Contemporary Media Stylistics: Contemporary Studies in Linguistics 9781350064089, 9781350064119, 9781350064096

Citation preview

CONTEMPORARY MEDIA STYLISTICS

Contemporary Studies in Linguistics Series Editor Li Wei, Chair of Applied Linguistics, University College London, UK The Contemporary Studies in Linguistics series presents state-of-the-art accounts of current research in all areas of linguistics. Written by internationally renowned linguists, the volumes provide a selection of the best scholarship in each area. Each of the chapters appears on the basis of its importance to the field, but also with regards to its wider significance either in terms of methodology, practical application or conclusions. The result is a stimulating contemporary snapshot of the field and a vibrant reader for each of the areas covered in the series.

Titles in the Series: Applying Linguistics in Illness and Healthcare Contexts, edited by Zsófia Demjén Contemporary Applied Linguistics Volume 1, edited by Li Wei and Vivian Cook Contemporary Applied Linguistics Volume 2, edited by Li Wei and Vivian Cook Contemporary Computer-Assisted Language Learning, edited by Michael Thomas, Hayo Reinders and Mark Warschauer Contemporary Corpus Linguistics, edited by Paul Baker Contemporary Critical Discourse Studies, edited by Christopher Hart and Piotr Cap Contemporary Linguistic Parameters, edited by Antonio Fabregas, Jaume Mateu and Michael Putnam Contemporary Media Stylistics, edited by Helen Ringrow and Stephen Pihlaja Contemporary Stylistics, edited by Marina Lambrou and Peter Stockwell Contemporary Task-Based Language Teaching in Asia, edited by Michael Thomas

CONTEMPORARY STUDIES IN LINGUISTICS

CONTEMPORARY MEDIA STYLISTICS Edited by Helen Ringrow and Stephen Pihlaja

BLOOMSBURY ACADEMIC Bloomsbury Publishing Plc 50 Bedford Square, London, WC1B 3DP, UK 1385 Broadway, New York, NY 10018, USA 29 Earlsfort Terrace, Dublin 2, Ireland BLOOMSBURY, BLOOMSBURY ACADEMIC and the Diana logo are trademarks of Bloomsbury Publishing Plc First published in Great Britain 2020 This paperback edition published in 2022 Copyright © Helen Ringrow, Stephen Pihlaja and Contributors, 2020 Helen Ringrow and Stephen Pihlaja have asserted their right under the Copyright, Designs and Patents Act, 1988, to be identified as Editors of this work. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. Bloomsbury Publishing Plc does not have any control over, or responsibility for, any third-party websites referred to or in this book. All internet addresses given in this book were correct at the time of going to press. The author and publisher regret any inconvenience caused if addresses have changed or sites have ceased to exist, but can accept no responsibility for any such changes. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. ISBN: HB: 978-1-3500-6408-9 PB: 978-1-3502-4714-7 ePDF: 978-1-3500-6409-6 eBook: 978-1-3500-6410-2 Series: Contemporary Studies in Linguistics Typeset by Deanta Global Publishing Services, Chennai, India

To find out more about our authors and books visit www.bloomsbury.com and sign up for our newsletters.

CONTENTS

LIST OF ILLUSTRATIONS NOTES ON CONTRIBUTORS 1 Introduction Helen Ringrow and Stephen Pihlaja

vii ix 1

2 ‘Beautiful masterpieces’: Metaphors of the female body in modest fashion blogs Helen Ringrow

15

3 Wolfing down the Twilight series: Metaphors for reading in online reviews Louise Nuttall and Chloe Harrison

35

4 The language of citizen science: Short strings and ‘we’ as a group marker Glenn Hadikin

61

5 The pragma-stylistics of ‘image macro’ internet memes Jane Lugea 6 The stylistics of emoji: An interactional approach Dwi Noverini Djenar and Michael C. Ewing

81

107

vi

CONTENTS

7 Rape victims and the law: Victim blaming and victimization in reports of rape in the British press Alessia Tranchese

141

8 Changing media representation of Gina-Lisa Lohfink as the icon of the ‘Nein heißt nein’ (no means no) movement in Germany Ulrike Tabbert

165

9 Child victims of human trafficking and modern slavery in British newspapers Ilse A. Ras

191

10 Reader comments and right-wing discourse in traditional news media websites Tayyiba Bruce

215

11 Straight-talking honest politics: Rhetorical style and ethos in the mediated politics of metamodernity Sam Browse

241

12 The aura of facticity: The ideological power of hidden voices in news reports Matt Davies

267

13 The style of online preachers Stephen Pihlaja

297

14 Conclusion: Contemporary media stylistics – The old, the remediated and the new Caroline Tagg

317

INDEX

329

ILLUSTRATIONS

FIGURES 3.1

Metaphorical construal 1 – READING IS A JOURNEY (READER AS AGENT)

46

3.2

Metaphorical construal 2 – READING IS A JOURNEY (TEXT AS AGENT)

46

4.1

Lexical selection process for any given string

65

5.1

Y U No

103

5.2

Futurama Fry

103

5.3

The Most Interesting Man in the World

104

5.4

Philosoraptor

104

5.5

Condescending Wonka

105

7.1

Victim(s) – process types and participant roles

150

7.2

Police and court – process types and participant roles

157

9.1

Target noun frequency over time (normalized)

201

11.1

Semiotic resources available to the cinematic narrator

254

11.2

Image and music in ‘For the many, not the few’

256

11.3

‘Interview’ shots in ‘For the many, not the few’

257

TABLES 2.1 3.1 3.2

Instagram followers of modest fashion bloggers as of 6 April 2018

23

Frequency of conceptual metaphor nodes coded in 1-star and 5-star reviews of Twilight on goodreads

44

‘Made me’ READING IS CONTROL constructions in the 5-star reviews

48

viii

3.3

ILLUSTRATIONS

‘Made me’ READING IS CONTROL constructions in the 1-star reviews

49

4.1

Concordance showing sample of dark matter (DM) corpus

68

4.2

Concordance showing sample of scienceblog (SB) corpus

69

4.3

Categories of meaning of we are data in DM and SB

70

4.4

All seven occurrences of that we are in the thirty-three lines of SB HUMAN data

72

All six occurrences of which we are in the fifty-seven lines of DM HUMAN data

72

Sample of we are data from fifty-four lines of DM AUDIENCE data

73

All four occurrences of that we are in SB AUDIENCE + SCIENTIFIC data

74

4.5 4.6 4.7 4.8

All four occurrences of we are told in SB AUDIENCE + SCIENTIFIC COMMUNITY

75

4.9

All four occurrences of we are in in DM AUDIENCE

76

4.10

All nine lines tagged as reader/writer in the data

77

5.1

The top five most popular image macro families, according to Memegenerator.net

89

Analysis of the top five families of image macros based on Shifman (2014) and Segev et al. (2015)

94

5.2 8.1

Naming strategies for lohfink

174

9.1

Articles, types and tokens per year

199

9.2

Gender of child victims, by noun

204

9.3

C-Collocates relating to youth and vulnerability

205

9.4

C-Collocates relating to abuse and exploitation

207

10.1

Representation of social actors in discourse, adopted from van Leeuwen’s framework (1996: 66)

219

Knowledge or beliefs about young Muslims in Telegraph comments

225

10.3

Knowledge or beliefs about Catholics in Daily Mail comments

231

10.4

Core Daily Mail knowledge and beliefs about Catholics

233

12.1

Editorial column headlines and regular straplines in the news data

276

A quantitative snapshot of the space given to paraphrasing and directly quoting Donald Trump in each lead news report

290

Headlines, subheadings and straplines in the news report data

292

10.2

12.2 12.3

CONTRIBUTORS

Sam Browse is Senior Lecturer in English Language at Sheffield Hallam University, UK. He is the author of Cognitive Rhetoric: The Cognitive Poetics of Political Discourse (John Benjamins, 2018). His work explores the stylistics of political discourse and the development of audience or reader-centred approaches to political text and talk. Tayyiba Bruce completed her PhD at Newman University, Birmingham, UK, in 2019. Her thesis focused on representations of Islam and Catholicism in two right-leaning British news websites using a corpus-assisted discourse study. Matt Davies is Senior Lecturer in English Language at the University of Chester, UK. His research interests include critical approaches to news discourse and the ideological function of syntactically triggered binary oppositions in news and other mass media texts. He is the author of Oppositions and Ideology in News Discourse (Bloomsbury, 2013). Dwi Noverini Djenar is Associate Professor in Indonesian Studies at the University of Sydney, Australia, with research interests in topics related to youth language practices, the stylistics of teen fiction, and language and place. Michael C. Ewing is Associate Professor in Indonesian Studies at the University of Melbourne, Australia. He researches and publishes on Indonesian and Javanese, focusing on the relationship between social action and grammatical structures.

x

CONTRIBUTORS

Glenn Hadikin is Senior Lecturer at the University of Portsmouth, UK. He is the author of Korean English (John Benjamins, 2014). He has published in a wide range of journals such as Corpora, English Today and the British Journal of Management. Chloe Harrison is Lecturer in English Language and Literature at Aston University, UK. She is the author of Cognitive Grammar in Contemporary Fiction (John Benjamins, 2017) and co-editor of Cognitive Grammar in Literature (John Benjamins, 2014). She is also the co-author of the textbook Cognitive Grammar in Stylistics: A Practical Guide (Bloomsbury, 2018). Jane Lugea is Lecturer of English Language and Stylistics at Queen’s University Belfast, UK. Her research interests lie in the stylistics of dialogue and narrative, not limited to written texts. She is the reviews editor for the main journal of stylistics, Language and Literature (Sage). Louise Nuttall is Senior Lecturer in English Language and Linguistics at the University of Huddersfield, UK, where she teaches stylistics and cognitive linguistics. She is the author of Mind Style and Cognitive Grammar (Bloomsbury, 2018) and co-editor of Cognitive Grammar in Literature (John Benjamins, 2014). Stephen Pihlaja is Reader in Stylistics at Newman University, Birmingham, UK. His research focuses on the dynamics of discourse, particularly in interaction around religious issues. Ilse A. Ras is Tutor in Criminology at Leiden University, Netherlands. She completed her PhD in English Language at the University of Leeds and is co-founder of the Poetics and Linguistics Association Special Interest Group on Crime Writing. Her work and teaching often crosses the boundaries between English language and criminology. Helen Ringrow is Senior Lecturer in Communication Studies and Applied Linguistics at the University of Portsmouth, UK. She leads the BA English Language and Linguistics programme. Her monograph, The Language of Cosmetics Advertising, was published with Palgrave in 2016. Ulrike Tabbert is Senior Public Prosecutor (Oberamtsanwältin) at a German prosecution office and Visiting Research Fellow at the University of Huddersfield, UK. She holds a PhD in Linguistics from Huddersfield and is the author of Crime and Corpus (John Benjamins, 2015) and Language and Crime (Palgrave, 2016).

CONTRIBUTORS

xi

Caroline Tagg is Lecturer in Applied Linguistics at the Open University, UK, with an interest in how social media constrains and enables different forms of communication. She is the author of Discourse of Text Messaging (Continuum 2012), Exploring Digital Communication (Routledge 2015) and Taking Offence on Social Media (Palgrave 2018, with Philip Seargeant and Amy Aisha Brown). Alessia Tranchese is Senior Lecturer in Communication and Applied Linguistics at the University of Portsmouth, UK, where she teaches a range of subjects, including gender, language and sexuality and analysing media discourse. Her research focuses on the representation of violence against women in the media.

CHAPTER ONE

Introduction HELEN RINGROW AND STEPHEN PIHLAJA

1 BACKGROUND The study of media discourse has long been an established sub-field within applied linguistics. With the development of the internet has come an increasing interest in the influence of different mediated contexts on the interaction among readers and text, including reading practices (Ensslin, 2012; Allington and Pihlaja, 2016), storytelling (Page, 2014, 2018), interaction of readers with texts in online spaces (Rowberry, 2016) and new forms of digital fiction (Shie, 2016; Bell, 2010; Bell, Ensslin and Rustad, 2013). This interest is due in part to the growing ubiquity of media in everyday life and also because the development of technology has created space for new ways in which text and audience interact, for example, with newspaper articles available online (Bruce, 2017). Contemporary media texts now exist in more formats and more platforms, offering stylisticians numerous opportunities to explore the changing shape and style of media discourse. Contemporary Media Stylistics focuses on the analysis of stylistic features of contemporary media texts, situating these features within their discursive and social contexts. For our purposes, we take a broad and inclusive view of the terms ‘contemporary’, ‘media’ and ‘stylistics’. From the outset, we recognize that ‘(the) media’ is, of course, not one monolithic structure owned and operated by a limited number of powerful actors but refers in its most basic sense to communication that is in some way mediated by technology. Increasingly, day-to-day communication among individuals takes place in technologically mediated contexts, from text messaging to social media and

2

CONTEMPORARY MEDIA STYLISTICS

video chats, that embed technology in many of our social relationships. This creates complex issues in how to treat technologically mediated interaction, particularly with increasing ambiguity between private communication and public broadcasting of messages. ‘Media’ is therefore used throughout the chapters in this book to refer to different collections of media types, including YouTube videos, blogs, Facebook pages, Twitter, online chat forums/message boards and newspapers. As the term ‘media’ can be a problematic category, differentiating between different kinds of media can also be problematic, particularly in the contemporary world. A distinction in the existing literature has been made between ‘old’ and ‘new’ media, with the ‘new’ referring broadly to more recent forms of communication on digital platforms (see Manovich, 2001, for an early discussion of how this term is used) compared with the older technologies that broadcast using analogue technologies (such as radio, television and newspapers). The differentiation between ‘new’ and ‘old’ media has been viewed as having an explanatory property: new media and old media represent not only different technologies but different ways of communicating. However, the ‘newness’ of this so-called new media, as Neary and Ringrow (2018) point out, is difficult to delineate from old media because the shift to digital technologies has meant that much of the old, legacy media (including radio and television) are now often accessed through digital technologies on ‘new media’ platforms on the internet, like the reading of newspapers online (Dovey et al., 2009). Differentiating between the two based on this criterion requires focusing primarily on the technology that is delivering the content to the audience; looking, for example, at a tabloid newspaper article online may be a materially different experience from holding a physical newspaper, but the content of the article could be identical. Our use of the term ‘contemporary’ attempts to capture this emerging media environment as it exists at the time this book is published. We recognize that the constant development of technology and its incumbent changes to language use means that differentiation between ‘old’ and ‘new’ media is increasingly problematic. We choose to view digital technology as providing new affordances – or ‘action possibilities’ (McGrenere and Ho, 2000) – for interacting with digital content and for interacting with others via mediated platforms. These new affordances can result in an almost identical experience to analogue technologies, or it may have some similarities and some differences, or it may be completely different, with new digital objects and new uses being created with these technologies. However, as Knobel and Lankshear (2014) point out in their discussion of new literacies, technological developments can embody new ‘ethos’ for how texts should be read and how readers interact with them. To that end, we are not primarily interested in how the data is classified but rather in considering how people communicate with one another

INTRODUCTION

3

in technologically mediated contexts. By examining a range of texts, this book explores issues surrounding blurred boundaries of media categorization, in a contemporary media landscape where texts, and their reading and writing practices, are in flux. Contemporary media, and particularly social media, contexts are often characterized by heightened interactivity and a disappearing distinction between producer and consumer (Seargeant and Tagg, 2014). Audiences both engage with texts in digital forms, such as commenting on online news sites or responding to media texts via Twitter, and generate their own media content on YouTube, Instagram, Periscope and so on. The increase in media interactivity can be seen to have led to ‘the transformation of a previously unidirectional broadcast mechanism to one which is bidirectional and increasingly dialogic’ (Neary and Ringrow, 2018: 305). The binary terms of ‘producer/consumer’ and ‘addresser/addressee’ can therefore not sufficiently capture how people interact with media and within technologically mediated contexts. Additionally, within these newer settings, pre-existing media audience models are not always sufficient for analysis, particularly given how contemporary media does not offer the audience a neutral picture of the world but rather provides representations with both implicit and explicit ideological positions (Baker, Gabrielatos and McEnery, 2013). This awareness of implicit and explicit ideology has been a long-standing tenet of critical discourse studies (Fairclough, 1995, 2001; Davies, 2012), and in new and social media contexts ideology has been made increasingly relevant in the filter bubbles (or echo chambers), wherein systems for providing content to users favour information that the system calculates as being most relevant for the user as well as increasing the use of the platform or site (Pariser, 2011). Filter bubbles further complicate how audiences and text producers (and those who often assume the role of both) interact and how that interaction should be analysed. The notion of a neutral reporter continues to be eroded in news reporting contexts that are increasingly and explicitly representing a particular view of the world. This book represents the state of the applied linguistic study of media stylistics at a point in the development of technologically mediated communication where mobile devices have become in many ways essential to contemporary life, but where the internet and being online is still recognizable as discrete from being offline. This book represents the breadth of approaches and conceptualizations of media in this context and highlights the diverse ways stylisticians make sense of the interaction in mediated spaces using close analysis of the interaction among texts, contexts, and text producers and consumers. Before introducing the work of this volume, a reflection on the current state of media stylistics, as well as where it has been and where it is going, will provide a framework for understanding the different studies included here and how the approaches relate to one another.

4

CONTEMPORARY MEDIA STYLISTICS

2 METHODS AND APPROACHES Because of the close relationship between media and technology, the analysis of text production and consumption in mediated contexts requires drawing on a variety of different tools to parse the influence of different contextual factors in the analysis of mediated language use. Media stylistics can be seen as a sub-field of stylistics, but it also has clear relationships with literary and media studies, with contributors to this volume drawing on different traditions in their theoretical, methodological and analytic frameworks. The nature of these interdisciplinary relationships results in the diversity of perspectives and authorities, often without an obvious lineage of shared authoritative scholars. At the same time, contemporary media stylistics methods, as this book shows, frequently have close relationships with methods for language analysis in other fields of linguistics and applied linguistics, and more generally, with stylisticians adapting methodologies and often then becoming leaders in the development of the tools they employ. Tools for stylistic analysis often include mixing of methods and approaches to describe and analyse different parts of media texts, particularly when considering the contexts in which they are produced and consumed. Stylisticians have also seen themselves as attempting to make tools for the empirical study of language relevant in literary studies and helping develop systematic, reliable methods that bridge the gap between two closely related, but often theoretically and methodically opposed, groups. Critical discourse analysis (CDA) (Fairclough, 1989, 1995) is perhaps the most influential analytic framework for language in media contexts from a linguistic perspective in the recent past. CDA is explicitly focused on how stylistic features, evidenced in lexicogrammatical structures, affect how power is represented and enacted in discourse. CDA studies regularly employ media texts in their analyses because of their ubiquity and importance in producing and sustaining political power in society. Studies of representation of power in media have historically focused on newspapers, both broadsheets and tabloids, in investigations of how different groups are favoured or marginalized by text producers. Critical stylistics (Jeffries, 2010) applies and adapts methods derived from critical analysis to look specifically at how CDA might be applied to ‘literary texts’. In a similar way, corpus stylistics (Semino and Short, 2004) has adapted methodologies from corpus linguistics (Biber, Conrad and Reppen, 1998). The focus of corpus linguistics began with using large collections of texts to study the regularities in lexicogrammatical structures in language. Early work in corpus linguistics had a profound effect on the ways in which patterns in language were perceived and studied. As methods have developed, the use of corpora and the adaptation of corpus linguistics tools in stylistics has evolved

INTRODUCTION

5

to become an important sub-field. Regularities in style choices can be observed over large collections of texts, giving evidence that can serve a variety of different topics in media discourse, including representations of processes and actors, recurring metaphorical imagery, and idiolects and dialects, as well as in the study of historical texts. Cognitive stylistics (Semino and Culpeper, 2002) applies the theories and approaches of cognitive linguistics to the study of text production and consumption. Cognitive stylistics investigates how the portrayal of events in different constructions affects the ways in which readers understand and perceive those events. Within the study of media texts, the focus on style choice and its effects has utility for almost any text or text type because it draws attention to the choices speakers make in representing their experience. Analysis of the portrayal of events, actors and processes shows how media texts do not necessarily depict a single reality but one person’s own experiences, which can be understood differently by readers and listeners. Recently, the application of text world theory (Gavins, 2007) to literary and media texts has similarly focused on the internal states of the author and the reader and has provided tools for describing and analysing these states. With the study of language and cognition more generally, particular interest has developed within stylistics for the study of metaphor. Motivated by the work of Lakoff and Johnson (1980), we can see an increased interest in conceptual metaphor theory or the idea that the use of metaphor is primarily a cognitive rather than linguistic phenomenon and is ubiquitous in thought and language. The study of metaphor in media texts can then provide further evidence about the relationship between cognition and the production of language with the potential to provide insights about style choices. Critical metaphor studies (Charteris-Black, 2004) combine the study of metaphor with elements of CDA and can be used to investigate the use of metaphor in establishing and maintaining power structures. Narrative and narratology again draw on a rich tradition within linguistic and discourse analysis, starting with the pioneering work of Labov and Waletzky (1967) on storytelling in the inner city. Storytelling is an important part of newspaper and broadcast texts and continues to be of interest to researchers in new and social media texts, where the representation of one’s experience in mediated contexts remains an important part of how users engage with one another on media platforms (Page, 2018). Analysis of storytelling, like the other approaches and methods outlined here, can be used to show how users understand themselves and others in the social world and how their experiences are constructed as reality on different platforms. Analysis of how users tell stories, the style choices they make when telling stories, and how they interact with the stories of others can show how social worlds are produced and developed in mediated contexts.

6

CONTEMPORARY MEDIA STYLISTICS

With further development of media technologies and the interaction of different text types, the application of linguistic tools to images can also be useful in multimodal analysis of media texts. Kress and van Leeuwen’s (2001) methods of multimodal discourse analysis draw on notions of (systemic) functional grammar, as with many of the other frameworks outlined here, to describe and analyse the choices that individual text producers make and their effects on readers. For media texts in online spaces, with mixtures of static and moving images, sound and text, the need to describe a variety of different style choices made by the text producer and their effect on the viewer/reader are likely to only grow in importance. Forensic linguistics and forensic stylistics (McMenamin, 2002) also have crucial media components, particularly in providing tools to investigate authorship in anonymous online, mediated contexts. The investigation of idiolect and style choices by comparing of different documents can provide evidence about text producers, their location and their identity, through the analysis of linguistic choices in texts. These methods have become particularly important in investigations of online harassment and abuse, in providing evidence for and against claims of authorship and shifts in the use and development of words (Grieve, Nini and Guo, 2018).

