eight largest clusters, the name we gave the cluster according to the typical themes that circulate among its most central users, the clusters' size in terms of connected users, their top tagged locations, and which businesses were found among the top ten users. Names of businesses related to specific informants have been changed for the sake of anonymity.

#### **4.3.2 Local Top Users: Navigating Hierarchies**

We located the ten most connected users in each cluster. Some of the top user profiles were businesses or hub accounts where several users share photos in specific genres (like in C1). Some users were 16 or younger (like most top users in C4), while others had since moved to other cities. We contacted three randomly selected top users with open profiles in each cluster. We also contacted some users who were most connected across the city as a whole, even if they were not prominent in any one cluster. Most responded, but some refused to be interviewed – though a few changed their mind when learning that they had been selected among local top users. In total, we ended up interviewing 12 Instagram users, 10 of which were randomly selected top users. Most were interviewed face to face, though two were interviewed via Skype. In the following their names (and the names of associated businesses, if any) have been changed.

*Number Cluster Size Top tagged locations Businesses among top users* C0 Unclustered users – Dyreparken; Scandic Kristiansand Bystranda; Kilden Teater- og Konserthus; Vågsbygd, Kristiansand; Sørlandssenteret Kristiansand Avis; Moods of Norway, Sorlandssenteret; NRK Sorlandet; C1 General cluster 762 Vågsbygd, Kristiansand; Hamresanden; Flekkerøya, Kristiansand; Flekkerøy; Lulu's C2 Event cluster 509 Vågsbygd Videregående Skole; Vågsbygd, Kristiansand; Flekkerøy; Dyreparken; Flekkerøya, Kristiansand C3 Fashion cluster I 307 Flekkerøy; VOLT Sandens; Berglihn Gullsmed; Flekkerøya, Kristiansand; Bønder i Byen C4 Church cluster I 304 Vågsbygd, Kristiansand; Vågsbygd Videregående Skole; Flekkerøy; Flekkerøya, Kristiansand; Fiskebrygga C5 Fitness cluster 301 Viking CrossFit; Spicheren Treningssenter; Emijoly Yoga & Pilates; Power CrossFit; Power CrossFit KRS Viking CrossFit C6 Fashion cluster II 295 Vågsbygd, Kristiansand; Fiskebrygga; Odderøya; Jordbærpikene Sandens; Baneheia C7 Outdoor cluster 248 Odderøya; UIA Universitetet i Agder; Hamresanden; Baneheia; Vågsbygd, Kristiansand C8 Church cluster II 233 Flekkerøy; Pentecostal Church; Vaya con Dios Coffee Shop; Bohem; Flekkerøya, Kristiansand

**Table 4.1:** Overview of the eight largest clusters obtained from community detection on relations among Instagram users in Kristiansand.

Local top users boast between 1,000–2,000 followers (though not all of these are local), while they themselves follow less than half this number. Most of them spend at least one hour editing photos (always taken with phone camera or digital camera, never the Instagram camera function) before they post them, and most of them use the VSCO app for editing. They also have some idea (more or less substantiated) about how or when one should post to generate higher levels of engagement.

We found that top users navigate two social worlds simultaneously that in some respects work according to different logics. On the one hand, there is the exclusive online-only community of Instagram. Here, they actively seek affirmation from influencers at the top of the hierarchy in specific lifestyle areas, such as physical exercise, fashion, foods, etc., whose presentation style they try to emulate, and whose affirmation they actively seek.

For instance, top users Erik (27) and Sonja (25) in the *Fitness cluster* (C5), both active at the *Viking CrossFit* center, seek out world leading CrossFit performers in order to follow them. On average, Erik's posts receive around 500 likes. But on a post featuring himself and a particular CrossFit influencer, he has received over 1,800 likes.

Erik: "I got 1,600 new followers in one day. That's pretty cool […] all their followers came over to me."

Similarly, Beate (28) uses the platform to showcase photos of architecture or natural landscapes. She deliberately seeks out influencers within these genres, and occasionally receives acclamation from them. Over time, she says, she has developed a sense of what specific influencers (and their followers) might affirm.

Beate: "Like this building […] I spent 20 minutes just getting the angle right. Here, [the influencer's] response means a lot [to me], because I put so much effort into it."

But this online-only exclusive "royal court" is not the only one the local top users navigate; there is also the world of personal friends, and the signs of affirmation that circulate in local on- and offline networks. Here, users police each other (and themselves) according to unwritten but strict codes of prestige and shame, where the fear of being ignored is more significant than the fear of explicit negative response.

Beate (28), a top user in the *Church cluster* (C8), engages in political work, but avoids politics on her Instagram profile because her friends might not be interested. As one of the most connected users across the city as a whole, she puts strict demands on herself regarding what she posts or likes and spends a lot of time finding the hashtags that she believes will bring influencers to her profile. Yet for local friends not part of the genre-specific community, she operates with a different set of criteria.

Beate: "[If] a friend posts a picture, then I like that picture regardless of what it is. Because I like my friends, and if they post it, then it's important to them, I reckon."

Other top users similarly speak of liking all their friends' posts, even the posts they dislike. Ragnhild (19) spoke of her own criteria for liking posts.

Ragnhild: "I like a lot of posts. To me, that's a way of supporting people [laughs], like 'Nice pic!,' even if I don't think it's a good picture, I… well, we're friends, so…"

Jon (38) feels his pastoral responsibilities in a local free church prevent him from unfollowing users that he knows offline, even if he dislikes their posts.

Jon: "I'm the kind of guy who's worried about hurting people's feelings. If I unfollow someone, I feel so awkward [laughs]. So… I've never done that."

Visibility in the respective worlds – online lifestyle networks and local personal friends – in both cases depends on engagement and active affirmation. If more engaged in one world, our informants will gradually become less visible in the other. For instance, engaging too much with friends' posts even if these are irrelevant to their online lifestyle communities means that their friends' posts will become visible to users in the more prestigious online communities, which might result in them unfollowing our informants, practically rejecting them from the hierarchy they are trying to ascend. However, ignoring friends' posts in order to preserve one's position in the online hierarchy means an increased risk of having to explain oneself to those friends offline.

Several informants have examples of instances where they, having initially "liked" a picture they would otherwise consider inappropriate, later confronted the friend face to face about it. Maria's (19) friend, for instance, posted a selfie from a party together with someone she barely knew, and afterwards did not remember taking the picture. Still, she was "scared of deleting it because of what others might think." Similarly, Beate tells of a friend who posted a selfie revealing what she (Beate) considers too much naked skin. This resulted in several friends confronting the girl.

Beate: "[W]hen we got together we were like 'why do you have to post pictures like that,' and then we had a nice talk about confidence and looks and stuff."

Elias describes the "courtly art of human observation" as a joint observation of the self and of others in relation to a specific social context (Elias, 1983, pp. 104–106). "This self-observation and the observation of other people are complementary," he writes, "[o]ne would be pointless without the other." We can observe this in how top users police their own presentation in light of how they expect others to respond. Maria (20, Kristiansand) speculates on why her efforts to secure followers occasionally fail.

Interviewer: "Has it happened that people you wish would follow you have unfollowed?"

Maria: "Yes [laughs] If I want to follow them, then it sucks when they don't want to follow me […] I put so much effort into posting good stuff, so when it's people I know, it really sucks when they unfollow me. I mean, it can't be because my pictures are bad, then I would get it, you know? I am really trying to post good pics."

Here, it is not receiving negative feedback that causes frustration, but the sense of being ignored. As in the pre-modern court, visibility is an existential necessity, and recognition of status in the hierarchy depends on being seen and affirmed. Several top users, such as Anne (20) has an app tracking who unfollows her and when.

Anne: "If someone unfollows me, then I unfollow them. I mean, unless they're, like, really big [on Instagram], then I totally get it."

The self-policing does not require actual feedback from followers, only a general fear of being ignored. Ragnhild (19) describes removing a picture she later regretted posting. Her insistence that she does not care what other people say behind her back, and that the decision to delete the post is entirely her own, go together with her explicit awareness that people might disapprove without telling her directly, and that she must maintain a positive attitude in order to be perceived as "inspiring."

Interviewer: "Did you receive any negative response on the post itself?"

Ragnhild: "No, it was my own decision. No one said anything […] not to my face anyway […] but probably behind my back, you know."

Interviewer: "What do you think they say behind your back?"

Ragnhild: "I don't know. People say so much. I don't really care anyway. I'm just trying to have a profile that's inspiring to other people, and I try to have a happy tone in what I'm writing. It's important to stay positive."

#### **4.4 Reproducing the Social Order**

In their seminal study of the town "Winston Parva," Norbert Elias and John L. Scotson describe how local family networks where characterized by unspoken criteria for how people were ranked in social hierarchies. While these criteria are known to everyone in the group, this is not explicit, but primarily on the level of practice (Elias, 1994, p. xxxviii). In our interviews with the top users in Kristiansand's local Instagram clusters, we find a similar dynamic. All our informants demonstrate a high awareness of how Instagram functions according to a logic of status displays, exchange of symbols of prestige, mutual policing, and stratification into offline and online hierarchies, even if they might have differing strategies for navigating these logics.9

The central question for the present purposes is whether what is associated with prestige in the local clusters corresponds to the symbols associated with "established" groups in Kristiansand. To answer this question, we might look to the symbols circulating in and characterizing particular clusters – such as C5 the *Fitness cluster* or C3 and C6, the *Fashion clusters*. In these typical Instagram genres, top users are typically sponsored by famous brands. Harald, for instance, claims not to care why people often like his posts, but at the same time reveals the degree to which he actively seeks the affirmation of particular brands and people in the fashion world.

Harald: "My friends say my Instagram feed is an 'inspofeed,' like an inspiration for others. I don't know about that. I don't think a lot about it […] But what I often do is tag famous fashion people or where I bought my clothes."

Interviewer: "In this picture you're enveloped in smoke."

Harald: "Yes, we bought a smoke bomb [...] and right here, the shop where I bought the trousers have commented on the picture. An emoji. And it's not the first time!"

The *Fitness Cluster* (C5) is another example. Sonja used to post both CrossFit and fashion pictures, but lately she has been "concentrating on the former." She is sponsored by various clothing brands and the *Viking CrossFit* center, and works out for free as long as she keeps tagging their username in her posts. Erik uses a professional photographer for most of his posts and insists that all his photos are "authentic" because they are taken during actual training sessions.

#### **4.4.1 Prestigious Consumption**

The clusters also indicate the socio-economic aspirations of the users. One indication is the shops and businesses found among top tagged locations. Here we find mediumto high-end jewelry or fashion outlets (such as Lulu's, Berglihn Goldsmiths, VOLT Sandens, and Bohem) targeting young adults. Other businesses found among the top users across the city as a whole (cluster association in parenthesis) include the music festival and event Palmesus (C0), Hansen & Co (C3), Bik Bok Sorlandssenteret (C3), Slottet Shopping (C1), and Image (C1). Cross-cluster tagged locations also

**<sup>9</sup>** This awareness also manifested in their being confused to be considered a "top user" in the first place; many were initially reluctant to be interviewed because they felt had no reason to be considered "big" on Instagram.

include dining places popular among middleclass consumers, such as the waterfront restaurants at Fiskebrygga (C6), Bønder i Byen (C3), and Jordbærpikene Sandens (C6).

We should note that when locations appear in various forms (such as Flekkerøy, Flekkerøya, Flekkerøy Kristiansand) it suggests that these are not established hashtags, but instead highlights the location itself and its connotations as worthy of display. Some informants speak of actively using Instagram to discover locations recommended by users they admire, whether it be shops, popular hang-outs, or weekend trip destinations. The areas Baneheia and Odderøya (both C7) are popular recreational areas where people go for runs or Sunday walks. In addition, Odderøya is a gentrifying area boasting cafés, museums, art galleries, and occasional music festivals (the Kilden Concert Hall (C0), as well as parts of Fiskebrygga (C6) are located in Odderøya). The area of Flekkerøy(a) is among the top tagged locations in C1, C2, C3, and C8. An island just off mainland Kristiansand, Flekkerøy generally has a more rural feel. It is a popular site for nature photo enthusiasts and is associated with conservative and somewhat charismatic Evangelicalism.

**Figure 4.1:** Density of clusters according to neighborhood.

The status and prestige associated with the neighborhoods represented in the clusters is further underscored by which areas are *not* represented. Table II shows the density of clusters in each city neighborhood. Darker colors mean that a higher proportion of posts in the particular cluster was posted from that neighborhood. As is evident, C0 is evenly spread out across the city's neighborhoods, reminding us that people use Instagram whichever part of the city they live in. The locations appearing prominently in the clusters, however, have important contrasts to the neighborhoods that do not. Flekkerøy, for instance, appears in official statistics as a neighborhood characterized by medium- to upper-medium income levels, relatively low numbers of single parents and unmarried people between 30–49, and with the city's lowest number of immigrant inhabitants (7%). Relatively low numbers of immigrants also characterize the larger neighborhood of Vågsbygd, which is tagged in clusters C1, C2, C4, C6, and C7.

By contrast, the neighborhoods of Grim, Hånes, and in particular Slettheia, all have very low density of Instagram clusters. According to the numbers from the municipality website, they also have the highest rate of visible ethnic-religious minorities, receivers of welfare, and typically medium- to low-median income levels. What this shows is that people living in neighborhoods with poorer households or a higher rate of visible ethnic-religious minorities might be on Instagram, but they are not included among the top users in the local clusters of online affirmation. Their neighborhoods remain invisible on social media. The point is underscored by the fact that out of all the top users across the clusters, only a handful appear (from their profile photos) to be nonwhite, and of these at least two identify as Christian (as signified in user name, profile picture, or post content), while none are (openly) Muslim.

#### **4.4.2 City-Wide Symbols of Prestige**

Some symbols of prestige are found across several of the clusters. Here, together with the manual examination of online profiles, the interviews help provide insight into symbols that are circulating more city-wide, at the interface of online and offline communities.

Sports and outdoor life are, as we have seen, associated with high status in Kristiansand's social order. In addition, several of the informants in clusters centering on landscape photography or fashion turn out to be active in various prominent fitness centers and free churches alike. Jon (37), for instance, a top user in the *Church cluster* is an active member of one of the CrossFit centers prominent in the *Fitness cluster*. Yet he never posts pictures of himself exercising, because he thinks this would contribute to an unhealthy focus on physical appearance. Still, he mainly follows CrossFit celebrities and instructors from the *Viking CrossFit* center, where he regularly works out. Jon's own posts feature pictures from the his church's conference center, a few natural landscapes, and him and his family on vacation.

Posts featuring nature and landscape photos are among the top shared and liked across the city – often intersecting with users emphasizing outdoor life as a form of healthy exercise. Some informants post semi-professional photos that have taken a long time to prepare, while others, like Cecilia, edit and post pictures from evening walks or weekend outdoor treks. Harald, a top user in one of the fashion clusters, says he enjoys going for walks and posting pictures of "nice scenery." For Beate, whose Instagram profile is prominent both across the city and in the *Church cluster*, it was nature photos that drew her to Instagram as a social media platform.

Beate: "You know… sunsets draw a lot of likes [laughs] […] That's what I like, seeing other people's lives and where they are, where they are travelling, and where they are heading."

Finally, the local social order's celebration of nuclear families is reflected and reinforced across the clusters, not primarily because there are many photos of families – though there is arguably a certain prestige associated with displaying a happy family situation – but also in that the family networks of the "established" seem to be reinforced in and across clusters; some of the families enjoying high status in the social order also enjoy high status in the local online hierarchies. For instance, top user Ragnhild belongs to one of Kristiansand's prolific Christian business families, and some of her younger family members have also built up prominent Instagram profiles. These are all found among the top users in the *Church cluster,* and their connections to recording artists and other influencers in various fields are visible both in post content, likes, and comments. The post that received most likes on Ragnhild's feed features her with some of these family members and famous influencers.

Interviewer: "Why do you think this picture received so many likes?"

Ragnhild: "I don't know. It's got [influencer] in it, and that's fun. And I notice that people like it when I show pictures with my family. That often generates a lot of likes […] most people know we have a strong family bond, that we are very close and love each other. So, I suppose people […] want to support that."

Finally, the Christian culture evident in Kristiansand's self-promotion and social order is likewise evident in and across local social media clusters. In terms of content, there is the Christian variation on a common Instagram theme: the inspirational quote on the background of a full teacup, and open landscape, and so on – only here, the quotes are Bible verses. Signs of users identifying with Christianity are also found in usernames, bios, or Christian symbol emojis. More importantly, the interviews reveal networks not immediately visible online. Several informants turn out to also belong to the offline church networks. Cecilia, a fashionista who is one of the most wellconnected users across the city as a whole, does social media work for a large free church, but deliberately avoids Christian symbolism on her own Instagram feed.

Cecilia: "I'd say I go to [church] maybe once a week [...] but [on Instagram] I'm kind of laying low… I mean, I have friends who go there too, but… I follow very few friends. So, I very rarely 'like' pictures of, say, coffee cups with a Bible verse on it [laughs], even if I'm sure I could post pictures like that myself. But, yeah, I don't tend to 'like' those anyway."

Harald is another case in point. He mainly posts pictures of fashion outfits and is proud of his "inspofeed." He also enjoys administrative access to one of the nature photo hubs in the *General cluster* which reposts landscape photos from around the region. At the same time, he occasionally volunteers in a large free church's media department – indeed, it was a youth pastor who first recommended Instagram to him. And it was Beate, top user in the *Church cluster*, who invited him to co-run the landscape photo account.

In summary, while the clusters to some extent make it possible to identify groups according to shared interests – fashionistas, church affiliates, and fitness enthusiasts – the clustering does not imply segregation. The *General cluster* C1, the largest cluster we have been able to locate, embraces all of these groups. Its top users include the former user profile of a free church's congress center, designers, nature photographers, and fitness enthusiasts. The most locally connected Instagram users in Kristiansand seek to distinguish themselves by displaying symbols of prestige that largely correspond to the ones we have found to characterize the local social order: they picture themselves with family and friends, in nice outfits, at shops or businesses associated with an aspiring middle class, exercising in picturesque natural surroundings, or on their way to church-related activities.

#### **4.5 Concluding Remarks**

Social media are often assumed to be vectors for social change, for challenging existing structures, and facilitating the forging of new and original identities unhampered by traditional norms. Our evidence points in the opposite direction. Using computational methods in combination with qualitative interviews and participating observation, we have located several social media clusters of users who affirm among themselves certain values and norms through public affirmation in the form of likes, comments, mentions, and so on. Indeed, without computational methods combined with local knowledge and ethnographic sensitivity, these local influencers would have been very difficult to locate, and the meaning of the symbols of prestige they circulate would be near impossible to interpret. Interviews show that top users are aware of the unspoken criteria for rising in the online as well as offline social ranks, and these criteria largely correspond to the symbols of prestige we were able to define in the social order of the city. Where one might have thought that social media would level the playing field between diverse groups in a medium-sized city, what we see is the reproduction of a social order where one group tends to dominate other groups in terms of visibility and influence.

#### **References**

*About Cultiva*. (2017, November 9). http://cultiva.no/english/


Stanley, L. (2016). The Established-Outsider Figuration, Race and Whiteness [Educational]. *Whites Writing Whiteness: Letters, Domestic Figurations & Representations of Whiteness in South Africa 1770s-1970s*. http://www.whiteswritingwhiteness.ed.ac.uk/thinking-with-elias/ the-established-outsider-figuration-race-and-whiteness/

## **5 Meme Collectives and Preferred Truths in Assam**

Sagorika Singha

## **5.1 Introduction**

This chapter focuses on memes and their rising popularity in the geopolitical region of Assam. It explores that which gives local memes their agency; in short, how diverse social media groups use memes to propagate or build public opinion among a new emerging public of internet users. Of late, social media has become a popular news source globally, and this is also true for Assam. But on social media, the news does not merely spread in a straight forward manner via numerous posts. Paratexts accompany these posts. In this context, paratexts – a term derived from literature – refers to the multitude of extraneous, ephemeral popular cultural by-products that populate our contemporary mediascape (Pesce & Noto, 2016). Memes these days act as a popular form of such paratexts, and they help put a spin on a news story or event through their movement in the digital space. This chapter considers certain instances where memes surface as vehicles propagating discussion around the socio-cultural and political situation in the form of a local situation or an inside joke and what this, in turn, tells us about the manifestation of the social order that exists in the social media space in local territories.

