Visual Citizenship

Front matter

VISUAL CITIZENSHIP

COMMUNICATING POLITICAL OPINIONS AND EMOTIONS ON SOCIAL MEDIA

Catherine Bouko

Contents

List of figures xii List of tables xiv Acknowledgements xvii Summaries of the chapters xviii

Introduction: Visual citizenship: communicating political opinions and emotions on social media 1 0.1 Concepts 4 0.2 Methods 6 0.3. Empirical insights 9 PART 1 Concepts 15 1 Everyday political expression as a citizenship practice 17 1.1 Broad definitions of citizenship, civic engagement and political participation 18 1.1.1 Dutiful versus self-actualising citizenship 18 1.1.2 Civic engagement 18 1.1.3 Political participation 19 1.2 Idealising political participation over (online) civic engagement 21 1.2.1 Civic engagement as latent participation 21 1.2.2. Three criticisms of the gateway model 21 1.3 Listening as an underestimated activity 23 1.4 Idealising rational deliberation and consensus 24

1.4.1. The anachronistic bourgeois public sphere 24 vi Contents

PART 2 Contents

5.1.2 Emotion and reason

5.1.3. The cultural practice of emotion

5.1.4 The social glue in a political context

Affective economies on social media

Visual content and emotions

5.3.1 Cognitive approaches to images and emotions

5.3.2 Iconic images and the circulation of emotions

5.3.3 Emotions and citizens’ visual imagery on social media

6.1

6.2

Challenges in content analysis

6.1.1 Quantitative and qualitative content analyses

6.1.2 Manifest and latent content

Challenges in discourse analysis

6.2.1 The analysis of discourse structures and patterns

6.2.2 Qualitative discourse analysis

6.2.3 Interviews and reverse image searches as complementary methods

7.1 7.2 7.3

The three functions of language O’Toole’s functional framework for paintings Kress and van Leeuwen’s three functions of visual communication 7.3.1 The representational function 7.3.1.1 Narrative structures 7.3.1.2 Conceptual structures 7.3.2 The interpersonal function 7.3.2.1 Interactivity 7.3.2.2 Modality: colour, perspective, background, representation, depth, illumination and brightness 7.3.3. The compositional function 7.3.3.1 Information value: left-right, top-bottom and centre-margin 7.3.3.2 Salience 7.3.3.3 Framing vii

73 74 75 76 79 79 80

92 92 93 95 96 97

103 105

107 107 107 114 120 120

125 128

128 136 137 viii Contents

7.4

Two challenges in analysing visual resources

7.4.1 Social semiotics and the “grammar” of visual content

7.4.2 Units of analysis in visual content

8 Methodological standards for quantitative content analysis of social media posts

8.1 8.2 8.3 8.4 8.5

A rich variety of existing research designs

Data collection and selection

Data contextualisation

The size of a valid sample

Intercoder and intracoder reliability and the size of the testing sample

9.1

9.2 9.3

Representational variables: connection with the

research question, standard and specific topics, frames

9.1.1. Preliminary variable: the connection with the research question

9.1.2 Standard variables: standard topics

9.1.3. Contextualised variables: specific genres and topics

9.1.4 Latent variables: frames based on manifest or latent content

Interpersonal variables

Compositional variables

10.1 Emotion, opinion and the issue of observability 10.2 Thematised emotions and opinions in text and image

10.2.1 Thematised attitude: affect, judgement and appreciation

10.2.2 Literal and figurative thematised emotions: major sets versus basic emotions

10.2.3 Literal and figurative opinions: inscribed or evoked judgement

10.3 Signal-like emotions and opinions in text 10.4 Supported emotions and opinions in text and image

10.4.1 Content patterns and related methodological questions

10.4.2 Text-image rhetoric: attitude in visual arguments

10.4.3 Specific emotions in supported attitude

138 140

147 150 151 153

163 165 167

175 181 187

194 194 200 201 204

209 213 217

218 222 228

Contents ix

Part 3

11 The Brexit vote and its aftermath: quantitative results 237 11.1 The connection with Brexit and tagging practices 238 11.2 A threefold concentration of topics 241 11.3 An unexpected variety of visual genres 246 11.4 Self-expression as the main social relation 249 11.5 Stance and the issue of polarisation 252