3 CURRENT TRENDS Because of the diversity of approaches to theory, methods and analysis, mapping current trends within media stylistics can be difficult. Depending on the particular perspective of the researcher and the focus of their analysis, different contexts and their incumbent debates and controversies will feature. The debates that are relevant to, for example, the analysis of tabloid reporting on the #metoo movement differ significantly from the debates on the analysis of the emoji in text communication. The backgrounds for every chapter in this volume are notably engaged in a diversity of different literatures, from a range of different theoretical, disciplinary and ideological perspectives. This is a marked strength of contemporary media stylistics, where diversity in approaches to texts is both accepted and celebrated. Within this diversity, overarching issues can be observed in the chapters in this volume and media stylistics more generally. The first primary trend is an emphasis on audience responses to texts, a trend that can also be seen as a dominant focus within contemporary media stylistics, particularly as the lines between audiences and authors continue to blur, and the responses to texts are complex and contradictory. This emphasis is in part a response to a tradition of previous work in stylistics which tended to focus on the form and style of texts themselves, as opposed to their reception. Stylisticians are progressively interested in exploring reader response in diverse contexts, including but not

INTRODUCTION

7

limited to contemporary British political rhetoric (Browse, 2018), the role of intertextuality in audience response (Mason, 2019), how readers engage with hyperlinks in digital fiction (Ensslin et al., 2019) and visitors’ responses to a ‘walk-through novel’ exhibition (Gibbons, forthcoming). A key theme running through much of this research is the myriad of possible and actual reader responses and the ways in which these different responses affect how texts exist in the world. Another related area of interest to stylisticians exploring contemporary media discourse is that of reliability and authority of texts, and more generally on questions about authenticity (see both Ringrow and Browse, this volume) and the accurate representation of events (see Tabbert, this volume). The prevalence of and debate around ‘fake news’ focuses on these issues, particularly how readers can determine whether authors or producers of a piece of news may deliberately be misleading or misinforming the audience (see Tandoc, Lim and Ling, 2018 for a full discussion of the difficulty in defining fake news). Although questions about accurate representations of events, and bias more generally, have been of interest in critical stylistics, the complications to context that social media introduce and the increased nefarious use of false representations by corporate and state actors make issues of authorship and authorial intent increasingly relevant. Discussions of the reliability of texts foreground questions about the position of readers and audiences in determining the value of texts and critical media literacy more generally (e.g. Funk, Kellner and Share, 2016). The increasing availability of, and interest in, differing forms of online data has led to applied linguists calling for a more carefully considered approach to the ethics of analysing online media texts (Spilioti and Tagg, 2017; Pihlaja, 2017; Mackenzie, 2017; Ringrow, this volume). Cutting across different disciplinary boundaries, there are debates over what constitutes the ‘public’ domain and whether, for example, non-password-protected internet discussion forums can be used and reproduced in research without explicit participant/ subject consent (Mackenzie, 2017). A good example of the complicated ethics surrounding online research can be seen in terms of the work (from a range of disciplines) on early pro-anorexia websites, which were regarded as complex and sensitive online communities, and researchers were thus advised to be extremely careful in conducting research in this domain (Buchanan and Zimmer, 2012). Early pro-anorexia websites did not (and arguably still do not) have an easily identifiable ‘offline’ equivalent; therefore, researchers argued it was difficult to ascertain how the concepts of informed consent, deception, subject recruitment, confidentiality of data and so on could easily apply (e.g. Brotsky and Giles, 2007; Giles, 2006). Stylisticians are increasingly analysing media texts online and recruiting participants through the internet for reader-response-style research. This raises questions around how data can be gathered, used and stored. A key guiding

8

CONTEMPORARY MEDIA STYLISTICS

principle should arguably be the minimization of harm to participants involved in the study, although there is not always agreement on what constitutes harm or indeed even what constitutes participants (e.g. if the commenters on an online discussion board are not informed that their comments are being used). The extent to which researchers can predict and mitigate possible harm in the collection and analysis of data remains an open debate and one with potentially serious consequences for how current research will be viewed in the future, particularly as social expectations around the use of personal data by corporations and researchers continue to evolve. The development of beliefs about privacy and personal data has not followed a linear path and arguments about what researchers should or shouldn’t do in terms of gaining consent before collecting media data and doing analysis continue. The role of English in contemporary media stylistics should also be considered. English, as the original language of the internet, has tended to be the default lingua franca of many global media texts and, consequently, the literature which analyses them (Danet and Herring, 2007). Within this volume, both Tabbert’s and Djenar and Ewing’s chapters do address languages other than English, and the English data explored in other chapters represent a range of different types of Englishes in their multiple and diverse contexts. In keeping with global trends in discourse analysis and stylistics, researchers must recognize the limitations of data choices by being careful not to generalize about ‘language’ without considering both non-English data and non-Anglophone settings, in addition to sign languages.

4 MEDIA FUTURES Considering these current trends, what media futures are likely to emerge in the coming years is, of course, difficult to forecast. Even a cursory analysis of the current media environment shows the unpredictability of the influence of media in day-to-day life. Indeed, few in 2012 would have predicted that the particular media environment would have led to Donald Trump’s improbable win in the 2016 US presidential election, much less the mechanisms which allowed him to do so. In hindsight, the trends can be clear, but at the same time, considering the arc of technology and the development of new technologies, platforms and practices, several issues of particular interest to media stylisticians are likely to remain salient in the coming decade. Throughout much of the history of the internet, discussions about public and private online spaces and how users conceived of themselves online was a regular theme in the analysis (boyd, 2011; Tagg, Seargeant and Brown, 2017). However, as technology and surveillance has developed, simply using technology to access information means that text production and consumption is passively monitored in almost every instance. Users who click on particular

INTRODUCTION

9

news stories provide information not only about what they have read but about how long they have read it, what they have skipped and where they have gone next in their reading. Internet service providers, websites (and companies that own them) and governments have access to this information and can use it in any number of ways. Physical technologies like phones, television, gaming consoles and virtual reality devices will continue to play an important role in how texts are produced and consumed. Some of these trends are predictable given their histories of development – we can assume computer processors will continue to become smaller and more powerful – but how they are used and the technologies they make possible are difficult to predict. One may have been able to predict the smartphone as a device, but the ways in which it has affected our production and consumption of texts would have been more difficult to forecast. The study of style in the development of media will then likely be tied to the development of technology and the particular affordances that those technologies provide, adapting the principles of linguistic analysis for different technologies. Social media platforms have developed into sites of surveillance (Trottier, 2016), where users voluntarily provide data about themselves and their use of technology in exchange for convenient access to information and others. The information that users provide allows for an increase in automation, and artificial intelligence (AI) has increasingly been able to accomplish tasks on user’s behalf. The use of AI in these contexts will likely have consequences for the development of text production and consumption going forward, particularly the development of technologies which monitor reading and writing online and tools that have been developed in relation to the style choices that people make. Now, AI is increasingly able to predict what writers are attempting to communicate through autocorrect and autocomplete functions in the production of texts. Programs like GPT-2 can now create texts based on machine learning of previous writing, and consumer products like Grammarly actively assist authors in the production of their own texts. As these technologies develop, the extent to which communication between interlocutors becomes automated in mediated environments will be an important trend, and equally important will be the potential for resistance to automation in the interest of authenticity. How users as text producers and consumers identify and value texts will determine how these technologies are implemented. These technologies also have consequences for the texts which users read and have access to. As noted earlier in this chapter, users are subject to a filter bubble (Pariser, 2011), wherein experiences of media and what information is presented to an individual differs depending their preferences and how automated technologies respond to these preferences, often without the user’s awareness. Machine learning algorithms now affect how ads are placed on particular webpages and what search results a particular user sees, with the

10

CONTEMPORARY MEDIA STYLISTICS

content increasingly being tailored to the individual in every context. The extent to which these technologies might continue to expand and not only direct traffic in different ways but adapt the presentation of the same material to different users will likely have consequences for how style develops. However, there is an increased visibility of these processes and a growing public scrutiny about the ways social media companies in particular operate. There is also the potential for intervention by technology producers, governments and users themselves to push back against technologies and resist these developments. Issues of privacy have been a perennial issue in the development of mediated communication in online spaces and are likely to continue into the near to medium future. Because AI relies on user information to develop and tailor effectiveness of surveillance on reading and writing, the extent to which this monitoring is either curtailed or expanded is subject to not only the ability of technology to monitor users but also the decisions of those producing, using and regulating those technologies. In response to this, media stylisticians will continue to adapt methods to take into account the complexities of text production and consumption in new environments, building on the work of those who have done so in the past. Moreover, the possibilities for media stylistics to provide insights on reading and writing in mediated contexts are endless. The chapters in this book represent a variety of different ways that media stylisticians have been approaching these questions in the recent past, and the insights that their investigations have shown provide a starting point for future work. The work is intended as an invitation to research, for readers to consider, adapt and apply to their own data.

5 BOOK CONTENT The goal of Contemporary Media Stylistics is to capture a range of different research methods and approaches to media texts in different contexts, from traditional media sources like newspaper reporting to online community contexts and discussions about reading practices. The range of approaches and topics aims to represent some of the most up-to-date methods and theories in the field while also attempting to inspire new research by readers of the book to look further at the topics covered. The book is organized with loose relationships between chapters meant to build on one another. The first two chapters discuss metaphor, showing how patterns in thinking and speaking demonstrate how text producers and consumers understand experience using metaphor. Chapters 3 through 5 look at stylistic patterns in different online contexts, including both interactive contexts like community forums and texts themselves such as memes. Chapters 6 through 9 look specifically at representations of marginalized groups and people in newspaper reporting. Finally, Chapters 10 through 12

INTRODUCTION

11

look at stylistic features in the ways in which texts and speakers represent particular opinions and positions in relatable and engaging ways, including editorials, political speeches and religious discourse. The book begins with Ringrow’s chapter (Chapter 2) which explores conceptual metaphors used to describe the female body in modest fashion blogs. Ringrow’s analysis shows that implicit ideologies about women and mothers are revealed in how bloggers write about modesty and the human body. Nuttall and Harrison (Chapter 3) similarly consider metaphor in their chapter, investigating how reading is conceived of using metaphorical language. They analyse data from the popular app and website Goodreads and look at 100 positive and 100 negative reviews on the site, considering how users frame positive and negative reading experiences using metaphorical language. Hadikin (Chapter 4) addresses ways communities of users in online spaces interact and the patterns that develop in that interaction. He uses corpus analysis to investigate ways in which group identity is constructed, particularly with the use of the pronoun ‘we’ in discussions on a citizen science forum. Lugea (Chapter 5) investigates stylistic features of memes, with a focus on structure in both the most liked and most disliked memes on the Memegenerator. net. The analysis employs principles of speech act theory to describe the ways in which adhering to and transgressing expectations about meme structure is reflected in what is considered a good or bad meme. Djenar and Ewing (Chapter 6) take a similar pragmatics-based approach to investigate the use of emojis in a Malaysian online community, focusing on the different ways emojis are used by users. Their analysis shows how emojis create dialogic resonance and intersubjective alignment, and the analysis suggests that the use of emoji needs to be understood contextually in discourse activity. Tranchese (Chapter 7) uses CDA to investigate the representation of rape in the British press, specifically looking at collocation and how social actors and processes are presented in this reporting. The analysis shows how reporting about rape reveals ideologies about gender roles and how ‘legitimacy’ is determined in reporting about rape. Tabbert (Chapter 8) similarly deals with the issue of sexual violence in newspaper reporting and in particular the case of Gina-Lisa Lohfink, a German reality television star, who was accused of lying about having been raped. The article uses discourse analysis of stylistic choices in German newspaper articles to show how Lohfink was negatively represented in the press and how those negative representations are linked to larger issues about ‘good’ victims. Ras (Chapter 9) also discusses issues around the representations of sexual violence using corpus analysis, and particularly how child victims of human trafficking are represented in newspapers. The analysis highlights how representations of these victims are used to legitimate governmental responses to trafficking.

12

CONTEMPORARY MEDIA STYLISTICS

Bruce (Chapter 10) similarly analyses print media, investigating right-leaning newspaper reporting on Islam and Catholicism. The study employs corpus linguistics in a socio-cognitive framework to analyse keywords in the articles and user comments on online articles, showing the link between article and comment content, and how stereotypical views are reproduced in response to news content. Browse (Chapter 11) uses rhetorical analysis of political speech to look at how ‘authenticity’ is created in the language use about the British Labour leader Jeremy Corbyn, particularly in campaign videos used to represent his position and the Labour Party. Browse shows how performance of authenticity is situated with respect to the contemporary metamodern political and media context. Davies (Chapter 12) presents an analysis of the different stylistic features in print newspapers, looking at the differences between the representation of ‘hard news’ and editorials. His analysis demonstrates how spoken discourse markers show stance taking and indicate the author taking an opinion rather than presenting hard news reporting. Pihlaja (Chapter 13) investigates stylistic features in religious discourse in mediated contexts, and how evangelical preaching is conducted in online contexts, particularly Facebook. The analysis shows how one preacher creates a sense of intimacy with the viewer, using positioning analysis to explore how preachers create and sustain relationships with audiences. Finally, Tagg (Chapter 14) draws together the themes in the different chapters and offers reflections on the topics, theories and methods drawn on in the book, in addition to considering how these ways of doing analysis may develop going forward.

REFERENCES Allington, D. and Pihlaja, S. (2016). ‘Reading in the age of the internet’. Language and Literature, 25(3), 201–10. doi:10.1177/0963947016652781 Baker, P., Gabrielatos, C. and McEnery, T. (2013). Discourse Analysis and Media Attitudes: The Representation of Islam in the British Press. Cambridge: Cambridge University Press. Bell, A. (2010). The Possible Worlds of Hypertext Fiction. London: Springer. Bell, A., Ensslin, A. and Rustad, H. (2013). Analyzing Digital Fiction. London: Routledge. Biber, D., Conrad, S. and Reppen, R. (1998). Corpus Linguistics: Investigating Language Structure and Use. Cambridge: Cambridge University Press. boyd, d. (2011). ‘Social network sites as networked publics: Affordances, dynamics, and implications’. In Z. Papacharissi (Ed.), A Networked Self: Identity, Community, and Culture on Social Network Sites (pp. 39–58). London: Routledge. Brotsky, S. R. and Giles, D. (2007). ‘Inside the “pro-ana” community: A covert online participant observation’. Eating Disorders, 15(2), 93–109.

INTRODUCTION

13

Browse, S. (2018). Cognitive Rhetoric: The Cognitive Poetics of Political Discourse. Amsterdam: John Benjamins. Bruce, T. (2017). ‘New technologies, continuing ideologies: Online reader comments as a support for media perspectives of minority religions’. Discourse, Context & Media, 24, 53–75. doi:10.1016/j.dcm.2017.10.001 Buchanan, Elizabeth A. and Zimmer, M. (2012). ‘Internet research ethics’. Retrieved from https://plato.stanford.edu/entries/ethics-internet-research/ (Accessed on 19 April 2019). Charteris-Black, J. (2004). Corpus Approaches to Critical Metaphor Analysis. Basingstoke: Palgrave Macmillan. Danet, B. and Herring, S. C. (2007). The Multilingual Internet: Language, Culture, and Communication Online. Oxford: Oxford University Press. Davies, M. (2012). Oppositions and Ideology in News Discourse. London: Bloomsbury. Dovey, J., Giddings, S., Grant, I., Kelly, K. and Lister, M. (2009). New Media: A Critical Introduction. London: Routledge. Ensslin, A. (2012). ‘“I want to say I may have seen my son die this morning”: Unintentional unreliable narration in digital fiction’. Language and Literature, 21(2), 136–49. doi:10.1177/0963947011435859 Ensslin, A., Bell, A., Smith, J., Van Der Bom, I. and Skains, L. (2019). ‘Immersion, digital fiction, and the switchboard metaphor’. Participations, 16(1), 320–42. Fairclough, N. (1989). Language and Power. London: Longman. Fairclough, N. (1995). Critical Discourse Analysis: The Critical Study of Language. London: Pearson. Fairclough, N. (2001). Language and Power. Essex: Pearson. Funk, S., Kellner, D. and Share, J. (2016). ‘Critical media literacy as transformative pedagogy’. In Handbook of Research on Media Literacy in the Digital Age (pp. 1–30). Hershey, PA: IGI Global. Gavins, J. (2007). Text World Theory: An Introduction. Edinburgh: Edinburgh University Press. Gibbons, A. (forthcoming). ‘“Why do you insist Alana is not real?”: Visitors’ perceptions of Andi and Lance Olsen’s “there’s no place like time” exhibition’. In A. Bell, S. Browse, A. Gibbons and D. Peplow (Eds), Style and Response: Minds, Media, and Methods. Amsterdam: John Benjamins. Giles, D. (2006). ‘Constructing identities in cyberspace: The case of eating disorders’. British Journal of Social Psychology, 45(3), 463–77. Grieve, J., Nini, A. and Guo, D. (2018). ‘Mapping lexical innovation on American social media’. Journal of English Linguistics, 46(4), 293–319. doi:10.1177/0075424218793191. Jeffries, L. (2010). Critical Stylistics: The Power of English. Basingstoke: Palgrave Macmillan. Knobel, M. and Lankshear, C. (2014). ‘Studying new literacies’. Journal of Adolescent & Adult Literacy, 58(2), 97–101. Kress, G. and Van Leeuwen, T. (2001). Multimodal Discourse: The Modes and Media of Contemporary Communication. London: Bloomsbury. Labov, W. and Waletzky, J. (1967). ‘Narrative analysis: Oral versions of personal experience’. In J. Helm (Ed.), Essays on the Verbal and Visual Arts (pp. 12–44). Seattle and London: University of Washington Press. Lakoff, G. and Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press.

14

CONTEMPORARY MEDIA STYLISTICS

Mackenzie, J. (2017). ‘Identifying informational norms in Mumsnet Talk: A reflexivelinguistic approach to internet research ethics’. Applied Linguistics Review, 8(2–3), 293–314. Manovich, L. (2001). ‘Post-media aesthetics’. In M. Kinder and T. McPherson (Eds), Transmedia Frictions, the Digital, the Arts, and the Humanities (pp. 34–44). Oakland, CA: University of California Press. Mason, J. (2019). Intertextuality in Practice. Amsterdam: John Benjamins. McGrenere, J. and Ho, W. (2000). Affordances: Clarifying and Evolving a Concept. Paper presented at the Graphics interface. McMenamin, G. R. (2002). Forensic Linguistics: Advances in Forensic Stylistics. London: CRC Press. Neary, C. and Ringrow, H. (2018). ‘Media, power and representation’. In A. Hewings, P. Seargeant and S. Pihlaja (Eds), The Routledge Handbook of English Language Studies (pp. 294–309). London: Routledge. Page, R. (2014). ‘Counter narratives and controversial crimes: The Wikipedia article for the “Murder of Meredith Kercher”’. Language and Literature, 23(1), 61–76. doi:10.1177/0963947013510648. Page, R. (2018). Narratives Online: Shared Stories in Social Media. Cambridge: Cambridge University Press. Pariser, E. (2011). The Filter Bubble: What the Internet Is Hiding from You. New York City: Penguin. Pihlaja, S. (2017). ‘More than fifty shades of grey: Copyright on social network sites’. Applied Linguistics Review, 8(2–3), 213–28. Ringrow, H. (this volume). ‘“Beautiful masterpieces”: Metaphors of the female body in modest fashion blogs’. In H. Ringrow and S. Pihlaja (Eds), Contemporary Media Stylistics. London: Bloomsbury. Rowberry, S. P. (2016). ‘Commonplacing the public domain: Reading the classics socially on the Kindle’. Language and Literature, 25(3), 211–25. doi:10.1177/0963947016652782. Seargeant, P. and Tagg, C. (2014). The Language of Social Media: Identity and Community on the Internet. Basingstoke: Palgrave Macmillan. Semino, E. and Culpeper, J. (2002). Cognitive Stylistics: Language and Cognition in Text Analysis (Vol. 1). Amsterdam: John Benjamins. Semino, E. and Short, M. (2004). Corpus Stylistics: Speech, Writing and Thought Presentation in a Corpus of English Writing. London: Routledge. Shie, J.-S. (2016). ‘Variations in the use of intertexts at the macro-contextual level: The case of English press news’. Language and Literature, 25(2), 95–112. doi: 10.1177/0963947015623614 Spilioti, T. and Tagg, C. (2017). ‘The ethics of online research methods in applied linguistics: Challenges, opportunities, and directions in ethical decision-making’. Applied Linguistics Review, 8(2–3), 163–7. Tagg, C., Seargeant, P. and Brown, A. A. (2017). Taking Offence on Social Media: Conviviality and Communication on Facebook. London: Springer. Tandoc, E. C., Jr, Lim, Z. W. and Ling, R. (2018). ‘Defining “fake news” a typology of scholarly definitions’. Digital Journalism, 6(2), 137–53. Trottier, D. (2016). Social Media as Surveillance: Rethinking Visibility in a Converging World. London: Routledge.

CHAPTER TWO

‘Beautiful masterpieces’: Metaphors of the female body in modest fashion blogs HELEN RINGROW

The genre of modest fashion blogs is growing among Evangelical Christian and Latter-Day Saint1 women, especially young wives and stay-at-home mothers (SAHMs) in North America. These blogs are often seen as appropriate (and even desirable) activities in which women develop and expand online communities in ways that are consistent with their religious beliefs. Blogs focusing on modest fashion (cf. Lewis, 2013) tend to emphasize the religious rationale for modest fashion choices and showcase particular modest outfits, aiming to encourage and inspire their readers who hold shared values. This chapter explores how the concept of modesty (and women’s bodies more generally) is represented in this online sphere. Following Knapton’s (2013) approach to metaphor in proanorexia blogs, the analysis in this chapter applies conceptual metaphor theory (CMT) (Lakoff and Johnson, 1980/2003) to the bio sections of twelve modest fashion blogs. The findings suggest common metaphors used in these blogs (such as LIFE IS A WORK OF ART [GOD IS AN ARTIST/WOMEN ARE DRAWINGS]) are used to emphasize a narrative of gender differences and to suggest faith-based views of the female body are radically different from those of mainstream society.

16

CONTEMPORARY MEDIA STYLISTICS

1 INTRODUCTION 1.1 Introducing modest fashion This chapter aims to explore discursive representations of the female body in modest fashion blogs, predominantly through the identification and examination of common metaphorical expressions used to describe both modest fashion and the female body. In so doing, this research hopes to add to an understanding of how metaphors can conceptualize religious values and to identify which metaphors are employed by women bloggers in this emerging genre of modest fashion blogs. Although modest fashion is not in itself a new concept, the increasing availability of modest fashion resources online (both in terms of fashion blogs and specialized internet retailers) is a more recent development (cf. Lewis, 2013). While proponents of modest fashion may welcome this change, the growing visibility of modest fashion online has also attracted attention from those more critical of dressing modestly (Lewis, 2013). Modest fashion blogs are predominantly (although not exclusively) produced by women who self-identify with particular faith groups (Lewis, 2013), and blogs focusing on modest fashion tend to emphasize the religious rationale for making modest fashion choices and showcasing certain outfits, aiming to encourage and inspire their readers who hold (or are presumed to hold) shared values. In discussing religiously motivated sartorial choices, modest fashion blogging can be contextualized within the wider domain of female-authored blogs as a way of narrativizing women’s personal experience in an online public sphere (Lopez, 2009) and as an expression of feminized popular culture (Levine, 2015). Exploring modest fashion blogs from a linguistic perspective can give insights into the ways in which language, religion, and gender intersect in this emerging internet genre, in addition to considering how this form of digital media ‘might enter into, punctuate, and impact women’s daily lives, tweaking, exacerbating, or revising how women move through and feel everyday experiences’ (Wilson and Yochim, 2015: 233). This chapter presents a discourse analysis of the bio sections of twelve contemporary blogs predominantly focused on modest fashion, looking specifically at the presence of metaphor (following Lakoff and Johnson, 1980/2003; Knapton, 2013). The findings suggest that metaphors help to (1) emphasize a narrative of gender differences and (2) suggest that faith-based views of the female body are radically different from those of mainstream society. 1.2 Controversies As one modest fashion blogger (Fresh Modesty) wryly comments, posting 500-plus posed photographs of oneself on the internet may not necessarily be

‘BEAUTIFUL MASTERPIECES’

17

considered ‘modest’, by any standard (see, for example, this post which discusses in more detail how this blogger grapples with the somewhat oxymoronic nature of modest fashion: http://freshmodesty.blogspot.com/2013/08/purpose-m e.html). This apparently paradoxical nature of modest fashion attracts frequent debate both in the academic literature (see, for example, Lewis, 2013; Cameron, 2013) and in the often much less forgiving online critiques of these blogs, such as on the GOMI (Get Off My Internets) forum, a so-called snark community, in which many religious fashion bloggers are criticized both for their apparent proselytizing and/or for their ostensible lack of consistency in implementing modest fashion ‘rules’. It is important to note, first, that there is no universally agreed definition of modest fashion and, second, that much reader response to the blogs tends to be encouraging in nature. Cameron (2013: 35) argues that although ‘judging’ is a normal part of fashion blogging, reader comments on modest fashion blogs do tend to be positive. Of course, comments are generally moderated by the blog owner: blogs are a form of participatory media, but they are not necessarily fully democratic (Neary and Ringrow, 2018: 303–14). In terms of the difficulty of defining modesty, Lewis (2013: 3) emphasizes that ‘modesty is a mutable concept that changes over time and is diversely adopted, rejected, altered by or in some cases imposed on different groups of women (and, to a lesser extent, men) in different times and places’. 1.3 Rationale for modest fashion? Not every modest fashion blogger dresses modestly for religious reasons, but all the modest fashion bloggers discussed in this chapter both identify with a faith group and cite this as the primary motivation for their fashion choices. This chapter focuses on personal fashion blogs written by Evangelical Christian and Latter-day Saint women.2 Although not all of these bloggers agree exactly on what modest dressing is or how best to implement it within and across these two religious communities, there seems to be a shared understanding of why modest dressing is important. Modest fashion for the Latter-day Saint bloggers has a particular religious context. Many members of the Church of Jesus Christ of Latter-day Saints wear a special kind of underwear known as ‘garments’ to symbolize their beliefs. Latter-day Saints who take part in ‘endowment rites’3 in the temple wear these items as part of the ceremony and then under their clothes during everyday life; as such, these garments are seen as a sacred and personal reminder of their religious commitments and beliefs (Davies, 2003: 217). Members of the Church of the Latter-day Saints are encouraged to wear these items of clothing respectfully and modestly, given their sacred nature, and not to show the garments to those outside of their faith (Davies, 2003). Modest dressing for many Latter-day Saint women therefore involves choosing items of clothing