#### **5.2 Memes as Digital Objects**

In order to critically discuss meme as a cultural artefact this chapter reads meme as a "digital object" and borrows from Yuk Hui's *On the Existence of Digital Objects* (Hui, 2012). In Hui's analysis, digital objects move beyond the realm of representation, hence he considers them to be material objects. Considering the materiality of digital objects helps us engage with the socio-economic systems that lead to their emergence. Borrowing from and extending Hui's methodology, this chapter maps out a framework to delve into local digital objects, and in doing so, reimagines cultural geography as an associated milieu wherein memes reveal mechanisms of local "truth-making." Digital objects are bereft of the definitive qualities of "objects." What happens then when people find themselves representatives of an era of digital objects – digital objects which serve as floating objects of communication drifting from one screen to another, shared by one user with another? In doing so, they form chains and networks of connections which involve not just the sharing of those objects but also the knowledge of the social order of each user.

Memes are a ubiquitous form of popular culture digital objects such as jokes, rumours, videos and websites which are circulated from person to person via the internet. People's interaction with media and infrastructure reveals their relation to the world and their respective nation-state. In accepting this, there is also the potential to foresee that the effect of individuation of digital objects is not only experienced by users and "machine agents" but is equally affected by the geographies they inhabit (Simondon, 1980). While objects have been historically understood as tangible things, here digital objects, memes, emerge as intangible artefacts that link users through their circulation. Further, memes serve a dual presence; they allow users to leave a digital trace online (as an extension of the self) while indicating an offline world (which our real selves inhabit).

#### **5.2.1 Digital Objects and Generation of Social Knowledge**

There is a distinctive difference between digital objects and their physical counterparts, including cultural records. While the latter is rooted in "solid, self-evident nature," the former remains evasive (Kallinikos, 2010). The signatory attributes of digital objects are editability, interactivity, openness, and distributedness (ibid.). While all of this holds true for memes, they are often evasive, ambivalent, paradoxical – features which affect not only their creation but also their circulation and meaning-making. Throughout this chapter, we witness these attributes at work with one another. As digital objects and despite being embedded "in local cultures and power structures," memes have allowed for a change in the way users interact and, more importantly, in the way social knowledge is generated online (Shifman, 2014). Most of the time, these digital artefacts or objects or things and their creators are themselves oblivious of the limit to their potency. This narration is an exercise in understanding how digital objects and artefacts have a life of their own, and how they bring forth an experience which pervades and affects various aspects of our lives.

Gilbert Simondon's notion of individuation deliberates that the genesis of an individual (either an individual or a social group standing for an individual) is a continuous process within an "associated milieu." As digital objects go through the process of concretisation, there is a presumption of improvement. Web 3.0 is supposed to be more enlightening than Web 2.0; GIFs are better than still images, and so on. However, following Simondon's idea of individuation, the never-ending process of concretisation allows for numerous possibilities. What we conceive at a point in the upgrade is just one of those multiple possibilities. In this chapter, local memes reveal different stages of apprehension or individuation, and not all of them engage with the same purpose. Within a regional milieu, as these digital objects grow, one of their potent objectives is the generation of social knowledge within the geography where they circulate. The caveat is that the social knowledge generated is also reflective of the composition and the order of society. This is revealed by the contesting mode of after-effects that the circulation of the digital objects leads to. While these artefacts are supposed to take on physical forms, what happens is that the instances triggered by these artefacts start reflecting the features of the objects themselves, namely, being evasive and ambivalent. This vagueness appears in the emergent post-truth era, marked by a rampant increase in the circulation of fake news, particularly in economies such as India. This is a transformation on a massive scale with far-reaching effects in the ways people interact, form opinions, and consume media information, as well as their experience of the "real" world which they inhabit.

The most fundamental way in which mobile media has transformed the contemporary world is by altering our relationships in social spaces. By demonstrating the sharing and creation of specific memes on social media platforms, mainly Facebook and WhatsApp, this chapter highlights the role such contemporary social digital objects play in driving this transformation in Assam. Users often create popular memes in the region in local languages such as Assamese, Bodo and Sylheti, among others. Language becomes a means for self-assertion, which is also a marker of "localisation" and the establishment of "cultural individualism." This development affects the tensions between the peripheral region of Assam (in northeast India) and the larger nation-state. The circulation of these memes leads to the emergence of a new vocabulary to discuss offline socio-political situations within specific social media groups. Additionally, they play a part in constructing conversations and events catering to a cluster or social group. The central concern here is how memes emerge as a newfound language and can develop platforms through their creators and users, in turn, transforming, asserting, and establishing new emerging publics.

Scholars have always looked at the northeast region of India as a politically unstable state with a history of secessionist movements and unrest (Mitra, 1993; Baruah, 2005; Bhaumik, 2009; Gill, 2013). However, there has been no significant attempt to read the region post the digital inroads, which have led to overwhelming changes. Propaganda and other pursuits go hand in hand with increased mobile internet penetration, and access to the internet ensures a new kind of landscape which no one has tried to gauge from a cultural study perspective. How are local citizens dealing with these changes and how are they affecting not only their appraisal of things and popular culture around them but also the production of these digital objects? Often, we find an incoherence in dealing with it. For example, on 18 May 2017, the Assam government issued an order announcing stricter monitoring of the use of social media (PTI, 2017). This order came after a series of unsavoury incidents10 concerning government officials went viral. In this context, a decision like this does

**<sup>10</sup>** A slew of such incidents included the leak of a video showing MLA Ramakanta Deuri in a compromising position with an unidentified women in a hotel room (The Asian Age, 2017). In another incident, MLA Aminul Islam broadcast his Assembly speech through Facebook Live (Roy, 2017).

not come as a surprise. It reflects poorly on the inability of both the users and the authorities to understand and control the online space.

Rumourmongering-instigated episodes and incidents riddle the current news landscape of the state, often accelerated by the use of memes on social media. In March 2017, there was a case of a *Fatwa* (that never was) issued against the popular reality show runners-up from Assam, Nahid Afrin (Saikia, 2017). The only evidence of any *Fatwa* was a leaflet signed by 46 clerics requesting residents not to attend musical performances or magic shows since they were against the Sharia.11 But the photo of the leaflet, written in the local Assamese language was shared extensively on social media and the non-existent threat on the life of the young singer was considered so real that the news received prime-time national TV and print media coverage (Saikia, 2017). As Sergio Sismondo remarks, "If the post-truth era starts by blowing up current knowledge structures, then it isn't very likely to be democratisation, and in fact most likely leads to authoritarianism" (Sismondo, 2018). There is a certain pattern to such propagations, as could be witnessed in the two cases mentioned above and many others which have cropped up in the past year in India. Most such incidents have taken place in tier II and III cities.12

There has been a definite shift in the way users access, interpret and consume information. The socio-economic and cultural milieu, along with users' interactions and adoption of the new technological dimension of information consumption affect this process. While the influence of social media and the behaviour of the youth has been chronicled (Tripathi, 2017; Arora & Rangaswamy, 2015), there are gaps in analysing these phenomena from the perspective of the formation of the user units and exploring how social media changes the interpretation of current events within a particular user unit. The primary question here is how the increasing presence of social media platforms affect users in realising their selves in the way social media stages an event within such cultural discourses. How does it, for instance, encourage their view of themselves as mediated users? In the online space, how does their participation affect the functioning of the social establishment? In Assam, political or institutional participation in the online space has been prominent, and employees of various government departments are encouraged to participate in WhatsApp groups

**<sup>11</sup>** Sharia law is part of the Islamic religious tradition. It is derived from the religious precepts of Islam, particularly the Quran and the Hadith.

**<sup>12</sup>** The Reserve Bank of India classifies cities based on their population. Tier II cities population (between 500,000 and 5 million) and category Z or tier III cities (population below 5 lakh). Police, when asked by journalists about the recent bout of mob lynching which took place in the country following rumours of child-lifters, mention how the places where such incidents took place had a high concentration of poor and illiterate people (Raina, 2019). Pamposh Raina, "New Media and Indian Elections 2019" (The Indian Express & Jindal School of Journalism and Communication's (JSJC) panel discussion, New Delhi March 29, 2019).

informally. Such institutions often use these groups to disseminate news and circulars within those sets of official participants.

Poe's law, first articulated in online forums, states that without a clear intention from the author, some readers might accept "obvious, exaggerated, parody of extreme views" to be true (Poe, 2005). Memes and GIFs take on a different value because they not only bring forward potent conversations in a casual and accepted way but are susceptible to Poe's law. It becomes all the more true for users who are becoming accustomed to this form of news dissemination. Once the user-base acknowledges this reality, it will be easier to understand the implications of the social makeup and to interpret political events through social media in a specific regional geography. This will bring evidence to the interpretation of culture and society among a relatively new kind of user-base and a new kind of political geography. Studies similar to this can ultimately help us understand in an ethnographic sense how events and popular culture deviate from textbook understanding when new media percolates into everyday plebeian discourse. When a user circulates a video or a meme online, an "apolitical" creator or propagator may remain "oblivious" that the readers and the receivers will interpret it in their own way. There is a lack of an established hierarchy in the comprehension of such information.

#### **5.2.2 Public Opinion Formation and the Meme-Machine**

Meme production is a machine exercise as it encourages an endless process of mining memes for their relevance and popularity. Memes work because they exhibit the classic attributes of the contemporary digital age – they are short, visual in nature, attention-grabbing, and users can discard them easily. They are easy catalysts for events of all nature. With the widening of the internet's reach, the relevance of memes has been growing, making the potency of the artefacts more visible and impending. Memes and discussions surrounding them act as crucial opinion generators. There lie two facets in the "humour element" that is intrinsic to the form – which often features a text tag overlaid on an image – while it grants the meme its momentum, the humour of a meme can often dilute the gravitas of an event.13

The asset of a meme lies in the latent potential of funny photoshopped images. The image referent with its instant political quality and easy shareability makes it impossible to ignore. Through memes, we witness the birth of "portable politics" for instant consumption on social media. In Assam, popular *memeable* characters are often local celebrities or figures in local media and politics. Posts which seem to be all

**<sup>13</sup>** In reference to this, it is interesting to think of how the Amul hoardings (which also get published in newspapers etc.) precede the idea of a meme in terms of the usage of image and text and allusions to prevailing news scene in the country.

the rage mainly involve caustic reactions to socio-political and cultural scenarios. The sarcasm-fuelled conversations provoked by such content become noteworthy when viewed through a lens of the "meme-synthesis." In essence, these conversations become significant because they reflect that the crises of the place, the region, are always a popular rhetoric and it comes as a small wonder that such content has the characteristics of vitriolic conversations and "strong" opinions.

Bjarneskans and his colleagues postulate that people strive to join a circle of individuals who "share the joke" (Bjarneskans et al., 1999). While this results in the creation of a group, often the participants strive to share individual opinions and issues which conform to the group's preferences merely because it seems cool to belong to the trendy crowd sharing similar beliefs. With regards to internet addiction, Jonathan J Kandell observed how a psychological dependence to being online could result in anxiety when one feels disconnected, thereby leading to a fear of missing out (Kandell, 1998). In a rush to belong, to find ever-popular and powerful memes, a user can inadvertently present an unpopular opinion which might prosper owing to the general propensity to "follow the crowd" on social media. We can see this in the contagious sharing of opinion that follows on social media.

#### **5.3 Network Economy and the Meme**

"Sharing" is the exhortation of the present generation, not just sharing one's private life and politics but also tastes, skills and capabilities. As Limor Shifman clarifies, sharing in this age stands both for distribution and consumption (Shifman, 2014). Sharing is also a way to network because when you share (and share widely) you also forge a network which spreads further depending on the amount of effort a user puts into that sharing. Sanjeev Goyal describes a network as a collection of nodes and links between them – individuals, firms, countries, or even a collection of such entities (Goyal, 2007). He also focuses on how structures of relationships assisted by such networks shape or influence individual behaviour in both a social and economic context. This reading is in opposition to another network approach, the Actor-Network Theory (ANT), which espouses a material semiotic method to understand networks (Latour, 2014). ANT focuses on the relational ties within a network – "the tracing of associations" (ibid.). With users sharing memes, naturally, the circulation occurs within specific networks of individuals, and most of the time we find groups affirming to a common interest in terms of acceptance of ideas and political voices in order to belong to the socially desirable networks.

Ryan Milner introduces the concept of "polyvocality," the possibility of a varied range of voices to be heard, which is enabled by memes and other digital objects (Milner, 2012). Even though as a concept it seems promising, in an economically weak state its probability is questionable. Even when the infrastructure is in place, it is not easy for the economically and socially marginalised to readily access information owing to illiteracy or a lack of basic familiarity with the technology. In such regions, a different form of digital literacy evolves. As late adopters, users often associate the internet with social media since they access the internet's possibilities only through social media and smartphones. Even though mainstream media still exists as the vital component of opinion formation, particularly in public discourse, new media and its tools, such as memes, exist to dent its influence, questioning and – in this case – even trolling it. For a new range of viewers, old media has become superfluous; an unavoidable, non-functioning necessity. In contrast, new media is the alternative; the substitute that they desire and one that they can customise.

Shifman observes how, in spite of the trend of a specific meme form, "the common ideas and forms shared by *many* internet memes might tell us something about digital culture" (Shifman, 2014). Memes do not exist in isolation. The sharing of images, memes, statuses, reviews to release emotional sentiments, among others, all contribute towards building the network. She considers this phenomenon as emergent of a central cultural logic which goes beyond the "sharing economies" (ibid.). In economic parlance, there is a value attached to such network formation. We can observe this pattern of value-making explicitly in the phenomenon of virality and which we often use as the barometer for measuring the success of online shared content. As Manual Castells muses, the new technological regime also results in a new networked, interdependent economy which ideally should encourage greater productivity and efficiency and which, in turn and with the right conditions, gives rise to equally dramatic organisational and institutional changes (Stark & Castells, 1997). However, memes, which distinctly rely on "networks of mediated cultural participation in their creation, circulation, and transformation," go beyond those organisational changes (Milner, 2012). Memes have the potential to work on social representations and public discourse owing to their informal engagement as an everyday pop-cultural artefact. Their propagation through personal social networks can simultaneously give birth to an alternate social voice, with both its advantages and disadvantages.

#### **5.3.1 Memes Experience and Existence: Offline Dynamics Acting on Memes**

TAM, a vaguely defined acronym standing for multiple terms such as *Trolls and Memes*, and *Trolling Assamese Media* among others, is a meme collective based in Guwahati, Assam. It started by targeting local Assamese media extensively, and they pointed out flaws in their shoddy reporting, factual incongruencies and grammatical errors. Ironically their own memes had typographical errors and were poorly written most of the time. There seemed to be a fascination with targeting individuals. Even when talking about the media, the target was usually a particular reporter, owners of specific news channels, activists and others. If Assamese journalist Hemen Rajbonshi14 was the most trolled person in their memes in the early period of the group's existence, he was later replaced by activist Akhil Gogoi. Every time they targeted someone, they garnered a group of users who shared the sentiment towards that person.

Furthermore, the group succeeded in creating an echo chamber of their own. The pop-cultural rehashing of the politico-cultural events via such groups links back to the question of trivialisation. However, it is this trivialisation that makes a post or a meme go viral, and we can consider that these meme groups curate such echo chambers based on their inherent politics and ideology. It further heightens the societal divide between those with access and those without. In an already contested terrain, access become the criteria for determining who is privileged and who is not.

#### **5.4 Akhil Gogoi, or How a Peasant Leader Became a Meme**

Memes certainly have a specific agenda and, not surprisingly, they thrive while promoting that agenda in their own informal ways. Nevertheless, the Facebook groups discussed in this chapter which create vernacular memes (hereon referred to as meme collectives) generally deny any overt support for a particular ideology or political preference. When questioned, they profess to stand for a more common community approach with the primary purpose of being humorous in dealing with everyday crises (Administrators, 2017). Treating social events through that prism, they have their agenda in place, and this is reflected clearly in the kind of memes they come up with. In this regard, there are numerous categories and subjects that such meme collectives seem to favour. Considering the scope and limits of this chapter, I am taking only one such category of meme as a case study – the category of memes featuring peasant activist Akhil Gogoi. Gogoi is one of the most in-demand memeable entities within this meme collective.

Akhil Gogoi memes are a special case because of what he is associated with and how the collectives exploit their understanding of his politics. Gogoi is a well-known peasant and Right To Information (RTI) activist in Assam, India. He is the president of the Krishak Mukti Sangram Samiti (KMSS) (Farmers' Freedom Struggle Committee) and a vehement critic of the right-leaning *Bharatiya Janata Party* (Indian People's Party) that leads the government at the centre. His critics have attacked him for his vitriolic speeches against the government's decisions, particularly regarding the

**<sup>14</sup>** In August 2015, Hemen Rajbongshi, a reporter with the local Assamese channel, Pratidin Time, aired a news capsule on "scantily clad women" in Guwahati and described them as a nuisance during summer time. He was referring to women who wore mini-skirts or any version of the skirts. The viewers harshly criticised the sexist reporting on social media resulting in the birth of the "I am so Assamese" memes.

granting of citizenship to Hindu Bangladeshis.15 The state has also accused him of instigating people to take up weapons to attack the government's "malicious intent." Considering the impact of Gogoi's activism on Assamese society and politics, it is not surprising that he is a regular feature in newspaper articles. What is interesting, however, is how an agrarian peasant leader, far removed from social media, has become one of the most popular memes in the region. Gogoi has been a memefriendly target of the *Troll Assamese Media* group since their inception in 2015. There is a self-vigilante sense of justice in the tone used by the people behind such memes. However, their notional value is contested owing to the collective's general, overt "ethno-nationalistic" leanings.

In this particular instance, the meme collective dissolves the public persona of a well-known and respected activist with the active use of topical memes which present a very different version of Gogoi – the activist. This section outlines the role such contemporary forms of social digital objects play in driving a construction of the image of specifically targeted individuals in Assam. This chapter assessed this by observing the sharing and creation of memes on social media platforms such as Facebook. In this way, apart from recognising the spatial and virtual connections that appear within the meme-world, the segment also reveals the inherent potency thriving in digital cultural objects and how they are employed to construct "preferred truths" to a cluster of social groups.

The TAM collective has two administrators. Owing to the rampant online threats they received, they asked to remain anonymous (Administrators, 2017). Both of them claim to work in print media but lack any formal training in the field. The convergence of print media and online media which led to the evolution of the collective could also highlight the patterns of access in the state itself. Even though internet access (primarily through mobile data) has increased, it is yet to replace print or other traditional media. When I asked one of the administrators what led him to create the group on Facebook, he replied, "During th(e) time Assamese media started yellow journalism. First, we started by trolling media, and then we trolled issues based on news. We have done it like a (sic) dynamism – to awaken people. To spread awareness among commonality (sic)" (Administrators, 2017). Their rationale for choosing Facebook as the platform and memes as the medium was that "Facebook is for common people and memes can attract (sic) easily in an attractive way … there was no such meme page in Assam then" (ibid.). What caught my attention, though, was the passionate tone they used to explain their objective behind founding the group

**<sup>15</sup>** In the 2014 General Election, the BJP government promised to amend the Citizenship Act, 1955 and introduced the Citizenship Amendment Bill of 2016 which will make illegal Hindus, Sikhs, Buddhists, Jains, Parsis and Christians migrants from Afghanistan, Bangladesh and Pakistan eligible for citizenship.

which was to tackle what they perceived as the misuse of journalism in victimising people in a prejudicial manner.