12.1 Ten patterns of attitude in multimodal social media posts 260

12.2 Five types of appraisers in the verbal elements 264 12.3 Eight types of verbal attitude in multimodal social media posts 267 12.3.1 Attitude in common-sense sayings, statements or proverbs 269 12.3.2 Indirect speech and endorsement of external voices 269 12.3.3 Anchorage of the visual content 270 12.3.3.1 Attitude through anchorage, no attitude in the visual content 272 12.3.3.2 No attitude through anchorage, attitude in the visual content 273 12.3.3.3 Attitude through both anchorage and in the visual content 274 12.3.3.4 No attitude neither through anchorage nor in the visual content 275 12.3.4 Diegetic relay of attitude 276 12.3.5 Verbal alignment and disalignment with attitudinal visual content 276

12.3.5.1 Verbal alignment or disalignment

with thematised emotions in the

visual content 276 12.3.5.2 Verbal alignment or disalignment

with thematised judgements in the

visual content 276 12.3.5.3 Verbal alignment or disalignment

with supported attitude in the visual content 277 12.3.6 Minimal attitudinal lexis 277 12.3.6.1 Minimal affect lexis 277

12.3.6.2 Minimal judgement lexis 277 x Contents

12.3.6.3 Minimal lexis of supported attitude 12.3.7 Attitude in general statements 12.3.7.1 Emotion-related general statements 12.3.7.2 Judgement-related general statements 12.3.7.3 General statements of supported attitude 12.3.8 Attitude in personal narratives 12.3.8.1 Thematised emotions in personal narratives 12.3.8.2 Thematised judgements in personal narratives 12.3.8.3 Supported attitude in personal narratives 12.4 Multimodal patterns and attitude in visual content 12.4.1 Convergence or divergence in the types of attitude 12.4.2 Literal or figurative nature of the verbal and visual content 12.4.3 Attitude and types of visual participants and processes 12.5 Contracting and expanding alternative voices: two patterns of engagement

13.1 Inferring multimodal metaphors 13.2 Emotion-related metaphors 13.3 Judgement-related metaphors in multimodal content 13.3.1 Quantitative findings: the prevalence of four metaphors 13.3.2 Scenarios and moral foundations in metaphors 13.3.2.1 Judgement-related metaphors without markers of moral foundations 13.3.2.2 Judgement-related metaphors with markers of moral foundations 13.4 Metaphoric creativity in image-based social media posts 13.4.1 Level zero: no metaphoric creativity at the conceptual, image or formulation level 13.4.2 Level one: non-metaphoric creativity at the formulation level 13.4.3 Level two: metaphoric creativity at the formulation level

Contents xi

13.4.4 Level three: double metaphoric creativity at the formulation level 311

13.4.5 Level four: metaphoric creativity at the conceptual level and image-schema creativity 312 13.5 Creativity in mundane experiences 314

Index 319

Figures

7.13 7.14

Demand gaze in “I want you for the US army” poster “Oh, I disagree with your politics?” meme featuring

Gene Wilder

Detail of the Bayeux tapestry (11th century)

“US survival rate” meme

Bernie Sanders in a meme interweaving politics and popular culture

“If Ted Cruz...” meme

Raising the flag on Iwo Jima by Joe Rosenthal (1945) “Let’s call her Monica” meme featuring the Clintons and their dog

“Oh, I disagree with your politics?” meme featuring Gene Wilder

“To those of you...” meme featuring Leonardo Di Caprio “Tell me again . . .” meme featuring former British Prime Minister Boris Johnson and Queen Elizabeth II

“This is fine” dog meme taken from KC Green’s webcomic series Gunshow

“Big penguin helps small penguin” picture

“The seven deadly chins” image macro

“2017: Russian election meddling” meme

“I am European...” meme

“They’d all be so much more electable” meme

“Canada r attacking” meme

“House of . . .” meme

“Pm tired of .. .” meme

“The founding generation .. .” meme

7.15, 7.16 and 7.17 Variations of the “disaster girl” meme, in

7.18 7.19 7.20 7.21 political and private contexts

Screenshot of the AOIR Facebook page, English version Screenshot of the AOIR Facebook page, Arabic version “Why are Trump...” meme