18

CONTEMPORARY MEDIA STYLISTICS

that cover their garments appropriately (although garments are generally removed for activities such as swimming and sex). Evangelical Christian women do not wear these garments, although some of their modest fashion choices (clothes covering the tops of the arms, knee-length skirts, and not wearing anything too tight or low-cut) may in practice be similar to their Latter-day Saint counterparts. With reference to Lewis’s (2013: 3) comments noted above about modesty not being an immutable concept, it is worth stating that the style of Latter-day Saint garments has changed over time, with some shortening of the length of sleeves, for example (Davies, 2003: 217), which may influence the clothing worn over them. The rationale given for Latter-day Saint and Evangelical Christian bloggers’ fashion choices tends to come largely from their interpretation of holy texts (notably, the Bible for both Evangelical Christian and Latter-day Saint women, in addition to the Doctrine and Covenants for the latter group) and, more broadly, from established norms within these religious communities. The reasons behind modest fashion given on the blogs discussed in this chapter are, on the whole, representative of modest fashion blogs produced by Latterday Saint and Evangelical Christian women. These motivations include: using fashion as a public display of personal convictions; not ‘leading men into sin’ (i.e. causing men to think or act in ‘sexually impure’ ways);4 emphasizing gender roles (bloggers often argue that clothing should reflect and even highlight what they deem to be traditional, God-given gender differences); differentiating oneself from others who do not share the same values; and showing respect both for one’s body and for other people. These views tend to echo those found in (predominantly Christian) religious literature especially ‘self-help’ books on this topic (see, for example, Beaver and Beaver, 2010; Franklin, 2006; Pollard, 2004; Peace and Keller, 2015; and Clark and Baird, 2016). Although there is some consistency in the justifications provided, it is worth nothing that, particularly within the Evangelical Christian tradition (perhaps slightly less so for Latter-day Saint women as there is more concrete guidance available; see earlier comments on garments), the ‘rules’ on modest dressing seem to be slightly more varied and, of course, biblical scripture is often open to interpretation. As Drury (1994: 30) argues, ‘It is on biblical texts that Christian attitudes to women are based, but this does not mean that there is any straightforward biblical teaching about women which everyone accepts’ – a proviso which is certainly applicable to modest dressing, among other things. With this caveat in mind, the following definition of modesty from Vaughan and Vaughan (2005: 92) does seem to be representative of the blogs explored in this chapter: Clothing is sensual or provocative when it is either ‘too little’, ‘too thin’, or ‘too tight’. By ‘too little’ we mean things like short skirts, low-cut hip-

‘BEAUTIFUL MASTERPIECES’

19

huggers, blouses that are either too short (halter tops), unbuttoned, too low-cut; gaping holes on sleeveless shirts that expose the undergarments or breasts, and things similar to this. There are too many examples to mention but the basic idea is that the article of clothing is leaving too much flesh exposed. There are valid and valuable critiques of the rationales behind modest dressing. For example, there is arguably a blurred line between the discourses of ‘purity culture’ and those of ‘victim blaming’ where focus is placed on what women are wearing in narratives around rape (cf. Benedict, 1992 on the media (mis)representation of women who have been raped). However, given historical gender imbalances in most major religions worldwide, the role of women in modest fashion blogs as spiritual interpreters and instructors is significant. As Lewis (2013) argues, ‘Online modesty discourse can indeed be regarded as a new form of religious discourse in which women are achieving recognition as religious interpreters and intermediaries’ (6–7), and this online modesty discourse also ‘creates a parallel world of fashion activity for groups of dressers usually ignored by the fashion media and extends discourse about modesty into arenas beyond conventional religious organisations or authorities’ (2). Modest fashion bloggers often try to address the question of ‘how to dress fashionably while adhering to one’s own interpretation of religious guidelines on dress, but also how others perceive or assume a religiously motivated modest dresser should dress or behave’ (Cameron, 2013: 24). How these women represent their faiths is also a key area of interest and may be perhaps more acute for Latter-day Saint modest fashion bloggers because the Church of Jesus Christ of Latter-day Saints has been subject to hostility and misrepresentation in the American public domain, something their increased official web presence is trying to counter (see Haws, 2013, and in particular pages 195 through 200 on high-profile Latter-day Saints in contemporary US society). Although fashion blogs, modest or otherwise, might be easily dismissed as trivial or mundane, the analysis of these digital ‘feminine’ spaces can potentially offer compelling insights into gender, culture and an online community in which peer-to-peer interactions may represent broader and potentially more rewarding experiences for women than traditional ‘top down’ media contexts (cf. Levine, 2015: 4–7). Fashion media, including fashion blogs, deserves serious academic attention as a space for both expression and interaction; a space which raises questions of the complicated connections between clothing, bodies and gender (Hunting, 2015: 116; 122). The competencies required for women to produce these blogs, in terms of fashion knowledge, technical capabilities, and personal branding and promotion (something which is key for this genre), should also not be overlooked (Hunting, 2015: 123).

20

CONTEMPORARY MEDIA STYLISTICS

2 METHODS 2.1 Research aims Metaphor has been selected for exploration in this chapter as metaphors are a useful way of researching language and ideology as they may potentially reflect and construct shared values of certain groups (Semino, 2008: 34). Metaphors can help to create and transmit knowledge and beliefs about the self and about experiences (Bates, 2015; Steen, 2010). Metaphor is also frequently used in the Bible and other religious, especially persuasive, texts (cf. Charteris-Black, 2004; Koller, 2017; Pihlaja, 2013), and some of the metaphorical expressions analysed in this chapter seem to draw on biblical metaphors. Charteris-Black (2004) found extensive use and diversity of metaphor in the Bible, with four times as many metaphors being found in the books of the Old Testament than the New Testament in his sample (this is likely to be due to the increased use of figurative language in the former and a more factual-style narrative account of Jesus’s life in the latter; see Charteris-Black, 2014: 181). Charteris-Black (2004: 173) argues that metaphor is particularly useful in religious texts because ‘it is a primary means by which the unknown can be conceptualised in terms of what is already known’, while Koller (2017: 6) describes religious discourse as ‘a prime candidate for metaphor’. The blogs analysed in this chapter can be described as religious texts in the sense that they are written by women who identify with particular faiths and discuss spiritual themes. The findings and discussion in this chapter aim to add to the growing body of studies which explore online gendered spaces, considering how these spaces play a role in terms of people’s understandings of themselves and gender (reproducing, challenging or subverting norms), although it is worth emphasizing these spaces do not represent a complete detachment from ‘offline’ discourses (Mackenzie, 2017b: 309). 2.2 Data analysed Twelve posts from contemporary modest fashion blogs produced by Evangelical Christian and Latter-day Saint women were chosen for analysis in this small-scale qualitative study. Blogging in general has a popularity among some Evangelical Christian and Latter-day Saint women, especially (although not exclusively) young wives and SAHMs (cf. Lopez, 2009). There are two broad streams in popular English language modest fashion blogs: Muslim modest fashion blogs, which tend to be from the UK, and Christian, Latter-day Saint and Jewish modest fashion blogs, which tend to be from North America (Cameron, 2013). This chapter predominantly addresses the latter, focusing on Christian and Latter-day Saint bloggers who seem to hold some shared approaches to modest dressing. As with other similar popular

‘BEAUTIFUL MASTERPIECES’

21

cultural texts (and perhaps unsurprising considering the religious context of these modest fashion blogs), these blogs represent femininity as a cisgendered construction (cf. Levine, 2015). The bloggers identify as women and, judging by the reader comments, their target audience appear to identify as women also. The conceptualization of gender on the content of these blogs is to all intents and purposes binary. Reflective of most popular Latter-day Saint and Evangelical Christian modest fashion blogs, the blogs in this chapter are US-centric (all bloggers featured here are currently based in North America) and written in English, predominantly (although not exclusively) by (what appear to be) white women who seem to identify as heterosexual. This is a rough impression based on how these bloggers portray themselves and how they discuss topics such as sexuality. However, judging from the readers’ comments and the popularity of the blogs (and, of course, their notoriety on internet discussion forms), their influence does extend beyond the borders of the United States, as is often the case with other popular feminized texts (cf. Levine, 2015: 11). The blogs analysed in this chapter are Chez Heidi, Girl Defined, Fresh Modesty,5 Modest Blondie, Courtney Toliver, Bramblewood Fashion/Plaid Lattes,6 The Boyer Sisters, Downtown Demure, Modest Goddess, Hello Modesty, Modest Style and Modestly Hot. These blogs all have a significant amount of modest fashion content: the majority are dedicated almost exclusively to modest fashion, with some additional occasional content on beauty and lifestyle more generally. The main exception to this is Girl Defined, a blog on ‘God’s beautiful design for womanhood’ (see https://www.girldefined.com/ and https://www. girldefined.com/meet-us) which devotes equal time to other topics deemed relevant to Evangelical Christian women as well as modesty. However, it has been included in this study as a focus on modesty is a recurring theme on this blog; as such, one of their ‘essay style’ posts on modesty7 has been selected for the purposes of this analysis (since their bio section reflects this more general remit of their blog). These blogs should be viewed as case studies of religious modest fashion blogs. Blogs vary greatly according to form and content, but, in keeping with most other internet texts which fall under the ‘social media’ umbrella, their use generally creates or maintains, even in a broad sense, some kind of social community (Myers, 2010). As Neary and Ringrow (2018: 303) argue, ‘This heterogeneity of social media genres, which includes a range of forums and platforms, reflects a corresponding hybridity of form, comprised as it is of both written and spoken discourse features.’ A majority of the blogs follow a personal fashion blog format, with typical posts including photographs of bloggers in nice settings (Hunting, 2015) plus commentary, for example, on why they chose this outfit, where they wore it and often tags (by season). Like other fashion blogs, there are links to the bloggers’ additional social media

22

CONTEMPORARY MEDIA STYLISTICS

accounts, including Instagram, YouTube, Pinterest and Twitter, and also site ‘badges’ if they belong to blogger networks. Latter-day Saint and Evangelical Christian modest fashion blogs have not typically been quite as heavily monetized as some of the other blogs by women, such as motherhood and lifestyle blogs (see Hunter, 2016); this does appear to be changing somewhat as there are an increasing number of sponsored posts and affiliate links, particularly from the more popular bloggers analysed here. Modest Goddess, for example, now has its own lingerie line and Girl Defined has published books on modest dressing and other Christian living themes. The features of blogs discussed so far mirror those identified by Cameron (2013) as common characteristics of modest fashion blogs. A final common aspect of these blogs identified by Cameron (2013: 33–4) is the ‘mixing’ of a perceived secular domain (fashion) with the spiritual: The blogs illustrate, through the design and images sourced along with the text and images produced, that the religious (beliefs manifest as in dress) and the secular (the fashion field) do not constitute fixed boundaries. The blog space, as a personal space and a site of identity construction, provides a place where such peripheries are explored personally and (re)articulated before and to a wider public. While that space between the sacred and secular might be unproblematic for the modest fashion bloggers existing in it, it is not always perceived as such by the viewing audience. With this in mind, for some of the blogs it is not necessarily immediately clear which religious faith (if any) the blogger associates with (Cameron, 2013). If the faith is ‘downplayed’, this could be an attempt to keep their readership as wide as possible (Cameron, 2013), in addition to their potential sponsorship deals. Some bloggers state their religious views somewhere on the homepage (often by linking to, for example, the official Church of Jesus Christ of Latter-day Saints website); others include them in the bio sections and introductory blog posts where their religious affiliation can sometimes become clearer (Cameron, 2013). Bloggers in general, especially those with large followings, can be considered a kind of contemporary ‘microcelebrity’ (Jerslev, 2016; Gamson, 2011; Marwick and boyd, 2010; Marwick, 2013), who become ‘internet famous’ (Tanz, 2008) through a bottom-up process. The performance of micro-celebrity is different from mainstream celebrity performance, certainly in terms of temporality, as bloggers and vloggers continually post and share with their followers or fans ( Jerslev, 2016). The modest fashion bloggers selected for this chapter demonstrate many of these aspects of micro-celebrity (Jerslev, 2016: 5238), especially their ‘continuous and multiple uploads of performances of a private self ’. Indeed, micro-celebrity can be aptly defined as ‘the presentation of oneself as a celebrity regardless of who is paying attention’ (Marwick, 2013: 144). Of course, the more popular

‘BEAUTIFUL MASTERPIECES’

23

TABLE 2.1 Instagram followers of modest fashion bloggers as of 6 April 2018 Number of Instagram Followers as on 6 April 2018

Blog

Faith Affiliation

Courtney Toliver

Evangelical Christian

47.5k

Modestly Hot

Evangelical Christian

36.2k

Girl Defined

Evangelical Christian

28.2k

Modest Style

LDS

25.9k

Modest Goddess

LDS

24.9k

Downtown Demure

Evangelical Christian

22.2k

Hello Modesty

Evangelical Christian

12.6k

Fresh Modesty

Evangelical Christian

7429

The Boyer Sisters

Evangelical Christian

2222

Bramblewood Fashion

Evangelical Christian

1719

Modest Blondie

Evangelical Christian

464

Chez Heidi

LDS

719

the blogger becomes, the more attractive they are to brands in terms of potential sponsorship deals (Jerslev, 2016). Personal fashion blogs are, in many respects, sites for public identity construction (Cameron, 2013), and reflect ‘ordinariness’ as a common theme in more recent North American celebrity culture (Gamson, 2011): women share their everyday fashion choices in a way that might seem more ‘democratized’ than traditional forms of celebrity. The modest fashion bloggers analysed in this chapter have between 719 and 47.5k Instagram followers (as of April 2018), as shown in Table 2.1. All the bloggers in this chapter have Instagram accounts, which give information on the number of followers that is not usually found on the blog. Of course, it is possible that their Instagram following may be larger or smaller than that of their blog, but nonetheless this gives some rough indication of their popularity. 2.3 Research ethics online In terms of selecting and analysing blogs, it should be emphasized that exploring data in increasingly complicated online contexts calls for a more nuanced approach to research ethics, going beyond the often blurry boundary between the public and the private. Just because something is online does not mean it should automatically be reproduced in research contexts without appropriate reflection, and, if applicable, seeking permission, in order to minimize risk of

24

CONTEMPORARY MEDIA STYLISTICS

harm to participants involved (Mackenzie, 2017a). It is also very plausible that differing voices online, even in the same immediate context (e.g. a particular thread on a parenting forum), may hold differing views on privacy, although we may be able to identify broad interactional norms which appear to be at play (Mackenzie, 2017a: 311). All of the blogs discussed in this chapter were freely available to any internet user (there was no requirement to sign up, register or pay for any of the content accessed). Not only were these posts publicly available but they also all had an identifiable, named author with some kind of internet profile; as noted above, they could be categorized as a microcelebrity. In this case, the bloggers appear to be ‘expecting to observed’ (Bates, 2015: 191; Eysenbach and Wyatt, 2002), and not always by their preferred target or ‘imagined’ audience (see Marwick and boyd, 2010 and Tagg et al., 2017 for more discussion of the audience in online communication). Some of the bloggers stated that excerpts could only be (re)produced with correct accreditation and links to the original provided; this has been done in-text and in the Appendix for all blogs. 2.4 Approaches to the data CMT (Lakoff and Johnson, 1980/2003) is the main analytical tool used in this study. The small-scale qualitative study discussed in this chapter can be seen as metaphor-led discourse analysis which explores how metaphor is used in different contexts of both language use and cognitive activity (Pihlaja, 2013; Cameron et al., 2009). A basic tenet of CMT is that metaphor is viewed as a mapping between two different domains: the more understandable, concrete domain (source) and the domain which is more abstract and difficult to comprehend (target). CMT also holds that metaphor is a way for us to think about key concepts (cognition) and not just a property of canonical literary texts (Lakoff and Johnson, 1980/2003). For identifying and analysing the metaphors in my blog data, I use a similar procedure to Knapton’s (2013) approach to metaphors found in pro-anorexia blogs. Her context is different to modest fashion blogs, but pro-anorexia blogs are another online genre in which ideas about the female body are constructed and circulated. Her methodology is based on Charteris-Black’s (2004) procedure for metaphor identification, which she argues is especially useful for non-literary metaphors. I have adapted Knapton’s (2013) process below for my smaller scale project (some of the steps were applicable to her larger corpus of data) and incorporated metaphor identification process (MIP) for the second step in the process: 1. Close reading of data; 2. Marked and identified metaphorical expressions using MIP from the Pragglejaz Group 2007 (an approach to help to systematically identify

‘BEAUTIFUL MASTERPIECES’

25

metaphors in which ‘basic’ versus ‘contextual’ meanings of terms are identified); 3. Metaphorical expressions grouped together according to their conceptual metaphors. For the purposes of this chapter, I focused on identifying the target domains for [MODEST] FASHION and the [FEMALE] BODY (see also Nuttall and Harrison, this volume). CMT is not without its critiques, one of which is that practitioners need to be more explicit about how they look for metaphors in their data (see Gibbs, 2009). While there is undoubtedly an element of subjectivity in the analysis I have conducted, the use of the process outlined above is an attempt to provide more systematicity and transparency in terms of how the analysis was carried out. The findings discussed in the following section identify features of my data which suggests metaphors have been employed in order to describe modest fashion, women’s bodies and related concepts. Of course, ‘no single theory may be capable of explaining all aspects of the complex phenomena that are metaphorical language and thought’ (Gibbs, 2009: 32, original emphasis), but the following analysis explores how metaphors are conceptualized and textually expressed in the online community context of modest fashion blogs. My findings are discussed in light of Charteris-Black’s (2004) work on biblical metaphor. Following Knapton (2013), quantification of the conceptual metaphors has not been undertaken in this study because a qualitative approach was deemed more appropriate for this small dataset (although there was a minimum threshold established as a starting point; see comments in the following section). The findings and discussion are not necessarily focused on the novelty of these metaphors but rather how familiar metaphors are refashioned for religious purposes.

3 FINDINGS The following sections discuss common metaphorical expressions found in the sample of blogs analysed. The discourse of modest fashion blogs demonstrates a merging of what are commonly viewed as ‘secular’ and ‘religious’ domains (Cameron, 2013) – although of course this common distinction is somewhat of a simplification of a complicated reality. As such, the metaphors used tend to be conventional metaphors (commonly used and understood in daily life, according to Lakoff and Johnson, 1980/2003), repurposed to apply to modest fashion and often echo those found in more canonical religious texts (cf. Charteris-Black, 2004). The genre of religious modest fashion blogs provides a space for women to articulate their beliefs, often through the use of these metaphors. Indeed, metaphors in religious discourse play an evangelical role as they make abstract concepts both more

26

CONTEMPORARY MEDIA STYLISTICS

meaningful and potentially easier to understand (see Charteris-Black, 2004: 175–6). The following subsections give some examples of metaphorical expressions found in the data under analysis, but it should be noted that this is not an exhaustive list. Following the procedure outlined in the methods section, the metaphors used to describe (women’s) bodies and (modest) fashion that were identified included MODEST FASHION IS A JOURNEY [TO A DESTINATION OF SPIRITUAL FULFILMENT], LIFE IS A WORK OF ART, and PEOPLE [WOMEN] ARE THINGS/OBJECTS OR COMMODITIES. These examples appeared a minimum of three times in the data analysed. 3.1

MODEST FASHION IS A JOURNEY [TO A DESTINATION OF SPIRITUAL FULFILMENT]

In the metaphor MODEST FASHION IS A JOURNEY [TO A DESTINATION OF SPIRITUAL FULFILMENT], features from the source domain of JOURNEY are mapped onto the target domain of DRESSING MODESTLY, as can be seen in the following examples: 1. So here’s to all you #skirtsgirls, THANK YOU for taking this journey with me and to continue to support me! [Modestly Hot] 2. I share here not as one who has arrived, but as one who wants to share her joy in the journey so that you might be encouraged. [Fresh Modesty] 3. [In sharing your story of modest dressing] touch a few lives along the way. [Modest Goddess] The source domain of JOURNEY is a conventional (i.e. frequently used as part of everyday language) metaphor, predominantly used to talk about life and relationships in many contexts (Lakoff and Johnson, 1980/2003). Journey metaphors are also prevalent in the Bible and other forms of Christian literature (Charteris-Black, 2004: 204; see also Koller, 2017: 10 on Quaker pamphlets). Although the journey metaphors used in the blogs analysed tend not to specify a particular type of journey (such as car, train or sea; see Lakoff and Johnson, 1980/2003: 44–5), there is a shared sense of travelling towards a destination; in these cases, the destination is SPIRITUAL FULFILMENT or IMPROVEMENT of some sort. As bloggers ‘travel’ towards modest fashion, they become more knowledgeable about and adept at modest dressing and thus improve their spiritual character and obedience in this area of their life. LIFE IS A JOURNEY is a popular conceptual metaphor in both political and biblical discourses, in addition TO PURPOSEFUL ACTIVITY IS TRAVELLING ALONG A PATH TOWARDS A DESTINATION (Charteris-Black, 2004). The destination, in religious contexts, tends to be spiritual progress (Charteris-Black, 2004: 205–6), encompassing the following formulae: SPIRITUAL LIFE IS A JOURNEY; SPIRITUAL ACTIVITY IS TRAVELLING ALONG A PATH TOWARDS A GOAL; and, more broadly, SPIRITUAL LIFE IS A STRUGGLE FOR SALVATION. In terms of struggle in particular, there may be hazards or difficulties on the journey towards spiritual goals in one’s life:

‘BEAUTIFUL MASTERPIECES’

27

[A]s in journeys we can encounter dangers and, at times, may stumble or get lost, so in the religious life, individuals may be tempted or choose not to ‘walk’ in the right direction but to diverge from the ‘straight and narrow path’ [...] knowledge of the types of hazard that we may encounter while travelling – such as getting lost and falling down – provide a metaphorical source for the sort of difficulties faced by those on a spiritual journey (Charteris-Black, 2004: 206). These obstacles are not always negative as they may build spiritual character and mean that effort is needed to attain goals, which is underpinned by a common conceptual metaphor formula in religious discourse: EFFORT MADE TOWARDS A SPIRITUAL GOAL IS WORTHWHILE (Charteris-Black, 2004: 206–8). Therefore, the modest fashion bloggers emphasize that spending time and effort on modest fashion is needed, and that they are always learning how to improve. Indeed, implying that one has already achieved their goals by arriving at their destination might be seen as boastful and not in keeping with Evangelical Christian and Latter-day Saint views that humans are imperfect and always striving to be more like Jesus Christ. This idea of continual learning and improving is explicitly referenced in the second example in which the blogger states she ‘has not arrived’ at her destination, but her ‘journey’ nonetheless provides ‘joy’. The idea that the journey is shared is key for these modest fashion bloggers, as their online communities provide opportunities for encouragement and support (e.g. reader comments), which reinforces Cameron’s (2013: 35) observations that reader response tends to be positive on many modest fashion blog posts. 3.2

LIFE IS A WORK OF ART

Another metaphor found in the blogs can be best described with the conceptual key, that is, the higher level metaphor which ‘houses’ other related metaphor formulae (see Charteris-Black, 2004: 16), LIFE IS A WORK OF ART, which then encompasses the following conceptual metaphors: GOD IS AN ARTIST/DESIGNER and WOMEN ARE DRAWINGS/CREATIONS. This can be seen in the following examples which emphasize how women have been created and designed by God and therefore have an inherent worth in their design: 1. A truly liberated female is one who knows her worth in God’s eyes and views her body as a handcrafted masterpiece designed by God for His glory. [Girl Defined] 2. Girls are [not objects to be consumed, but] are precious human beings designed in the image of their loving Creator. [Girl Defined] 3. God designed you as a beautiful masterpiece and His plans for you are far greater than anything our raunch culture has to offer. [Girl Defined]