From the example of the rise of the anti-Gogoi memes, what we can realise is that the activities of such meme collectives are disturbing, especially considering their trolling agenda. Their agenda seems innocuous and fun, but it is their political underpinnings that actually drive the conversations. In the case of this group in particular, the issues they project have a strong "Assamese identity, neglected state" sentiment to it. They discuss the poor infrastructure in the state, parental pressure, joblessness and unemployment and the underdeveloped media industry in general; they vocally display their united hatred against Bangladeshis and their shared love for Zubeen Garg16 and sports. All of these issues magnify the united identity of the "typical" Assamese youth and also vouch for a right-wing inspired political ideology. While the TAM collective refrains from claiming any political affiliation as such, they have no qualms in attacking others because of their political affiliations. Meanwhile, the popularity of the Akhil Gogoi memes which gradually grew to become a genre of their own in the meme collective's page reflects the value attached to what Gogoibashing stands for and how it is successful as a perspective.

The increasing attacks on him in real life reflect the infamy of Akhil Gogoi on the page. For those witnessing this phenomenon, the alarming question is whether one is affecting the other or is it because Gogoi ceases to be of interest in real politics that he becomes a staple on the online group's page. Talking to the young people in the group, it is interesting how the anarchism otherwise attributed to Gogoi become tepid in comparison to the hatred they shower on him. Memes often appropriate dominant discourse. However, Gogoi's support base, which lies among the liberal demography, draws a contradictory picture lauding him as "Assam's new voice of dissent" (Barbora, 2019). Diversity in any form of artefact circulation is important. However, when the means to create memes is accessible only to a specific group of people, the lack of diversity becomes troubling. It is a reminder of what Milner writes: that in the competition of meme creation, even though there is scope for a variety of perspectives to participate, it is only certain perspectives, values and references which gain ground (Milner, 2012). Here the digital transformation has occurred at a fast pace. However, users have difficulty coping with these transformations since the development in access is incongruous to their literacy and economic growth. Meanwhile, the creators of these online collectives have access to more resources and knowledge, allowing them to direct the conversation in the way they want because the majority of the consumers would not have the same agency and would be inclined to accept the creator's version as truth.

**<sup>16</sup>** Zubeen Garg is a popular music artist from Assam. Garg is one of the most influential contemporary representatives of Assamese culture and identity.

**62** Meme Collectives and Preferred Truths in Assam

**Figure 5.1:** Akhil Gogoi Meme.

The original image in the meme (see figure 5.1) appeared for the first and only time on Twitter when Goutam Baruah, a self-proclaimed Sanatan Hindu, posted it in September 2016 right after the Kaziranga eviction incident. In his tweet he writes "#Kaziranga Eviction Thnx AKHIL GOGOI 4 Uniting Assamese Ppl redardless (sic) of Political Differences..Great Job Assam Govt" (sic) (Baruah, 2016). The TAM meme collective rehashed the context of the image and used it numerous times on their page to propagate Gogoi's support for the illegal Bangladeshis. A sexual, more intimate pretext is used to underline the image as a visible form of the meme. Two other memes also establish a similar viewpoint of Gogoi demonstrating a deep, even sexual, attraction towards Bangladeshi immigrants. According to Milner, "Memes were a means to transform established cultural texts into new ones, to negotiate the worth of diverse identities, and to engage in unconventional arguments about public policy and current events. Memes were a mix of old inequalities and new participation" (Milner, 2012). Others though, including Liesbet van Zoonen, have questioned this position. Van Zoonen outlines how powerful institutions use populist forms of public discourse to influence the public (Van Zoonen, 2006). With cases such as these, we observe how emerging alternative media can also be used by individuals and groups to manipulate people into spreading their preferred view of the world.

Geert Lovink points out that since most memes spread within specific bubbles of people who share a viewpoint, the primary challenge lies in creating "bubblebreaking memes" (Lovink, 2017). Lovink rightly observes the method of using memes as a weapon, something exhibited in the US with the alt-right. According to him, memes are weaponised "as in shitposting17 on Twitter, a form of cognitive denial-ofservice attack." As he concludes, this sense of thematic identification is something that glues the sharers of the memes together. The localisation of memes, thus, inertly hints at the transformation of an original form, but a form which is nevertheless malleable. Thus, we can assume societies favouring certain memes also believe in their transformative potential. For example, sharing memes about consistent power cuts in the region can perhaps help when they reach the right officials who can improve the existing conditions so that the jokes about the problem no longer exist. Alternatively, vilifying and belittling a peasant activist and his stances can ultimately make him redundant and unnecessary in the eyes of the public. In a way, we can interpret this as the users' attempt to focus on issues which they imagine will improve when they are repeatedly discussed or pointed out on social media.

#### **5.4.1 The Media Event and the Meme**

Does a meme function as an event itself? In May 2017, following a series of WhatsApp circulated rumours, three men were lynched in Jharkhand's Singbhum district on suspicion of being child abductors (Special Correspondent, 2017). In Assam too, there have been numerous similar reports where information shared via social media was used to trigger incidents, including one identical to the Jharkhand incident (Karmakar, 2018; Anon., 2018; Anon., 2017). It is thus imperative to assess the effectiveness with which memes may trigger events by taking up a fake issue or news and plastering it over related images.

Moreover, its viral circulation makes its all the more probable for an event to become bigger in scope and consequently reach a larger group of people. Even though social media seems like a popular means to instigate or further prioritise a current issue, holding memes single-handedly responsible for creating an event will be too farfetched. Memes, though, act as indicators that gauge the importance of a current issue. In the case of the Akhil Gogoi trolling, one can see it as propaganda; an emotional, political issue is taken up and even mocked and associated with an assortment of other agendas, to make the memes surface repeatedly. When trolling him with that dose of humour and sarcasm, the desire is not only to make the collective hatred escalate but also to make it palatable and far-reaching.

Just like memes, we can also study how a personal form of exchange can be made responsible for a media event which is more public and widespread. It can thus be seen as an atomic response accumulating to form an event. So, can a meme be simply an agent of a media event and without the potential to be an event itself? To what

**<sup>17</sup>** Shitpost is the internet slang for a worthless post on a message board, newsgroup, or other online discussion platform.

extent is a meme central to the mobilisation of issues? If designed and received in the right way, a meme has immense potential to provide required value to a cause. The importance of the event does not have to be grand or its scale large. It can be local, or niche and its potency has to do with the impact it has – sometimes even at an individual level in the real world as witnessed by the lynchings instigated by WhatsApp rumours.

While a meme's potential is difficult to fathom, and it is questionable whether one meme alone can have a significant effect – sometimes its potential depends on the ambition of its creator and the user-base. There is every possibility of players misusing memes to a great extent, so we need to organise the meaning of the event itself if we want to understand the meme as a media event. Currently, with the examples at hand, a meme and its function are quite medium-specific. In other words, it transforms into something depending on the desires of its sharer and user-base. To the rumourmonger it is a vicious tool, to the exasperated armchair activist it is a cathartic rejoinder, to the humour-loving general public it is more of alternative entertainment and a valid way to contemporise their recreational media. Memes thus fall into the strange category of malleable objects, just like their malleable selves – photoshopped countless times, in numerous languages, the same image and the context may stand for one idea but might be loaned to many.

#### **5.5 Conclusion**

This chapter discusses the potency of apparently mundane digital artefacts to come together to reveal newer ways of expression and to emerge as the driving force behind such participation. We can conclude that the ingrained socio-political notions of online users steer obsessively popular meme topics and issues that ultimately ended up as inconsequential material for consumption by online users. Through meme collectives, the youth have found ways and means of fuelling old sentiments in a new garb. In many other instances (not discussed in this particular chapter), memes have highlighted burning issues in a way which makes them comprehensible and far more accessible to a new user-base. In the process, it is also evident that contemporary civil issues are morphed into their corresponding digital counterparts to suit current times. However, while it acts as a critical stance to whatever seems to be the problem at hand, it is also disturbing when the reins only seem to be available to a select few and issues can be equally subsumed as made prominent. The chapter contemplated how new media technologies bring change in the way small towns communicate, particularly in the way in which they discuss crises. While the traditional way of participation gets gradually but invariably transformed, particularly in the methods chosen for communicating ideas and events, it is clear that only particular sentiments and ideologies spread through these newer platforms. The creation of these platforms helped in building a cultural autonomy for the people of the region and also allowed for other controversial stances to exist and breed.

The growth of a new digital userbase participating, collaborating and building a sense of presence allowed for peripheral voices to thrive through their own energy. However, this occurred simultaneously with the proliferation of preferential views and the ease in creating factions and in circulating them. Those with keyboards at their fingertips had more say in this emergence. In this gradual transformation of the socio-cultural space in Assam, memes exist as a potent pursuer, as an easy catalyst extending the new media platform and encouraging the development of a contemporary local vocabulary which is aspirational but also has a succinct personality. While we cannot necessarily call memes tools of resistance, users do employ them as an effective intervention, and on the global stage they present a window for the local meme collectives' agenda-setting capabilities.

#### **References**

Administrators, T. (2017). *Facebook Messenger* [Interview] (25 October 2017).

Anon. (2017). *BBC News.* [Online] Retrieved March, 8, 2018 from https://www.bbc.com/news/ world-asia-india-39769172

Anon. (2018). *News18.com.* [Online] Retrieved July, 10, 2018 from https://www.news18.com/videos/ india/mob-beats-up-purported-holy-men-in-assam-1676687.html

Arora, P. & Rangaswamy, N. (2015). Digital romance in the Indian city. *Journal of Communication*, 49(3), 84–105.

Barbora, S. (2019). Assam ' s New Voice of Dissent. *Economic and Political Weekly*, 46(28), 19–22.

Baruah, G. (2016). *Tweet.* Guwahati: Twitter.

Baruah, S. (2005). India and its Northeast: A new politics of race. *IIC Quarterly,* 32(2 & 3), pp. 165–76.

Baruah, S. (2010). Society versus State in Assam. In *Durable Disorder: Understanding the Politics of Northeast India* (pp. 123–144)*.* New York: Oxford University Press.

Bhaumik, S. (2009). *Troubled Periphery: Crisis of India's Northeast.* New Delhi: Sage Publication. Bjarneskans, H., Grønnevik, B., & Sandberg, A. (1999). *The Lifecycle of Memes.* [Online] Retrieved

September, 23, 2018 from http://www.aleph.se/Trans/Cultural/Memetics/memecycle.html

Gill, P. (Ed.). (2013). *The Peripheral Centre: Voices from India's Northeast.* New Delhi(Delhi): Zubaan.

Goyal, S. (2007). *Connections: An Introduction to the Economics of Networks*. Princeton: Princeton University Press.

Hui, Y. (2012). What is a Digital Object?, *Metaphilosophy*, *43*(4), 380–395. doi: 10.1111/j.1467- 9973.2012.01761.x.

Kallinikos, J. (2010). Digital objects: Definitions and attributes, *First Monday*, *15*(6), 1–16.

Kandell, J. J. (1998). Internet Addiction on Campus: The Vulnerability of College Students', *CyberPsychology & Behavior*, *1*(1), 11–17. doi: 10.1089/cpb.1998.1.11.

Karmakar, R. (2018). *The Hindu.* [Online] Retrieved June, 10, 2018 from https://www.thehindu.com/ news/national/other-states/two-men-lynched-on-suspicion-of-being-child-lifters-in-assam/ article24122413.ece


## **6 Insurgent Ways of Looking: Gendering the Witness and the Land in the Visuality of Israel-Palestine**

#### Liat Berdugo

In 2014, I was introduced to a special unit of exclusively female soldiers called *Tatzpitaniot* ("The Watchers") on a visit to an Israeli Defense Forces (IDF) army base in the occupied Palestinian Territories. As a woman and Israeli citizen myself, I have long been interested in the Israeli military and its gendered roles: under Israeli law I would have been conscripted into service, had my status as an American citizen and resident not exempt me. My visit thus evoked in me a vision of a life that might have been mine, but which I had, for better or worse, dodged.

It was vision, itself, that comprised the key task of these female soldiers whom I visited. Their sole job was to watch live video streams of the IDF's network of 1700 security cameras mounted along key sites in the West Bank and Gaza, in an effort to surveil Palestinians and secure Israel's borders (Harel, 2017). Often, these women watch multiple screens at once, staring at video feeds for hours at a time as they scan for movement. They report any "suspicious" visual activity to their army superiors, who then order on-the-ground action when deemed necessary (Gross, 2015). Put differently, the sole task of a *Tazpitaniah*, a "watcher," is to weaponize her sense of sight.

The assignment to the *Tazpitaniot* brigade is deeply unpopular among Israeli conscripts. When a conscript asked on an online messaging board, "Is Tazpitanit an intersting [sic] army role?" others replied with comments such as "If you hate yourself then go there by all means," and "It is [as] interesting as is watching paint dry in real time... Get something else... Anything!" ("Is Tazpitanit an Intersting [sic] Army Role?") Young Israeli women share strategies for 'failing' the *Tazpitaniot*'s correlated vision examinations, encouraging each other to pretend not to see the computerized blinking cursors that test the accuracy and alacrity of their peripheral vision (Yom Hame'ah, 2015). These potential *Tazpitaniot* feign a kind of momentary blindness – a failure of vision that, I argue, has become a norm for Jewish Israelis wishing to ignore the unsightly monstrosity that is the occupation (Berdugo, 2017).

To combat these negative stereotypes, the IDF recently produced a promotional video titled "With Her Eyes She Defends Israel's Borders" (2019), valorizing the work of a watcher, a *Tazpitaniah*. The video features a single young woman in army fatigues, appearing an ordinary Israeli teenage soldier in every respect. The video declares, "She may look like any other soldier, but she has a secret weapon" – the camera zooms in dramatically on her face before the narrator continues – "her eyes" (see Figure 6.1). In this way, the video propagandizes her sense of sight, fully laminating together vision with weaponry. If "winning is keeping the target in constant sight" as Paul Virilio writes, then a *Tazpitaniah* must keep her eyes wide open (Virilio, 2009, p. 2). One former soldier described that "Tazpitaniot are forbidden from looking away from our screens, for even a second" (Silkoff, 2018).

Perhaps predictably, the video concludes with a close up of the soldier's eyes. The narrator says, "It's thanks to women like her, who keep their eyes open, that millions of Israelis can close their eyes safely every night." The *Tazpitaniah* then winks knowingly at the camera (see Figure 6.2). A wink is a momentary closure of the eye, a flirt – a colluding gesture that accentuates the weapon she has sharpened to protect her nation. Her wink, we are to understand, is not a blink; instead she keeps one eye open at all times, training her watchful gaze on her target.

**Figure 6.1:** Video still from the Israeli Defense Forces' video, "With Her Eyes She Defends Israel's Borders," 2019.

**Figure 6.2:** Video still from the Israeli Defense Forces' video, "With Her Eyes She Defends Israel's Borders," 2019.

The women-only IDF unit relies on troubling stereotypes of the exclusively feminized eye as a maternal force of protection: a stereotype that extends responsibilities of an essentialized feminine, watchful gaze over children to a gaze that watches over territory. In Israel-Palestine, in a conflict that weaponizes and structures gazes (Azoulay, 2008, 2011a; Hochberg, 2015; Maimon & Grinbaum, 2016; Berdugo, 2021), I argue that there has been a particular focus on the vision of and visual documentation produced by women as a problematically gendered force of protection. A woman's vision is celebrated as a militarized extension of the dominant Israeli scopic regime, as with the *Tazpitaniot*; as a Palestinian counter-visual tactic celebrated by the human rights group B'Tselem, which distributes cameras to Palestinians living in highconflict zones and gathers the footage. I show that in Israel-Palestine in particular, this focus on women's sight essentializes a long history of relating Zionism to the land, with all the feminized ideas associated with the land-based focus on "mother earth" or "mother nature" as a sustainer of people and the celebration of women's roles in Zionism as *Haluzot* ("female pioneers"). I critique such celebrations of women by turning towards ecofeminist critiques, which teach us to consider the ways in which women and the land have both been historically subjugated by a shared history of oppression. Ultimately, I argue in favor of de-gendering vision and de-gendering the land, instead towards a kind of sightline that celebrates a disobedient, insurgent way of looking: one that visibilizes the very frame of sight, and of the camera, as a means towards new kinds of resistance in conflict.

#### **6.1 A "One-Way Hierarchy of Vision"**

There is a striking similarity in looking down sightlines of a camera, and of a gun, in what many scholars have described as the joint and mutually entwined history of visuality and weaponry (Feldman, 1997; Sontag, 2002; Lebow, 2012). Virilio teaches that violence and visuality are concomitant, stating, "For men at war, the function of the weapon is the function of the eye" (2009, p. 26) This lethal mutuality has been cemented for video cameras in particular, as the first motion picture camera was invented to mimic the design of semi-automatic, revolving rifles (Lebow, 2012).

Drawing from this history, we can understand the role of "watchers" such as the IDF's *Tazpitaniot* not as an unexpected weaponization of sight, but a natural evolution of the militarization of the gaze. In Israel-Palestine, this is a militarization that extends to Israeli civilian bodies as well in an effort to surveil and control Palestinians. Architect Eyal Weizman has noted that Jewish settlements are often placed on hilltops or mountains, overlooking Palestinian villages that largely reside in the fertile valleys below, and thus creating a distinctly vertical separation between populations even as they are horizontally mixed (2007). In these settlements, the directions of roads, plots, the houses, and windows within the houses all direct an Israeli civilian gaze "out and down" over the Palestinian residents below, enlisting a civilian population to act as a watchful eye to monitor Palestinian action (Ibid., p. 132). Such settlements have transformed the West Bank into a network of visual monitoring stations, staffed by civilians who might also be "enjoying the view" (Ibid., p. 133).

However, Palestinians may not look at these Israeli Jewish settlements. According to the rules of engagement as of 2003, IDF soldiers may shoot-to-kill any Palestinian caught observing IDF activities near Israeli settlements with binoculars or in any other "suspicious manner" (Harel, 2003) Jewish Israelis thus maintain what Eyal Weizman has called a "one-way hierarchy of vision" over their Palestinian counterparts. Put differently, Israelis dominate what Nicholas Mirzoeff has called the "right to look" (2011), as a subset of a host of other visual rights: the right to see and to be seen; rights to look and to surveil; rights to be out of sight (of surveillance, for instance); and rights to have one's image trusted (rather than subject to a "digital suspicion" (Kuntsman & Stein, 2015) through claims of photoshopping, cropping, or falsification in post-production). If "the gaze that sees is the gaze that dominates," as Foucault has written, then the seer has power over the seen (1994, p. 39). An uneven distribution of power in the realms of economics, politics, natural resources, equality, and justice within a conflict zone also reaches the realm of visuality itself.

It is against this backdrop that, in 2015 – the same year that YouTube was founded and video was gaining prominence on the Internet – an Israeli NGO called B'Tselem welcomed video into its repertoire by hiring a video coordinator and publicly including its videos as a main navigational component on its website (B'Tselem, 2005a). B'Tselem, also called The Israeli Information Center for Human Rights in the Occupied Territories, was established in 1989 by a prominent group of attorneys, academics, journalists, and Knesset members to document human rights violations in the Occupied Territories (B'Tselem, 2005b). Before adopting video as a tactic, B'Tselem published statistics, testimonies and eyewitness accounts, and reports in an effort to create a human rights culture in Israel.