“His brain? ...” fake ad

Figures

“Remember when . . .” meme

“Runs out of . . .” meme

“Quick! Bring a shovel . . .” tweet

“Brexit is like . . .” meme

Boris Johnson as a persecutor

Soldiers as heroes

Abraham Lincoln as a COVID-19 victim

Boris Johnson as a fool in “Boris Johnson . . .” meme Example of attitudinal prosody in a tweet that combines thematised, signal-like and supported attitude Brexit-related Flickr post featuring South Park character Stan and Boris Johnson

Flickr post with visual creativity for signal-like emotion Marian Kamensky’s cartoon shared in a citizen’s Flickr post after the Brexit vote in 2016

Patrick Chappatte’s cartoon shared in a citizen’s Flickr post after the Brexit vote in 2016

Boris Johnson on a zip wire on the day of Britain’s first 2012 Olympic gold

Picture of a bird with open beak coupled with verbal elements that fulfil the anchorage and relay functions in a Flickr post

Multimodal metaphor of the UK and the EU going in opposite directions after the Brexit vote

“The domino effect of Brexit” multimodal metaphor Orientational visual metaphor and ontological multimodal metaphor in a Flickr post

Literal visual representation of the “hole” metaphor in the Brexit context

The 27 parachutes metaphor

The “gentle curve” Brexit metaphor xiii

207 216

288 289

291 310

312 313

Tables

1.1

The changing citizenry: the emerging youth ideal of self-actualising citizenship (AC) versus the traditional civic education ideal of the dutiful citizenship (DC)

The strengths of quantitative and qualitative research Sites and modalities for interpreting visual materials

SFL model for paintings adapted from O’Toole

Summary of Kress and van Leeuwen’s SFL-based framework for reading images, with slight adaptations from the original

Snapshot of the conceptual structures (see Table 7.2) Snapshot of interactivity in conceptual structures

(see Table 7.2)

Image-text configurations in social media posts on Facebook, Flickr, Instagram and Twitter, for newsfeed and when users click on specific posts (January 2022)

Units of analysis in O’Toole’s, Kress and van Leeuwen’s and Booriis and Hoslanova’s frameworks

Overview of corpus-based studies of image-based social media posts

Sample size: least likely units and significance levels Split-half technique to measure the internal consistency of results with a Chi-Square test

Krippendorff’s required numbers of values for testing samples Taxonomy of media topics elaborated by the International Press Telecommunications Council, version as of February 2022

Marchal et al.’s adapted categories of visual genres Mahoney et al.’s adapted code hierarchy of themes Butkowski’s classification of displays of voting bodies in ballot selfies

Wodak’s discursive strategies for referential/nomination and predication of social actors, objects, phenomena, events and processes, and actions

148 154

156 158

166 168 170

Tables

Butkowski’s displays of relationality

Bell’s variables of behaviour and social distance

Types of social relations inferred from informational, relational and entertainment content

Marchal et al.’s adapted typology of visual format,

Thematised, signal-like and supported attitude framework Thematised attitude framework

Comparison of affect types between Martin and White’s “old” model on the left and Bednarek’s “new” model on the right

Facial and bodily expressions that thematise emotions Van Leeuwen’s (de)legimisation strategies in discourse (De)legitimisation strategies and prompt questions for the analysis of inscribed and evoked judgement in text-image content

Framework of signal-like attitude

Eight appraisal criteria and questions to elicit supported attitude in discourse

Corpora of Flickr and Twitter text-image posts with direct, indirect or no connection with Brexit

Distribution of the corpus of text-image Flickr and Twitter posts in the month of analysis

Classification of Brexit-related topics and their salience in Flickr and Twitter text-image posts

Types of visual genres in the Flickr and Twitter text—image posts

Types of social relations inferred from informational, relational and entertainment content in the Flickr and Twitter text-image posts