28

CONTEMPORARY MEDIA STYLISTICS

4. God calls you, ‘His workmanship …’ (Ephesians 2:10) and says you are ‘… fearfully and wonderfully made’ (Psalm 139: 14). Those are words you will never hear from mainstream society. [Girl Defined] 5. We dove into God’s word to study what He had to say about our design as women. [Girl Defined] 6. God himself created and prepared your body especially for you! [Modest Goddess] 7. Our sinful world has taken what God designed to be sacred and precious [female bodies; female sexuality] and turned it into something common and cheap. [Girl Defined] 8. I write about the importance of modesty because I believe that one of the ways we show respect for ourselves and our Father in Heaven who made us, [sic] is by clothing our bodies both respectfully and beautifully. [Modest Goddess] The examples above point to a gendered design for men and women, as cis, binary constructions, with an emphasis here on how women are specifically created. These examples emphasize a difference between men and women in terms of their creation by a divine being. These metaphorical expressions help to highlight the role modest dressing has in both reflecting and respecting this divine creation, with the WORK OF ART metaphor particularly highlighting the apparent beauty and intentionality of this design (e.g. beautiful masterpiece; fearfully and wonderfully made; and handcrafted masterpiece). The bloggers of Girl Defined elaborate on this idea of gender-specific design in one of their books (a spin-off venture from their blog): Instead of promoting our beautiful and unique design as women, we’re told it’s not really that special. We’re really no different from men. And that mind-set offends the two of us. As women, we’re proud of who God made us to be and don’t appreciate being forced into a neutral mould. [...] We know God designed females and males to be equally valuable in worth but purposefully different in roles. (Clark and Baird, 2016: 66) In the quotation above, the authors suggest there exists a clear distinction between secular and sacred approaches to male and female bodies. This distinction maps onto an idea found particularly within some strands of Evangelical Christianity: the idea that Christians should be ‘counter cultural’ by differentiating themselves from prevailing societal views and practices which may not be compatible with those of their faith, often supported with biblical references such as 2 Corinthians 6:14; Romans 12:2; 1 John 2:15-17; 1 John 2:15; and John 17:14-15. Modest dressing, therefore, may be a form

‘BEAUTIFUL MASTERPIECES’

29

of ‘counter cultural’ action, as some women use modest fashion to distinguish themselves from the ‘outside world’ (see, for example, Cameron, 2013: 27). However, the scared/secular distinction is not always so clear cut, as Fader (2009) highlights in her study of how Hasidic Jewish women engage with the secular world to achieve spiritual goals. However, modest fashion can be worn by women who are not religious, and the increasing availability of modest fashion retailers online could potentially be augmenting its popularity among non-religious groups (Lewis, 2013), which does complicate the ‘counter cultural’ aspect: if more women are wearing modest-style clothes for fashion reasons alone, how can religious modest dressers identify themselves as such? It is worth pointing out that, unlike, for example, Muslim dressers who choose to wear hijab, it is not always obvious that all Evangelical Christian and Latter-day Saint modest dressers are, first, dressing modestly (sometimes their outfits are those that are popular within contemporary North American fashion) and, second, that they may be doing so for religious reasons. Nonetheless, even if not immediately visible from looking at some of the outfits, the metaphors of creation and design used in these blogs stress this rationale for dressing modestly: to appropriately cover what these women believe to be God-given characteristics (i.e. the design of the female body). 3.3

PEOPLE [WOMEN] ARE THINGS/OBJECTS OR COMMODITIES

Women are described in these blogs as things, objects or commodities, with the following conceptual metaphors: PEOPLE [WOMEN] ARE THINGS/OBJECTS and PEOPLE [WOMEN] ARE COMMODITIES. The references to consumption can also lead us to the more specific conceptual metaphor formula of WOMEN ARE FOOD. Some examples from the blogs are provided below: 1. However, I believe that by dressing modestly, you are having the world see you as a person, not as property or as a ‘piece of meat’. [Modest Style] 2. I hated being viewed as an object for others to consume. [Girl Defined] 3. Girls are not objects to be consumed, but are precious human beings designed in the image of their loving Creator. [Girl Defined] Significantly, when this conceptual metaphor formula is used in the blogs, it tends to be negated, with the argument being that it is society that views women in this objectifying way while God and/or certain faiths do not, which fits into discussions in the previous section of ‘counter cultural’ beliefs and practices. These metaphorical expressions emphasize the objectification of women as things for consumption is in fact contrary to the beliefs of religious faiths. The HUMAN IS FOOD metaphor (Lakoff, 1987: 409) is a conventional

30

CONTEMPORARY MEDIA STYLISTICS

one, especially WOMEN ARE FOOD (Goatly, 2007: 89–90) which tends to be used to reference women’s sexual desirability and/or appearance (sweetie, tart, etc.), often in relation to satisfying men’s appetites. Here, the WOMEN ARE FOOD metaphors are used by these bloggers to argue that society’s views on women are objectifying and disrespectful. This aspect can be seen most clearly in the ‘piece of meat’ example cited above, in which the bloggers’ views seem to echo those of 1 Samuel 16:7 in the Old Testament (NIV): ‘The Lord does not look at the things people look at. People look at the outward appearance, but the Lord looks at the heart.’ The WOMEN ARE OBJECTS metaphorical expressions are negated to suggest God’s emphasis on anti-commodification of women’s bodies, which is seen as running counter to current mainstream North American society. Additionally, it is worth nothing that some of the examples from both LIFE IS A WORK OF ART and WOMEN ARE FOOD could suggest a further conceptual metaphor of BODIES ARE MONEY. Society is seen to be ‘cheapening’ women’s ‘value’ and ‘worth’ through not appropriately viewing or dressing female bodies; again this highlights the ‘counter cultural’ element. In the final example of this section, there is a constructed opposition (cf. Jeffries, 2010) drawn between ‘objects to be consumed’ and ‘precious human beings’, with the latter therefore presented as not objectifying – although of course ‘precious’ might be objectifying in other contexts, if used as an adjective to describe gemstones, for example. Here, ‘precious’ is used to underline God’s design for women: as something which has inherent worth and value, and as something which contradicts the prevailing cultural norms.

4 DISCUSSION AND CONCLUSION The previous sections explored the results of a small-scale qualitative study which identified some key metaphorical expressions in Latter-day Saint and Evangelical Christian modest fashion blogs. These included MODEST FASHION IS A JOURNEY [TO A DESTINATION OF SPIRITUAL FULFILMENT], LIFE IS A WORK OF ART and PEOPLE [WOMEN] ARE THINGS/OBJECTS OR COMMODITIES. It has been argued that metaphors are used in these blogs to emphasize that the differences between women and men are sacrosanct, and to suggest faith-based views of women’s bodies differ substantively from those held by society in a North American context. Future directions for this research topic would involve expanding the data beyond the bio sections to blog posts proper, to see if the common metaphorical expressions found here are borne out in a larger collection of data. The blogs explored in this chapter suggested similar, shared metaphor usage in Evangelical Christian and Latter-day Saint blogs, but a larger collection

‘BEAUTIFUL MASTERPIECES’

31

of blog data would facilitate specific comparisons and contrasts to be drawn between these two faith groups. In addition, analysis of North American Jewish modest fashion blogs would also provide an interesting point of reference, and would ascertain if metaphors of modesty are consistent across the Jewish and Christian traditions. Additionally, it should be made clear that not all Latter-day Saint and Evangelical Christian women dress modestly; women of faith(s) are not one homogenous group. Many will disagree with or challenge some of the blog content discussed in this chapter. However, the concept of modest fashion is one which does appear to be significant for many Evangelical Christian and Latter-day Saint women, and the analysis so far suggests that metaphors are a way for bloggers to express their views on this topic in an online context. The medium of a blog does encourage a certain form of self-presentation, both textually and visually, on pertinent subjects which may hold personal significance for the blogger. Modesty is one part of a much bigger gendered picture for the bloggers discussed in this chapter. These women situate themselves in faiths which emphasize the importance of different roles for men and women within familial and societal contexts. As Clark and Baird, from the Girl Defined blog, argue in their 2016 book, ‘The results of throwing out God-defined gender roles have been devastating to our society. Men don’t know how to be men anymore. Women don’t know how to be women. Marriages are in shambles. Relationships are spiralling downward. Homes are more like hotels’ (2016: 67). For religious modest fashion bloggers, educating themselves and others on how to best comply with modest dressing fits into an overall goal of improvement of self on one’s spiritual journey.

APPENDIX: LIST OF BLOGS ANALYSED Courtney Toliver http://www.courtneytoliver.com/ Modestly Hot https://modestlyhot.wordpress.com/ Girl Defined https://www.girldefined.com/ Modest Style http://modest-style.com/ Modest Goddess http://www.ourlovelydeseret.com/ Downtown Demure https://downtowndemure.com/ Hello Modesty http://hellomodesty.com/ Fresh Modesty http://freshmodesty.blogspot.com/ The Boyer Sisters http://boyersisters.com/ Bramblewood Fashion http://www.bramblewoodfashion.com/ Modest Blondie http://www.modestblondie.com/ Chez Heidi http://www.chezheidi.com/

32

CONTEMPORARY MEDIA STYLISTICS

NOTES 1. The Church of Jesus Christ of Latter-day Saints have official discouraged the use of ‘Mormon’ and ‘Mormonism’; see: https://www.churchofjesuschrist.org/church/ne ws/mormon-is-out-church-releases-statement-on-how-to-refer-to-the-organization?l ang=eng 2. It is beyond the scope of this chapter to provide a comprehensive overview of the main theological differences between the Church of Jesus Christ of Latter-day Saints and (varying strands of) Evangelical Christianity (but for an introduction, see Davies, 2003: 5–6, 144–77, 240–4). The term ‘Evangelical’ is used in this chapter to denote a particular brand of North American Christianity which focuses on biblical scripture, salvation and revivalism (cf. Bowler, 2019: xv). 3. Davies (2003: 211–15) explains the significance of endowments as follows: ‘Nowhere do LDS views of divine wonders, eternal states and ritual interconnect more than in endowment rites, rites that explain, prepare for, and confer upon select members both an identity as divine individuals set within eternal relationships and the capacity to conquer death. [...] Having been sealed for time and eternity, a married couple who meet the moral, doctrinal and economic demands of the Church may take their endowments. [...] Various parts of the body are set apart for their service to God prior to the individual being vested with the temple garment, a form of undergarment covering chest and lower body.’ 4. This rationale feeds into a wider ‘purity culture’, commonly found in North American Evangelical Christian circles, in which pre- and extramarital sex is strongly discouraged, and the onus is generally on women to avoid dressing in ways which may be considered ‘provocative’ to men. For a contemporary feminist critique of purity culture see Valenti (2009). 5. Fresh Modesty has recently moved over to a new blog platform at freshmodesty. com, but her new site has, so far, not been updated frequently and so her original site (see Appendix) has been selected for analysis in this chapter. 6. Bramblewood Fashion has now rebranded as Plaid Lattes and has more of a beauty focus; for the purposes of this chapter I therefore included the bio section from the original website. 7. See https://www.girldefined.com/modesty-dignity-over-sexual-compromise

REFERENCES Bates, C. F. (2015). ‘“I am a waste of breath, of space, of time”: Metaphors of self in a pro-anorexia group’. Qualitative Health Research, 25(2), 189–204. Beaver, J. and Beaver, H. (2010). Christian Modesty: A Matter of the Heart. Pleasant Prairie: Baptist Growth Publications. Benedict, H. (1992). Virgin or Vamp: How The Press Covers Sex Crimes. New York: Oxford University Press. Bowler, K. (2019). The Preacher’s Wife: The Precarious Power of Evangelical Women Celebrities. Princeton: Princeton University Press. Cameron, J. (2013). ‘Sartorially sacred or Fashion Faux Pas: Visual interpretations of modesty online’. In A. Day, G. Vincett and C. R. Cotter (Eds), Social Identities between the Sacred and the Secular (pp. 23–38). Farnham: Ashgate.

‘BEAUTIFUL MASTERPIECES’

33

Cameron, L., Maslen, R. and Todd, Z. (2009). ‘The discourse dynamics approach to metaphor and metaphor-led discourse analysis’. Metaphor and Symbol, 24(2), 63–89. Charteris-Black, J. (2004). Corpus Approaches to Critical Metaphor Analysis. Basingstoke: Palgrave Macmillan. Charteris-Black, J. (2014). Analysing Political Speeches: Rhetoric, Discourse and Metaphor. Basingstoke: Palgrave Macmillan. Clark, K. and Baird, B. (2016). Girl Defined: God’s Radical Design for Beauty, Femininity, and Identity. Grand Rapids: Baker Books. Davies, D. (2003). An Introduction to Mormonism. Cambridge: Cambridge University Press. Drury, C. (1994). ‘Christianity’. In J. Holm and J. Bowker (Eds), Women in Religion (pp. 30–58). London: Continuum. Eysenbach, G. and Wyatt, J. (2002). ‘Using the internet for surveys and health research’. Journal of Medical Internet Research, 4(2, e13.) Retrieved from http://www.jmir.org/2002/2/e13/ Fader, A. (2009). Mitzvah Girls: Bringing Up the Next Generation of Hasidic Jews in Brooklyn. Princeton: Princeton University Press. Franklin, R. (2006). Designed By God: Honest Talk about Beauty, Modesty and Self-Image. Grand Rapids: Discovery House. Gamson, J. (2011). ‘The unwatched life is not worth living: The elevation of the ordinary in celebrity culture’. MPLA, 126(4), 1061–9. Gibbs, R. (2009). ‘Why do some people dislike conceptual metaphor theory?’ Cognitive Semiotics, 5(1–2), 14–36. Goatly, A. (2007). Washing the Brain: Metaphor and Hidden Ideology. Amsterdam: John Benjamins. Haws, J. B. (2013). The Mormon Image in the American Mind: Fifty Years of Public Perception. New York: Oxford University Press. Hunter, A. (2016). ‘Monetizing the mommy: Mommy blogs and the audience commodity’. Information, Communication & Society, 19(9), 1306–20. Hunting, K. (2015). ‘Fashioning feminine fandom: Fashion blogging and the expression of mediated identity’. In E. Levine (Ed.), Cupcakes, Pinterest, and Ladyporn: Feminized Popular Culture in the Early Twenty-First Century (pp. 116–36). Illinois: University of Illinois Press. Jeffries, L. (2010). Opposition in Discourse: The Construction of Oppositional Meaning. London: Continuum. Jerslev, A. (2016). ‘In the time of the microcelebrity: Celebrification and the YouTuber Zoella’. International Journal of Communication, 10, 5233–51. Knapton, O. (2013). ‘Pro-anorexia: Extensions of ingrained concepts’. Discourse & Society, 24(4), 461–77. Koller, V. (2017). ‘The light within: Metaphor consistency in Quaker pamphlets, 1659–2010’. Metaphor and the Social World, 7(1), 5–25. Lakoff, G. (1987). Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. Lakoff, G. and Johnson, M. (1980). Metaphors We Live By (1st edn). Chicago: University of Chicago Press. Lakoff, G. and Johnson, M. (2003). Metaphors We Live By (2nd edn). Chicago: University of Chicago Press. Levine, E. (2015). ‘Introduction: Feminized popular culture in the early twenty-first century’. In E. Levine (Ed.), Cupcakes, Pinterest, and Ladyporn: Feminized Popular

34

CONTEMPORARY MEDIA STYLISTICS

Culture in the Early Twenty-First Century (pp. 1–12). Illinois: University of Illinois Press. Lewis, R. (Ed.) (2013). Modest Fashion: Styling Bodies, Mediating Faith. London and New York: I.B. Tauris. Lopez L. K. (2009). ‘The radical act of “mommy blogging”: Redefining motherhood through the blogosphere’. New Media & Society, 11(5), 729–47. Mackenzie, J. (2017a). ‘Identifying informational norms in Mumsnet Talk: A reflexivelinguistic approach to internet research ethics’. Applied Linguistics Review, 8(2–3), 293–314. Mackenzie, J. (2017b). ‘“Can we have a child exchange?” Constructing and subverting the “good mother” through play in Mumsnet Talk’. Discourse & Society, 28(3), 296–312. Marwick, A. (2013). Status Update: Celebrity, Publicity and Branding in the Social Media Age. New Haven, CT: Yale University Press. Marwick, A. and boyd, d. (2011). ‘I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience’. New Media and Society, 13, 96–113. Myers, G. (2010). The Discourse of Blogs and Wikis. London: Continuum. Neary, C. and Ringrow, H. (2018). ‘Media, power and representation’. In A. Hewings, P. Seargeant and S. Pihlaja (Eds), The Routledge Handbook of English Language Studies (pp. 294–309). London: Routledge. Peace, M. and Keller, K. (2015). Modesty: More than a Change of Clothes. Harmony, NJ: P and R Publishing Company. Pihlaja, S. (2013). ‘“It’s all red ink”: The interpretation of biblical metaphor among Evangelical Christian YouTube users’. Language and Literature, 22(2), 103–17. Pollard, J. (2004). Christian Modesty and the Public Undressing of America. San Antonio: Vision Forum. Semino, E. (2008). Metaphor in Discourse. Cambridge: Cambridge University Press. Steen, G. J. (2010). ‘Researching and applying metaphor’. Toegepaste Taalwetenschap in Artikelen (TTwiA), 83, 91–102. Tagg, C., Seargeant, P. and Brown, A. A. (2017). Taking Offence on Social Media: Conviviality and Communication on Facebook. London: Springer. Tanz, J. (2008). ‘Internet famous: Julia Allison and the secrets of self-promotion’. Wired. Retrieved from http://www.wired.com/2008/07/howto-allison/ Valenti, J. (2009). The Purity Myth: How America’s Obsession with Virginity Is Hurting Young Women. Berkeley: Seal Press. Vaughan, D. and Vaughan, D. (2005). The Beauty of Modesty: Cultivating Virtue in the Face of a Vulgar Culture. Nashville: Cumberland House Publishing. Wilson, J. and Yochim, E. C. (2015) ‘Pinning happiness: Affect, social media, and the work of mothers’. In E. Levine (Ed.), Cupcakes, Pinterest, and Ladyporn: Feminized Popular Culture in the Early Twenty-First Century (pp. 232–48). Illinois: University of Illinois Press.

CHAPTER THREE

Wolfing down the Twilight series: Metaphors for reading in online reviews LOUISE NUTTALL AND CHLOE HARRISON

The development of social media platforms devoted to the discussion of books provides a source of insights into how readers interact with texts in their daily lives and, as such, offers a growing source of data for stylistics. Popular fiction such as Stephenie Meyer’s Twilight series (2005–2008) attracts thousands (in some cases millions) of ratings and reviews by readers, which are often highly polarized. Recent work in stylistics has used such data as a source of insights into felt, experiential aspects of reading, applying the same stylistic frameworks to the reviews as those applied to the texts themselves (e.g. Harrison, 2017; Nuttall, 2017). In this chapter, we analyse the range of metaphors used by readers to describe contrasting experiences of Twilight (Book One), and the embodied experiences which contribute to both its popularity and rejection among readers. Drawing on cognitive metaphor theory (Lakoff and Johnson, 1980, 1999), previous research has identified three main conceptual metaphors as reflecting readers’ engagement with texts: READING IS TRANSPORTATION, READING IS CONTROL and READING IS INVESTMENT (Stockwell, 2009). Here, we test and develop these observations by examining a sample of 200 reader reviews collected from the online forum Goodreads. Comprising 100 of the most positive (5-star) and most negative (1-star) reviews of Twilight, these responses to the text are submitted

36

CONTEMPORARY MEDIA STYLISTICS

to qualitative analysis using NVivo software, and metaphors for reading are grouped and analysed using concepts from cognitive linguistics. Applying cognitive grammar (Langacker, 2008), we explore the creativity with which readers extend, combine and elaborate conventional metaphors for reading in this discourse context, and identify further recurring metaphors such as READING IS EATING for this text. Comparison of our positive and negative reviews reveals differences in the mappings of these metaphors, specifically the framing of the reader, which reflects the varying quality of the embodied experiences being described. We argue that, when contextualized in relation to a particular work of fiction and a particular online discourse context, the language produced by readers can offer insights into polarized reading experiences such as immersion and resistance.

1 INTRODUCTION: THE LANGUAGE OF ONLINE REVIEW FORUMS Founded in 2007, Goodreads has become the largest reading community in the world. At the time of writing (May 2017), the website reports 65 million members, 68 million reviews and 2 billion books that have been added to the virtual bookshelves. The site is itself ‘a social network for book reviews’ (Fay, 2012), where members are given the opportunity to fill bookshelves with books they have read or are yet to read, and which they recommend, rate and review for fellow readers. For stylisticians, the book reviews found on this site reflect reading in a non-academic and non-professional context, since the website is based on ‘a book-club model rather than a journalistic one’ (Fay, 2012; see also Gavins, 2013: 8). This is a point expounded by co-founder Otis Chandler in the opening message on the site, where he identifies his motivation for developing the platform: ‘When I want to know what books to read, I’d rather turn to a friend than a random person or bestseller list.’ The rise in popularity of online book reviews – and of Goodreads as a platform for reviewing books – reflects changing contemporary reading practices. With advances in digital technologies, not only do we discover, read and access texts in different ways, but the environments in which we discuss them are also changing (Allington and Pihlaja, 2016). In her online article in The Atlantic that advocates Goodreads as a platform for reviving and renovating the format of the book review, Fay (2012) argues, ‘For now, Goodreads is basically Facebook with books, but if enough contributors set the bar high with creative, funny, and smart reviews it might become a force of its own.’ In short, she suggests that the specific discourse environment offered by Goodreads and other book recommendation services ‘will alter journalistic literary criticism’ (Fay, 2012). Significantly, these changing practices also mean that different types of reading experiences are emerging, as the distinction between ‘solitary’ and ‘social’

WOLFING DOWN THE TWILIGHT SERIES

37

readings (Peplow et al., 2016) becomes increasingly blurred. This pluralism can be seen, for example, in the title of another recent article about Goodreads: ‘Millions of People Reading Alone, Together’ (Narula, 2014). The nature of such reading experiences and the effects of this new online reading environment are topics for cognitive stylistics. While ‘reading’ has been studied from a range of perspectives in cultural studies, sociology, anthropology and book history (see Allington and Pihlaja, 2016, for a review), cognitive stylistics is specifically concerned with the felt, experiential quality of reading, or ‘texture’ (Stockwell, 2009). Informed by theories of embodiment in the cognitive sciences, cognitive stylistic research seeks explanations of reading phenomena such as immersion and empathy, along with the ways in which these experiences may vary, or be actively resisted by readers (e.g. Stockwell, 2005; Gavins, 2007; Whiteley, 2014; Nuttall, 2018). A recent development in (cognitive) stylistics has seen these objectives combined with the methodologies observed in reader response research. ‘Naturalistic’ approaches to the study of reader responses (Swann and Allington, 2009), or analyses of the uncontrolled responses of readers in everyday reading environments such as book clubs, classrooms or online forums, have been seen to offer a source of insights into reading experiences (Whiteley, 2011; Gregoriou, 2012; Gavins, 2013; Harrison, 2017; Nuttall, 2017). Readers’ discursive choices in describing their thoughts and feelings towards a text have been said to offer a useful basis for modelling the processing which underlies them (Stockwell, 2009: 79; Allington, 2011: 318). However, the extent to which such responses offer access to the online reading experiences of readers is mediated by the other communicative functions at work in these contexts, which may include a desire for social affirmation (or the opposite – a desire to provoke conflict) and the construction of identity within a social hierarchy (Swann and Allington, 2009; Peplow, 2011; Gregoriou, 2012). In the case of online forums such as Goodreads, the ‘affinity spaces’ (Gee, 2004; Curwood, 2013; Vlieghe, Muls and Rutten, 2016) created by reviewers based on a single shared interest (a book) can be seen to form a specific context which motivates, and is influenced by, the language of its participants. In this chapter, we explore the nature of the reading experiences observed in this context from a cognitive stylistic perspective. Combining concepts from cognitive linguistics with a naturalistic reader response method, we compare the language which characterizes the reviews of one book on Goodreads in terms of metaphor. We investigate the insights offered by this data with regard to readers’ varied embodied experiences of the text, as well as the different communicative, interpersonal functions which co-exist on this social media network. In doing so, we aim to uncover some of the affordances of this new media environment for cognitive stylistic understandings of immersion and resistance, and the nature of ‘reading’ more broadly.