At first, B'Tselem's videos were largely comprised of videotaped testimonies of Palestinian victims, which bolstered the organization's long-standing initiative to collect written testimonies from across the Occupied Territories (B'Tselem, 2005c). B'Tselem also produced short documentary videos that catalogued the hardships that the occupation placed Palestinians (B'Tselem, 2006). B'Tselem housed the role of its new video coordinator within its public relations department, signaling the organization's belief in the communicative potential of moving images.

In 2007, B'Tselem staff toured Hebron to see the conditions of Palestinians living adjacent to Israeli settlements (Tarabieh, 2019). Due to frequent settler attacks and stonings, Palestinian residents living in close proximity to the settlements were forced to erect metal grates around their houses for protection. As a result of its tour, B'Tselem resolved to give a video camera to one family who lived inside a "cage house", the Abu 'Ayeshas.

In January of 2007, a member of the Abu 'Ayesha family filmed the "Sharmuta" or "whore" video, a short clip in which an Israeli settler attempted to shut her into her house ("Sit here, in the cage!") and then bullies her by calling her a "whore" nine times in a hissing, menacing tone (B'Tselem, 2007). The Sharmuta video became the first ever citizen-recorded video in B'Tselem's camera distribution project and circulated widely within Israeli and international media outlets, to the point that it has become part of the Israeli lexicon (B'Tselem, 2009).

Thus the B'Tselem Camera Project was born, three years before the Arab Spring brought vast attention to the power of technology in the hands of the oppressed. Within its first year and a half, B'Tselem's distributed over one hundred cameras to Palestinian families in the Occupied Territories and hired two more staff members to support its growth (B'Tselem, 2008a and 2008b). Today there are around two hundred B'Tselem-issued cameras in the field, and B'Tselem has amassed an archive of over 4,500 hours of raw footage, which it keeps on a handful of different servers and in rows of original magnetic tapes behind glass doors in its Jerusalem headquarters (Tarabieh, 2019). B'Tselem's Camera Project is notable for the consistency of its video publication: the project has published an average of one video per week since its conception, often adding subtitles and other demarcations for the public.18 Today, B'Tselem's website prominantly features a carousel containing video footage shot by its volunteer videographers. On YouTube, B'Tselem's channel has over 43,000 subscribers, and its most popular video has been viewed over 4 million times (B'Tselem, 2017).

Notably, one thing that distinguishes the Israeli-Palestinian conflict is the freedom to record. Israeli law is more lenient on recording in public than is the US state of Massachusetts. In a letter from a Public Inquiries Officer, the IDF Central Command states, "filming in Judea and Samaria [the West Bank] is permitted, including filming of IDF soldiers, so long as nothing about the filming interferes with the forces' operations or serves to collect classified information" (2009). Moreover, the volunteers who film for B'Tselem sign an explicit legal document with the Israeli Defense Forces. Certain things are off limits to the camera's eye, such as court proceedings, army facilities, any persons working for Israel's internal security services ("Shabak" – similar to America's FBI), and checkpoints.

B'Tselem's Camera Project was launched with the conviction that a Palestinian with a camera reverses the normal order of domination under the Israeli occupation. Likewise, the citizen with the cell phone video of the police reverses the normal hierarchy of vision, which flows from state to people, from white to black. The reversal of the typical one-way hierarchy of vision allows for a view into the slow, suspended violence of the Israeli occupation (Azoulay & Ophir, 2005). Yet, importantly, the majority of B'Tselem videos are not leveraged as proof of Israeli criminality. Instead,

**<sup>18</sup>** As of August 2019, B'Tselem had uploaded 631 videos to its YouTube channel, for an average of 1.01 videos per week in the twelve years since its founding. The remainder of the unpublished footage is viewable in B'Tselem headquarters with prior permission.

they are mobilized as exculpatory evidence on behalf of Palestinians, themselves, who seek to prove their own assumed criminal bodies to be innocent (see also Finn, 2009). "When you are a Palestinian living in a place like Hebron, you are considered by the Israelis to be guilty unless proven innocent," said former B'Tselem volunteer, Issa Amro. He continued: "… for us, the cameras are not only a way to document events but also to protect ourselves when false complaints are made against us by Israeli soldiers." Amro summed up what I heard from many B'Tselem volunteers in the field: the Palestinian video camera serves as a digital alibi.19

#### **6.2 A Woman's Camera**

B'Tselem volunteer videographers frequently referenced the exculpatory, defensive power of their cameras. Palestinian women, especially, speak of the camera as giving them strength, as if it were a shield or a piece of armor to deflect Israeli state violence. Among the B'Tselem volunteers, women have been growing in number in recent years, with 49% of the Camera Project's training sessions including women as of B'Tselem's most recent annual report (B'Tselem, 2018). B'Tselem makes an intentional effort to recruit women as volunteers by offering female-only training courses, and by appointing women as high-level paid employees for the Camera Project (Ginsburg, in press). Traditional Palestinian culture, which retains a rather conservative and patriarchal stance on women's roles (Baxter, 2007), was at first resistant to women's involvement as B'Tselem volunteer videographers.20 Yet recently, as these roles have been shifting, women with cameras have reported that their activities with cameras have become accepted and even praised within their communities (M. al-Ja'bri, 2019; A. al-Ja'bri, 2019; Jaber, 2019).

A B'Tselem volunteer named Khadrah 'Abd al-Karim from the Palestinian village of 'Asirah al-Qibliyah near Nablus, frequently videotapes the attacks from the nearby extremist Jewish settlement of *Yitzhar*.21 Al-Karim has been filming for B'Tselem since 2008, just one year after the Camera Project first launched. She said:

**<sup>19</sup>** For more on the exculpatory role of the B'Tselem camera, see Berdugo 2021, and specifically the 2012 case of Palestinian teenager 'Abd al-'Aziz Fakhouri.

**<sup>20</sup>** The women videographers who film for B'Tselem told me that these traditional gender roles vary from city to city and region to region. For instance, the videographers in Hebron feel these roles limit them more strongly than their counterparts in the Nablus area.

**<sup>21</sup>** Yizhar is particularly known for its retaliatory *Tag Mechir* or Price Tag attacks, in which extremist Israeli settlers seek to harm and vandalize Palestinians and their property as a 'price' for Palestinian violence or for anti-settlement activity more generally.

Filming the images is helpful to us. It has given me inner strength. I'm no longer scared. When they [Israeli settlers or soldiers] come, I go out. I used to hide when the soldiers came by. Now, I go outside with my husband. It has helped me very much on a personal level (B'Tselem, 2014b).

Likewise, a volunteer named Lubna Saleh, began filming with a B'Tselem camera in 2007, after Israeli settlers set her family's car on fire. While she had previously felt a kind of fear of the settlers that kept her sealed within the domestic realm of her home, her camera has liberated her. She said:

I used to be scared to go out and face them [the settlers]. Now I'm not scared to go out and film…. True, I worry about getting injured, but I am more concerned about my children…. Sometimes, men ask me why I film. I tell them: to protect my home and my children (B'Tselem, 2014c).

To Saleh, the purpose of filming is to protect her domestic sphere (her home, her children). The camera has become a shield that safeguards the very realm over which she, as a woman, is traditionally held responsible. B'Tselem has made a point to call out women as the bearers of a particularly significant and unique burden of the Occupation's human rights violations:

Women are the ones who must find a way to run a household without regular or sufficient water. Women are the ones in charge of caring for children and they are the ones who must obtain food if their homes are demolished. Women are also usually the ones at home during settler attacks, so they are the ones who must shield their children (B'Tselem, 2014a).

In a traditional, patriarchal society, the realm of the woman is the home; yet the women from the village of 'Asirah al-Qibliyah describe exiting the home in order to film with their B'Tselem cameras. They speak of "going out" or "going outside" the home as a marked difference from their previous actions of "hiding" inside their domestic spaces, feeling "scared." Khadrah 'Abd al-Karim even intimated the outside realm as a distinctly masculine location when she said, "Now, I go outside with my husband." 'Outside' is the public sphere in which her husband appears, and in which she now joins because of her camera.

Many women volunteers who film for B'Tselem chose not to leave the home. Scholar Ruthie Ginsburg has researched the phenomenon of Palestinian women who film out their windows, making their own very private spheres sites for anti-colonial activism (Ginsburg, 2016). Of course, their homes are never quite private; the Palestinian home is constantly subject to the intrusion of the Israeli colonizing power, which 'makes its presence felt' through house searches, seizures, and demolition threats. But when women film out their windows, they achieve a superior perspective of 'looking down' at activity on the street, as if with the powerful view of Israeli hilltop settlements, just for a moment. Moreover women who film out their windows remain in their own distinct space, separate from the action 'out there'. Their acts of documentation therefore cannot be inhibited by physical assaults on the camera (such as grabbing, hitting, or breaking) because the camera resides in a separate space from the action. As Ginsburg has noted, a woman's "distinct space functions as a *camera obscura* – a darkroom where she sees and is unseen, watching the event without being a part of it" (Ibid., p. 49). In these cases, the Palestinian woman recording activates a mode of visual documentation in which, to put it crudely, her gender "protects" her footage.

The director of B'Tselem's video department, Ehab Tarabieh, reported that women currently comprise 40% of the Camera Project's volunteer videographers (2019). When asked of the growing number of women who volunteer, Tarabieh said, "Recently we want this," explaining that he believed these women to be "more calm" and not feeling the (implied masculine) urge to get involved in the altercations they record, therefore producing what he deemed "better" footage (Ibid.). Tarabieh said that most of B'Tselem's published footage has been recorded by women. B'Tselem celebrates women's participation in the Camera Project by featuring them in "campaigns, in articles and broadcasts that appear in the media, on the organization's blog, and in film festivals" (Ginsburg, in press, p.2) Tarabieh said to me, "Our dream [with the Camera Project] is to have only women," as they film much "better" than their male counterparts (2019).

Taken together, these personal comments and institutional celebration of the female videographer gesture towards something essential in women that produce the superior visual documentation of the Israeli occupation. While perhaps appearing feminist on the surface, this stance of celebration serves to gender the role of the witness – even as she wields the powerful video camera as a tool and weapon. While we might think B'Tselem's radical, counter-visual practices work "against the grain of normative representation," Ginsburg notes that the celebration and essentialization of women's visual documentation as providing a superior or unique perspective entails a problematic gendering that functions within – and problematically encourages – normative representations of women in zones of conflict (in press, p. 18).

#### **6.3 Women's Ethical Agency in Conflict**

Conflict, it is often said, is a man's fault (as in, "Men start wars"). While sounding trite, scholars have long researched the ways in which women and men are treated differently in war. R. Charli Carpenter's research shows that women are more likely to be deemed civilians than men (2006). To the category of women, she adds other populations that are likely to be considered civilians, no matter their context: children, the elderly, the sick, and the disabled. Put differently, in conflict "femininity and masculinity are often coded: the former as civilians, the latter as combatants" (Ginsburg, in press, pp. 7–8).

Moreover, women in conflict are more likely to be portrayed as victims than their male counterparts. Wendy Hesford has noted that the Palestinian sufferer is visibilized as a feminized victim with whom a viewer should sympathize (2011). Hesford has noted that this feminization of the victim is problematic for its re-inscription of the spectator's Western and neoliberal values on the subject, thus confining her (often non-Western) body to a violently partialized set of comprehensions. Indeed, B'Tselem's portrayal of women as innocent parties who, as Ginsburg notes, "should be shielded from violence and saved" (Ginsburg, in press, pp. 10–11) further problematizes the gendered portrayal.

I'd like to focus on the notion of a shield, for a moment, as an object of protection which the feminized victim lacks, but deserves as a civilian wishing to remain safe in times of conflict. A shield is, by definition, an object of protection. The classic shield is a large piece of metal that one wears to protect against bodily harm (from a blow, or a projectile weapon). Shields therefore function in a necessarily preventative manner. Shields act as defenses, not offenses: they assume a kind of danger 'out there' from which the body requires protection. B'Tselem volunteer Shuruk Saleh describes how she "goes out" to the public realm of appearances to "defend [her] rights" with her camera. Indeed, the English word *shield* derives from the proto-Indo-European word, *\*(s)kelH-,* meaning to "split" or "divide," as if to separate the interior vulnerability of the body from the violence of the exterior world. It is notably radical to portray the danger as 'out there', outside of Palestinian homes and instead in the space of appearances dominated and controlled by Jewish Israelis: such a portrayal reverses the occupation's narrative that it is *Palestinians* who are alien and violent intruders into Jewish Israeli space, and must, for instance, be kept in their place by an eight meter tall separation wall. Yet the very idea of a woman's need for a shield or defense plays into the idea that women are innocent civilians, victims, and generally lacking in political agency within the Israeli-Palestinian conflict.

#### **6.4 Watching the Land**

Allow me a brief detour to another zone of conflict: the contentious border between the United States and Mexico, where, in 2006, a program was founded to allow remote, civilian monitoring of the border in an American effort to curtail illegal immigration. A network of 200 security cameras were mounted along Texas's Rio Grande river, and were live-streamed online via a website called BlueServo.net (Moll, 2014). On BlueServo, citizens could flag and report suspicious activity, which was then sent to the corresponding Texas sheriff's office. The project transformed everyday civilians into a "Virtual Border Patrol" (or as the website called them, "virtual Texas deputies") who aided US law enforcement agents (Burnett, 2009). While the program ended in 2012 due to lack of funding, its website is still active and accepting new user registrations, even though its "live" camera feeds are now reruns (Texas Border Watchers BlueServo). At its height, though, BlueServo boasted involvement by over 203,000 volunteer watchers, representing an estimated 1 million hours in free labor for the Texas Sheriff (Moll, 2014).

Popular reporting on BlueServo focused on the role of women as watchers of the US border. Lori Andrews reported on an Oklahoma woman "who visits BlueServo each night after work and also tracks bald eagles online" and proclaimed, "'I watch eagles and illegals. That's a fun thing to do'" (2012, p. 7). CNN focused its coverage on a "suburban stay-at-home mom" from Rochester, New York named Sarah Andrews, who spent at least four hours a day watching BlueServo (Suttler, 2009). According to CNN's reporting – which was quoted and circulated in major US news outlets such as The New York Times – Andrews used to watch the border "when her baby girl takes an afternoon nap," shifting her maternal, watchful gaze away from her daughter and directly towards the land, in what can only be described as a stunning feminized understanding of watching that echoes the Israeli Defense Force's *Tazpitaniot* brigade (Ibid.).22 Sarah Andrews and the all-female *Tazpitaniot* unit do not "guard" like soldiers, nor do they "defend" like Border Patrol agents; instead they "watch," like babysitters.23

Palestinian women who film for B'Tselem also watch over the land with the eye of their lenses. In the more rural villages surrounding Nablus and South Hebron Hills, B'Tselem cameras are given to families living at the fringes of their villages who often are the first ones attacked by Israeli settlers nearby (a-Deb'I, 2019). Unlike B'Tselem volunteers who film in dense urban areas such as Hebron, these volunteers use tripods to stabilize their cameras as they record distant, sweeping, and often zoomed-in shots of their villages' land, its school, and agricultural or grazing fields that comprise their livelihood. Their use of tripods transforms their cameras into something more akin to mounted security cameras, aimed to watch the land and violations of it.

In unpublished footage from May of 2015, a B'Tselem volunteer named Thawra 'Eid filmed young Israeli settler children who had descended from a nearby Jewish settlement to launch an attack on her Palestinian village of Burin. Thawra films the wide expansive landscape between her home and the hill above as one of the settler children slingshots rocks in her direction (see Figure 6.3). As another example,

**<sup>22</sup>** When asked why she watches the US-Mexican border, Andrews cited two reasons: first, she hoped her watchful eye would prevent drugs from reaching New York via Mexico; second, she describes herself as "nosy about what's going on" along the border.

**<sup>23</sup>** A discussion of feminized watchers in Israel-Palestine would be incomplete without mention of Machsom Watch ("Checkpoint Watch"), a volunteer organization of Israeli women who have been observing the "West Bank checkpoints, the separation fences, the agricultural gates, the military courts and Palestinian villages"# since 2001 as anti-occupation peace activists aiming to influence public opinion against the occupation and to curtail bad action by the Israeli soldiers they watch. The group is comprised of primarily female retirees, meaning a generation of women who could be the soldiers' mothers. They say, "When we look at the soldiers as our sons, part of our flesh and blood, the next generation – we fear the mental experience they have to go through and the ethical values they trample on during the military service in the Occupied Territories" (Machsom Watch, cited in Ginsburg, 2011). In this way, their watching ties explicitly to babysitting. For more on Machsom Watch, see Halperin (2007), Carter Hallward (2008), and Kutz-Flamenbaum (2012).

a B'Tselem volunteer named Wydian Zaben filmed Israeli settlers as they torch the fields around Burin (see Figure 6.4). She documents the land to protect it, and to incriminate the settlers who aim to annihilate it. In both clips, we hear the women's children speaking in the background, reminding us that their gaze has not drifted far from their domestic life or indeed that they may be watching the land and their children, simultaneously.

I visited the homes of both Thawra and Wydian in 2019, accompanied by B'Tselem's field coordinator for the Nablus region named Salma a-Deb'i. Both women live at the outskirts of their Palestinian village that occupies a small fertile valley near Huwwara. As land became more expensive in the village, families like theirs began settling the hillsides of the valley, closer to the Jewish settlements and illegal Jewish outposts that sprung up on the hilltops above their village. Both Thawra and Wydian suffer from close proximity to the settler attacks, but also enjoy the scopic benefit of sweeping views of the village's land – views that are not beautiful nor enjoyable, but instead dangerous, heartbreaking, and rife with violence. Moreover, as mothers, both women frequently film attacks on Burin's single school which both their children attend, and which their homes look down upon with a clear unobstructed line of sight. Both women record Israeli settler attacks on the school from their homes.

**Figure 6.3:** B'Tselem volunteer Thawra 'Eid zooms in on Israeli settler children who are attacking her Palestinian village of Burin by throwing stones with slingshots, May 20, 2015, © B'Tselem.

**Figure 6.4:** B'Tselem volunteer Wydian Zaben films Israeli settlers torching the fields around her Palestinian village of Burin, Aug 15, 2015, © B'Tselem.

The settler violence in Burin is characteristic of a conflict entailing history of land disputes (Gerner, 1994; Caplan, 2011; Gelvin, 2014). Early Zionist efforts focused on land ownership and cultivation as a key feature of a new Jewish State. As Emanuela Rubenstein noted, "in the eyes of Herzl and his contemporaries, productivity meant one thing: engaging in agriculture" (Rubinstein, 2015). Agriculture was considered the key activity that transformed the character of the weak, urban Jew into a 'Sabra', a 'new Jew' who became strong, self-sufficient, and productive.

The shift towards agriculture entailed a shift in gender roles, as well. Eran Kaplan has written that mainstream Zionism entailed the "creation of a new society that would challenge traditional (diaspora) social divisions, including gender" instead fostering the images of "*halutzim* (male pioneers) and *halutzot* (female pioneers, see Figure 6.5) who together conquered the Palestinian wilderness" – a wilderness that was characterized as unproductive and barren by a Jewish colonial gaze (Kaplan, 2001, p. 12 ) The most radical experiment in gender equality was the Labor Zionist agricultural communes, or *Kibbutzim*, which boasted a view of women as equal to their male counterparts. Revisionist Zionists, on the other hand, continued to view the land as a distinctly "masculine sphere of play" whose boundary was extradomestic, while women remained relegated to the home and the family (Ibid., p. 14). However, even then, women were called upon to participate in "building the land" to serve the growing needs of the Jewish state (Kark, 2004, p. 139).

**Figure 6.5:** "The Po'alot (female workers) in the agricultural settlement Ayanot," Jan 2, 1940. Source: Schwartz Tel Aviv. From the The Jewish National Fund (JNF) Photographic Archive.