Stance on Brexit in the Flickr and Twitter text—image posts Ten patterns of attitude coded in Brexit-related Flickr posts that include a picture and the social relation of self-expression

Twenty combinations of patterns of attitude, in the Flickr corpus

Martin and White’s engagement system

Visual and verbal elements of four emotion-related multimodal metaphors

Source and target domains in multimodal Brexit-related metaphors, in descending order

Types of Brexit-related metaphors

Schematic and taxonomic metaphoric interpretation Charteris-Black’s summary and adaptation of Haidt’s six moral foundations

XV

182 183

184 189

199 202

205 206 210

263 282

13.6 Realisations of the “country is a bounded container”

metaphor to express caring about the UK and its unity

on Flickr 301 13.7 Two metaphor scenarios based on the container metaphor,

by a pro-Remain and a pro-Leave supporter, respectively,

on Flickr 305 13.8 Research design to analyse metaphoric creativity in

multimodal content 309

Acknowledgements

Many people have contributed to the writing of this book, for which I am very grateful.

My first thanks goes to David Garcia, Professor of Social and Behavioural Data Science at the University of Konstanz (Germany) who, after meeting me very briefly at the Association of Internet Researchers Congress in 2016, kindly agreed to help me collect Flickr and Twitter data for this project. I am enormously indebted to him for his generosity of time and energy, which I also experienced in our numerous e-mail exchanges about data science and statistics.

I also feel extraordinarily fortunate to have benefited from the insightful comments and advice offered by my PhD student Jorge Disseldorp and Veronika Koller, Reader in discourse studies at Lancaster University (UK). Their comments, both on content and form, considerably improved the quality of all chapters. Working with such committed reviewers was a great pleasure. I am extremely grateful to them. I cannot thank Veronika enough for taking time out of her busy schedule to review my book.

Special thanks also goes to my colleagues Anneleen Spiessens and July De Wilde in the Department of Translation, Interpreting and Communication at Ghent University (Belgium) for their assistance. Heartfelt thanks also goes to my colleague Kristien Temperville, who taught me a lot of tricks about academic English along the way.

Furthermore, I wish to warmly thank the authors of the drawings or photographs for their permission to include them in the book: Walter A. Aue, Patrick Chappatte, Didier Darrigand, KC Green, Ibdilillah Ishak and Cosimo Matteini.

Lastly, Iam very grateful to the Editorial Team at Routledge, in particular Suzanne Richardson and Stuti Goel, for their assistance in the preparation of the manuscript.

Summaries of the chapters

Introduction

Grounded in the visual turn, visual political expression on social media is a form of everyday citizenship that is highly personal, creative and affective. Part 1 of the book provides entry points to the key terms and phrases in italics. The discussions and guidelines that I propose in Part 2 are intended to help researchers develop methodological designs for quantitative and qualitative analyses of image-based content. In Part 3, I flesh out the conceptual considerations in Part 1 and the methodological propositions in Part 2 in an empirical case study, namely the Brexit vote, by which the British people decided, on 23 June 2016, to leave the European Union.

Chapter 1

This chapter addresses why everyday political expression on social media can be considered a citizenship practice, whether it is image-based or not. Combining attention to societal issues with the activity of speaking out, it constitutes a form of civic engagement. However, everyday political expression on social media is to be distinguished from political participation, as it does not seek to directly target political, economic or social actors with the aim of influencing actual outcomes. While the societal value of everyday political expression as such should not be overestimated, it should not be neglected either. Its value lies in the potential role it can play as a gateway to political participation, but also in peer-to-peer influence and in complementarity agency, by which citizens share their views and, in doing so, help each other improve their collective knowledge and understanding. Furthermore, this chapter also discusses that citizenship is not a static concept; it is reified by each society and is contingent on time and place. Therefore, considering everyday political expression on social media as a citizenship practice requires one to move away from essentialised and idealised approaches to citizenship that are disconnected from reality.