38

CONTEMPORARY MEDIA STYLISTICS

2 METAPHORS FOR READING Cognitive stylistic studies have identified a number of metaphors on which we draw to describe our engagement with fictional worlds. Developing work by Gerrig (1993), Stockwell (2009: 80) states that any scan of the online reading group data that is currently available reveals very evidently that there are three main organising discourse metaphors being used: ●

reading as transportation

●

reading as control

●

reading as investment.

Stockwell notes that the first of these, the transportation metaphor, is the most popular, giving examples such as ‘We follow the boy on his journey around the world’ and ‘There are quite a few battle scenes which I found slightly heavy going’. Notably, Stockwell’s examples include both positive and negative evaluations of a reading experience, along with those that combine the two: ‘It takes a while to get into but eventually you get taken along with the action.’ Further, he goes on to observe that, of the three metaphors, only ‘reading as transportation’ and ‘reading as control’ are ‘bidirectional’: ‘in both cases, the reader is either the one controlling or the one actively doing the travelling, or alternatively is the one controlled or the one who is carried off on the journey’ (Stockwell, 2009: 81). These observations raise two questions: (1) In what different forms are these underlying metaphors expressed by readers? and (2) What kinds of experiences, attitudes and feelings are they used to describe? With regard to the second question, such ‘bidirectional’ metaphorical expressions, and the active or passive role in which they cast the reader, might be seen to reflect reader experiences of being ‘immersed’ in a narrative (Gerrig, 1993; Stockwell, 2009). In cognitive stylistics, the process of immersion in a fictional world is modelled as a process of ‘recentering’ (Ryan, 2001), ‘deictic shift’ (Galbraith, 1995; Gavins, 2007) or ‘psychological projection’ of the self into a new spatio-temporal context (Whiteley, 2011), which is facilitated by the language of a text. An open question for such accounts concerns the extent to which the reader is an active or passive participant in this process. The active reader suggested (implicitly or explicitly) by such worlds-based models reflects Gerrig’s (1993: 185) rejection of Coleridge’s notion of a passive ‘suspension of disbelief ’ in favour of a strategic ‘willing construction of disbelief ’, or what he describes as an active ‘performance’ of the text. On the other hand, work by Stockwell (2005, 2013) describes the ‘positioning’ of a reader by a text in an interaction between ‘readerly disposition and textual imposition’ (2013: 263), or text-driven experiences of immersion as ones that readers can choose

WOLFING DOWN THE TWILIGHT SERIES

39

to engage with or resist. While seeking to test and specify Stockwell’s (2009) observations in a general sense, therefore, this chapter also aims specifically to explore the ways in which experiences of immersion (both positive and negative/successful and unsuccessful) are represented by readers themselves, and the active or passive role they take in this process. The framework that we will use to address these questions is cognitive (or conceptual) metaphor theory (CMT). Originating in the work of Lakoff and Johnson (1980, 1999), CMT’s core premise is that the metaphorical expressions we encounter in discourse reflect and shape entrenched ways of thinking about abstract concepts in terms of other more concrete, and hence readily comprehensible, concepts. In this account, the ‘reading as transportation’ metaphor reflects a conceptual metaphor READING IS A JOURNEY, whereby a target domain (READING) is understood through a mapping of content from a source domain (JOURNEY) containing our knowledge of this familiar aspect of bodily experience. Common source domains such as JOURNEY are understood to have an embodied basis in our shared experience as humans, which makes them suitable means through which to understand and communicate complex or subtle experiences such as reading fiction (compare Gibbs, 2005; Frank, Ziemke and Zlatev, 2008). In this sense, CMT offers one explanation for the pervasiveness of metaphor in interpersonal contexts like Goodreads, in which reviewers seek to share psychological, emotional and ethical responses to texts, and to do so in terms accessible to others. However, CMT has faced a number of criticisms, many of which are particularly relevant to stylisticians. First, the very notion of a generic ‘embodiment’ shared by all human beings appears to be at odds with the attention to individual or culturally distinct readings which underpins (cognitive) stylistics, and the creative ability of writers to communicate extraordinary or idiosyncratic experiences through language (see, for example, Senkbeil and Hoppe, 2016). Responding to the wider issue of ‘embodiment’ as defined in cognitive linguistics, Kövecses (2008, pp. 177–8) notes that different aspects of our universal embodiment may be focused upon by particular cultures (and by extension, perhaps, individuals) in what he terms ‘differential experiential focus’ (see also source domains as ‘view-pointed’ in Dancygier and Sweetser, 2014: 2). Related to this issue, Kövecses (2008: 180) also observes that, in addition to embodiment, metaphor use is often influenced by context-specific motivations, such as ‘the setting and topic of the situation in which the metaphorical conceptualization takes place’. The selection of source domains in discourse about reading, therefore, might be seen to reflect not only the embodied experiences which underpin it but also the topic (the specific book in question), the setting (the online forum), and the personal interests and motivations of the reviewers themselves. Another significant criticism of CMT is its tendency to pay most attention to the underlying conceptual metaphors at the expense of their varied linguistic

40

CONTEMPORARY MEDIA STYLISTICS

manifestations in discourse, and their development through interaction (Pragglejaz Group, 2007; compare Kövecses [2002] on ‘subindividual’ vs. ‘individual’ levels of analysis; see also Cameron, 2003). From a stylistic perspective, neither CMT nor the related theory of conceptual integration (Fauconnier and Turner, 2002) fully account for the contextualized effects of the specific language choices made by speakers/writers for conceptualization, or ‘the linguistic dimension of creativity’ (Semino, 2008: 52; also Browse, 2014). Such variation in metaphor use is often discussed in terms of ‘framing’ in subsequent metaphor research, defined by Semino et al. (2016: 1) as the ability of metaphors to ‘express, reflect, and reinforce different ways of making sense of particular aspects of our lives’. Combining cognitive and discoursebased approaches, Semino et al.’s (2016) account of framing includes the specific entities, roles and relations that are focused in the source domain, or specific ‘scenario’ (Musolff, 2006) mapped by a conceptual metaphor, as well as the relative ‘agency, (dis)empowerment, evaluations, and emotions’ (2016: 18) which arise from the discursive use of this metaphor in context. The framing of metaphors can show distinctive patterns, and deviations from patterns, in how particular topics are conceptualized and linguistically expressed. Senkbeil and Hoppe (2016), for example, consider the macrometaphors present in Hornbacher’s memoir Wasted, which is a personal account of the author’s experience with anorexia and bulimia. This study discusses the conventional metaphor DEALING WITH A DISEASE IS WAR but notes how, in this memoir, the agentive role of the writer in this war scenario is altered: individuals with anorexia/bulimia are seen to fight for the illness rather than against it. This can be seen, for example, in the sentence ‘you’re a friggin’ unstoppable dieting army and you’ll all go down together’ (Senkbeil and Hoppe, 2016: 8). One means of analysing such localized framings, set out more fully in Browse (2014) and Nuttall (2018), is to enrich CMT with concepts drawn from cognitive grammar (Langacker, 2008). Viewed in this light, metaphors can be regarded simply as instances of the wider processes of ‘construal’ that underpin all linguistic expressions (see Langacker, 2008: 55–85). In this framework, construal describes our ability to linguistically portray – and so conceptualize – the same experience in alternate ways. Like all construals, metaphorical expressions cue a ‘matrix’ of domains in memory (in the case of metaphors, a source and target domain), content from which may be variably ‘focused’ in attention, conceived at different levels of ‘specificity’ and allocated different amounts of ‘prominence’. Further, the specific participants (e.g. anorexia/bulimia sufferers) conceptualized within such domains can be analysed in terms of their variable role within an ‘action chain’ (Langacker, 2008: 355), or a force-dynamic interaction between an active participant (the

WOLFING DOWN THE TWILIGHT SERIES

41

AGENT,

or ‘energy source’) and a passive participant (the PATIENT, or ‘energy sink’). Combined with discussion of the ‘mappings’ between domains as part of a CMT analysis, discussion of the nature of the resulting conceptualization in these terms allows us to explore the effects of specific language choices for the meaning of metaphors in context. In the analysis that follows, this model of ‘metaphorical construal’ (Nuttall, 2018) is used to explore the framings of reading and readers which characterize the metaphors in our reviews.

3 TWILIGHT: READERS AND RECEPTION Stephenie Meyer’s Twilight collection (2005–2008) is a bestselling young adult series, which was adapted into a series of films (2008–12) and which is identified as the source of inspiration for the hugely successful fan fiction series Fifty Shades (James, 2015–2018). This study examines reader reviews of the first book in the series, Twilight (2005). At the time of writing, this novel has received 96,914 reviews on Goodreads along with just over 4 million ratings. Of these ratings, 1,441,592 are 5-stars (the highest rating possible), 479,241 are 1-star (the lowest rating), with the remaining 2,229,974 reviewers giving the novel 2, 3 or 4 stars. With an average rating of 3.58 stars, the novel’s popularity on the site, together with its highly polarized responses, is reflective of its wider societal reception and influence. A cursory internet search reveals numerous fan sites devoted to the series, including www.twilightlexicon.com, en.twilightpoison.com and www.edwardsmeadow.com, visited by self-proclaimed ‘Twihards’ or members of ‘Team Edward’ or ‘Team Jacob’. Meanwhile, first-time author Meyer has since been named one of Time magazine’s most influential people (Time, 2008). Twilight (2005) is centred on protagonist Isabella (‘Bella’) Swan, who moves from Arizona to live with her father in Forks, Washington, where she meets and falls in love with a vampire, Edward Cullen. Much of the action of the novel takes place in the high school in which Bella and Edward meet and concerns their developing romance. Accident-prone Bella is saved from several dangerous situations by Edward, who simultaneously attempts to save her from himself and the danger he poses to her as a 100-year-old member of a secret clan of blood-thirsty (but ‘vegetarian’) vampires. Critics and readers alike have expressed surprise at its success, voicing concerns about the troubling messages it sends to girls and women about romance and gender norms (see Silver, 2010, for a review). For many, Edward’s controlling behaviour and Bella’s vulnerability promote outdated ideas about male dominance and female submission to a young audience who may lack the critical resources necessary to reject its underlying ideology. However, other researchers have argued against the view of (young) readers as passive recipients of ideologies, identifying a more active, critical engagement with

42

CONTEMPORARY MEDIA STYLISTICS

fiction which appropriates and negotiates it for their own purposes (Sarland, 2005). Citing evidence from fan sites, Facebook and classroom discussions, Silver (2010: 137) describes a ‘give and take in girls’ readings’, whereby they ‘allow themselves to swoon into Meyer’s fantasy of everlasting passion and devotion’ while simultaneously resisting engagement with Bella and specific aspects of her behaviour. Reflected in the Goodreads reviews of Twilight, these critical concerns are ones that have yet to be addressed from a cognitive stylistic perspective. In the remainder of this chapter, we outline an analytical approach that aims to offer insight into this ‘give and take’ during reading, or the experiences of immersion and resistance which underpin this novel’s success.

4 METHODOLOGY The top 100 community reviews for both the 1-star and 5-star ratings were collected from the Goodreads website, using the site’s default sorting algorithm for the ‘most interesting’ reviews.1 From the 5-star reviews, six were discarded (either because they written in a language other than English or due to inappropriate content2), and from the 1-star reviews four were discarded for the same reasons. In each case, the next listed reviews on the site were collected in their place. The resulting data sets of 100 reviews were then uploaded to NVivo Pro 11 and metaphors were coded following the metaphor identification procedure (MIP), as outlined by the Pragglejaz Group (2007). This procedure involves reading the whole text in order to ascertain a general understanding of the discourse and then marking up those lexical units that are being used metaphorically, by comparing the ‘basic’ (the historical or more concrete definition of a term) against the ‘contextual’ meaning (how the term functions in this discourse context) of the lexical item in question (see Pragglejaz, 2007, for a more detailed breakdown of the steps involved in this approach). For the purposes of our analysis, only those metaphors for which the target was the act of reading (or a closely related target: writing, reviewing or the book itself) were coded. The coded metaphorical expressions were grouped into ‘nodes’, which in this case reflected our sense of their underlying conceptual metaphor, taking into account their basic meaning and their function within the context of the review. The researchers coded the data sets for the 1-star and 5-star reviews separately, before swapping to check the other’s coding. Any instances in which the metaphor coded was unclear were highlighted and accepted only if inter-coder agreement was reached. At the end of this process, a total of 637 individual expressions for the 1-star reviews and 419 expressions for the 5-star data set were coded, respectively.3

WOLFING DOWN THE TWILIGHT SERIES

43

Taking into account the complex ethical considerations involved in research using online discourse as data (Spilioti and Tagg, 2017), and the potential vulnerability of reviewers of Twilight in terms of age, the extracted reviews (all in the public domain) have been anonymized, with each reviewer given an identifier (R1 or R5 according to the respective negative/positive data sets, followed by a number between 1 and 100) (see also Pihlaja, 2017, for discussion of the decision to acknowledge or anonymize in social media contexts). In the analysis which follows, all quotations from reviews are reproduced with original spelling, emphasis, grammar and punctuation.

5 ANALYSIS Table 3.1 shows the results of the initial coding, with each of the ‘nodes’ (or conceptual metaphors) ordered by frequency of coded instances in each data set (1-star, 5-star) and the total 200 reviews.4 Only the ten most frequent metaphors are recorded here, and those with other related target domains are not included in order to restrict our focus. Where appropriate, these related metaphors (e.g. WRITING IS REGURGITATING and BOOKS ARE EXCREMENT) will be discussed at relevant points in the discussion that follows. Comparison of these rankings to Stockwell’s (2009) predictions generates some interesting observations. Stockwell’s estimation that ‘reading as transportation’ is the most popular metaphor used by readers, making up approximately half of the metaphors used to describe reading, is borne out to some extent in our data, though not consistently. Expressions coded as READING IS A JOURNEY are prevalent and include ‘I loved everything about it and flew through it’ (R5.5) and ‘having to follow your narrator through 24-fucking-hours of a day …’ (R1.3), though notably these make up only a quarter of the total metaphorical descriptions of reading coded in our data. Further, instances of READING IS CONTROL such as ‘the language and writing keep the reader hooked’ (R5.4); ‘I’ll admit I was roped in by this book at first’ (R1.16); and READING IS INVESTMENT, for example, ‘it’s worth the time and energy, and the tug on your emotions’ (R5.14); ‘you totally ripped off your readers there’ (R1.2), lend support to Stockwell’s discussion of these pervasive metaphors for reading. However, his estimation that these make up, in equal parts, the other half of metaphors used to describe reading in such discourse is not supported. Significantly, a number of other conceptual metaphors were frequently observed across our data. READING IS INVESTMENT, described by Stockwell (2009: 80) as the most ‘overlooked’ metaphor used by readers, appears further down our lists than another metaphor, coded as READING IS EATING, and a range of other novel metaphors are observed besides. These summative findings indicate that the range of metaphors used

44

CONTEMPORARY MEDIA STYLISTICS

TABLE 3.1 Frequency of conceptual metaphor nodes coded in 1-star and 5-star reviews of Twilight on goodreads 1-Star Reviews

5-Star Reviews

READING IS A JOURNEY READING IS EATING

(64)

(51)

READING IS CONTROL

READING IS CONTROL

(30)

READING IS EATING

READING IS PHYSICAL

READING IS A JOURNEY

(57)

READING IS CONTROL

(26)

READING IS EATING

READING IS A (ROMANTIC) RELATIONSHIP

CONTACT

(63)

READING IS A JOURNEY

(46)

READING IS INVESTMENT

Total Reviews in Sample

READING IS SEEING

(13)

ILLNESS

(10)

READING IS INVESTMENT

(12)

READING IS A (ROMANTIC) RELATIONSHIP

(10)

READING IS A RELIGIOUS EXPERIENCE

READING IS PHYSICAL CONTACT

(34)

READING IS A (ROMANTIC)

COMPETITION

READING IS SEEING

READING IS A COMPETITION

(29)

READING IS A

(10)

READING IS MENTAL ILLNESS

COMPETITION

(3)

(13)

READING IS MENTAL ILLNESS

(10) READING IS A RELIGIOUS

(3)

(31)

(12)

(7)

READING IS A

(42)

READING IS PHYSICAL

RELATIONSHIP READING IS MENTAL

(77)

READING IS INVESTMENT

CONTACT

(16)

(109)

(21)

(22)

READING IS SEEING

(121)

EXPERIENCE

(2)

READING IS A RELIGIOUS EXPERIENCE

(9)

by readers to describe their experiences is broader and more complex than previously suggested, and that other pervasive metaphors may have been overlooked in stylistic research. However, based on subjective judgements of metaphorical groupings, these rankings are limited in what they can tell us about the use of metaphor in the reviews. More interesting patterns can be seen when the metaphorical expressions themselves are examined. As noted previously in Section 2, Stockwell (2009) described the variable ‘directionality’ of metaphors for reading, and their resulting representation of the reader as an active or passive participant in the events or actions conceptualized. Based on this observation, we hypothesize that the linguistic manifestations of conceptual metaphors will differ, depending on whether the review is describing a positive or negative reading experience. Comparing our 1-star and 5-star data sets, and drawing on the view of ‘framing’ discussed in Section 3, we expect to see variation in the ‘agency, (dis)empowerment, evaluations, and emotions’ (Semino et al., 2016: 18) encoded in readers’ use of these metaphors in context. In the following sections, we examine the three most pervasive conceptual metaphors in our data individually in these terms.

WOLFING DOWN THE TWILIGHT SERIES

45

5.1 Reading is a journey As suggested by Stockwell (2009), the linguistic expressions of the JOURNEY metaphor in our data vary widely in terms of their representation of the agency of the reader, the text and its writer in the experience of reading Twilight. In our data, such differences of framing can be seen in the various types of journey, or specific scenarios that act as the source domain in different conceptualizations of the reading experience. A recurring scenario seen in our data is that of DIVING. The following examples, taken from the 5-star reviews, all describe the experience of reading the novel using this metaphor (lexical items identified as metaphorical have been underlined): I enjoyed this Young Adult, Paranormal Romance and glad that I finally took the plunge. (R5.29) Accept the fact that you won’t get any sleep until you’ve finished this whole, thick book. Then dive in and enjoy! (R5.76) Whatever it may be, young adult novel, vampire novel, or romance novel, each reader will furrow into the depths of Meyer’s tight prose. (R5.82) I had read Like Water for Chocolate by Laura Esquivel because my mom had insisted I could get into it. I do hand it to her I finished and loved it and chatted with her endlessly about it. (R5.27) I haven’t stopped reading since and don’t expect to emerge from Twilightland till mid next week. (R5.95) Her story is one that any romantic girl who has loved or desires to love can fall into and be held to the pages as the story unfolds. (R5.32) From a cognitive grammatical perspective (see Section 2), this particular metaphorical construal focuses a specific set of participants and processes as part of the DIVING scenario it invites us to mentally represent. In the construal underpinning these expressions, the reader is the agentive participant in this action chain, focused in the role of energy source, which enters, unimpeded, into the text. This construal is represented in Figure 3.1. Notably, the readers’ behaviour in these examples can be intentional, as in ‘dive’, or unintentional, as in ‘fall’. Meanwhile, the text is construed as the PATIENT in this action chain – the passive recipient of the action (R5.82, 27), or the backgrounded location for the movement (R5.29, 76, 95, 32). As suggested by Stockwell (2009), examples in which the agency, or ‘directionality’, of this metaphor is reversed can also be seen in our data. Descriptions such as those below, again from our 5-star data set, can be seen to construe this same movement ‘into’ the text, but this time focusing the reader

46

CONTEMPORARY MEDIA STYLISTICS

FIGURE 3.1: Metaphorical construal 1 – READING IS A JOURNEY (READER AS AGENT).

FIGURE 3.2: Metaphorical construal 2 – READING IS A JOURNEY (TEXT AS AGENT).

as the passive participant (or PATIENT), who is affected by the actions of the text/its writer (the AGENT). This construal is represented in Figure 3.2. In these terms, then, Stockwell’s ‘bidirectionality’ can be understood as a reversal in the energy transfer between participants focused in the source domain. I was pulled into this love story immediately. (R5.55) I read Twilight a couple years ago 4 times I pretty much read it every year but I never put in my reviews until now and I really like the fact … Stephanie Myers sucks you into her books. (R5.67) I was happy to suspend belief and while intellectually I know that I should be more critical of it, on an emotional level I was just sucked in. (R5.96) Notably, this DIVING scenario appears to be largely constrained to the 5-star reviews in our data set. Instances of the first construal (which we sub-categorized in our data as READER AS AGENT) are absent entirely from the 1-star review data, while instances of the second construal (TEXT AS AGENT) are used only by reviewers here in a negated context (e.g. ‘I didn’t get so absorbed in it that I couldn’t imagine reading any other book alongside it’ [R1.71]) or to comment on the responses of other readers to the text (e.g. ‘it’s hard to imagine how so many people got suckered into this book’ [R1.10]; ‘Even those who tried desperately to ignore the hype somehow got sucked into the huge marketing machine that was Twilight’ [R1.61]). Instead, this underlying source domain

WOLFING DOWN THE TWILIGHT SERIES

47

manifests here in descriptions of the text as ‘shallow’ (R1.68, 79, 83) or the lack of ‘depth’ to the text’s characters and meaning (R1.8, 16). With such distinct patterns characterizing the 1-star and 5-star reviews, these different construals of this same underlying conceptual metaphor might be seen to reflect the contrasting embodied experiences of the text being described. In our data, the embodied sense of a (downwards) movement ‘into’ the text characterizes only successful, immersive engagements with the novel, in an interesting subversion of the conventional conceptual metaphors GOOD IS UP, BAD IS DOWN (Lakoff and Johnson, 1980, 1999). Within such positive descriptions, the two different metaphorical construals observed in our data could suggest a degree of variation in the extent to which this immersive experience is felt by the reader to be an active or passive engagement. Interacting with this embodied motivation, of course, are the specific interpersonal, rhetorical functions that metaphor carries in this discourse context. As exemplified by the examples given above, use of the construction ‘suck[er/ed] in’ in this online forum appears to carry a negative evaluation of the text, either as self-defence for a lack of critical perspective during reading (R5.96) or as a criticism of others for their acceptance of its ideology (R1.10, 61). 5.2 Reading is control The READING IS CONTROL metaphors used by the reviewers in the Twilight data sets similarly display variations of framing in terms of agency. Like READING IS A JOURNEY, READING IS CONTROL is also bidirectional, which, as we have observed, reflects two alternative construals of the source domain in terms of a forcedynamic energy transfer between participants. Stockwell (2009: 80) argues that the most common framing of this metaphor involves ‘readers being the entity controlled’. One particular manifestation of the READING IS CONTROL metaphor that positions the reader as the entity being controlled is the ‘made me X’ syntactic construction, which appeared in both of our data sets. The six instances of this construction in the 5-star reviews are listed in Table 3.2. Though the CONTROL here evidently comes from the text, the exact cause varies, with the energy source or AGENT in the underlying action chain identified as the narrative voice (R5.31, 39), the book as a whole (R5.16, 75, 92) or the author specifically (R5.83). Although use of the force-dynamic expression ‘made me’ (see Talmy, 1988, 2000) indicates that the reader feels they have no control over the state that ensues, the resulting impact on them as PATIENT is in each case an emotional experience (‘grin’, ‘care’, ‘feel’, ‘fall for’, ‘fall in love’ or ‘desperate’) that is framed positively by the reviewer. This contrasts with the occurrences of the ‘made me’ constructions that appear in the 1-star reviews. Compared to the 5-star data set, the ‘made me’ constructions in the 1-star reviews occurred more frequently (twenty-five

48

CONTEMPORARY MEDIA STYLISTICS

TABLE 3.2 ‘Made me’ READING IS CONTROL constructions in the 5-star reviews It’s fascinating and

makes me

desperate to get the book when it finally comes out. (R5.16)

I forgot how snarky Bella was. And of course the sweet parts still

made me grin like a loon. (R5.31)