Yet if Zionism's inclusion of women in the public sphere could be considered feminist, it was only conditionally so. Zionism's early focus was wedded to land, and women were portrayed as essential to Zionism; yet, women were still characterized as having an essentially feminine connection to the land: a connection that served to sustain the next generation of the Israeli state through the dual responsibilities of motherhood and subsistence farming. Like the feminized conceptions of "mother earth" or "mother nature," Zionism fetishied women for their unique and essentialized relationship to children in the same way that it fetishised their unique and essentialized relationship to the land. This view was mirrored by NGOs in the 1980s, teaches Melissa Leach, who "put forward the view that women were the primary users and managers of the environment at the local level" (Leach, 2007, p. 69; see also Dankelman & Davidson, 1988; Rodda, 1991). At this time, NGOs around the world argued that women's work tied them intimately to the environment, especially in "reproductive and subsistencefocused activities" (Leach, 2007, p. 69).

At the same time as early Zionist factions were establishing themselves, Jewish organizations – most notably, the Jewish National Fund (JNF) – were founded to buy and develop land for Jewish settlement in Israel-Palestine. The JNF was founded in 1901 in what was then Ottoman Palestine, later British Mandate for Palestine. Together with its parent organization, today the JNF owns 93% of the land of Israel, which cannot be leased to foreigners, non-Jews, or Palestinians (Israel Land Authority, 2013). Notably, the JNF executed large campaigns of *afforestation* – distinct from the more common *reforestation*, which is planting trees where they once grew – instead introducing "trees to sites that never supported forests, or had no forest cover for a long period of ecological time" (Nyland, 1996). The JNF solicited donations from diasporic Jewry for its campaigns via its iconic blue donation boxes, to great effect and financial solvency. Today, many Jews donate money so that trees may be planted in Israel-Palestine in honor of loved ones or familial lifecycle events, and many even travel to Israel to plant the trees with their own hands (Jewish National Fund, n.d.).

In the early years, the JNF planted olive and fruit trees because of their romantic biblical associations (Stemple, 1998). But these trees required a lot of care. Therefore in the 1920s, the JNF switched to Aleppo pine trees (or *Pinus halepensis*), which became the main tree planted because it grew quickly and "suited the European image of a proper forest" (Shoham, 2017, p. 80). Others claim that the Pine was chosen for the acidic deposits its needles leave on the ground, which prevents undergrowth and deprives Palestinian shepherds of pasture (Weizman, 2007). The problematic focus on monoculture has created massive "Pine deserts" that now dominate Israeli forests, making them ecologically susceptible to devastating wildfires (Osem et al., 2011). Moreover, as of research released in 2014, more than two-thirds of the JNF forest sites are located on ruins of former Palestinian villages. In other words, the trees have been structurally planted to efface the visual traces of prior Palestinian inhabitation within the State of Israel (Ibid.; see also Weizman, 2007 and Berdugo, 2020).

What is striking about the violent colonialism of these JNF forests is the very innocence – and generally presumed *goodness* – of the simple act of planting a tree. This optimism is epitomized by the dedication of the book, *Trees as Good Citizens,* which is offered up in celebration of "every man, woman and child who plants a tree" (Pack, 1922).24 Likewise the US Secretary of Agriculture once proclaimed, "Every tree is beautiful, every grove is pleasant, and every forest is grand; the planting and care of trees is exhilarating and a pledge of faith in the future" (U.S. Department of Agriculture Forest Service, 1905, p. 10). Trees are held to be uncomplicatedly positive, and those who plant them are good. In this way they map cleanly on to the gendering of civilians in war: like a tree, a woman is presumed to be an innocent and vulnerable bystander – a victim just like the very scars left on a landscape itself. Certainly plants hold scars and other memories of damage. In the former village of Deir Yassin, twenty Palestinians were shot to death by Israeli paramilitary forces, with the fatal bullets piercing the prickly pear hedges behind them (see Figure 6.6). The cactus lived on, bearing the marks of this violence like a shadow, like a ghost.

**<sup>24</sup>** This book even opens one of its chapters with the following poem: "What does he plant who plants a tree? / He plants, in sap and lead and wood, / In love of home and loyalty, / And far-cast thought of civic good / His blessing on the neighborhood" (p. 23).

**Figure 6.6:** Deir Yassin (also transliterated as Dayr Yasin), Associated Press, April 1948.

#### **6.5 An Ecofeminist Critique of the Feminized Witness**

The feminized land, the feminized witness: ecofeminists would argue that both have been subjected to what Melissa Leach calls "a shared history of oppression by patriarchal institutions and dominant western culture" (2007, p. 70). From an over forty-year history of ecofeminist scholarship we are taught that women and nature share a mutual state of oppression from male-dominated society: women and nature are both seen as property, for instance; moreover, in the same way that men dominate women, humans dominate nature. In this way, the vast afforestation campaigns led by the JNF are part of an effort to dominate nature. Similarly, the subjugation of women to the role of passive "watchers" are part of systematic efforts to dominate or contain them under the patriarchy. Put differently, both women and nature have been subjected to what feminist philosopher Val Plumwood has called "the standpoint of mastery" of the self to the other, involving "seeing the other as radically separate and inferior, the background to the self as foreground." (Plumwood, 1993; Plumwood, 1996; Griffin, 2001, p. 284).

Against the backdrop of an ecofeminist critique, we can see that notions of women as possessing a natural or special relationship with the environment must be challenged alongside notions that women possess a special or uniquely innate ability to watch – especially over the land, itself – and to produce visual documentation of what they see. Instead, what might it look like to treat Palestinian women camerapersons as frontline defenders of human rights, who, like their male counterparts, record to oppose the occupation? In a 2016 panel that followed a screening of women-recoded B'Tselem videos at the Tel Aviv Cinematheque, one of the woman videographers was asked whether her 'womanness' uniquely affected her videos. Ginsburg described that this woman answered, without hesitation, *"No, I just was there"* (Ginsburg, in press, pp. 12–13). Indeed, what if Palestinian women videotape spectacular footage of the Israeli occupation because it is a horrific, monstrous spectacle of an occupation? Ecofeminists would argue that the notion of a uniquely feminized witness, who produces uniquely superior documentation, must be decommissioned as a mere social construction. An ecofeminist critique would further argue that such a conception of a woman B'tselem videographer is intricately and intersectionally linked to the subjugation of the land as a feminized and controllable commodity – a thing that can be afforested, sculpted, bought, owned, and fought over.

#### **6.6 Towards an Insurgent Way of Seeing**

In San Francisco's Prelinger Library, I came across a book titled *Should Trees Have A Standing?*, published by legal scholar Christopher Stone in 1974, the same year that French writer Françoise d'Eaubonne coined the term *ecofeminism*. Stone's book argues that trees and other natural objects should have their own legal rights, just as people and even corporations do. Trees could then claim damages when they were cut down; streams could demand reparations for pollution; and natural objects in general would have rights to "seek redress on their own behalf" through lawyers (Stone, 1974, p. 17). Perhaps this sounds absurd, but "each time there is a movement to confer rights onto some new 'entity' the proposal is bound to sound odd or laughable," wrote Stone (Ibid., p. 8). Certainly at one point it was laughable to bestow rights onto women, slaves, persons of color, and the incarcerated25; certainly, it is considered laughable under the current Israeli government to bestow rights to occupied Palestinians. Stone explains that this "is partly because until the rightless thing receives its rights, we

**<sup>25</sup>** One could argue that those rights have never been properly bestowed.

cannot see it as anything but a *thing* for those of us who are holding rights at the time" (Ibid.). Rights give 'things' agency to be more than things. Rights remove 'thingness', transforming things into bodies, entities, and states.

Palestinian women who film for B'Tselem are made into *things* twice over. First, they are 'thing-ed' by the Israeli regime by their status as Palestinians, who lack rights generally considered to be universally human under the UN's Universal Declaration of Human Rights (UN General Assembly, 1948). Secondly, these women are 'thing-ed' by the valorization of their position as feminized watchers, as feminized watchers and feminized producers of visual documentation. Contemporary Palestinian feminists remind us, though, that the primary cause of Palestinian women's subjugation – and Palestinian society more broadly – is the Israeli occupation, not their status as women (Sayigh, 1981; Souad, 1994; Jad, 1995; Allabadi, 2008; Rought-Brooks et al., 2010). These feminists – together with ecofeminists – beg the question: what might it look like, visually, to treat the Israeli occupation itself as the preeminent subjugator of women B'Tselem volunteers? How might it appear to pay attention, with a sincere attention to vision and visuality, to Palestinian visual documentation of the Israeli occupation?

Following scholars like Judith Butler (2004, 2009) and Ariella Azoulay (2008, 2011b), I argue that what is needed is an insurgent way of seeing: a kind of vision that acts contrary to the dominant political regime by visibilizing the frames of images themselves – or to use Azoulay's terminology, that considers the photographic event as a whole, not merely the image that results from a photographic act. This kind of seeing is rebellious in its opposition to established norms and dominant scopic and political paradigms. Indeed, it is a shift in comprehension of what injustices these images seek to remedy, from what legal scholar Nancy Fraser defines as *ordinarypolitical misrepresentation* to the much more severe *metapolitical injustice* (2008). Injustices of ordinary-political misrepresentation are injustices in which a civil society and its government denies people who are generally agreed to be members of that society the opportunity to participate in decision making, as peers or equals. As Fraser describes, these political injustices occur within political societies "whose boundaries and membership are widely assumed to be settled" (Ibid., p. 407). For instance, ordinary-political misrepresentational injustice occurs when a citizen is denied a fair trial under law, for instance. Such injustices occur "when a polity's rules for decision making deny some who are counted in principle as members the chance to participate fully…" (Ibid.). The citizenship or membership of such a person into a political community has not been called into question, as it is generally agreed. Instead, the injustice occurs from the mistreatment of that citizen by the polity.

The second and more severe level on injustice is one in which civil society and its government wrongly draw the boundaries of citizenship. This kind of injustice, called *metapolitical injustice*, entails the denial of civil membership to a population, or the denial of its opportunity to participate in what Fraser calls "authorized contests over justice" such as elections (Ibid., p. 408).26 Metapolitical injustice arises "as a result of the division of political space into bounded polities" or when "a polity's boundaries are drawn in such a way as to wrongly deny some people the chance to participate at all in its authorized contests over justice" (Ibid.). In such cases, a person is not simply denied the right to ordinary-political representation, but denied the very right of constituting a body that is entitled to political representation at all. Here I think of the American civil rights movement, and its work against the mischaracterization of African-Americans as non-Americans who were not entitled to the right to vote. This kind of justice is a second-order injustice. It is *meta* political, of a higher order than the realm of the political, itself. This injustice is perpetrated by an act of misframing persons outside of the edges of the political, as if pushing them outside the boundaries of a frame, a photograph.

Just as ecofeminists teach us that what is needed to protect ecology is not a feminized conception of nature, I have shown that what is needed is not a celebrated feminized witness. Instead, what is needed is an insurgent, disobedient way of seeing that conceptualizes events captured as metapolitically unjust. Put differently, what is demanded by this way of seeing is not a celebration of the female witness, the female gaze, or the feminized connection to nature; instead, this way of seeing consider the frame of the image to be the problem, itself. This is, ultimately, a way of seeing that elevates citizen videography in zones of conflict to engage in radically new possibilities of resistance.

#### **Acknowledgements**

I thank the staff and volunteers at B'Tselem for their generosity and accomodation, which has made this research possible. Thank you to my Arabic language translator, Yuval Orr; and to the Dorot Foundation whose funding has enabled portions of this research.

#### **References**


**<sup>26</sup>** A citizen can be denied of the right to vote, or denied the right to elect equal representation by, say, district gerrymandering.


Azoulay, A. (2008). *The civil contract of photography* (R. Mazali & R. Danieli, Trans.). Zone Books.


**86** Insurgent Ways of Looking


**88** Insurgent Ways of Looking


## **7 Gender and Race in the Digital Town Hall: Identity-Based Attacks Against US Legislators on Twitter**

Sarah Sobieraj and Shaan Merchant

#### **7.1 Introduction**

In August 2018, amidst an unprecedented number of women running for public office in the US, the New York Times released a video of current and former female candidates talking about their experiences with harassment and sexism, much via social media, during their campaigns (Kerr, Tiefenthäler, & Fineman, 2018). Throughout the video, the women describe abuse that is gendered and racialized. Among the women in the video is Iowa Democrat Kim Weaver, who pulled out of her 2016 congressional race amid a torrent of sexist and anti-Semitic abuse (Astor, 2018). There is ample reason to believe racist and sexist abuse plague white women and people of color once they hold office as well. In 2017, Senator Kirsten Gillibrand received a throng of sexist tweets after she called for Senator Al Franken's resignation. And as recently as June 2018, after Representative Maxine Waters called for lay people to confront Trump administration representatives in public, she faced such intimidation on- and offline that she cancelled two speaking engagements out of concern for her safety. This is not unique to politicians in the US. In 2016, British Member of Parliament Jess Phillips received 600 rape threats on Twitter in one day in response to strongly worded remarks she made about sexism from the left (Elliot & Turner, 2017).

Previous research has identified abuse of politicians on social media but has not identified – and in some cases not looked for – racial or gender differences. The anecdotal evidence above, coupled with recent research documenting widespread digital harassment of people of color and white women in arenas ranging from gaming (Gray, 2012; Fox & Tend, 2017) and academia (Ferber, 2018; Veletsianos et al., 2018) to journalism (Gardiner, 2018; Adams, 2018; Chen et al., 2018), suggests systematic research is needed to explore how participation in digital publics differs for political figures from historically underrepresented demographic groups. This research mines the content of tweets containing @mentions to 16 US legislators, revealing that male and female politicians on social media navigate very different digital worlds. What's more, we find particularly egregious hostility directed at female legislators of color, much of it explicitly identity-based, drawing on racist and sexist stereotypes and epithets to discredit, intimidate and shame the recipients.

These unevenly distributed consequences of public service represent tangible evidence of patterned resistance to political voice and visibility of those from historically underrepresented groups. This abuse undoubtedly has a variety of personal and political costs for the legislators targeted for attack, but we must also consider the social costs: as a result of this blowback, we lose qualified candidates and hear a narrower range of perspectives. The macro and micro implications of this hostile work environment will need to be explored more fully, as leaving Twitter, YouTube, and Facebook are not practical options in a political context where the populace increasingly gets information via social media platforms and evidence mounts that a robust social media presence helps candidates win elections.

#### **7.2 Literature**

There is a burgeoning body of literature that explores politicians' use of social media, particularly its role in campaigning (e.g. Bimber, 2014; Evans et al., 2014; Jungherr, 2015; Sweetser & Lariscy, 2008) and constituent interaction (e.g. Jackson & Lilleker, 2011; Larsson & Øyvind, 2013; Waisbrod & Amado, 2017). Research shows that most political leaders prefer to use these platforms to broadcast, showing greater interest in one-way communication than in engaging voters or constituents (Gibson et al., 2014; Graham et al., 2013; Hoffmann & Suphan, 2016; Lev-On, 2011; Sweetser & Lariscy, 2008). This is not particularly surprising given that members of the general public often respond to politicians in hostile, even abusive, ways. Most politicians in the US have experience with combative town hall meetings, uncomfortable encounters while canvassing, and being on the receiving end of rancorous phone calls. Lambasting and jeering politicians is not a new phenomenon – there is a long tradition of lashing out at politicians, with tactics ranging from heckling and hate mail to shoe and tomato throwing (Bennett, 1979; Ibrahim, 2009; Temkin & Yanay, 1988).

While social media were not required to give politicians a hard time, platforms such as Twitter, Facebook, YouTube, and Instagram create new points of access for the public and offer opportunities for feedback that require little time, skill or expense. These new pathways for communication between political leaders and the public create new venues for negative and potentially embarrassing or harmful flak (in addition to more constructive engagement), while reducing the personal investment and the risk of repercussions for critics, particularly for those hiding behind the veil of anonymity.

Recent research documents harassment of political figures extending into the social media space (James et al., 2016; McLoughlin & Ward, 2017). Yet, looking at the findings, one might conclude that the digital abuse is evenly distributed, something that seems empirically unlikely given existing identity-based hierarchies. James, Farnham, Sukhwal, Jones, Carlisle, and Henley (2016) conducted a four-site, crossnational survey of MP experiences with digital harassment in the UK, Australia, New Zealand and Norway and found a tremendous amount of abuse, but did not examine gender or racial patterns in their data. And, in 2017, McLoughlin and Ward conducted an analysis of tweets to MPs and found that the gender of the MPs didn't matter (they did not examine race). The most extensive academic work on harassment of political figures (online and offline) comes from the threat assessment literature, which focuses on the perpetrators and evaluates risk in nearly universal psychological terms without attending to the broader patterns of victimization (Dietz et al., 1991; Adams et al., 2006; Hoffman & Sheridan, 2008; Meloy, 2014).

We have little research about how the digital aspects of public life vary for political leaders with different social locations, save an Amnesty International finding that among women, black, Asian, and minority ethnic MPs received 41% of the abusive tweets, even though white female MPs outnumber them nearly 8 to 1 (Stambolieva, 2017). This gap in our understanding exists even though there are reasons to suggest representatives of color and white women find digital publics more hostile than their white male counterparts. For example, evidence suggests that women in politics experience disproportionate amounts of violence offline (Dalton, 2017; Krook, 2017). Recently, Krook (2017) looked at journalistic accounts and research reports from nongovernmental organizations such as the National Democratic Institute and concluded that violence against women in politics is rampant, particularly in African countries. And as the first African American President, Barack Obama entered office faced with an unprecedented number of death threats (Parks and Heard, 2009). It seems counterintuitive that digital spaces would diverge from these offline patterns.

The fact that differential digital abuse of political leaders has not been studied is also peculiar, given the noteworthy gendered and racialized patterns to digital abuse in other arenas (Chen et al, 2018; Gardiner, 2018; Ferber, 2018; Nakamura, 2002; Daniels, 2009; Herring, 1999; Citron, 2014; Gray, 2012; Veletsianos et al., 2018; Sian, 2018). Watching politics – at least in the US – certainly suggests that gender and race matter. Consider the backlash against Florida Representative Frederica Wilson; After Wilson criticized President Donald Trump's phone call with Myeshia Johnson, the widow of La David T. Johnson who had been killed in Niger, she received a deluge of abuse via social media, much of which took on overtly racial and gendered dimensions. Figures 7.1–7.5 offer illustrations. The comments are pointedly racialized in several ways, including the disturbing suggestion that Wilson be lynched and the reference to her and Representative Maxine Waters as "race hustler pimps." The references to Waters' physical attractiveness and the description of her as a "hooker" show the way that the pushback can be gendered. These examples also offer insight into the ways abuse can be intersectional, as seen in the reference to welfare – an implicit connection to one of the most pernicious stereotypes of black women in the US: the lazy, entitled welfare queen. Such vitriol is an example of what Moya Bailey calls **"**misogynoir**,"** hatred directed towards black women where race and gender intersect and play a pivotal role in the discrimination, abuse and bias.27 US Representatives Ilhan Omar (D-MN) and Alexandria Ocasio-Cortez (D-NY) have also endured venomous identity-based abuse online. For example, the comments directed at Omar, a Somali-American and

**<sup>27</sup>** See Bailey (2018) for a discussion of the origin of the term.

observant Muslim, include references to her as a "rag head cunt," a "gorilla looking bitch," and a "sand nigger," as well as outlandish accusations of her participating in cannibalism, an incestuous relationship with her brother and terrorist activity (Sobieraj, 2019).