Summaries of the chapters xix

Chapter 2

The development of new technologies in the post-war period of the 20th century has both democratised and enhanced visual practices. This in turn has led to a gradual re-evaluation of visual communication, which has gained added momentum at the start of the 21st century with the co-emergence of smartphones and social media. Sharing pictures online has enabled people to enter into predominantly pictorial dialogues. This chapter takes a brief look at how the visual turn occurring in the second half of the previous century constitutes a re-appreciation of visual communication. This new approach relates to a “new visual literacy” which can be contrasted with the preceding structuralist and modernist “old visual literacy” approach, in which the visual is both subservient to, and fully dependent on, linguistic systems. Flickr, Twitter, Facebook and Instagram were largely created to focus on visual content or adapted at a later stage to accommodate such content to attract higher rates of user engagement. Avoiding a techno-deterministic approach to these platforms, this chapter instead touches upon the contexts in which users interact with platform affordances, leading to specific conventional practices on different social media and to platform-specific vernaculars. Political actors and social movements have long since recognised the power of the visual as a communicative tool, as evident at present in their strong visually based social media presence. While citizens’ expressions in the context of social movements and concrete political events have gained much scholarly attention, the practices of citizens’ civic expression elsewhere remain largely understudied.

Chapter 3

Citizenship has become both liquid and monitorial, which, embedded into a consumerist context, has transformed the everyday into a legitimate site of political expression, as encapsulated in the notion of lifestyle politics. Liquid citizens are connected to all sorts of places and perform a large variety of multi-directional and episodic activities; monitorial citizens voice their concerns and claims on an episodic basis, when their personal centres of interest are at stake (see Chapter 1). This chapter discusses how boundaries between public and private spheres have started to gradually blur under the influence of enhanced mobile technologies and established social media practices. In everyday conversation, the topic of politics often arises from the mundane and, as with most other topics, primarily serves a social, rather than primarily political, function, that is to maintain good relations and one’s social status. Political self-expression in the everyday as such emanates from the private sphere and is thus highly reliant on personal experiences. Due to blurred boundaries online, these experiences subsequently come to constitute shared common spheres. As an example of how established (visual) practices online may take on a political dimension and how public and private spheres merge, this chapter briefly discusses how citizens’ selfies as a private xx Summaries of the chapters expression in political contexts (e.g. during elections) may perform both social and communicative functions and create performative and spectatorial intimacies through a sense of proximity. Finally, we will see how a form of eye-witnessing when shared online in political contexts may become a form of saying, and how bearing witness to political events may become a form of civic participation.

Chapter 4

In this chapter, we will see how image-based play and creativity fulfil an important role in contemporary participatory culture. First, it discusses how the idea of media citizenship incorporates the importance of popular culture and entertainment in creating cultural public spheres. In these spheres, reflexive citizens, through their emotional engagement, enter into performative actions, such as discussion, critique and imagination; actions which may impact on identity, social values and political orientation. DIY citizenship further emphasises the importance of creativity and play, as in a digital world citizens become content creators. Visual content online, such as political memes, is often heavily intertextual and plays with both conventional and creative metaphors, in which especially humour plays an important role as a social lubricant. A world full of stimuli, in which complex political ideas create a distance between citizens and political spheres, relies on heuristic devices to facilitate understanding, but which also carry the risk of oversimplification and trigger cognitive biases leading to cultural stereotyping.

Chapter 5

Citizens’ expressions on social media often revolve around affective and emotional engagements. This chapter discusses how the affective turn and affective citizenship shape contemporary citizens’ engagements, which in online environments may lead to particularised emotional practices and alignments. It shows how affect as an unactualised potential can be thought of as a precursor to more concrete, embodied emotional and cognitive processes. Both the rational and emotional are mutually intertwined dimensions within our information processing systems. Affective citizenship thus highlights how any engagement (political or not) entails a degree of affective investment, which through its expression can foster a collective dimension and aid a sense of belonging. Emotion can be seen as a social expression and thus as a social practice. In this respect, emotional repertoires are built collectively. As with any social structure, this entails a degree of normativity and power dynamics. On social media, affective proximity and emotional expressivity become means to overcome physical distance, boosting the importance of these affective dimensions. In practical terms, we will see how emotion is a vital part of visual processing and how seeing as doing images stimulate an emotional response based on internalised norms and values. This chapter also

Summaries of the chapters xxi addresses how iconicity, achieved through familiarity, plays a role in emotional exchanges, highlighting the social dimension of affective communities.