I liked Bella’s wry, self-deprecating narrative voice from the get-go; and both the two main characters and many of the secondary ones really came to life for me in a way that

made me care about them. (R5.39)

It gutted me, but

made me feel alive at the same time. (R5.74)

Despite what Stephen King said, I think Stephenie Meyer did an amazing job (she

made me , and millions of other girls, fall for a fictional character. Come on!) and I like her writing style. (R5.83)

This is one of those books that

made me fall in love with reading again. (R5.92)

instances as compared to six). As expected, the 1-star ‘made me’ constructions are similarly evaluative, but framed much more negatively, as the examples in Table 3.3 demonstrate. The source of the control, or the agent in these action chains, can again be traced to different parts of the book or the book as a whole (R1.34; R1.62; R1.96), specific stylistic features (R1.1) or ideas within it (R1.3; R1.75). Another pattern apparent in these 1-star examples is the inducement of a physiological reaction (‘laugh uncontrollably’, ‘feel total and complete revulsion’, ‘kind of want to spew’, ‘gag’, ‘cringe’ or ‘boggle’) as well as the severity of emotional feeling (‘very, very angry’). Like the examples of the READING IS A JOURNEY metaphor in the previous section, in all of these descriptions the reader conceptualizes themselves as the energy sink; the entity at the end of the action chain that is affected by the text. It is also notable that in these examples the readers are then positioned as EXPERIENCERS (Langacker, 2008: 356); being at the end of the action chain has brought about a change in their perceptual, emotional or physiological state. Unlike the 5-star examples, however, the 1-star manifestations of the metaphor frame these experiences as ones that are unwelcome and that impact upon the body negatively (see also Section 5.3 for conceptualizations of the physicality of the reading experience). Another distinctive framing of this metaphor in our data is the focusing of the specific scenario of ADDICTION. This construal appears in both 1-star and

WOLFING DOWN THE TWILIGHT SERIES

49

TABLE 3.3 ‘Made me’ READING IS CONTROL constructions in the 1-star reviews but I wish that Meyer had come up with a better idea that didn’t

make me

laugh uncontrollably at the thought. (R1.1)

Purple Prose – Ew ... to this ... seriously, all the purple prose

made me

want to throw the book across the room. (R1.1)

All the high school/teenage stuff honestly

made me

boggle. (R1.3)

If I had actually finished the book I might have had more to say, but every new page

made me

cringe. (R1.34)

But I have to talk about this book, because it

made me

very, very angry. (R1.62)

But I might be suddenly channeling Andrea Dworkin, because this book

made me

feel total and complete revulsion. (R1.62)

Bella and Edward’s relationship is based entirely on physical attraction (he’s beautiful, she smells good), so it

made me

gag everytime Bella/Meyer tries to forcefeed you the idea that it’s the greatest, most loving romance of all time. (R1.75)

The first half of this book

made me

kind of want to spew. (R1.96)

Positively everything she did

made me

grind my teeth. (R1.85)

5-star reviews but more frequently in the 5-star reviews (twenty instances as compared with only three instances in the 1-star reviews). As an instantiation of the CONTROL metaphor, READING IS ADDICTION represents a specification of the source domain: in particular, the absence of control on the part of the PATIENT (the reader). The increase in the text’s control over the reader is explicitly marked through the word ‘addiction’, which appears in different forms nine times in the 5-star reviews and three times in the 1-star reviews. In the three examples in the 1-star reviews, the references to addiction described a previous reading stance which has now altered (‘When I first read the book, I was addicted to it’ [R1.15; R1.97]), or a compulsion to read in order to criticize it more soundly (R1.65). The book is additionally compared to ‘drugs’ more superordinately (R5.94), and two reviewers compared the book to ‘crack’ (R1.97; R5.89) in particular, whereas others talk about ‘craving’ (R5.1) the text, getting a ‘temporary fix’ between books (R5.16) and about being a ‘bookjunkie’ (R5.32). This ADDICTION metaphor can be seen as an ‘elaboration’ (Lakoff and Turner, 1989; Semino, 2008: 45) of the CONTROL metaphor, through specification of a

50

CONTEMPORARY MEDIA STYLISTICS

particular part of this more generic metaphor: in this case, giving up further control to the text so much so that readers construe themselves as being dependent upon it. While mapping both the dangerous and pleasurable aspects of this source domain onto the reading experience, in this affinity space, readers’ creative use of this conventional figurative template is a mark of enjoyment, predominantly, for this text, whether that enjoyment be ongoing (as in the 5-star reviews) or in the past (in the 1-star reviews). For some readers, then, a positive immersive reading experience of this book is one which not only is passive but also generates a feeling of dependency. In all the instantiations of the CONTROL metaphor described so far, the reader is construed as the entity with less agency. This can also be seen in many other manifestations of the metaphor in the data, with reviewers describing being ‘hooked’ (R5.41, 69, 83), being taken in the ‘grip’ of the book (R5.48), the ‘gripping’ storyline (R5.3), being ‘caught up’ (R3: 13) or ‘roped in’ by the story (R1.16; R5.73). At the same time, however, there are a smaller number of manifestations of READING IS CONTROL in the data sets that construe the control as coming from the reader (just five instances in 1-star reviews, and three instances in the 5-star reviews), as the following examples demonstrate: But, I really regret ever buying and forcing myself to finish it (I hate not finishing books, even if I hate them). (R1.1) And yet despite all of this I could not stop reading these books. It was like a train wreck that I could not pull myself away from. (R1.58) After seeing mixed reviews on it and reading that many people complained about the writing and characters, I bypassed it for awhile, but I had to read it. (R5.5) Young adult novels do not normally hold much interest for me, and even after my daughter and niece begged me to read this one, I resisted. ... There was something very compelling about this book, partly because of the appeal of the sweet young love in the story. (R5.21) In all these examples, the control is in fact represented as arising from multiple directions at once. The second example, for instance, construes the reader as an AGENT in that they are attempting (unsuccessfully) to perform actions (‘I could not stop reading’; ‘I could not pull myself away’ [R1.58]). At the same time, the book maintains agency in the implied energy and activity of the ‘train wreck’ that competes with the agency of the reader, acting as an obstacle in this action chain, or what could be described in force-dynamic terms as an opposition between ‘agonist’ and ‘antagonist’ (Talmy, 1988, 2000). This tug-of-war for control can also be seen in the other reviews, in which

WOLFING DOWN THE TWILIGHT SERIES

51

an interaction between competing forces is apparent in lexical choices such as ‘forcing myself ’, ‘bypassed it’ and ‘resisted’. Examined at a finer level of detail, these instances of this metaphorical construal (READING IS CONTROL; READER AS AGENT) might be further differentiated in terms of whether the reader’s agency manages to overcome that exerted by the text. Considered altogether, the different manifestations of the READING IS CONTROL metaphor might tell us something about the experience of reading this text. As predicted by Stockwell (2009), both the 1-star and 5-star reviews reveal a preference for a construal of the text as controller (the AGENT), and the reader as controlled (the PATIENT), with 81 instances coded for TEXT AS AGENT in our overall sample of 200 reviews, compared to just 8 coded for READER AS AGENT. In those limited instances in which the reader was construed in an agentive role, this was often represented as competing with an opposing force presented by the text. This finding is interesting when considered in light of stylistic accounts of narrative immersion (Section 2). While the process of becoming immersed in a fictional world is often modelled as an active process of ‘projection’ (Whiteley, 2011) or ‘performance’ (Gerrig, 1993) in response to textual cues on the part of a reader, in our data, this sense of agency does not appear. Instead, across both positive and negative responses to the text, readers represent their experience of this particular narrative and resulting emotions – as driven, or controlled, by the text itself – with their own agency most apparent in the form of resistance to this control. 5.3 Reading is eating So far this chapter has observed the ways in which the linguistic manifestations of two prominent reading metaphors, READING IS A JOURNEY and READING IS CONTROL, vary across the 1-star and 5-star conditions, as well as the ways in which they remain consistent. Another notable pattern in the data was the appearance of metaphors that have not previously been explored in stylistic research. Some of these include examples of pairing the target READING with the source domains of A RELATIONSHIP, SEX, A RELIGIOUS EXPERIENCE, MENTAL ILLNESS and SEEING, among others. READING IS EATING is one of the most frequently used metaphors, appearing in both data sets (twenty-six occurrences in the 5-star reviews and fifty-one in the 1-star reviews). This metaphor for reading has hitherto not been researched or discussed extensively (cf. Radway, 1986) despite arguably having a conventional usage in English (instantiated, for example, in the phrase ‘bookworm’ which can be used to describe an avid reader, or in how a person might describe their ‘taste’ in books). Building on Kövecses’s (2008) discussion of the context-specific motivations for metaphor (see Section 2) the prevalence of this metaphor in the Twilight reviews may also be primed by the book itself, since the cover of the

52

CONTEMPORARY MEDIA STYLISTICS

2007 paperback edition features two hands offering a red apple. In addition, the text itself features many food analogies to describe Edward’s appetite for blood, and for Bella’s blood in particular: ‘You know how everyone enjoys different flavors?’ he began, ‘Some people love chocolate ice cream, others prefer strawberry?’ I nodded. ‘Sorry about the food analogy – I couldn’t think of another way to explain.’ (Meyer, 2005: 85) The prominence of food as a theme is noted by critical literature written on the text (see, for example, Dunn, 2009), and is also acknowledged by some readers as a recurring (and to some extent ‘foregrounded’ [Mukařovský, 1964]) feature of the narrative. R1.32, for instance, complains that ‘the motions of eating ravioli and looking at the mushrooms on the plate are painstakingly dragged out’, and R1.96 observes that, ‘seriously, there’s a description of how Bella rearranges the fridge’. R5.20, meanwhile, has assigned this book to the Goodreads shelf category ‘food-drinks’. More generally, readers in the data sets talk about ‘devouring’ the book (R5.4, 63), its ‘cheesy’ tone (R5.8; R1.8) and ‘sugar-coated’ themes (R5. 11), and their particular reading ‘taste’ (R5.4) and ‘insatiable hunger’ (R5.43) for the series. In the 5-star reviews, the READING IS EATING metaphor frequently manifests in a more specific construal, with reviewers comparing reading Twilight with eating junk food, in particular. These reviewers liken reading the book to eating specific items, such as ‘chocolate’ (R5.1), ‘McDonalds’ (R5.4), ‘carnival food’ such as ‘deep fried corn dogs, greasy pizza, funnel cakes, and elephant ears’ and ‘cotton candy’ (R5.7), ‘pizza’ (R5.7, 42) and to drinking ‘wine’ (R5.69). While the READING IS EATING examples in the 1-star data also name specific junk food items (a ‘marshmallow’ [R1.10] and a ‘Twinkie’ [R1.12]), different aspects of this shared JUNK FOOD source domain are mapped across the two data sets. For the 5-star reviewers, the junk food is ‘refreshing’ (R5.6), ‘delicious, sinful’ and an ‘indulgence’ (R5.1), but for the 1-star reviewers the emphasis is on the ‘unhealthy’ (R1.62, 79) aspect: ‘it’s addictive to many and yet lacks any kind of substance/nutritional value besides empty calories’ (R1.12). In addition to focusing on different aspects of the same domain, in some manifestations of this metaphor, there is also evidence of ‘conceptual blends’ with other domains (Fauconnier and Turner, 2002). The influence of ADDICTION, for example, is still present in some of the evocations of EATING: for example, ‘it left me with that same craving for more that Harry Potter did’ (R5.1); ‘giving you a sweet addiction’ (R5.63); ‘this book was addictive, I couldn’t enough of it, like pizza (can’t get enough of that … obviously if Edward thought Bellas scent

WOLFING DOWN THE TWILIGHT SERIES

53

was empowering, he never smelt pizza … here’s me getting sidetracked with food again 😂)’ (R5.19). Such creative ‘combinations’ of conceptual metaphors (Lakoff and Turner, 1989; Semino, 2008: 48) focus a matrix of domains, various aspects of which collectively enrich our understanding of the target. Cameron (2003) refers to this process as a type of ‘metaphor construction’, where a target becomes ‘re-juxtaposed’ with a ‘developed or contextualized’ source domain in order to ‘reformulate the metaphor’. In the last example given here, the pleasurable experience of eating ‘pizza’ is reformulated with the sense of a lack of control suggested by ‘addiction’. Consequently, this blend of domains emphasizes the simultaneously pleasurable and overpowering experience of reading the book for this reader. 5.4 Reading is choking/regurgitating; books are excrement Alongside the specific food-substances referenced in the READING IS EATING 1-star reviews, there are also some unique instantiations of this metaphor that do not appear in the 5-star data set. A number of reviewers liken reading Twilight to eating something that causes choking (‘gag worthy fluff ’ [R1.1]; ‘choke it down’ [R1.42]; ‘the romance was gag worthy’ [R1.68]; ‘gag!’ [R1.11]). Like the CONTROL metaphors discussed in Section 5.2, this metaphorical construal reflects an act of resistance to the text’s agency on the part of the reader: for example, ‘it made me gag everytime Bella/Meyer tries to forcefeed you the idea that it’s the greatest, most loving romance of all time’ (R1.75). Here, the source domains EATING and CONTROL are combined, and again, the reader is conceptualized as the PATIENT in the action chain, while ‘Bella/Meyer’ adopt the AGENT role. This construal of such resistant EATING expresses a reading experience which is not smooth, easy or even a voluntary activity; a complete contrast to the action chains construed in the 5-star data which describe wilfully ‘devouring’ the book (R5.4, 63). Related to this latter particular scenario of CHOKING/ GAGGING on food, another mapping of this metaphor appears in expressions that conceptualize the aftermath of the eating process, and specifically ‘feeling sick’ (R1.85, 1, 35, 83) or vomiting (‘It was a edible [sic] I’d vomit it back up’ [R1.37]; ‘VOM VOM VOM’ [R1.58]); or the unpleasant ‘taste that lingers’ (R1.32) after reading. These construals focus an additional scenario within the general source domain of EATING: REGURGITATING. In doing so, they can be said to reflect an ‘extension’ of this metaphor (Lakoff and Turner, 1989; Semino, 2008: 44), or an extension of the underlying action chain conceptualized for EATING, to include a new slot or action once the eating is complete. Finally, BOOKS ARE EXCREMENT is one metaphor that appears only in the 1-star reviews. Other than explicit references to ‘shit’ and ‘bullshit’ (R1.21, 59, 92), reviewers who use this metaphor also make a number of comparisons between the book and ‘a turd’ (R1.10), ‘crap’ (R1.1, 27, 32, 42, 65, 75), or more

54

CONTEMPORARY MEDIA STYLISTICS

tangentially related descriptions such as ‘Toilette Twilight’ (R1.2), the ‘buttawful sequels’ (R1.16), ‘toilet paper’ and ‘piss-poor’ writing (R1.84). Further, there is evidence of additional metaphor combinations: R1.2 describes the text as ‘a shit storm in action’, blending this source domain with that of a NATURAL FORCE. BOOKS ARE EXCREMENT differs from the previous metaphors discussed in this analysis, however, in that it focuses a different target domain: BOOKS rather than READING. This focusing of the target domain makes the critical nature of these reviews more directed; the negative evaluation is not based on an experience to which the reader contributes but rather is solely attributed to the book as an artefact. It could be argued that the metaphors mentioned in this section – READING IS EATING and examples that map CHOKING and REGURGITATING, specifically, and BOOKS ARE EXCREMENT – relate to the same generic-level metaphor of READING IS EATING, but that in each manifestation different aspects of this process (understood in a broader sense – its ‘maximal scope’ [Langacker, 2008]) are profiled. Specifically, it is a matter of which part of the EATING process is being attended to: with emphasis on the ‘initial’ (ingesting the food/book), the ‘medial’ (how that food gets digested) or the ‘final’ (the outcome) sections of the same action chain (Talmy, 2000: 255–309). Indeed, instances in which the target domain of WRITING is described in terms of REGURGITATING (e.g. ‘Twilight reads like Meyer has read a lot of mediocre novels and regurgitated the same kind of language onto the page’ [R1:1]) demonstrate a further extension of this source domain and its profiled action chain. Finally, an interesting example of readers’ creative extensions of this metaphor is seen in a review that advises other readers to ‘steer clear of this book other than for the humour value and the shock that American teenagers are not only being shovelled this shit, but are lapping it up in droves’ (R1.88). This reviewer’s creative construal profiles both the initial and final parts of this EATING action chain, so that these activities are construed as part of a cyclical process. The clustering of lexis surrounding a specific TOILET scenario, along with those seen for JUNK FOOD and ADDICTION earlier in this analysis, highlights how readers in such online discourse environments can influence others’ language and actively shape the distinctive style features of the affinity space; this is shown, in particular, through the extension of the same running joke (e.g. R1.2 acknowledges that ‘Toilette Twilight’ has been previously used by another reviewer on the site); and in using creative lexical blends for adapting the title to this joke: ‘Twishite’ (R1.42) and ‘Twifart’ (R1.16). Such recurring discourse features form part of the ‘shared language’ of this discourse community (Swann and Allington, 2009; Peplow, 2011) and allow readers to demonstrate their ‘shared affinity’ (Vlieghe et al., 2016: 28) within this environment. Given the dysphemistic language and the on-record taunts by reviewers towards those in the 5-star camp, the use of this metaphor can be seen as a performative tool

WOLFING DOWN THE TWILIGHT SERIES

55

used by 1-star reviewers to help position themselves more firmly within the group.

6 CONCLUSION This chapter has investigated readers’ varied and creative uses of metaphor to describe experiences of fiction in an online media context. The findings of this study offer a number of significant insights for stylistic accounts of reader experiences of narrative fiction, and for critical accounts of Twilight in particular. First, building on observations by Stockwell (2009), this analysis has provided support for three prominent metaphors for reading – READING IS A JOURNEY, READING IS CONTROL and READING IS INVESTMENT – noted in previous work as underpinning readers’ descriptions of their own experiences. By analysing the range of linguistic forms that these metaphors take in discourse, and their specific ‘framing’ of reading in terms of ‘agency, (dis) empowerment, evaluations and emotions’ (Semino et al., 2016: 18), or metaphorical construal of this process in terms of an ‘action chain’ (Langacker, 2008), this chapter has revealed a number of patterns in their use in relation to this particular text. One such pattern is the tendency to describe successful, immersive experiences of reading in terms of a DIVING scenario, or a downwards movement ‘into’ the text. Indeed, this metaphorical understanding of textual engagement is one that underpins the stylistic term ‘immersion’ itself. Restricted to positive responses to this novel (in the 5-star reviews, or in a negated context in 1-star reviews) this embodied experience of immersive reading was one in which readers variably conceptualized themselves as active or passive participants. Instantiations of the READING IS CONTROL metaphor, however, across both positive and negative reviews, consistently framed the reader in a passive role – as the recipient of the text’s agency, with their agency limited to an act of resistance to the text. Notable in this respect, was the recurrence of another specific scenario ADDICTION to describe the experience of being controlled, as a feeling of dependency that was simultaneously dangerous and pleasurable. Viewed in light of existing cognitive stylistic accounts of immersion and resistance, these accounts of everyday reading experiences offer a useful perspective on the acts of ‘recentering’ (Ryan, 2001), ‘projection’ (Whiteley, 2011) or ‘positioning’ (Stockwell, 2013) set out in such models. Specifically, they suggest a ‘give and take’ (Silver, 2010) in readers’ engagement with this novel, or a degree of awareness in these readers of the passivity or resistance with which they accept its invitation for engagement. In addition, our data has revealed a range of other conceptual metaphors for reading, which have not been discussed in previous stylistic accounts. The most salient of these: READING IS EATING, we have proposed, is one that is primed by, and developed within, this specific online media context. Another such

56

CONTEMPORARY MEDIA STYLISTICS

metaphor in the Twilight data, not addressed in this chapter, is READING IS A COMPETITION. One reviewer (R1.69), for example, takes a step away from the text entirely and cites ‘the fans’, rather than any textual feature within the book, as the justification for their hatred: ‘It’s the fans that made me hate [emphasis in original] the book.’ This acknowledgement of the ‘other’ group of fans forms another distinctive feature of the Twilight reviews, and possibly other reviews for texts that are renowned for their polarized reception. Such an explicit reference to ‘the fans’ could be seen as a sign of direct provocation in this discourse environment, as well as a sign of in-group membership (Swann and Allington, 2009; Peplow, 2011): here, R1.69 is clearly positioning themselves as a ‘Twihater’. Such markers of in-group and out-group membership are explicitly noted across both review sets in the lexical blends that reference the ‘Twilighters’ (R1.22), ‘Twitards’ (R1.42), ‘Twihards’ (R1.61; R5.8, 18, 42, 77) and ‘Twihaters’ or simply ‘the haters’ (R5.11, 63). In the case of the 5-star data, the reviewers are further subdivided according to ‘teams’: you are either ‘Team Edward’ or ‘Team Jacob’ (R5.3, 8, 14, 30, 77). The scenario of a COMPETITION appears in other metaphors used by both sets of reviewers: one of the 1-star reviewers talks of giving the book ‘a fighting chance’ (R1.85), for example, whereas a 5-star reviewer describes the book as a ‘winner’ (R5.46). Again, there is evidence of taking an antagonistic stance: R5.14 states, ‘I challenge you to read the book.’ In these instantiations, reading is understood as a competition which goes beyond the text itself; here, the underlying action chain is represented as being between groups of readers, and the opposing 1-star and 5-star teams, rather than a relationship that exists solely between the reader and the text. The extent to which metaphors newly identified in this study (READING IS EATING, READING IS A COMPETITION, etc.) are conventional in discourse about reading requires further research. Future studies might consider the impact of genre on metaphor use (is it the case that specific uses of metaphor are conventional within the fantasy or Young Adult romance genres, for instance, or localized to individual series or texts?). The analysis here has identified the features of our data which suggest a context-specific motivation for the ways in which particular metaphors are creatively employed. These features include the social factors which influence this online community, the famously polarized reception of this text in the media and foregrounded themes within the text itself. Such creative metaphor use, we propose, presents an exciting area of research for stylisticians interested in changing reading practices and the socially oriented reading experiences that new media affords.

NOTES 1. Goodreads’ default sorting algorithm orders the reviews presented on the site on the basis of the following factors: ‘length of the review, number of people who

WOLFING DOWN THE TWILIGHT SERIES

57

liked it, recency of the review, popularity of the reviewer (i.e. number of people who have liked reviews by that person across all books)’ (Goodreads, 2018). Since this is how the website presents its reviews by default, these reviews are also those that are most likely to be read by users of the site. 2. Content deemed inappropriate for inclusion in our data set comprised reviews which publically named and criticized other reviewers on the site. The fact that such reviews appeared in our data is an interesting feature of this discourse context, but one that we have decided not to focus on here given their relative scarcity. 3. This difference in the number of coded metaphors in our data sets reflects a difference in the length of the 1-star and 5-star reviews. The top 100 reviews for the 5-star condition yielded 32,733 words in total, while the 1-star data set was significantly larger at 57,788 total words. See Mason (2015: 66) for discussion of the tendency for negative reviews to be greater in length. 4. Multiple, separate instantiations of the same metaphor across individual reviews were coded discretely. Hence, the frequencies given indicate the number of coded instances of that node across the data set, as opposed to the number of reviews in which they appeared.

REFERENCES Allington, D. (2011). ‘“It actually painted a picture of the village and the sea and the bottom of the sea”: Reading groups, cultural legitimacy, and description in narrative (with particular reference to John Steinbeck’s The Pearl)’. Language and Literature, 20(4), 317–22. Allington, D. and Pihlaja, S. (2016). ‘Reading in the age of the internet’. Language and Literature, 25(3), 201–10. Browse, S. (2014). ‘Resonant metaphor in Never Let Me Go’. In C. Harrison, L. Nuttall, P. Stockwell and W. Yuan (Eds), Cognitive Grammar in Literature (pp. 69–82). Amsterdam: John Benjamins. Cameron, L. (2003). Metaphor in Educational Discourse. London: Continuum. Curwood, J. S. (2013). ‘The Hunger Games: Literature, literacy and online affinity spaces’. Language Arts, 90(6), 417–27. Dancygier, B. and Sweetser, E. (2014). Figurative Language. Cambridge: Cambridge University Press. Dunn, G. A. (2009). ‘You look good enough to eat: Love, madness, and the food analogy’. In R. Housel and J. J. Wisnewski (Eds), Twilight and Philosophy: Vampires, Vegetarians and the Pursuit of Mortality (pp. 7–24). Hoboken, NJ: John Wiley & Sons Inc. Fauconnier, G. and Turner, M. (2002). The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. Plymouth: Basic Books. Fay, S. (7 May 2012). ‘Could the internet save book reviews?’ The Atlantic. Retrieved 6 April 2018 from https://www.theatlantic.com/entertainment/archive/2012/05/ could-the-internet-save-book-reviews/256802/ Frank, R. M., Ziemke, T. and Zlatev, J. (Eds) (2008). Body, Language and Mind. Vol. 1: Embodiment. Berlin: Mouton de Gruyter. Galbraith, M. (1995). ‘Deictic shift theory and the poetics of involvement in narrative’. In J. F. Duchan, G. A. Bruder and L. Hewitt (Eds), Deixis in Narrative: A Cognitive Science Perspective (pp. 19–59). Hillsdale: Lawrence Erlbaum.