**Figure 7.1:** Tweet 1.

**Figure 7.2:** Tweet 2.

**Figure 7.3:** Tweet 3.

**Figure 7.4:** Tweet 4.

#### **94** Gender and Race in the Digital Town Hall

Sobieraj (2017, 2020) highlights the ways in which such abuse is often heightened in male-dominated spheres such as science and technology, gaming and sports. It makes little sense intuitively that politics – another male-dominated space – would be different.

In light of this, McLoughlin and Ward's (2017) finding that gender is not a significant predictor in the abuse of MPs on Twitter seems suspect. We suspect the finding reflects a validity issue, generated as an artifact of the combination of the sentiment analysis and key-word coding process, which – while excellent at capturing obscenity and epithets –misses a great deal of hostile content. As a result, we revisit Twitter to explore how the experiences of legislators differ based on their racial and gender identities, but with a more interpretive approach and coding attuned to context.

#### **7.3 Methods**

To assess gender and racial differences in the treatment of politicians on Twitter, we examine tweets directed at legislators that incorporate three modes of harassment identified in Sobieraj (2017): attempts to discredit, intimidate, or shame. These kinds of abuse are more complicated to assess than the presence of particular words, but this reliability challenge is far outweighed by an enhanced ability to capture digital hostility in a meaningful way. We bolstered reliability through pre-testing and revision of the codebook, which resulted in lengthy operational definitions and examples to increase coder confidence and accuracy. In addition to checking for the presence of these modes of abuse, coders also assessed tweets for specific content, such as: the invocation of the legislators' race and gender and references to physical appearance, sex and the body, as Sobieraj (2017) identifies these as common in digital attacks against women, particularly women of color. Thumbnail descriptions of the variables are provided in Appendix A and the complete codebook is available upon request. We use these data to test the following hypotheses:

H1a: Attempts to discredit will be more common in tweets targeting female legislators than in those targeting men.

H1b: Attempts to discredit will be more common in tweets targeting legislators of color than in those targeting white legislators.

H1c: Attempts to discredit will be more common in tweets targeting female legislators of color than in those targeting other groups.

H2a: Attempts to intimidate will be more common in tweets targeting female legislators than in those targeting men.

H2b: Attempts to intimidate will be more common in tweets targeting legislators of color than in those targeting white legislators.

H2c: Attempts to intimidate will be more common in tweets targeting female legislators of color than in those targeting other groups.

H3a: Attempts to shame will be more common in tweets targeting female legislators than in those targeting men.

H3b: Attempts to shame will be more common in tweets targeting legislators of color than in those targeting white legislators.

H3c: Attempts to shame will be more common in tweets targeting female legislators of color than in those targeting other groups.

In terms of tweet content, we hypothesize:

H4a: The race (of the legislator) will be invoked more in tweets directed toward legislators of color than in those directed toward white legislators.

H4b: The race (of the legislator) will be invoked more in tweets directed toward female legislators of color than in those directed toward men of color.

H5a: The gender (of the legislator) will be invoked more in tweets directed toward female legislators than in those directed toward male legislators.

H5b: The gender (of the legislator) will be invoked more in tweets directed toward female legislators of color than in those directed toward white women.

#### **96** Gender and Race in the Digital Town Hall

H6a: References to physical appearance, sexual situations and the body will be more common in tweets directed toward female legislators than in those directed toward male legislators.

H6b: References to physical appearance, sexual situations and the body will be more common in tweets directed toward female legislators of color than in those directed toward white women.

To isolate the effects of race and gender, we compiled a list of 16 US senators and representatives: four are white male politicians, four are white female politicians, four are male politicians of color and four are female politicians of color. Lack of diversity in the House and Senate prevented us from exploring more fine-grained racial distinctions or from taking sexual orientation and gender identity into account. Legislators were sampled purposively based on race, gender, visibility on Twitter, national name recognition and party. A nearly equal number of Democrats and Republicans are represented across groups; however, due to the low number of Republican women of color in congress, we were unable to achieve parity in that group (three Democrats (D) and one Republican (R) were selected). The 16 legislators included are: Leader Mitch McConnell (R), Speaker Paul Ryan (R), Sen. Bernie Sanders (D), Sen. Chuck Schumer (D), Sen. Cory Booker (D), Rep. Keith Ellison (D), Sen. Marco Rubio (R), Sen. Tim Scott (R), Sen. Susan Collins (R), Sen. Lisa Murkowski (R), Rep. Nancy Pelosi (D), Sen. Elizabeth Warren (D), Sen. Tammy Duckworth (D), Sen. Kamala Harris (D), Rep. Mia Love (R), Rep. Maxine Waters (D).

We amassed a database of tweets directed toward these 16 legislators via @ mentions (in the form of @replies and original tweets) using Twitter Archiver. We then conducted a census of the first 152 tweets directed at each legislator beginning at a specified time on March 6, 2018. For higher profile legislators, these 152 tweets were accrued in a matter of hours, while for others they spanned several days. When two or more politicians from our list were @mentioned in the same tweet, the tweet was discarded, unless the named politicians shared the same race and gender. In addition, a number of tweets were discarded because they had been deleted by the time of coding (in late March 2018), their links or images no longer existed, or they were identified as spam. Of the 2,432 possible tweets, 2,216 usable tweets were coded. Legislators averaged 138.5 usable tweets. The census (n = 2,216) consisted of 548 tweets @mentioning women of color, 566 @mentioning white women, 560 @ mentioning men of color, and 542 @mentioning white men. After training and pilot testing, a team of three researchers manually coded the tweets. Two of the coders identify as non-Hispanic white women, and the third coder identifies as an Asian-American man. Inter-rater reliability testing indicates the level of agreement among coders to be very good, with Cohen's Kappa = 0.925 (SE = 0.021, with 95% confidence interval: from 0.883 to 0.967).

#### **7.4 Analysis**

As described above, this research involves a census of each of the first 152 @mentions positioned toward each of the 16 legislators at the time of our investigation, eliminating the variation that would have been introduced had we relied on random selection to generate our sample. Because the study is based on a population, inference is not needed to interpret the findings, and can, in fact, be misleading (Gibbs et al., 2017; Gorard, 2013). We, therefore, follow Kenski, Filer, and Conway-Silver's 2018 analysis of campaign tweets and restrict our analysis to descriptive statistics.

#### **7.5 Findings**

Our first set of hypotheses looked at three common strategies used to limit the speakers' impact in digital publics as identified by Sobieraj (2017): discrediting, intimidating and shaming. Tweets that were coded as discrediting attempts include insinuations or accusations that the targeted legislator is not qualified, capable, well-informed, trustworthy or deserving of respect, as seen in this tweet: "@RepMaxineWaters Ms. Waters you should get ready for the asylum. You are disoriented, angry, and paranoid," which contains the not-so-subtle subtext that Waters should not be taken seriously because she is mentally unstable and overly-emotional. We hypothesized that these types of comments would be more common in tweets directed at female legislators (regardless of race) than male legislators, more common in tweets directed toward legislators of color (regardless of gender) than in those directed toward white legislators and most common in tweets directed at female legislators of color.

In terms of gender, our hypothesis was confirmed: 58.44% of tweets directed toward female legislators attempted to discredit them, while "only" 37.84% of tweets directed toward male legislators did so (see Table 7.1 and Figure 7.6). Tweets containing discrediting efforts are 2.31 times more likely to be directed at female legislators (odds ratio: 2.3097, 95% CI: 1.9479 to 2.7386, p < 0.0001). In contrast, legislators of color (48.65%) and white legislators (47.74%) received a roughly equivalent amount of such criticism (odds ratio: 0.9645, 95% CI: 0.8164 to 1.1394, p = 0.6707). Figure six shows the percentage of tweets directed at each subgroup of legislators that contained discrediting attempts. The result prevents us from confirming hypothesis 1c; the highest percentage of discrediting tweets were directed at white women (68.37%) rather than women of color (57.12%) as we predicted. Attempts to discredit Nancy Pelosi were particularly common; a remarkable 82.7% of all tweets she received attempted to discredit her.

**Figure 7.6:** @mentions containing attempts to discredit.


**Table 7.1:** @mentions containing attempts to discredit.

Tweets @mentioning legislators that contain direct or indirect threats of reputational, political or physical harm to the elected official, those close to them or their party are relatively infrequent (see Figure 7.7 and Table 7.2),28 but the intimidation attempts that

**<sup>28</sup>** Importantly, while attempts to intimidate are not abundant, they are more common than this

do appear are three and a half times as likely to be directed at female legislators than male legislators (odds ratio: 3.5258, 95% CI: 1.5998 to 7.7704, p = 0.0018). However, while the percent of intimidating tweets positioned at legislators of color is higher (1.81%) than those of white legislators (1.44%), this difference is not statistically significant (odds ratio: 1.2546, 95% CI: 0.6467 to 2.4341, p = 0.5024). Stepping back to look at all four subgroups (Figure 7.7), we see intersectional effects at play. As hypothesized, attempts to intimidate are more common in tweets targeting female legislators of color than in those targeting other groups. Among legislators of color: female legislators of color received a higher proportion of intimidating tweets (2.92%) than their male counterparts (0.71%). Women of color also received a higher proportion of @mentions using intimidation tactics than white women, but the difference is smaller: 2.92% in contrast to 2.12%.

**Figure 7.7:** @mentions containing attempts to intimidate.

dataset suggests, as most overtly threatening tweets directed at public figures are caught relatively quickly by Twitter. Indeed, there were a number of tweets no longer available for coding and accounts that has been suspended at the time of coding. While these tweets may be missing for a variety of reasons, in all likelihood a subset contained content that would qualify as intimidation.


**Table 7.2:** @mentions containing attempts to intimidate.

For the purposes of this study, shaming involves efforts to taint the public perception of the legislators by exposing something personal, rather than political, that is meant to be hidden or private, taking a personal, rather than political, action that was initially public and recast it in a markedly negative light, or fabricating personal, rather than political, stories intended to be "gotcha" type moments that purport to reveal something about the person's character. "@NancyPelosi still hitting the booze pretty hard @NancyPelosi?! Either that or Botox is rotting your brain!" We hypothesized that attempts to shame would be more common in tweets directed at female legislators than male legislators as well as for legislators of color than white legislators. Both hypotheses were confirmed, as shown in Table 7.3. We found 7.81% of @mentions directed toward female leaders contained personal shaming, in contrast to 2.81% of those directed at male leaders (odds ratio: 2.3097, 95% CI: 1.9479 to 2.7386, p < 0.0001). In terms of race, 3.88% of @mentions directed at white legislators contained attempts to shame, in comparison to 6.14% of those directed at legislators of color (odds ratio: 1.6194, 95% CI: 1.0950 to 2.3950, p = 0.0158). And, as predicted, intersectional abuse is compounding; women of color are by far the most likely to receive @mentions that target them with personal shame (see Figure 7.8). One out of every ten tweets directed at women of color involves shaming, about 40% more than white women and almost 80% more than men of color. We are therefore able to confirm the third related hypothesis; attempts to shame are more common in tweets targeting female legislators of color than in those targeting other groups.

**Figure 7.8:** @mentions containing attempts to shame.


**Table 7.3:** @mentions containing attempts to shame.

Often when tweets contain attempts to discredit, intimidate or shame their targets, attackers use identity-based pushback as leverage. The next set of hypotheses were intended to test the prevalence of @mentions containing identity-based content. First, we hypothesized that legislators' racial, ethnic and/or religious identity would be invoked more often for people of color than for whites, and that among people of color, women would receive a higher proportion of this content than men. One such example of invoking racial identity: "@SenatorTimScott Hey Timmy boy..i see you the black lackey at the white house.. did you take @LindseyGrahamSC coffee this morning. ? you obviously dont like people of your color..you chose a white man over a black man..Afro-Americans going to drop you good.. your career is over." As seen in Table 7.4, both hypotheses were confirmed; tweets directed at legislators of color are almost three times as likely to include references to their race or perceived race (7.04%) than those positioned toward white legislators (2.53%) (odds ratio: 2.9238, 95% CI: 1.88827 to 4.5405, p < 0.0001). Gender compounds the abuse; 10.40% of the tweets @mentioning women of color include race-based commentary in contrast to 2.14% of those @mentioning men of color (odds ratio: 2.9796, 95% CI: 1.7802 to 4.9871, p < 0.0001).

The heavy burden on female legislators of color reappears when we explore @ mentions that include comments about elected officials' gender or gender-identity. These kinds of remarks often use gender as a means to devalue or discredit the legislators or their contributions (e.g. "@NancyPelosi Pelousy you're just a commie hag," "@NancyPelosi they are not Americans, sweetheart" and "@NancyPelosi shut up bitch"). We were able to confirm our hypothesis that gender is invoked more in tweets directed toward female legislators than male legislators; Table 7.5 shows that 7.99% of tweets @mentioning female leaders reference or comment on their gender, in contrast with 2.90% of the tweets directed at male leaders (odds ratio: 2.9034, 95% CI: 1.9211 to 4.3879, p < 0.0001). Once again, gender-talk is disproportionately directed at female legislators of color; tweets including this kind of commentary are 1.66 times more likely to target female legislators of color than white female legislators (odds ratio: 1.6584, 96% CI: 1.0653 to 2.5818, p = 0.0251).


**Table 7.4:** Percent of @mentions invoking race (of legislator).


**Table 7.5:** Percent of @mentions invoking gender (of legislator).

Our final hypotheses involved direct and indirect comments about elected officials' physical appearance, bodies and/or sexual behavior. For example, one such tweet reads, "@RepMaxineWaters @realDonaldTrump I swear I think James Brown faked his own death and returned in drag as Maxine Waters." We predicted that tweets directed at female legislators would more often include body-based commentary than those directed at their male counterparts. This was confirmed: 4.04% of @mentions directed at women included such comments, while only 2.18% of at mentions directed at men did so (odds ratio: 1.9808, 95% CI: 1.1439 to 3.1255, p = 0.0130), as visible in Table 7.6. We can also confirm our hypothesis that body-based comments are more common for female legislators of color than for white women (odds ratio: 2.0334, 95% CI: 1.0293 to 4.0168, p = 0.0410).


**Table 7.6:** Percent of @mentions containing body commentary.



#### **7.6 Discussion and Conclusions**

As visible in Table 7.7, 12 of the 15 hypotheses were confirmed, revealing that in the case of US legislators, identity plays a central role in determining who receives negative pushback on social media as well as the ways they are attacked. In other words, people from historically underrepresented groups – particularly women of color – deal with more shaming, discrediting and intimidation, and they also manage a barrage of messages suggesting that they have little of value to contribute because of their gender and race. Attackers draw on racial and gender stereotypes as a way to invalidate the ideas and efforts of people from these underrepresented groups. Attackers are particularly aggressive toward women of color, who receive gendered abuse, racialized abuse and abuse that is intersectional in nature.

While this research documents some of the uneven terrain navigated by US legislators, we suspect the differences are of a greater magnitude than we have been able to capture. Future research would be well served to examine variation in tone and intensity of the messages directed toward political figures from different groups. In looking at the tweets, it is not uncommon to find that even when body-based commentary, for example, is present for white male legislators, it is relatively mild. Mitch McConnell, for example, is often called "turtle," a long-running reference to Jon Stewart saying that McConnell looked like a turtle on The Daily Show. While this is rude and may be hurtful, it feels substantively different from a tweet @mentioning Tammy Duckworth that says, "Bitch please go fuck yourself with a cactus." We attempted summary measures for overall tone, but in examining the interrater evaluations, determined that they were not adequately reliable and were forced to discard them. Future research should work on capturing these critically important distinctions.

Much of this story is bleak, indeed, even white male legislators are unlikely to find Twitter a fully-comfortable environment – almost 40% of the tweets directed at them include attempts to discredit them. But we were looking for nastiness and did not code for prosocial interaction or substantive political engagement. Had we done so, we might have been able to capture the tweets we read that addressed policy preferences, raised questions and revealed that amid tweets telling Maxine Waters that she is, "a loud mouth neger liberal pile of shit," or telling her to "TAKE OFF THAT SKANKY WIG," there are also tweets that point legislators to attend to issues they feel deserve attention, some that ask questions about how representatives' stated goals would be accomplished, those that work through stories in the news, and those rife with gratitude and celebration. There is value here, at least for some.

These unsettling attempts to limit the political voice and visibility of those from historically underrepresented groups remind us that winning an election does not mean the victors have overcome discrimination. Indeed, the near mandatory presence in digital town halls may place elected officials into closer contact with misogyny/misogynoir and racism than they have had to deal with in the past. This abuse has a variety of personal and political costs for those targeted for attack, but it also has social costs: if we lose qualified candidates or hear from people of color and white women less often or more guardedly as a result of this abuse, our digital public spaces contain a narrower range of perspectives. The macro and micro implications of this hostile work environment will need to be explored more fully, as leaving Twitter, YouTube and Facebook are not practical options in a political context where the populace increasingly gets information via social media platforms and evidence mounts that a robust social media presence helps candidates win elections.

#### **References**


## **Appendix A: Thumbnail Descriptions of Key Variables**

## **ATTEMPTS TO DISCREDIT**

This variable is intended to measure whether the tweet suggests that the elected official is not qualified, capable, well-informed, trustworthy, or deserving of respect. Attempts to discredit generally work to deflate or undercut a person's status (see codebook for detailed examples).

## **INTIMIDATION**

This variable is intended to measure whether the tweet contains direct or indirect threats of reputational, political, or physical harm to the elected official, those close to them, or their party. This code was not be used for dispassionate political forecasting (i.e. "this could hurt their chances for reelection"); it was be reserved for attempts to intimidate (see codebook for detailed examples).

## **PERSONAL SHAMING**

This variable is intended to measure whether the tweet includes efforts to publicly humiliate or contaminate the public perception of the target as a person. Personal shame**:** 1) exposes something personal that is meant to be hidden or private (or pile on when someone else does so), 2) takes a personal action that was initially public and recast it in a markedly negative light, or 3) fabricates personal stories intended to be "gotcha" type moments that purport to reveal something about the person's character. Note that political shame, such as suggesting someone's political decisions have been shaped by financial interests, were counted as "discredit," not personal shame (see codebook for detailed examples).

## **RACE/RACIALIZATION**

This variable is intended to measure whether the tweet contains direct or indirect comments about the elected official's perceived racial, ethnic, or religious identity. The variable is – at its heart – intended to capture how often and for whom racial / ethnic / religious identity is invoked. (see codebook for detailed examples).

#### **GENDER TALK**

This variable is intended to measure whether the tweet contains direct or indirect comments about the elected official's stated or perceived gender identity. The variable is – at its heart – intended to capture how often and for whom gender identity is invoked (see codebook for detailed examples).

#### **BODY-BASED COMMENTARY**

This variable is intended to measure whether the tweet contains direct or indirect comments about the elected official's physical appearance or sexual behavior. (see codebook for detailed examples).

## **8 Participatory Propaganda: The Engagement of Audiences in the Spread of Persuasive Communications**

#### Alicia Wanless and Michael Berk

Rapidly evolving information communications technologies (ICTs) and increased connectivity to online sources have drastically altered the ways individuals engage in the public information domain and, in turn, become influenced by it. As citizens become increasingly plugged-in, a savvy propagandist can acquire enhanced means of swaying opinions around the world by combining tech-enabled formats of persuasive content, automated dissemination and audience engagement in content propagation. In addition, by obfuscating the origins of propagandistic content through audience participation via the internet and social networks, the propagandist can also increase its receptivity and influence effects. For example, people tend to find recommendations from their personal social network more credible than others (Nielson, 2015). Such subtle mass persuasion, through and by means of personal networks, is problematic in liberal democracies founded on the premise that freedom of choice by citizens on political matters is expected to inform public decision-making and power structures (Lippman, 1922; Irwin, 1919; Marlin, 2011).