Chapter 6

This chapter highlights the particular qualities and limitations of content analysis and discourse analysis, especially when applied to image-based datasets. For content analysis, this chapter focuses on the importance to establish and distinguish the right type of variables, while simultaneously showing, through practical examples, the challenges in doing so successfully. For discourse analysis, it discusses the importance of recognising structures and patterns both in text and images, as well as the social circumstances of the production, distribution and reception of such artefacts. In the practical context of social media, this chapter finally addresses the particular difficulty of the “context collapse” that occurs online; multiple audiences congregate in virtual places, and social media users and researchers alike are left to imagine the contexts of production, distribution and reception.

Chapter 7

Halliday’s systemic functional linguistics (SFL) was designed to analyse the relationships between texts and aspects of social life. The model acknowledges the functionality of language both for its communicative purpose to communicate ideas as well as its social and compositional aspects. SFL has found a broader applicability beyond linguistic systems and has become predominant in semiotics. This chapter discusses two SFL-based models that can be applied to the analysis of visual content. O’Toole’s functional framework for painting (1990) adapts Halliday’s representational, interpersonal and textual functions to relate to visual art. Kress and van Leeuwen’s (2006) model similarly assumes three functions, but it rather focuses on the meaning-making potential as expressed through subdivisions within functional patterns and processes. While the model establishes some more-or-less conventionalised patterns of meaning, it also, at least in theory, stresses the importance of a contextualised approach. This chapter pays extra attention to some of the challenges that the analysis of visual content poses. There is, for example, reason to doubt if the conventions of Kress and van Leeuwen or O’Toole’s visual markers would be applied similarly by different interpreters. For visual content, much more so than for linguistic systems, it is a valid question how we process images, as a whole or through their component parts.

Chapter 8

Chapters 8 and 9 focus on the research designs of 26 corpus-based studies on visual citizenship. Together, they offer a comprehensive view of the scientific publications in the field. This chapter discusses the studies that rely xxii Summaries of the chapters on quantitative methods, as it aims to provide the most concrete schemes possible with an eye on replicability, systematicity and reliability. When studying social media content, it is first important to acknowledge that the potential population or the number of memes, tweets and posts to study is practically unlimited, continuously changing and unknowable. It thus becomes central to explicitly define the parameters in which the data collection and selection takes place. This chapter touches upon different methods that try to address the optimal corpus size and discusses the issue of coding reliability by looking at some statistical measures that are used to calculate reliability and at recommended sizes for the samples on which reliability is tested.

Chapter 9

This chapter discusses and suggests additions to typologies and taxonomies that were designed for the quantitative categorisation and coding of concepts related to the representational, interpersonal and compositional functions of image-based content. For the representational function, I distinguish preliminary, standard, contextualised and latent variables. In doing so, I pay attention to (1) the direct relation of the data to the research aims and subject in terms of noise, (2) the relative levels of generality and specificity that research contexts require in relation to the categorisation of variables and (3) the distinction between manifest and latent content. In discussing concepts such as visual genres, themes, topics and frames, this chapter stresses the importance of identifying what one is coding for, that is, what the visual content consists of, what content is represented in images, what the content is about, and which perspectives can be inferred from social media posts. In relation to interpersonal variables, this chapter discusses key concepts such as affiliation and relationality, and introduces the variable of play. We will see how incongruity and exaggeration constitute play, and how they can be related to common strategies on social media, such as delegitimisation. Lastly, this chapter looks at how socio-cultural regularities in the use of compositional patterns can lead to the emergence of visual and multimodal genres.