58

CONTEMPORARY MEDIA STYLISTICS

Gavins, J. (2007). Text World Theory: An Introduction. Edinburgh: Edinburgh University Press. Gavins, J. (2013). Reading the Absurd. Edinburgh: Edinburgh University Press. Gee, J. P. (2004). Situated Language and Learning: A Critique of Traditional Schooling. London: Routledge. Gerrig, R. J. (1993). Experiencing Narrative Worlds: On the Psychological Activities of Reading. New Haven: Yale University Press. Gibbs, R. W. (2005). Embodiment and Cognitive Science. Cambridge: Cambridge University Press. Goodreads. (2018). ‘Twilight reviews’. Retrieved 6 April 2018, from https://www.goo dreads.com/book/show/41865.Twilight Gregoriou, C. (2012). ‘“Times like these, I wish there was a real Dexter”: Unpacking serial murder ideologies and metaphors from TV’s Dexter internet forum’. Language and Literature, 21(3), 274–85. Harrison, C. (2017). Cognitive Grammar in Contemporary Fiction. Amsterdam: John Benjamins. James, E. L. (2011–2017). Fifty Shades (series). London: Vintage Books. Kövecses, Z. (2002). Metaphor: A Practical Introduction. Oxford: Oxford University Press. Kövecses, Z. (2008). ‘Conceptual metaphor theory: Some criticisms and alternative proposals’. Annual Review of Cognitive Linguistics, 6, 168–84. Lakoff, G. and Johnson, M. (1980). Metaphors We Live By. Chicago: Chicago University Press. Lakoff, G. and Johnson, M. (1999). Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought. New York: Basic Books. Lakoff, G. and Turner, M. (1989). More than Cool Reason: A Field Guide to Poetic Metaphor. Chicago: University of Chicago Press. Langacker, R. W. (2008). Cognitive Grammar: A Basic Introduction. Oxford: Oxford University Press. Mason, J. (2015). Narrative Interrelation: A Cognitive Model of Intertextuality and Its Application to the Study of Literature. Unpublished PhD thesis, University of Nottingham. Meyer, S. (2005). Twilight. New York: Little, Brown. Mukařovský, J. (1964). ‘Standard language and poetic language’. In P. L. Garvin (Ed.), A Prague School Reader on Esthetics, Literary Structure and Style (pp. 17–30). Washington: Georgetown University Press. Musolff, A. (2006). ‘Metaphor scenarios in public discourse’. Metaphor and Symbol, 21(1), 23–38. Narula, S. K. (12 February 2014). ‘Millions of people reading alone, together: The rise of Goodreads’. The Atlantic. Retrieved 6 April 2018, from https://www.theatlan tic.com/entertainment/archive/2014/02/millions-of-people-reading-alone-togetherthe-rise-of-goodreads/283662/. Nuttall, L. (2017). ‘Online readers between the camps: A text world theory analysis of ethical positioning in We Need to Talk About Kevin’. Language and Literature, 26 (2), 153–71. Nuttall, L. (2018). Mind Style and Cognitive Grammar: Language and Worldview in Speculative Fiction. London: Bloomsbury. Peplow, D. (2011). ‘“Oh, I’ve known a lot of Irish people”: Reading groups and the negotiation of literary interpretation’. Language and Literature, 20(4), 295–315.

WOLFING DOWN THE TWILIGHT SERIES

59

Peplow, D., Swann, J., Trimarco, P. and Whiteley, S. (2016). The Discourse of Reading Groups: Integrating Cognitive and Sociocultural Perspectives. London: Routledge. Pihlaja, S. (2017). ‘More than fifty shades of grey: Copyright on social network sights’. Applied Linguistics Review, 8(2–3), 213–28. Pragglejaz Group. (2007). ‘MIP: A method for identifying metaphorically used words in discourse’. Metaphor and Symbol, 22(1), 1–39. Radway, J. A. (1986). ‘Reading is not eating: Mass-produced literature and theoretical, methodological, and political consequences of a metaphor’. Book Research Quarterly, 2(3), 7–29. Ryan, M. L. (2001). Narrative as Virtual Reality: Immersion and Interactivity in Literature and Electronic Media. Baltimore: John Hopkins University Press. Sarland, C. (2005). ‘Critical tradition and ideological positioning’. In P. Hunt (Ed.), Understanding Children’s Literature (pp. 30–49). London: Routledge. Semino, E. (2008). Metaphor in Discourse. Cambridge: Cambridge University Press. Semino, E., Demjén, Z. and Demmen, J. (2016). ‘An integrated approach to metaphor and framing in cognition, discourse and practice, with an application to metaphors for cancer’. Applied Linguistics, 1–22. Senkbeil, K. and Hoppe, N. (2016). ‘“The sickness stands at your shoulder ...”: Embodiment and cognitive metaphor in Hornbacher’s Wasted: A memoir of anorexia and bulimia’. Language and Literature, 25(1), 3–17. Silver, A. (2010). ‘Twilight is not good for maidens: Gender, sexuality, and the family in Stephenie Meyer’s Twilight series’. Studies in the Novel, 42, 121–38. Spilioti, T. and Tagg, C. (2017). ‘The ethics of online research methods in applied linguistics: Challenges, opportunities, and directions in ethical decision-making’. Applied Linguistics Review, 8(2–3): 163–7. Stockwell, P. (2005). ‘Texture and identification’. European Journal of English Studies, 9(2), 143–53. Stockwell, P. (2009). Texture: A Cognitive Aesthetics of Reading. Edinburgh: Edinburgh University Press. Stockwell, P. (2013). ‘The positioned reader’. Language and Literature, 22(3), 263–77. Swann, J. and Allington, D. (2009). ‘Reading groups and the language of literary texts: A case study in social reading’. Language and Literature, 18(3), 247–64. Talmy, L. (1988). ‘Force dynamics in language and cognition’. Cognitive Science, 12, 49–100. Talmy, L. (2000). Towards a Cognitive Semantics (Vol. 1): Concept Structuring Systems. Cambridge: Cambridge University Press. Time (2008, 12 May). ‘The 2008 time 100’. Time. Retrieved 29 May 2018 from http://content.time.com/time/specials/2007/article/0,28804,1733748_1733752_ 1736282,00.html. Vlieghe, J., Muls, J. and Rutten, K. (2016). ‘Everybody reads: Reader engagement with literature in social media environments’. Poetics, 54, 25–37. Whiteley, S. (2011). ‘Text world theory, real readers and emotional responses to The Remains of the Day’. Language and Literature, 20(1), 23–42. Whiteley, S. (2014). ‘Ethics’. In P. Stockwell and S. Whiteley (Eds), The Cambridge Handbook of Stylistics (pp. 39–47). Cambridge: Cambridge University Press.

CHAPTER FOUR

The language of citizen science: Short strings and ‘we’ as a group marker GLENN HADIKIN

Citizen science is a form of science crowdsourcing where members of the public offer their time to help scientists tag, classify and collect data in the field. The concept arguably began when the US Audubon society decided to count birds rather than shoot them as a Christmas day activity in 1900 (Silvertown 2009). Over the last eight years Zooniverse.org has grown to be the world’s largest online platform offering volunteers the opportunity to classify galaxies based on NASA images, work with camera traps in the Serengeti and monitor penguins in the wild. This chapter introduces discourse of citizen science discussion forums and asks what role short strings of words such as that we are and we are in play in terms of group identity. I compare the linguistic culture of Zooniverse forums to that of a popular blog, ScienceBlogs. The corpus analysis software WordSmith Tools (Scott, 2012) is used to focus on strings containing the word we and the data is tagged according to the referent of we in each line such as we meaning human, the scientific community or a particular team of researchers. Analysis shows the tendency for one thread of Zooniverse to focus on its shared membership of the group human rather than split up into smaller groups, whereas the results from ScienceBlogs are more consistent with an atmosphere of debate. The posters, for example, use we to identify with subgroups such as the United States or Christians. Certain strings also seem

62

CONTEMPORARY MEDIA STYLISTICS

to be associated with particular contexts in unexpected ways, such as a high frequency of that we are when we refers to all humans; results are discussed in terms of Hoey’s lexical priming (Hoey, 2005) and my previous work on lexical selection (Hadikin, 2015). Overall, the chapter highlights the need for careful examination of co-text. It also shows how lexical selection can provide insights about how online communities construct themselves in discourse.

1 INTRODUCTION The term ‘citizen science’ suggests a community of committed amateur scientists around the world who engage in projects for various reasons such as enjoyment and the desire to feel they are helping science or their local community. The aim of this chapter is to introduce lexical selection as a theoretical concept and to explore structures that may play a role in the process in a citizen science forum; a science forum that is not considered citizen science is also used for comparison. Lexical selection (Hadikin, 2015) sees language use as the sharing and modification of short strings of words that change over time in ways comparable to natural selection. The development of lexical selection is based on Hoey’s theory of lexical priming (2005) which is also discussed in this chapter. The chapter focusses on the short string we are and asks the following questions: 1) How does the meaning of we in context affect lexicogrammatical patterns? 2) What can we learn about an online community by focussing on the use of structures containing we are? 3) To what extent can lexical selection and lexical priming be used to explain patterns associated with the use of we are in two online science communities? No minimum occurrence of any structure is set for this analysis but most of the discussion is based on structures that occur four times or more. A single line can, however, tell us something about the priming of a writer. Two sets of data are used in the chapter. Citizen science data is taken from Zooniverse.org. It is currently the world’s largest citizen science website (Simpson, Page and Roure, 2014) with over 1.3 million users at the time of writing. The site consists of over ninety research projects that invite the public to engage in relatively simple tasks, such as identifying an animal or marking the shape of a galaxy; each task takes between 2 and 10 minutes for volunteers. Data sets range from images of galaxies and captures from motion-sensitive cameras in Africa to First World War diaries, and specimen images from museum collections. When a user reaches the homepage they see projects with names such as Chimp & See,

THE LANGUAGE OF CITIZEN SCIENCE

63

Science Gossip and Higgs Hunters – these are examples of the range of topics that might attract new site users. The citizen science data used for this chapter was drawn from a long forum discussion about dark matter. A second website, ScienceBlogs.com (hereafter referred to as SB), was selected for comparison; a corpus of SB data was prepared by Akshay Minocha and contains blog posts on various science topics published between 2006 and 2014. The data is readily available to subscribers as a 103 million word corpus in Sketch Engine (Kilgarriff et al., 2004). SB contains blogs with subject areas ranging from life science and physical science to brain and behaviour and technology. As well as the blogs themselves, data is taken from the comments sections. Blog posts on the site include ‘Greater than the Sum of Its Parts: Why Being the Parent of Many Isn’t as Hard as You Think’; ‘The Alleged Slowdown in Global Warming Didn’t Happen, New Research Shows’; and ‘On Teaching Pseudoscientific Controversies in Universities …’.

2 LEXICAL SELECTION AND LEXICAL PRIMING This section introduces lexical selection and lexical priming as the main theoretical concepts in the chapter, and it highlights the kind of language structure that may play a role in lexical selection. It begins by introducing Hoey’s theory of lexical priming (Hoey, 2005) before going on to discuss how the idea of lexical selection developed from it. Hoey calls for lexis to be considered more centrally than previous theories of language and begins his argument with two sentences: In winter Hammerfest is a thirty-hour ride by bus from Oslo, though why anyone would want to go there in winter is a question worth considering. Through winter, rides between Oslo and Hammerfest use thirty hours up in a bus, though why travellers would select to ride there then might be pondered. Hoey (2005: 5) The first is from a travel book Neither Here nor There (Bryson, 1991) and represents the natural collocations and grammatical patterning we might expect from any educated writer; the second was constructed by Hoey to highlight the effects of unusual collocations. He rightly points out that many, if not all, readers see the second version as clumsy but that no linguistic theory of the last 200 years fully accounts for this psychological role in our judgement. Hoey goes on to suggest ten detailed priming hypotheses (Hoey, 2005: 13) that suggest that the psychological role of collocation in our language use also manifests in the form of semantic association, pragmatic association and colligation:

64

CONTEMPORARY MEDIA STYLISTICS

the tendency any word has to occur in particular grammatical positions or functions. Based on Hoey’s (2005) work, Wray’s (2002) model of formulaic language and Dawkins’s (1976) concept of the meme, I began to explore how shared primings in a community lead to the way certain strings of words are used. In Hadikin (2014), for example, I show that the string of four words (known in corpus linguistics as a 4-gram) that kind of thing was significantly more frequent in a corpus of Korean spoken English collected from speakers in Liverpool than a comparable corpus of language collected from speakers in Seoul. It is likely that a tendency to select the item kind over sort in expressions like kind of has combined with an increased tendency to use vague language (such as kind of and sort of) in the forms used by Korean adults living in Liverpool. This results in what appears to be a ‘new string’ that is getting used and reproduced by the Liverpool Korean community more frequently than either the Korean community in Seoul or the British communities. The potential for uncommon strings to be adopted by a certain community and selected over alternatives – perhaps even over strings that would be judged as more natural by British listeners – led me to consider the role of language strings as memes in the sense used by Dawkins (1976). Dawkins suggests that there could be a unit of culture comparable to the gene in biology and that, similar to a gene, minor variations to the unit – the meme – and competition between memes would lead to a form of cultural evolution. In lexical selection, I focus more narrowly on human language. Figure 4.1 illustrates the model I introduced in Hadikin (2015). The stages marked string memorized in identical form and string memorized in altered form in figure one partially reflect Wray’s (2002) concept of a formulaic string so that a language user who processes that kind of thing as a formulaic chunk is represented by the left side of the figure. It could also be a person who learns the structure and is then primed to reproduce the original in a form very close to the original but does not necessarily store the term as a formulaic chunk. Note here that memorized includes conscious memorization as well as subconscious storage of the linguistic information necessary to reproduce a very similar form.1 This includes both psychologically stored formulaic strings and strings of language that are generated online by the language user and thus would be seen as speaker-external strings in Myles and Cordier’s (2016) terminology. A speaker or writer with a psychologically formulaic store of the string can then either produce the original string – that kind of thing – in a future, comparable context or choose to vary the string. This gives two variants that can potentially prime both the language producer and their interlocutors to reproduce one or both forms. The case of reproducing the original form is labelled a standard output in Figure 4.1; the variant is a positive mutation and

THE LANGUAGE OF CITIZEN SCIENCE

65

Input form

String memorized in identical form

Standard output

Positive mutation

String memorized in altered form

Negative mutation

Social pressure

FIGURE 4.1: Lexical selection process for any given string reproduced from Hadikin (2015: 461).

the next generation of the string would be the next time the language user or an interlocutor is in a comparable situation and could reasonably select from the two. A negative mutation in Figure 4.1 comes about via the string memorized in altered form stage so the original string – that kind of thing – is not stored as a psychological formulaic sequence and the speaker is not strongly primed to produce the original form during online processing. This would likely result in a form traditionally seen as a mistake and social pressure would reduce the chance of the language user or their interlocutors reproducing the form in similar contexts in future. It would essentially be removed from the lexical ‘gene pool’. Note, however, that both positive and negative mutations could spread depending on local circumstances and that social pressure (as well as other factors such as memory and an individual’s grammar) influences all potential output forms. One cannot tell from an existing corpus which forms have developed via which of the three routes, so such differentiation will not be attempted in this chapter. Mathematical and computer modelling techniques such as Baxter et al. (2009) provide evidence that evolution-like models can solve language problems and confirm that this is a fruitful area of research. In Hadikin (2015), I argue for the need to introduce the line of thought into other sub-fields of linguistics such as corpus linguistics and research in digital communication. One way of exploring this is to ask what the unit of selection might look like

66

CONTEMPORARY MEDIA STYLISTICS

and how it moves around different communities. In Hadikin (2015) I discuss the word and Sinclair’s (1996) extended lexical unit as potential candidates and suggest the term ‘extended selection unit’ (ESU) to describe the theoretical replicator in the model (comparable with a gene in the sense of being the larger chain of information units that can vary in terms of sub-units). I do not claim to describe exactly what an ESU looks like in this chapter – a detailed description may take decades. The aim of this chapter is to explore structures that may play a role in the process in a citizen science forum. In Section 3 I introduce the ‘we environment’ – the co-text and context surrounding the lexical item we and how it likely plays an unusual role in the selection of lexical units.

3 CONTEXT AND THE WE ENVIRONMENT The lexical item we was chosen as a way into the online data because of its potential for marking members of a group. Also, if a selection model of language proves to be a useful way to explain the language issues that influence data – especially a model that is closely related to Hoey’s lexical priming (2005) – one must consider the potential influence of semantic processing effects such as the self-reference effect (SRE) and the group-reference effect (GRE). These effects are likely to play a role across a range of online media. The self-reference effect is the name given to the phenomenon that people can remember language strings better when they refer to the self rather than other semantic areas (Johnson et al., 2002). Similarly Johnson et al. (2002) describe a group-reference effect. In one study of the recall of adjectives, students were asked questions such as ‘Does this word describe you?’ (encoded as a selfreference question), ‘Does this word describe Hofstra2 students?’ (encoded as a group-reference question) and ‘Does this word mean the same as XXXX?’ (encoded as a semantic question). After fifty-four trials, which involved students taking a card and then answering one of the above questions with various words and distractor tasks, the students recalled words from the self-reference tests and the group-reference tests significantly better than words from the semantic tests. The effect was different for different groups, with smaller groups such as family showing a greater effect than large groups such as gender and religion. With an effect such as this in mind, we should be very cautious about assuming that lexical priming and selection processes around group-marking pronouns will be directly comparable to other noun-like entities (various lexical items acting as a noun or noun group). To explore lexical priming and lexical selection in online forums it is essential to pay careful attention to local context. In the remaining sections I offer more detail about the two data sets and explore how the meaning of we affects its behaviour as a component of larger units in this context.

THE LANGUAGE OF CITIZEN SCIENCE

67

4 DATA AND METHOD Zooniverse.org is a citizen science forum consisting of volunteers participating in short tasks such as labelling the shape of a galaxy. Typically when a Zooniverse user – or zooite as they often call themselves – has finished classifying a single item such as a galaxy or a photograph they are invited to ask questions or discuss the item or similar items further in a forum. A single forum discussion about dark matter (DM) – a hypothetical kind of matter that was first proposed in the early 1930s to explain the movement of galaxies (Zwicky, 1933) – was selected as the citizen science data for this chapter. An in-house script was used to scrape the data from the site as part of a larger project about language use in citizen science (Wright et al., 2016). It is one of the largest threads in the discussion part of the site with 545 posts and was initially chosen by the project team simply as one of the largest threads to test the script and for initial explorations of the data. With 1,227 occurrences of the item we it was confirmed as a suitable site to explore in terms of groups, group membership and the possible existence of selection units containing we in this online community. There are a total 200,526 tokens in the corpus when calculated by WordSmith Tools 6 (Scott, 2012) – different software will give slightly different figures based on their definition of a token. One piece of software, for example, may count a number such as twenty-two as two tokens whereas another would count it as one. The second data set from SB was introduced as a point of comparison for the behaviour of we are units in the DM corpus; it was judged to be comparable in terms of subject material and as online media. The SB corpus is made up of 103 million tokens and made available through Sketch Engine (Kilgarriff, Rychly, Smrz and Tugwell, 2004). Table 4.1 shows a sample of the data to illustrate the sort of language used in the citizen science forum. There are 124 occurrences of we are in the DM corpus or 618 occurrences per million words which were tagged using WordSmith Tools 6 (Scott, 2012) to indicate the meaning of we in terms of the people or group getting included in each case. A tag such a HUMAN, AUDIENCE or SCIENTIFIC COMMUNITY was added to every usage of we are based on close reading to see which group of people was included as ‘we’. The string we are was selected for this study as it is the most frequent we * 2-gram in DM and the second most frequent in SB with 31,860 occurrences or 259 occurrences per million. * we 2-grams would be important for further studies in both lexical priming and lexical selection but I begin with this structure because of an interest in how units containing verbs might be selected by the language users. Table 4.2 shows a sample concordance of we are in the SB corpus. A concordance of 124 random lines of the total 31,860 occurrences of we are

68

CONTEMPORARY MEDIA STYLISTICS

TABLE 4.1 Concordance showing sample of dark matter (DM) corpus 1 thread shortly but as yet

we are

living in an interest vacuum

2 needs collisions yet

we are

going in the wrong direction

3 masses of neutrinos. Yet

we are

close. It’s not too long a

4 50 billion light years.

We are

all holograms. Great news

5 departed this world.

we are

already at the midpoint of

6 Earth n>=3 and that’s why

we are

genetically>99% R. However

7 the right hand side. While

we are

digging I need to introduce

8 ordinary matter, of which

we are

made. Because it isn’t equi

9 of the proton from which

we are

all made. The effect is to

10 and neutrons from which

we are

assembled had to come from

11 allomatter around of which

we are

unaware with the same charges

12 out with experiments which

we are

still performing in our time

13 matter forms from which

we are

all made sings a tune almost

14 ordinary matter of which

we are

also made. The conservation

15 antineutrino of which

we are

most familiar, is associated

16 irrespective of whether

we are

here or not, now or then,

were selected for comparison. Similarly, each case of we are was tagged to mark which group of people was getting included as ‘we’.