This paper is broken into three sections. In the first, a multidisciplinary literature review aggregates individual studies published recently that analyze the known digital, behavioral and psychological tactics available to propagandists aiming to engage target audiences online. Extensive research conducted by scholars over recent years on new ICT tools, social networks, influence tactics and their manifested effects on consumers of online information, including their political choices, has been instrumental in acquiring the first appreciation for the scale, complexity and social repercussions of modern persuasive communications – what we called "participatory propaganda." In the second section, the paper draws on original research conducted during the 2016 US presidential elections to analyze how Trump supporters applied these tactics to engage Facebook followers in the promotion of persuasive content, thus encouraging them to become propagandists themselves. In the conclusion, the research results are placed into the broad context of the emerging information environment describing possible repercussions for citizens' political engagement and arguing that further modelling of digital propaganda is required to better understand the risks to liberal democracy associated with using such techniques.

#### **8.1 Modern Propaganda and the Evolution of Its Participatory Model**

Propaganda is a much-contested term applied in many contexts under different names either to explain and justify more acceptable forms of public influence, or to denigrate and dismiss similar efforts done by perceived opponents. This difficulty in defining propaganda stems in part from its complicated relationship with liberal democracies where public opinion is expected to influence political decision-making and the act of manipulating it calls into question the agency of voters and the democratic system itself.

Traditionally, propaganda has been described as the use of persuasive information to manipulate a target audience into a behaviour desired by the propagandist (Bernays, 1928; Lasswell, 1948; Ellul, 1965; Marlin, 2013; Jowett & O'Donnell, 2015). In this topdown communications model, the sender-receiver roles were typically static with the propagandist (government, corporate, military) issuing persuasive messaging to achieve a specific outcome among the target audience (general public). Early pioneers in the field of public relations, who also led American propaganda activities during World War I, saw how easily public opinion could be swayed and, yet, saw domestic propaganda as an acceptable tool for the management of popular views (Bernays, 1928; Lasswell, 1934; Lippmann, 1922). Around the same time, propaganda began to acquire a pejorative connotation as it was "associated mainly with totalitarian regimes and war efforts" and "was perceived as a threat to liberal democracies" (Ross, 2002, p. 17). To differentiate between acceptable forms of influence, domestic- and foreignbound, other terms such as public affairs or public relations (Moloney, 2006), or public diplomacy or information operations (Garrison, 1999) were introduced.

This classic understanding of propaganda, however, must be adapted for the Digital Age. With the internet and social media, the traditional separation between "the propagandist" and "target audience" is rapidly blurring, with the latter beginning to play a more significant role in spreading propagandistic content and influencing others through personal networks – a more dangerous development since people are more likely to believe those familiar to them (Garrett & Weeks, 2013) or those they view as influential (Turcotte et al., 2015).

In this context, a new concept for the understanding of modern propaganda is suggested – participatory propaganda – that adapts the traditional definition proposed by Jowett & O'Donnell29 to the Digital Age's technological capabilities allowing modern propagandists, at least in theory, to exert a qualitatively greater influence.

**<sup>29</sup>** The original definition of propaganda by Jowett & O'Donnell reads "Propaganda is the deliberate and systematic attempt to shape perceptions, manipulate cognitions, and direct behavior to achieve a response that furthers the desired intent of the propagandist." See Jowett & O'Donnell, 2015, p. 7.

Participatory propaganda is the deliberate and systematic attempt to shape perceptions, manipulate cognitions, and direct behaviour of a target audience while seeking to co-opt its members to actively engage in the spread of persuasive communications, to achieve a response that furthers the desired intent of the propagandist.

The nuanced difference from the traditional interpretation of propaganda lies in the fact that in the Digital Age modern technologies allow propagandists to not just push a message, but get audience "buy-in" through content that triggers engagement with it.30 The ease and convenience of social media offer users a psychologically rewarding and socially acceptable opportunity to "take action" online, empowering them to influence their reality with a few mouse clicks (at least to the extent the size of their network allows it). This superficial and illusory influence, which nonetheless may appear quite real to users, transforms them from passive consumers of propaganda into active campaigners. Through this engagement, propagandists can amplify their message, and by obfuscating its provenance, increase its receptivity among wider populations.

From a methodological perspective, participatory propaganda transcends a traditional, vertical and unidirectional form of persuasive communication where a clearer distinction, but also connection, between a content originator and target audience existed. The diffusion of interconnected networks abetted by users' interconnectivity to and across various online platforms creates a borderless network for distribution and amplification of persuasive content where each "object" of influence (a target individual or group) can, and does, also become the new "subject" of content production and distribution in a "snowball" fashion. In this dynamic environment, an original message would continue to trigger, reinforce, or exacerbate pre-existing sentiments associated with it, prompting its consumers to actively engage in its propagation, both on and offline, for as long as it continues to reflect their entrenched values and perceptions. Whether modified or not by the consumer, the core message often remains intact, acquiring a "new life" in each new wave of content dissemination. Easy access to online monitoring tools allows the original propagandist to follow and assess the spread of their messaging, adapting strategies in a constant feedback loop and inserting additional content, as and if required.

Participatory propaganda offers the ability to truly dominate the information space through volume of messaging, delivered through a mix of real people and automated accounts, effectively making it difficult to discern where fake ends and authenticity begins.

**<sup>30</sup>** In modern practice of campaign management, the evaluation of a campaign's effectiveness commonly relies on engagement statistics offered by social media platforms. These statistics are presented as a manifestation of the message's effectiveness, which of course could be misleading since it often demonstrates the potential spread of messaging but not its actual effect on an audience, if it has any. Sharing of content and commenting, on the other hand, could be perceived as a truer engagement, unless it is accomplished through unauthentic behaviour (e.g. bots).

The analysis of modern political campaigns with their increasing reliance on social networks demonstrates the case in point. While many campaigns still push messages as traditional propaganda, as defined by Jowett and O'Donnell, namely the "deliberate, systematic attempt to shape perceptions" (e.g. popular opinions of Trump supporters) such that it "directs behaviour to achieve a response" (e.g. support for Trump in the form of online participation and voting) furthering "the desired intent of the propagandist" (e.g. the Trump campaign), they increasingly acquire the characteristics of a *participatory propaganda model*. This research, based on analyses of recent academic studies and the US presidential campaign by Donald J. Trump, identified six digital tactics for engaging a target audience online to draw them into active dissemination of persuasive messaging. The order in which these tactics are presented below corresponds to the order of steps a propagandist would take to develop and disseminate the original persuasive messaging.

#### **8.1.1 Hyper-Targeted Audience Analysis**

Using behavioural advertising methods, propagandists acquire information about their target audience's activities online, which enables them to position highly targeted ads in front of users (Matthew, 2017). Trackers that facilitate the collection of this information were found on 114 websites supporting Trump during the 2016 election (Albright, 2016a). This information can be used to segment target audiences by psychographic categories (Psychometric Centre, 2017), providing a propagandist with an extremely accurate assessment of user preferences and what might provoke them into action (Cohen, 2017). Social networks facilitate the application of this knowledge to targeted ad placement (Solon, 2017), as was demonstrated by the Trump campaign (Nix, 2016). After targeted audience analysis is conducted and profiles created, the creation of provocative content ensues.

#### **8.1.2 Provocative Content**

Provocative content aims to appeal to pre-existing user beliefs and perceptions, triggering an emotional reaction in return, which is then expected to manifest itself through engagement both on and offline. At least three types of content aimed at provoking a response among target audiences were used by Trump supporters: fake news, memes, and data leaks.

Fake News. Facebook defined fake news as "articles that purport to be factual, but which contain intentional misstatements of fact with the intention to arouse passions, attract viewership, or deceive." (Weedon et al., 2017). Since lies spread faster online than the truth (Silverman, 2015), fake news has become a global problem (Connolly et al., 2016). Conspiracy theories, often a feature of fake news, reduce complex issues to "binary opposition, simplifying – and misrepresenting – the political space," (Moore, 2015, p. 9) and a person's degree of partisanship is linked to their likelihood of believing conspiracy theories or fake news (Frankovic, 2016). Governments and non-state actors alike are spreading disinformation online (Weedon et al., 2017), including Trump, who spread fake news during the campaign (Maheshwari, 2017). When such content was shared by known and trusted opinion-leaders on Facebook, in essence legitimizing it, they tended to influence audience perspectives (Turcotte et al., 2015). Indeed, stories favouring Trump were shared nearly four times more than those supporting Clinton (Allcott & Gentzkow, 2017).

Memes are often humorous phrases, images or videos that are copied or adapted with slight variations and then shared online (Blackmore, 2000). During the 2016 election, Facebook groups sprang up dedicated to sharing "dank memes" (Hsu, 2016) and a controversial Silicon Valley tech entrepreneur funded a "meme factory" to support Trump (Hern, 2016). So-called "meme battalions" created visual content that "relentlessly drew attention to the tawdriest and most sensational accusations against Clinton, forcing mainstream media outlets to address topics – like conspiracy theories about Clinton's health – that they would otherwise ignore" (Schreckender, 2017).

Memes reduce the public policy debate to shallow sound bites and ridicule stripped of contextualized understanding of available political choices (McClure, 2016). This contributes to 'media endarkenment' reducing complex political issues to simplified entertainment and misinformation (Lazitski, 2014).

Leaks have long played a role in American political propaganda (Castronovo, 2014). According to the Oxford Dictionary, a leak is the "intentional disclosure of secret information" (2017). During the 2016 election, Clinton was dogged by several leaks, which could have been one factor affecting her standing in public opinion polls (Enten, 2016). These included the hacking of her Chief of Staff John Podesta's emails (Frank, 2016), the leaking of comments she made about Bernie Sanders supporters (Democracy Now, 2016), and the continued FBI investigations around the private email server Clinton used while serving as Secretary of State (exposed through a hack), which hampered her campaign (Williams, 2016).

Eliciting an emotional response among a target audience, provoking them into active participation, is an effective method of target audience engagement. In particular, this is true if the content is fed through existing channels where an audience already receives associated information, such as an online echo chamber.

#### **8.1.3 Echo Chambers**

Drawing from the insights gained in hyper-targeted content analysis, a propagandist can identify online echo chambers with specific audiences who can then be targeted with provocative content to which they are most likely to react. An online echo chamber is a digital space where content reflecting a specific point of view reverberates, exposing those within it to only that one prevailing perspective. Digital technologies enable the quick creation of echo chambers or filter bubbles (Breitenbach, 2017), in part through algorithms that sort information (Bakshy et al., 2015), but more so by the choices individuals make about content consumption (Bessi et al., 2016; Grömping, 2014). Once inside an echo chamber, a user is fed content fitting pre-existing views and preferences, such as political party affiliation (Wall Street Journal, 2016).

Echo chambers identified during the 2016 election were strengthened by a growing animosity between political camps (Thompson, 2016), as well as a lack of media trusted by both Republicans and Democrats (Pew, 2016a), and thus information exchange was hindered across party lines (Mitchell et al., 2016). Moreover "political echo chambers not only isolate one from opposing views, but also help to create incubation chambers for blatantly false (but highly salient and politicized) fake news stories" (Pennycook et al., 2017).

Echo chambers supporting Trump shared fakes news during the election (Dreyfus, 2017; BBC, 2016), with some hyper-partisan, right-wing Facebook communities feeding followers 38% fake content (Silverman et al., 2016).

#### **8.1.4 Manipulating Feed and Search Algorithms**

This step relates to positioning and boosting provocative content in front of users by manipulating important online algorithms. Internet giants, such as Facebook (Facebook, 2017a, 2017b) and Google (Google, 2017), use algorithms to provide users with content they think is wanted. Search returns have been found to sway voter decisions (Epstein & Robertson, 2015) and algorithms enable echo chamber development (Barret, 2016). Learning and understanding how these platforms operate both through openly available information and experimentation may allow a propagandist to ensure their provocative content appears higher and more often in returns, or in front of a user.

Algorithms had a role in the 2016 elections. Fake news supporting Trump trended on Facebook through algorithms (Lee, 2016), whereas Google search autocompletes and returns favoured Trump, spreading false information with a far-right bias (Solon & Levin, 2016).

Google Search algorithms can be gamed in at least two ways:

Hyperlinking and Seeding of Content: Posting content, such as fake news, on multiple websites and linking back and forth between sources helps boost content in Google search returns (Moz, 2107), and if nothing else, can bury opposing information from appearing in the first pages of returns. Indeed, in one study using hyperlink network analysis pro-Trump websites were found to be choking out mainstream media (Albright, 2016b).

Botnets and Automated Posting: Lobby groups (Monbiot, 2011), governments (Dhami, 2011), and businesses (Kabin, 2013) are among the many who attempt to create the illusion of grassroots support by using fake social media accounts to distort the information space for strategic purposes. Posting fake comments and reviews aims to harness the cognitive bias of "social proof" (Ambled & Bui, 2011), whereas botnets (and heavily automated posting) can manipulate algorithms. For example, Twitter bots gamed Google's algorithm for displaying "real time news" into promoting disinformation during a 2010 senate election in Massachusetts (Mustafaraj & Metaxas, 2010).

During the 2016 election, pro-Trump automated Twitter accounts dominated discussion about the US election 5 to 1 over pro-Clinton messaging, and "strategically colonized pro-Clinton hashtags," according to Oxford Internet Institute research (Kollanyi & Howard, 2016). Bots also accounted for nearly one-fifth of online discussion about the election (Bessi & Ferrara, 2016), negatively affecting political discourse by drowning opposing views and elevating Trump-related returns in Google search ranking.31

#### **8.1.5 Encouraging Followers to Action**

Once inside echo chambers, followers can be encouraged through posts and email distribution lists (Albright, 2016a; Plouffe, 2010) to participate in the spread of propaganda, including: sharing messages; co-opting or borrowing influencer accounts to share content (Katalenas, 2016); or encouraging trolling (Cheng et al., 2017; Buckels et al., 2014) to stifle debate.

To many, Trump is a troll (Silver, 2015; Offman, 2016; Lapowsky & Marshall, 2017), but he was also supported by a legion of online trolls during the election (Marantz, 2016), spreading disinformation (Kang, 2016; Gallucci, 2016) and attacking Clinton supporters online (Chmielewski, 2016). Some online communities, such as the *United States Freedom Army* (who believes the left is engaging the right in a civil war) offered its members a monthly directive on actions to take on Twitter, and elsewhere in the spread of their content and support for Trump (Lotan, 2016).

**<sup>31</sup>** This online domination through automation led researchers to coin a new term for explaining computer-assisted propaganda techniques – the computational propaganda. See, Woolley, S.C. & Howard, P.N. eds., 2018. *Computational propaganda: political parties, politicians, and political manipulation on social media*. Oxford: Oxford University Press.

#### **8.1.6 Using Traditional Media**

Media play a critical role in furthering political agendas (Wodak, 2013; Engel & Wodak, 2009; Engel & Wodak, 2012); after all, "the media are a key element in the construction of public understanding" (Philo, 2008, p. 539). Rates of a politician's media coverage correlate to popular support levels (Vliegenthart et al., 2012) and Trump was consistently mentioned more on television, online, and social media (Wanless, 2016). By the start of the primary election campaign in early 2016, Trump had been enjoying "more nightly news coverage than the entire Democratic field combined" (Borchers, 2015). Media coverage can be earned in at least three ways:


#### **8.1.7 Assess, Modify and Repeat**

Strategic communicators can rely on multiple technological means to monitor, assess and evaluate the quality and extent of reach acquired in relation to each original message. The data analysis is used to modify messaging to ensure greater uptake or fine-tune the approach to specific target audiences.

As these studies demonstrate, many of these techniques when combined could be used to encourage followers and co-opt audiences into active participation, becoming propagandists for a cause and thus deliberately working to persuade their own personal networks too. To evaluate the extent of their application during the US presidential elections, an original study of Facebook pro-Trump pages was conducted to assess how audiences might have been engaged in the creation and distribution of persuasive political messaging.

#### **8.2 Modelling Participatory Propaganda**

#### **8.2.1 Methodology and Data**

The study included social network and content analyses undertaken on 17 Facebook pages, using data related to a month-long period leading up to the 2016 election (7 October to 7 November). These pages included three that supported Trump during the election, as well as seven conservative-leaning and seven liberal-leaning media outlets. The digital tactics outlined in the previous section were used as a frame for investigation.

Social Network Analysis (SNA) has shown to be an effective method to study online group dynamics, information diffusion processes, and political polarisation in social media (Gruzd & Roy, 2014; Gruzd & Tsyganova, 2015; Scott, 1988, 2011). Facebook pages and groups have been analysed to identify echo chambers (Grömping, 2014; Bakshy et al., 2015; Del Vicario et al., 2016). And content analysis has been used to assess right-wing populist rhetoric in media (Bos et al., 2010, 2011; Sheets et al., 2016).

Facebook was selected for this research as 79% of American adults who use the internet also use this social network (Pew, 2016b), making it the most popular and thus representative social media for studying politics in the US.

The data collection process was executed using Facebook Graph API through Netvizz (Rieder, 2013), and only publicly available data was used. Network visualisations were created using the open-source SNA software, Gephi (Bastion et al., 2009).

The pages analysed are as follows:


**Table 8.1:** Facebook pages analysed.

The three pro-Trump pages were chosen as a sample of those supporting his candidacy, with one showing its open support through the name (*Citizens for Trump*), another having been found spreading fake news (Silverman et al., 2016) supporting Trump (*Eagle Rising*), and a third standing out as a node in initial, exploratory network analysis (*Wake Up & Reclaim America*).

Drawing from a Pew Research Centre survey on Political Polarization and Media Habits (Mitchell et al., 2016), seven media outlets trusted consistently by respondents who self-identified as liberal or conservative were selected. One substitute was made on the conservative-leaning side, which was *Infowars*, given the role it played in the election (Finnegan, 2016). Media pages were used to assess how pro-Trump pages were engaging with news outlets.

#### **8.2.2 Facebook Page Like Networks**

An initial "seed" page liking other pages revealed a directed network of pages linked to it on Facebook, which was visualised using Gephi. The data in this pull also included information regarding page categories, follower numbers, and rates of engagement.

#### **8.2.3 Facebook Page Posts**

All of the posts made by these pages during the month leading up to the election were also collected, including information regarding the type of post, engagement rates, and embedded links.

This data was analysed to answer the following questions:


#### **8.3 Findings**

#### **8.3.1 Pro-Trump Pages Shared Different Kinds of Provocative Content, Such as Fake News, Memes, and Data Leaks**

Fake News. The links shared to the three Trump-supporting pages were mostly nonmainstream media. On average, link posts comprised 53.22% of updates made by the pro-Trump pages. *Eagle Rising* shared more links than the other two (83.25% of posts), with nearly half of those links (45.4%) pointing to the page's own website eaglerising. com, which contains coverage speculating on connections between Clinton, terrorists and Nazis,32 for example, and the Clinton campaign's alleged use of psychological warfare (which in turn points back to another site shared by these pages called ipatriot.com).33

After Breitbart, the most shared domain to *Citizens for Trump* was gatewaypundit. com, a blog that has posted many questionable articles on Hillary Clinton, including

**<sup>32</sup>** See: https://eaglerising.com/23441/the-real-link-between-hillary-clinton-terrorists-and-nazis/

**<sup>33</sup>** See: http://eaglerising.com/36390/how-the-clinton-campaign-is-using-psychological-warfare/

that she secretly called for Trump's assassination,34 had suffered a brain seizure,35 and that she had a gum and immune disorder.36 During the period between 7 October to 7 November 2016, *Citizens for Trump* shared 13 *Gateway Pundit* articles, accounting for 4.32% of all link posts, including one speculating on Clinton's health that enjoyed 319 shares on Facebook.37 *Wake Up & Reclaim America* also shared 14 *Gateway Pundit* articles, including a post suggesting Clinton was involved in having Supreme Court Justice Scalia assassinated.38

Memes. Drawing from a similar study of *Breitbart* posts (Renner, 2017), memes were counted by the total number of photo posts made by the pro-Trump pages. Memes account for a considerable number of posts on community Facebook pages such as *Wake Up & Reclaim America*. In analysis of 1,330 posts made by *Wake Up & Reclaim America* in the month leading up to the 2016 election, nearly half were image posts. Nearly two-thirds of those photo posts were shared by the page administrator from other Facebook user posts, pages or groups, such as *Liberal Wackadoodles*, indicating spread through a wider community. Memes were also shared by *Eagle Rising* (14.79%) and *Citizens for Trump* (26.62%).