Chapter 10

In this chapter, the focus is on appraisal in text-image content, and more specifically on attitude, which is subdivided into affect, judgement and appreciation in Martin and White’s framework (2005). Affect and judgement in discourse (roughly emotions and opinions, respectively) are often tightly interwoven and their expression relies on a broad range of devices, cues and related discourse patterns and strategies that are not always easy to distinguish. The discussion is structured around three broad modes of the semiotisation of attitude in discourse: (1) thematised, (2) signal-like and

Summaries of the chapters xxiii

(3) supported attitude. First, in discussing thematised attitude, we will see how emotions and opinions become the object of discourse either through direct denotative processes or through more figurative and more — or less conventionalised, connotative expressions. For visual content, rather than trying to identify basic emotions from facial and bodily expressions, I follow a more cautious approach according to which minimal patterns of thematised emotions, namely valence and arousal, can be identified in facial and bodily expressions. However, higher-level divisions between specific emotions are very challenging through visual analysis alone. Second, for signal-like attitude, specific patterns signal the presence of emotions and/or opinions in discourse, especially when they are combined with each other (e.g. exclamation marks can signal emotions and/or opinions). Unlike thematised attitude, which concerns discourse about emotions and opinions, signal-like attitude refers to discourse as emotion or opinion.

Lastly, supported attitude is informed by cognitive schematisations rather than by specific linguistic patterns, as in the case of thematised and signal-like attitude. I am concerned here with the clues of what causes certain attitudes rather than their consequences in discourse (i.e. signal-like attitude). Analytically, I rely on a set of eight appraisal criteria that help me to infer emotions and opinions from manifest content, like the proximity in time and space between a situation that is schematised in discourse and its writer/speaker. In addition, I will discuss how visual arguments can reveal attitude and how they can be constructed and analysed.

Chapter 11

This chapter explores visual citizenship through the textual and visual analysis of Brexit-related Flickr and Twitter posts published in the month that followed the momentous vote. It does so by focusing on the posts of ordinary citizens and by quantitatively examining how topics, visual genres, social relations and stances are incorporated in these posts. However, even when Brexit (hash)tags were included in these posts, this did not automatically guarantee a direct discernible link to the referendum. As Flickr in particular allows for a high number of tags per post, and users additionally use hashtags very broadly, only three out of four posts in the Flickr corpus were directly Brexit-related. Largely the same topics were discussed on Twitter and Flickr, although compared to findings from other research, immigration, sovereignty and the NHS only play a minor role in the corpus. Brexit itself is more prominent, as a high number of posts tend to share general attitudes without addressing more specific topics. While both Twitter and Flickr posts mostly perform an additional function of self-expression, Twitter posts with only one social function are predominantly used for information sharing, while single-function Flickr posts more often contain a form of eye-witnessing. Lastly, the posts that articulated a stance mostly came out against Brexit. However, a remarkably large proportion of posts did not take a stance either xxiv Summaries of the chapters way, as the information-sharing on Twitter and the eye-witnessing on Flickr often occurred without it.

Chapter 12

The previous chapter presented the results of the quantitative analyses of Flickr and Twitter posts. Specifically, almost half of the Flickr posts include self-expression, making it the dominant social function, ahead of eye-witnessing. The proportion of visual content that is not photography (more than a quarter) is also relatively surprising for a social media platform that caters to photo lovers. These two results show the richness and diversity of that social media platform. Chapter 12 will explore further the issue of self-expression by focusing on posts that include a photograph, related or unrelated to Brexit, and self-expression. In doing so, I do not consider posts based on cartoons, memes, etc., which have already been the focus of a large body of research. Photographs and self-expression are included in 468 posts, accounting for 29% of the Flickr corpus. In other words, a third of the Brexit-related Flickr posts comprise a picture and markers of self-expression. In Section 12.1, I code the presence of thematised, signal-like and supported attitude in image and text and outline patterns of text—image relations used when citizens voice their emotions and/or opinions. In the second section, I address the issue of subjectivity in self-expression when I outline five types of appraisers in discourse. I provide a typology of eight types of verbal attitude in Section 12.3 and discuss visual patterns in Section 12.4. While these frameworks are suited for quantitative findings, they also allow for the writer’s attitudinal engagement to be analysed in fine-grained detail. In this respect, I briefly introduce Martin and White’s (2005) system of engagement in Section 12.5 and illustrate it with examples from the Flickr corpus.