#### **Figure 8.1:** Post type by pro-Trump Facebook pages.


**<sup>34</sup>** See: http://www.thegatewaypundit.com/2016/09/hillary-clinton-dog-whistle-call-assassinatingtrump-press-conference/

**<sup>35</sup>** See: http://www.thegatewaypundit.com/2016/08/expert-analysis-finds-hillary-clintons-recentseizures-sign-brain-damage/

**<sup>36</sup>** See: http://www.thegatewaypundit.com/2016/09/dental-expert-hillary-clinton-suffering-seriousgum-infection-immune-disorder/

Data Leaks. Hacks and leaks were certainly discussed online. All of the pro-Trump pages made mention of "Wikileaks,"39 a non-profit that aims to "open governments," which in that period had shared more of the leaked Podesta emails to its website. Of the three pro-Trump pages, 65 posts mentioned "Wikileaks" during the month leading up to the 8 November election, accounting on average for 2.75% of all posts during that period. Both the conservative- and liberal-leaning media outlets analysed made mention of "Wikileaks" in this timeframe too: the seven right-leaning pages mentioned "Wikileaks" 131 times, accounting for 2.72% of all posts made on average, whereas the left-leaning pages referenced it 47 times, or in just 0.46% of all posts. The pages for *InfoWars*, *Sean Hannity*, and *Wake Up & Reclaim America* referenced "Wikileaks" on average more than the others, accounting for 44% of all mentions found.

#### **8.3.2 Pro-Trump Pages Constitute an Echo Chamber – A Like-Minded Community, which Shared Similar Content**

A manual categorization of pages based on names and content reveals that nearly all (94.1%) of the *Citizens for Trump* network are right-leaning, pro-Trump pages, while 82.7% of those within the *Eagle Rising* network are. In the *Wake Up & Reclaim America* network of over 5,000 pages, a sample of 1,000 pages, representing 18.8% of the total, revealed that 67.8% of these were right-leaning, pro-Trump pages covering topics reflected in Trump's campaign rhetoric, such as pro-Christian, anti-Muslim, pro-military, pro-police, anti-immigration, and pro-life views.

As noted earlier, the three pro-Trump pages shared more alternative media sources than mainstream links in the month leading up to the 2016 election. Of those links shared to the pro-Trump pages and pointing to the conservative- and liberal-leaning pages also analysed, most were from either *Fox* or *Breitbart*. The page *Eagle Rising*  shared none of the 14 media pages analysed, and the 1,143 links posted between 7 October and 7 November 2016 pointed to just 14 websites, including eaglerising.com.

#### **8.3.3 Pro-Trump Pages Posted Across Multiple Websites Affected Google Search Returns**

A Google search of article titles posted on pro-Trump pages sheds some light on how such networks function. For example, *Eagle Rising* shared an article from the blog the blacksphere.net entitled "Hillary Clinton: Calls Blacks Professional Never Do Wells." This post garnered 157 shares on Facebook.

**<sup>39</sup>** See: https://wikileaks.org

A Google search using the article's exact title returned the original post, as well as several nearly exact reprints on other sites, with some linking back to *The Blacksphere* article. A search for *The Blacksphere* returned 734 results, including posts from rightwingnews.com, teapartytribune.com, and thegatewaypundit.com. Some of these links were posted by other users in comment sections and online forums, and Sharescount40 suggested the URL was shared 12,500 times across social networks. The article was also picked up by online trend aggregators like Trendolizer,41 indicating the efforts to spread this content had some impact. The apparent domination of these sites' content among first pages of Google returns may create an impression that Hilary Clinton has a clear bias against African Americans.

#### **8.3.4 Followers of Pro-Trump Pages Were Encouraged to Action**

All three pro-Trump pages encouraged their audiences to participate in spreading content and voting for Trump. As rates of follower shares demonstrated, *Citizens for Trump* and *Eagle Rising* were more successful than *Wake Up & Reclaim America*, likely since they asked followers to share and spread messages more often.

**Figure 8.2:** Average shares on posts by page.

**<sup>40</sup>** See: https://sharescount.com/

**<sup>41</sup>** See: http://bit.ly/2q9RYHs

Depending on one's own echo chamber, the size of pro-Trump networks might come as a surprise. To some media pundits, Trump rode to the White House on a wave of fringe support (Coppins, 2015) – but that would be a mistake, as analysis of the networks shows.

Each of the networks were visualised using Gephi, presenting a total of 5,416 nodes with 100,208 edges between them. To put that into perspective, similar data pulls were made on two media page groups. The three pro-Trump pages had 16.3 times more nodes and 55.86 times more edges than the liberal-leaning media group.

Each network contained a considerable percentage of pages that have selfcategorized on Facebook as "Community," but also "Public Figure," "Politician" and some form of "News/Media" (See Figure 8.3).

**Figure 8.3:** Facebook page categories selected by pages within pro-Trump page like network.

The pro-Trump network was then analysed using Gephi (Figure 8.4). Additional statistical analysis was conducted, using Modularity, which identifies the various communities within a network, marked in the data visualization below by colours. The pro-Trump network wasn't just bigger in comparison; it was also more closely integrated between pages with an Average Weighted Degree of 18.502 compared to that of the conservative-leaning media group at 9.01 or the liberal-leaning at 5.404 (the higher the number, the greater the average number of edges that touch a node in the network).

**Figure 8.4:** Three pro-Trump Facebook page like networks.

Pages liking each other demonstrate a possible channel for the spread of information. To investigate further, Netvizz was used to pull all posts made by each page from 7 October to 7 November 2016. These posts where analysed using Excel to count the mentions of specific terms (such as Clinton, Trump, and Wikileaks), how many posts were shared from other accounts, and what web domains were shared to the page, for example. The same investigative process was then applied to analysing the two media page groups.

Around one-third of the posts made by *Wake Up & Reclaim America* (34.1%) and *Citizens for Trump* (28.7%) were shares from other Facebook accounts or pages, indicating community-like behaviour on these two pages.

**Figure 8.5:** Number of posts shared from other Facebook pages or accounts.

Some pages, such as *Occupy Libtards*,42 enjoyed repeated shares to *Wake Up & Reclaim America*, including *The Deplorables*.43 This Facebook group has 472,297 Members (as of 18 April 2017) and takes its name from a comment made by Hillary Clinton during the election about Trump supporters.

These pro-Trump pages are not operating in isolation. Of note, as bigger nodes in the pro-Trump Page Like network visualisation are *Fox News*, *Sean Hannity*, *The Blaze*, and *Glenn Beck* (see the darker orange community in the upper left of the

**<sup>42</sup>** See: https://www.facebook.com/Occupy-Libtards-5-670970859684203/

**<sup>43</sup>** See: https://www.facebook.com/groups/309472556081534/

network) – not to mention the *NRA Institute for Legislative Action* and *The Heritage Foundation* (Figure 8.6).

**Figure 8.6:** Zoomed in screenshot of pro-Trump Facebook page like network using Gephi.

Beyond the official political campaign Facebook pages, hundreds if not thousands of other pages pumped content supporting Trump to sympathetic users of that social network. Indeed, within the *Wake Up & Reclaim America* network, 207 page names contain the word "Trump" – with many more that are pro-Trump that do not, making them much more difficult to track. Together, these Facebook pages support each other with reciprocal Page Likes and sharing of posts, while also mobilizing users to not just spread the message but also support Trump. In so doing, these online communities are also tapping into bigger organisations, such as media outlets, lobby groups, and think tanks – hinting at a much more systemic participatory propaganda effort.

#### **8.3.5 Pro-Trump Pages Heavily Engage with Media Outlets**

The pro-Trump Pages certainly followed right-leaning media outlets, as can be identified in the Page Like networks visualised above and featuring *Fox News* and *Sean Hannity*, among others.

The liberal-leaning media group, visualised in Figure 8.7, comprised seven almost entirely independent communities. The visualisation below uses Gephi's stronger gravity function to keep the communities closer together for ease of viewing; however, they are not linked so closely in reality. What's more, the Facebook pages tend to be grouped into 'ego networks,' meaning any given media outlet tends to only like pages related to that network, such as its own TV shows or journalists.

**Figure 8.7:** Left-leaning media Facebook page like (or "ego") network.

The conservative-leaning media group is quite different (Figure 8.8). The massive *Infowars* community dominates the visualisation, represented here below by the large yellow section, running into the *Alex Jones* network in blue, which comes with it.44 While nodes connect the *Infowars* monolith to *Fox*, the key connector page is *Judge Andrew Napolitano*. This is interesting in itself, as in past analysis of media Facebook Page Like networks, *Fox* stood out from outlets such as the BBC for its connecting to personalities, both their own journalists as well as US politicians, suggesting that some media outlets aren't just covering the news, but engaging directly with the subjects making the news (Wanless, 2015). This form of engagement could be considered alarming, if the notion of impartial news is accepted as crucial to a functioning democracy.

**Figure 8.8:** Right-leaning media page like networks.

**<sup>44</sup>** Since then, Facebook has banned Alex Jones and Infowars blog from its pages.

#### **130** Participatory Propaganda

When these two media groups are combined with the pro-Trump network (Figure 8.9), the liberal-leaning outlets become islands unto themselves almost entirely disconnected (the blue communities at the bottom left), while the conservativeleaning media are absorbed into the overall community, and as noted above, in some cases becoming influential nodes.

**Figure 8.9:** All Facebook page like networks combined.

In short, the conservative-leaning media network is more of an ecosystem that stretches beyond news outlet borders, blending into each other and pages beyond just media and journalists, into communities.

#### **8.3.6 Pro-Trump Pages in a Participatory Propaganda Model**

While the analysis presented above is based on a very limited number of pages, the degree of engagement and inter-connectivity, both inside the network and with supportive media, demonstrate the existence of a systematic and coordinated attempt to influence the US voters to support the Trump campaign. All but two of the critical tactics identified in the participatory propaganda model were used by Trump supporters to achieve this goal, including sharing provocative content (fake news, memes, and data leaks), feeding such content into an echo chamber, reposting the same content, encouraging followers to do the same, and connecting with media and larger organisations supportive of Trump. The only two critical elements missing were hyper-targeted audience analysis (Step 1) and monitoring and evaluation (Step 7), as these activities are typically conducted through in-house research during a campaign and are not traceable through open sources. However, Cambridge Analytica has openly claimed to have used such tactics for the Trump campaign (Nix, 2016).

#### **8.4 Conclusion: Participatory Propaganda in Liberal Democracies**

As this chapter demonstrates, the organized deployment of various emerging technological and manipulative techniques in a digital era creates conditions for the emergence of an interactive form of engagement online where followers (target audience) are drawn into participating in the creation and spread of persuasive messaging. The example of the Trump 2016 presidential election campaign was used to demonstrate how these new tactics were deployed in combination with traditional media coverage to draw a considerable online following. This follower engagement constitutes a qualitatively more enhanced form of propaganda that is much more 'invasive' in nature – not to mention potentially very dangerous for liberal democracies, especially if cases of foreign interference in the electoral processes, such as the alleged Russian influence on the 2016 US election results, continue to mount.

As internet penetration rates in democratic countries surpasses 80%,45 with many others in tow, nearly half of those populations finished high school before the web was even invented.46 As of 2015, 21% of American survey respondents indicated they were online "almost constantly" (Perrin, 2015), and by the end of the first quarter

**<sup>45</sup>** For more data, see 2016 CIRA report for Canada, the 2016 Office for National Statistics report for UK, and the 2017 Pew Research Center report for the USA.

**<sup>46</sup>** Similar models of participatory propaganda have been identified through subsequent research on the 2017 U.K. general election (https://lageneralista.com/anti-establishment-blues-2017-u-k-election/) and Canadian political Facebook pages (https://lageneralista.com/polarising-politics-in-canada-afacebook-study/).

in 2016, the average American was consuming 10:39 hours (Nielsen, 2016) of media across devices each day. Unlike radio and television before it, the internet has people constantly connected to information. Americans are at the vanguard of these changes – and as such are among the most vulnerable populations to information warfare, be it in the form of participatory propaganda, social engineering or cyber-attacks.

 With such levels of exposure to a constant barrage of information, the ability of any one individual to discern its veracity and relevance in a broader context of daily life is constantly challenged. Furthermore, the effects of continuous online exposure on individual mental health or general perceptions of the world are still too poorly understood, and as such, are not yet part of mainstream knowledge or incorporated into national education curriculums at the level required to cope. The negative and long-lasting repercussions of such limited understanding are perhaps nowhere as serious as in national politics.

In 2014, the World Economic Forum listed "the spread of misinformation online" as one of the top ten trends facing the world (WEF, 2014). By 2016, Reporters Without Borders declared that we "have reached the age of post-truth, propaganda, and suppression of freedoms – especially in democracies" (2017). As demonstrated, modern propagandists have a considerable arsenal of methods at their disposal to manipulate populations, influence their opinions or engage them in active propagation of the desired content that go well beyond the creation and distribution of 'fake news' alone. Unfortunately, the tools and methods discussed in this article can be used by savvy propagandists everywhere, regardless of their political clout or country of provenance. What perhaps stands out most in this participatory propaganda model is its perpetuation. Through the use of online communities, such participatory propaganda campaigns run as long as the cause driving it matters to its members – or rather, those administrating such groups are able to produce content that engages and provokes followers. Finding ways to identify and measure engagement within these networks to understand the driving rationale, as opposed to blocking them, should be a priority for those studying liberal democracies.

#### **References**


#### **136** Participatory Propaganda


Plouffe, D. (2010). *The Audacity to Win: How Obama Won and How We Can Beat the Party of Limbaugh, Beck, and Palin*. Updated ed. Penguin Books.

Renner, N., (2017, January 30). Memes trump articles on Breitbart's Facebook page. *Colombia Journalism Review*. [Online] Retrieved May, 11, 2017 from http://bit.ly/2pCf3Rs

Reporters Without Borders. (2017). *2017 World Press Freedom Index – tipping point*. [Online] Retrieved May, 11, 2017 from http://bit.ly/2pyXh61

Rieder, B. (2013, May). Studying Facebook via data extraction: the Netvizz application. In *Proceedings of the 5th Annual ACM Web Science Conference* (pp. 346–355). ACM.

Ross, S. T. (2002). Understanding propaganda: The epistemic merit model and its application to art. *Journal of Aesthetic Education*, 36(1), 16–30.


Solon, O. (2017, 27 October). Russia's RT Reveals Twitter's Pitch to Sell Millions of Dollars in 2016 Election Ads. *The Guardian.* [Online] Retrieved November, 1, 2017 from http://bit.ly/2ycyh9a


Trump, D. (2016, 4 July). Dishonest media is trying their absolute best to depict a star in a tweet as the Star of David rather than a Sheriff's Star, or plain star! [Post] *Twitter.* Retrieved November, 1, 2017 from http://bit.ly/2xJrKP5


Wanless, A. (2015). *You've Got a Friend in News.* [Blog] La Generalista. Retrieved November, 1, 2017 from https://lageneralista.com/youve-got-a-friend-in-news/

Wanless, A. (2016). *Facebook and the U.S. Elections*. [Blog] La Generalista. Retrieved November, 1, 2017 from https://lageneralista.com/facebook-stats/

Weedon, J., Nuland, W., & Stamos, A. (2017). *Information Operations and Facebook*. [Online] Facebook. Retrieved May, 11, 2017 from http://bit.ly/2oOOS9s

WEF. (2013). *Outlook on the Global Agenda 2014*. [Online] World Economic Forum. Retrieved May, 11, 2017 from at http://bit.ly/2q8UIqy

Williams, P. (2016, May 25). Guccifer, Hacker Who Says He Breached Clinton Server, Pleads Guilty. *NBC News*. [Online] Retrieved January, 21, 2017 from http://nbcnews.to/1szUR3X

Wodak, R. (2013). *Right-wing populism in Europe: politics and discourse.* A&C Black Woolley, S. C. & Howard, P. N. (2018). (Eds). *Computational propaganda: political parties,* 

*politicians, and political manipulation on social media*. Oxford: Oxford University Press

## **List of Figures**

**Figure 4.1:** Density of clusters according to neighborhood **— 46**

**Figure 5.1:** Akhil Gogoi Meme **— 62**

**Figure 6.1:** Video still from the Israeli Defense Forces' video, "With Her Eyes She Defends Israel's Borders," 2019 **— 68**

**Figure 6.2:** Video still from the Israeli Defense Forces' video, "With Her Eyes She Defends Israel's Borders," 2019 **— 68**

**Figure 6.3:** B'Tselem volunteer Thawra 'Eid zooms in on Israeli settler children who are attacking her Palestinian village of Burin by throwing stones with slingshots, May 20, 2015, © B'Tselem **— 77**

**Figure 6.4:** B'Tselem volunteer Wydian Zaben films Israeli settlers torching the fields around her Palestinian village of Burin, Aug 15, 2015, © B'Tselem **— 78**

**Figure 6.5:** "The Po'alot (female workers) in the agricultural settlement Ayanot," Jan 2, 1940. Source: Schwartz Tel Aviv. From the The Jewish National Fund (JNF) Photographic Archive **— 79**

**Figure 6.6:** Deir Yassin (also transliterated as Dayr Yasin), Associated Press, April 1948 **— 81**

**Figure 7.1:** Tweet 1 **— 92**

**Figure 7.2:** Tweet 2 **— 92**

**Figure 7.3:** Tweet 3 **— 93**

**Figure 7.4:** Tweet 4 **— 93**

**Figure 7.5:** Tweet 5 **— 94**

**Figure 7.6:** @mentions containing attempts to discredit **— 98**

**Figure 7.7:** @mentions containing attempts to intimidate **— 99**

**Figure 7.8:** @mentions containing attempts to shame **— 101**

**Figure 8.1:** Post type by pro-Trump Facebook pages **— 121**

**Figure 8.2:** Average shares on posts by page **— 123**

**Figure 8.3:** Facebook page categories selected by pages within pro-Trump page like network **— 124**

**Figure 8.4:** Three pro-Trump Facebook page like networks **— 125**

**Figure 8.5:** Number of posts shared from other Facebook pages or accounts **— 126**

**Figure 8.6:** Zoomed in screenshot of pro-Trump Facebook page like network using Gephi **— 127**

**Figure 8.7:** Left-leaning media Facebook page like (or "ego") network **— 128**

**Figure 8.8:** Right-leaning media page like networks **— 129**

**Figure 8.9:** All Facebook page like networks combined **— 130**

## **List of Tables**

**Table 4.1:** Overview of the eight largest clusters obtained from community detection on relations among Instagram users in Kristiansand **— 41**

**Table 7.1:** @mentions containing attempts to discredit **— 98**

**Table 7.2:** @mentions containing attempts to intimidate **— 100**

**Table 7.3:** @mentions containing attempts to shame **— 101**

**Table 7.4:** Percent of @mentions invoking race (of legislator) **— 102**

**Table 7.5:** Percent of @mentions invoking gender (of legislator) **— 103**

**Table 7.6:** Percent of @mentions containing body commentary **— 103**

**Table 7.7:** Summary of hypothesis test results **— 104**

**Table 8.1:** Facebook pages analysed **— 119**