5 USE OF WE ARE IN DM AND SB The referent of we was tagged in each line of data in the two corpora and in this section I present the main findings. I begin by highlighting the range of referents used in the two data sets before discussing three particular structures from the data we are HUMAN, we are AUDIENCE and reader/writer. We are HUMAN refers to structures where we means all people; we are AUDIENCE is used when we is used rhetorically to construct the writer and their audience as a group. Example (1) below is an example of we are AUDIENCE. In these cases there is no evidence that the writer is referring to all humans – the co-text suggests they are referring to themselves and a particular group of readers. The category reader/ writer is used when ‘we’ is used to only refer to the writer and a named second person in the group. Table 4.3 shows the results of labelling the data for the meaning of we in the two corpora. Arguably the most obvious difference is the number of categories

THE LANGUAGE OF CITIZEN SCIENCE

69

TABLE 4.2 Concordance showing sample of scienceblog (SB) corpus 1

Christians … I do not think

we are

majority so I appologize [sic]

2

and to simply believe what

we are

told. In this way, science

3

science is what we use when

we are

concerned that we might

4

the dollar is ‘struggling’

we are

sacrificing our liberties

5

I find it ironic that

we are

not on a Catholic site

6

So, since

we are

trying to teach our kids

7

we are carbon, we are ether We are saints, we are

kings

8

exemplar. Well, his god who

we are

told remains in control

9

of those consequences

we are

indeed gaining knowledge

10 considering he made us as

we are

) and needed to have a

11 to find the ‘ley lines’

we are

not all loons over here

12 quality of life. And here

we are

talking about a measurable

13 I should clarify that

we are

actually arguing over a

14 slaughtered): remember that

we are

dealing with individuals

15 we show to Evan that

we are

really quite nice people

16 the strong point that

we are

not special or unique,

listed. The close reading of many of the discussions and the decisions involved in working out the appropriate labels gave an overall impression that the users posting in DM were constructing themselves as a close group who knew other posters by name and were sharing knowledge rather than competing. This contrasts with the SB data which appears to have a clearer culture of debate with different groups presenting arguments and seemingly aiming to defeat their opponents. The data in Table 4.3 reflect this initial impression with a total of 111 lines of DM data either referring to humans as a species or using we are to include the poster and audience as a group as a rhetorical device, such as example (1) (1) Just as I am beginning to feel that we are getting to a point where at least some aspects of the discussion is … These are labelled audience. The two categories of HUMAN plus AUDIENCE come to a total frequency of just fifty-eight in SB with smaller categories such as geographical regions – the United Kingdom, the United States, Chile and others – and groups positioning themselves on either side of big debates such as abortion (one case of we refers to pro-choice people), global warming (one case refers to people who accept global warming is a problem caused by humans)

70

CONTEMPORARY MEDIA STYLISTICS

TABLE 4.3 Categories of meaning of we are data in DM and SB DM Groups and Occurrences

SB Groups and Occurrences

human

57

human

33

audience

54

audience

25 12

reader and writer

7

scientific community

private sector

1

USA

6

friends

1

group of posters

5

family

1

idiomatic

3

normal matter

1

family

3

French people

1

Christians

2

reader and writer

2

UK

2

NB: Twenty-eight additional categories are not shown in the SB list because they only have one member. Omitted categories include people with schizophrenia, people who support science, atheists, the northern hemisphere and Israel

and religion making up the remainder (six cases comprise groups labelling themselves as Christians, secularists, freethinkers, atheists and anti-theists). This supports a sense of debate and conflict that I got from close reading. The differences shown in Table 4.3 will not come as a complete shock because SB covers a much wider range of scientific subjects than the DM data. My point is to show that this difference in group formation clearly shows during referenttagging analysis, and that the different social and semantic environment can influence language patterns. It is also not obvious that such a difference would show clearly from 248 occurrences of the word we. Of the DM’s 200,526 tokens, we are and we’re occur 124 and 4 times, respectively, or at a ratio of 31:1. By contrast, of the SB’s 103 million tokens, there is a much lower ratio of we are to we’re, with the two occurring 31,860 and 25,730 times, respectively, or at a ratio of 1.24:1. This suggests a rather unusual note of formality for online communication. One possible explanation could be that one or two influential users started selecting the we are form over we’re and it spread across the DM thread or there is a wider tendency across Zooniverse for users to project a more formal image. This could be seen as a marker of education and expertise. The explanation may also be a combination of these two factors. Follow-up work on this question will prove useful in terms of getting a better understanding of how powerful members of a forum influence others in terms of language choice.

THE LANGUAGE OF CITIZEN SCIENCE

71

5.1 We are HUMAN The most frequent category in both corpora is HUMAN, that is, the posters are constructing their main group as the whole of humanity when they use strings containing we are. There are fifty-seven lines of HUMAN data in DM compared to just thirty-three in SB – recall that we began with 124 lines from each corpus so it would have been reasonable to predict similar numbers. For example: (2) From a biological viewpoint we are created from the present (3) mace to the head. In practice we are in the ancient world, a (4) a common ancestor for apes, we are one species of them, as These lines suggest that in DM the posters see themselves and their audience as a single group, and selecting we are over alternative forms such as humans are may reinforce in-group communality. SB appears very similar in the way posters are using we are to refer to the human species in discussions about our place in the universe and evolution but a small number of lines such as (5), (6) and (7) reflect the culture of debate with some posters expressing caution about science that is not seen in DM. (5) my assertion is that, if we are to rely upon science for answers (6) to spew out. Sanford said we are the result of two mosquitoes hooking up (7) The whole theory of Darwin, that we are descended from monkeys, has been finally debunked The most notable repetition of a 3-gram in either HUMAN category is the set of seven occurrences of that we are in SB (see Table 4.4); there are just two in DM. This makes it a good candidate for a selection unit or part of a selection unit that appears to be favoured in the SB environment. Hadikin (2015) highlights Sinclair’s extended lexical unit (Sinclair, 1996) as a possible candidate for a selection unit so, with this in mind, it makes sense to reflect on the status of that we are. Six out of the seven occurrences refer to an abstract thought in positions to the left of that, such as point, idea, reason and theory, so it is reasonable to suggest this semantic area is part of the unit as shared by these writers in this context. One 3-gram (three-word expression) that stands out in the DM HUMAN data is shown in Table 4.5 – which we are. This is a small set of data, of course, but there is not a single line of which we are in the thirty-three lines of comparable SB HUMAN data. We can note a fairly simple colligation in that three lines of the data show which we are preceded by of and three preceded by from. Further to the left the semantic area of MATTER is appearing with references to matter, protons and neutrons, allomatter, and proton. To the right of

72

CONTEMPORARY MEDIA STYLISTICS

TABLE 4.4 All seven occurrences of that we are in the thirty-three lines of SB HUMAN data 1

the strong point

that we are

not special or unique,

2

it is the idea

that we are

all more alike than we a

3

about the idea

that we are

supposed to know in advance

4

The reason

that we are

so unique among primates

5

have all the toxins

that we are

exposed to on a daily basis

6

whole theory of Darwin,

that we are

descended from monkeys,

7

embraced theory

that we are

created to respond to vibrational

TABLE 4.5 All six occurrences of which we are in the fifty-seven lines of DM HUMAN data 1

ordinary matter, of

which we are

made. Because it isn’t equi abundant

2

ordinary matter of

which we are

also made. The conservation

3

and neutrons from

which we are

assembled had to come from

4

allomatter around of

which we are

unaware with the same charges

5

matter forms from

which we are

all made sings a tune almost

6

of the proton from

which we are

all made. The effect is to

which we are the semantic area of MADE appears in five out of the six lines. An exploration of the right side of we are in DM HUMAN shows six cases of we are all which includes two cases of from which we are all made – each one may reflect the individual being aware of at least two units in the discourse – MATTER from/of which we are MADE and we are all. Seven occurrences of we are made occur in the data but they contain the string that’s simple as we are made of ordinary matter being repeated three times as a single poster refers to experiencing an editing glitch. This is clearly important information in terms of the dispersion of the unit but note it is also quite a normal part of computermediated communication and is likely to have a priming effect on readers the more frequently they are exposed to the structure. 5.2 We are AUDIENCE We are AUDIENCE is the second largest category in both data sets but it raised more tagging challenges than the we are HUMAN set. In DM the speakers were generally addressing people they knew in the DM community or using we are in a rhetorical function such as line (8) from DM AUDIENCE:

THE LANGUAGE OF CITIZEN SCIENCE

73

(8) Were GUT unification at 10^17GeV where U(1) might intersect in some other universe then there would be no observed preferences, everything annihilates and we are left with a universe of photons and no matter particles at all, neither ordinary nor allomatter! Table 4.6 shows a sample of the DM AUDIENCE data. In all cases the writers appear to be addressing an imagined site audience but there is no evidence that we refers to all humans. The situation was slightly different for a small number of cases in SB. Consider the difference between example (9) and example (10): (9) also find it very odd that we are being expected to accept blogposts as scientific publications Cases such as (9) were classed in the AUDIENCE category for SB. As with (8) the poster appears to be addressing the imagined site audience. Case (10) was labelled differently because the we group is also marked as being members of the scientific community. These cases were labelled SCIENTIFIC COMMUNITY.

TABLE 4.6 Sample of we are data from fifty-four lines of DM AUDIENCE data 1

Without a common ancestor

we are

left with either, DM

2

everything annihilates and

we are

left with a universe of photons

3

quantum fermion repulsion and

we are

assuming it is perfect,

4

In their interactions and

we are

moving into the weak force

5

On states of matter and as

we are

concerned with hydrogen,

6

which we wouldn’t be aware.

We are

of course aware of the

7

take a long time, basically

we are

on course, steady as you go

8

particles, in this case

we are

talking about the B^0photon

9

stars or galactic centres.

We are

searching for something quite

10

wait long but we have days.

We are

going to look at two atoms

11

before, then by definition

we are

not in the earliest Early

12

its gravitational effects.

We are

dealing with light years

13

more than the 92 elements

we are

familiar with. They aren’t

14

Atoms being energy,

we are

in effect talking of energy

15

of Kadesh. Fair enough,

we are

all aware of PR eg. the war

16

plug it into an equation.

We are

still dealing with life scientists

74

CONTEMPORARY MEDIA STYLISTICS

(10) can not believe how advanced in science we are. I work in a drug rehab facility (In most of these cases I had to read a considerable amount of the full blog post to understand context.) In the rest of this section AUDIENCE has been combined with SCIENTIFIC COMMUNITY for SB data and compared against the AUDIENCE category in DM. The use of that we are as a structure is less frequent in the SB AUDIENCE + SCIENTIFIC COMMUNITY data set with just four out of thirty-six occurrences (11 per cent) compared with seven out of thirty-three (21 per cent) in the SB HUMAN set (see Table 4.7). This difference in the meaning of we appears to break the semantic link with abstract thought that was discussed in Section 5.1, although strictly speaking the difference in meaning is associated with the lack of semantic association and there is no evidence of one causing the other. A difference in the grammatical patterning is also apparent; in the SB HUMAN data all seven occurrences saw that we are functioning as a relative clause with structures such as the toxins that we are exposed to, the strong point that we are not special and the idea that we are all more alike. None of the four occurrences in this data set involve that we are in such a pattern. Analysis of the R1 position (i.e. the word that follows we are in the right slot) in SB AUDIENCE + SCIENTIFIC COMMUNITY raises a curious aspect of lexical selection due to the presence of four cases of we are told. This is the only 3-gram (the only three-word expression) in the set with four or more occurrences. What makes it curious is the real-world issue that arguably prevents it occurring in the SB HUMAN set; it is difficult to imagine a situation where the writer would use the structure with we to mean all of humanity because there would be no one left outside the group to do the telling. If one were to write ‘we are told that if you eat too much’, for example, but we means all humans there would need to be another species advising us on our eating habits. Of the four data sets discussed so far only SB AUDIENCE + SCIENTIFIC COMMUNITY has any occurrences of we are told, so this may suggest a community cautiously TABLE 4.7 All four occurrences of that we are in SB AUDIENCE + SCIENTIFIC data 1

efforts and be sure

that we are

not producing more MD/PhDs

2

so find it very odd

that we are

being expected to accept blogposts

3

everything, rather,

that we are

pretty damn sure about some things

4

slaughtered): remember

that we are

dealing with individuals

THE LANGUAGE OF CITIZEN SCIENCE

75

TABLE 4.8 All four occurrences of we are told in SB AUDIENCE + SCIENTIFIC COMMUNITY

1

Clergy and scientists,

we are told,

‘warned that it was as impious

2

blood pH homeostasis, yet

we are told

that if you eat too much

3

exemplar. Well, his god who

we are told

remains in control

4

to simply believe what

we are told.

In this way, science and religion

reporting claims they are aware of in society but, looking at the context shown in Table 4.8, they appear to be distancing themselves from the claims. With references to Clergy and scientists, his god and science and religion in the cotext in three cases we again see tensions between science and religion showing in the SB corpus which is not apparent in DM. The 3-gram we are in occurs in all four data sets we have looked at in this chapter – the two HUMAN data sets and the two AUDIENCE data sets. Table 4.9 shows the four occurrences in DM AUDIENCE. It is highlighted here as a very familiar priming that does not show any significant variation based on the meaning of we. At this point in particular some readers may begin to think ‘this is just English grammar’, some may argue that it is ‘natural’ for a prepositional phrase to follow a verb such as are. As linguists, however, we do not know for sure that the structure in contact, for example, is stored as a separate unit for this writer from the structure we are (see the top line of Table 4.9). The idea that this prepositional phrase operates as a full selection unit for this particular writer and his or her community or that they have no primings to add particular structures to the left would be an assumption that I am hoping to make more explicit in this chapter and future work on lexical selection. The writer may be primed to produce we are in contact or even we are in contact with as a unit. All one can say for certain is that the writers have selected the forms in Table 4.9 because they were primed in two ways – they are likely primed to accept are as an appropriate form to follow we – and they were not motivated to change the wording after any reflection or self-editing before pressing ‘post’. In these cases there can be no further claims made about the writer’s priming without other texts by the same writer. The form we are in effect stands out in Table 4.9 as the prepositional phrase in effect breaks up a verb phrase. Though not an area of focus for this chapter it would be interesting for future researchers to explore details of the statistical chance of different verbs associating with such an adverbial and what this tells us about the nature of selection units. The data in Table 4.9 also suggests a potential collocational relationship with the as we see we are in the bottom left corner and we are in the ball park. The DM HUMAN set has three occurrences

76

CONTEMPORARY MEDIA STYLISTICS

TABLE 4.9 All four occurrences of we are in in DM AUDIENCE 1

electromagnetic receptors.

We are in

contact with a generation

2

Atoms being energy,

we are in

effect talking of energy

3

later subtle argument since

we are in

the bottom left corner

4

rest mass at 511000 eV so

we are in

the ball park for low energy

of we are in – we are in a vacuum, we are in the ancient world and we are in a manner stuff which suggests a possible colligation with articles in the R1 position following we are in. SB AUDIENCE + SCIENTIFIC COMMUNITY contains just a single structure we are in the United States which supports the idea of the colligation with articles. SB HUMAN has two cases – we are in love and we are in control of. The structures we are in and we are in ARTICLE are candidates as selection units or partial selection units that can survive in all four environments. Perhaps it makes more sense to suggest that, as a structure becomes common in a wide range of diverse contextual environments, we say ‘this is just English grammar’ as shorthand rather than using our everyday knowledge of grammar as an argument against such detailed descriptions of the behaviour of units in context. 5.3 Reader/writer The smaller categories in Table 4.3 include cases where we refers to a single reader/writer in both DM and SB. In these cases close reading of the whole thread showed there were just two people having a discussion. All data tagged as reader/writer are shown in Table 4.10. Note that Table 4.10 differs from other tables in Sections 5.1 and 5.2 in that it does not show a subset of a large category such as HUMAN or AUDIENCE. The table shows all the examples from this category, so it is important to note that this is an unusually small set. This category shows data in which we are was getting used in a discussion with two people or the writer explicitly says they are visualizing two people to make a point and nothing in the context or co-text suggests the writer is considering the wider audience as part of the group in the sense of the meaning of we. The AUDIENCE category, by contrast, contains either situations that suggest the writer is addressing a larger group or ambiguous situations. In DM, a single conversation dominates these data as one writer asks the reader to imagine both himself and the reader sitting atop two helium atoms as he makes several points based on this imagined journey of two particles moving through the atmosphere. Recall that we see SREs and GREs in Johnson et al. (2002) – participants have a better memory of lexical items that refer to themselves or to small groups such as their family. Similarly, Hoey (2005) argues that lexical priming

THE LANGUAGE OF CITIZEN SCIENCE

77

TABLE 4.10 All nine lines tagged as reader/writer in the data 1

centre. Density is a r^3 relation,

we are

What about pressure. Imagine another

2

the right hand side. While

we are

digging I need to introduce

3

it’s not gravitation is it?

We are

in freefall. Gravity attracts!

4

your unlucky number unless

we are

born on the same date of

5

greater the air pressure?

We are

now going to dig an even deeper

6

pressure needs collisions yet

we are

going in the wrong directions

7

but I am at 64ft (1/2ft^2).

We are

both accelerating under gravity

8

would you. As you pointed

we are

used to their sites/individuals

9

origin becomes valid. Since

we are

bandying about quotes

is context specific, it is reasonable to suggest the specific context of a writer visualizing himself and a reader in this situation creates a space where certain lexical forms thrive; in Table 4.9 we see a possible priming for the continuous aspect of verbs following we are. We are both accelerating, we are digging, we are going and we are now going comprise the colligation with we are in freefall adding to a semantic association of movement that affects five out of seven lines in DM. This is not a complete surprise in a discussion of movement but recall the data was not selected because of references to movement. The data were tagged as reader/writer and the colligation and semantic association came along with the meaning of we in a way that would not easily have been predicted before this study. The presence of since we are bandying about quotes in SB also has a continuous form in R1 position and a metaphor of movement which provides further support that this may form part of a we are CONTINUOUS MOVEMENT unit getting shared in science forums in situations where two interlocutors are involved or being imagined. As a final note about the reader/writer set, the lines from DM show no signs of tension or disagreement between any writers or forum members but in SB – in just two lines of data – we see (11) we are used to their sites/individuals churning out crap and (12) since we are bandying about quotes which both suggest an atmosphere of tension between the two interlocutors and an outside group in (11) and between interlocutors in (12). It would be interesting to note the extent to which this evaluative information is carried along with the unit in further work on the SB corpus or similar websites.

78

CONTEMPORARY MEDIA STYLISTICS

6 CONCLUSION This chapter has examined the extent to which the meaning of we in context might influence the lexical and grammatical patterns that surround the structure we are. Introducing the lexical selection model (Hadikin, 2015) and using the Zooniverse DM and SB data, I explored how these two different linguistic environments and the two different communities may allow for different linguistic units to emerge and be utilized in similar or divergent ways. The analysis showed the DM corpus has thirty-one occurrences of we are for each occurrence of the contracted form we’re. There are 1.24 occurrences of we are for each case of we’re in SB. Their uses include the following: ●

●

●

The most frequent 3-gram (three-word expression) in the SB HUMAN data is that we are which appears to form part of a selection unit in this context comprising the 3-gram plus the semantic area IDEA in L1 position. It is reasonable to suggest an atmosphere of debate contributes to the use of this structure. The most frequent 3-gram in the DM HUMAN data is which we are. This appears to form part of a larger unit MATTER from/of which we are made in five out of six cases. The semantic part of a possible IDEA that we are structure is absent in all four occurrences of that we are in SB AUDIENCE +SCIENTIFIC suggesting a link between the referent of we are and nearby co-text. The data in the reader/writer category appears primed to attract continuous forms such as we are both accelerating and we are digging in DM providing further support for the claim that the referent of we affects co-text.

The chapter focused on four categories of we uses in the corpora: DM HUMAN, SB HUMAN, DM AUDIENCE and SB AUDIENCE + SCIENTIFIC. Research on the GRE and SRE (Johnson et al., 2002) shows that people have better memory of lexical items that refer to a group they belong to (group-reference) or themselves (selfreference) so it is reasonable to argue the group being referred to will also affect lexical priming (Hoey, 2005) in a way that has not been previously explored. Analysis has shown that the two corpora contain very different kinds of data – the DM corpus suggests a supportive atmosphere where group members tend to support one another and get along whereas SB has a more striking atmosphere of tension and debate. Papacharissi’s (2004) suggestion that ‘the majority of the participants appreciated these online debates, because they provided them with the opportunity to hone their argumentation skills’ (277) is more in line with the SB corpus. The participants’ language in DM suggests they were there to learn physics from each other and share ideas rather than hone their argumentation skills. A low-tension environment might appear more conducive to Zooniverse research and the overall success of a citizen science

THE LANGUAGE OF CITIZEN SCIENCE

79

website, but a combative atmosphere could also, in theory, produce desirable results. More research is needed to fully understand the effect of this variable. A number of three-word structures were selected as starting points for analysis based on relatively high frequency in each of the data sets – these include that we are, which we are, we are told and we are in. Each of these structures occurred in the data four times or more. I suggest these are good candidates as selection units (or partial selection units) and the patterns they occur in – colligation patterns and semantic association – suggest the linguistic environment supports replication of the structure. The string which we are in the DM HUMAN category, for example, suggests the movement of a MATTER from/of which we are made unit that does not occur in the other three main data sets. (A single case of electron anti-neutrino of which we are most familiar in DM AUDIENCE suggests the same basic structure but with space for variation on the right side.) The strings discussed in this chapter have likely moved via the ‘standard output’ route shown in figure one. A positive mutation would likely come across as a creative use of language in the context and a negative mutation would be seen as a mistake or poor use of the structure by others in the community. The key finding from this chapter is the value of marking the meaning of we in short structures and the insights that can be gained from looking at this with lexical priming and lexical selection in mind. Although this chapter has focused on the lexical item that follows we, there is nothing special about this lexical material other than telling the analyst how we are might be used to construct group identity. For a more complete understanding of how lexical priming and lexical selection explain online forum interaction, one could focus on the lexical items that precede we. That would likely pick up data from the end of clauses and lead to a discussion of selection processes affecting a writer at the point of text creation before they produce we. One might also carefully track an individual forum user over time, because although every text contains evidence of the user’s primings and selection processes, a single forum user may have a tendency to produce structures such as we are in and we are used to as high frequency structures across forums. Further work with individual language users would provide a valuable new insight on lexical priming. Although this chapter highlights similarities between language and evolution, one should not expect a selection model of language to be identical to the movement and mutation of DNA. Lexical selection encourages researchers to look closely at the context and reflect on the linguistic options available to a language producer at any given moment. Analysis of language use in forums like Zooinverse.com can provide insight about sub-genres of online communication, like talk about science in online forums, but also more broadly about how online communities are subtly constructed in language.

80

CONTEMPORARY MEDIA STYLISTICS

NOTES 1. Very similar in this chapter is used in the same way as Hadikin (2015: 3) in the sense that the string is perceived as being within the range of possible phonological realizations so that most other language users in the speech community would generally see it as being the same form and not notice a difference. 2. The study took place at Hofstra University, USA.

REFERENCES Baxter, G., Blythe, R., Croft, W. and McKane, A. (2009). ‘Modeling language change: An evaluation of Trudgill’s theory of the emergence of New Zealand English’. Language Variation and Change, 21(2), 157–96. Bryson, B. (1991). Neither Here nor There: Travels in Europe. London: Secker and Warburg. Dawkins, R. (1976). The Selfish Gene. Oxford: Oxford University Press. Hadikin, G. (2014). Korean English: A Corpus-Driven Study of a New English. Amsterdam and Philadelphia: John Benjamins. Hadikin, G. (2015). ‘Lexical selection and the evolution of language units’. Open Linguistics, 1, 458–66. Hoey, M. (2005). Lexical Priming: A New Theory of Words and Language. London: Routledge. Johnson, C., Gadon, O., Carlson, D., Southwick, S., Faith, M. and Chalfin, J. (2002). ‘Self-reference and group membership: Evidence for a group-reference effect’. European Journal of Social Psychology, 32(2), 261–74. Kilgarriff, A., Rychly, P., Smrz, P. and Tugwell, D. (2004). ‘The sketch engine’. In Proceedings of Euralex 2004, Lorient, France, pp. 105–16. Myles, F. and Cordier, C. (2016). The Psycholinguistic Construct of Formulaicity in Second Language Learners: Issues of Conceptualisation and Identification. Retrieved from http://files.flarn.org.uk/formulaicity.mp4. Papacharissi, Z. (2004). ‘Democracy online: Civility, politeness, and the democratic potential of online political discussion groups’. New Media & Society, 6(2), 259–83. Scott, M. (2012). WordSmith Tools Version 6. Stroud: Lexical Analysis Software. Simpson, R., Page, K. and De Roure, D. (2014). ‘Zooniverse: Observing the world’s largest citizen science platform’. In Proceedings of the 23rd International Conference on World Wide Web Companion, Seoul, Korea, 2014, pp. 1049–54. Sinclair, J. (1996). ‘The search for units of meaning’. Textus, 9, 75–106. Wray, A. (2002). Formulaic Language and the Lexicon. Cambridge: Cambridge University Press. Wright, S., Clarke, B., Hadikin, G., Saraceni, M., Viggiano, C. and Williams, J. (2016). Language of Citizen Science. Retrieved from http://www.port.ac.uk/centre-for-eu ropean-and-international-studies-research/research-projects/language-of-citizenscience/ Zwicky, F. (1933). ‘Die Rotverschiebung von extragalaktischen Nebeln’. Helvetica Physica Acta, 6, 110–27.

CHAPTER FIVE

The pragma-stylistics of ‘image macro’ internet memes JANE LUGEA

Internet memes are a discourse type emblematic of participatory media; they are user-generated, multimodal texts which are shared, replicated and reconfigured across national, linguistic and cultural boundaries. Adopting an existing framework for the analysis of internet memes (Shifman, 2014), this chapter begins by revealing the stylistic features of popular ‘image macro’ families, that is, patterned combinations of image and text which, through widespread (re-)production, have become an established sub-genre, cohering around a certain fixed set of features, called ‘cohesive quiddities’ (Segev et al., 2015). First, Shifman’s model is elaborated with reference to the work of Erving Goffman and applied to the top five most popular image macro families in general. Once their core characteristics have been established, I examine the most liked and disliked versions of each, employing user evaluations as a measure of their success and in order to better understand the rules governing the creation of image macros. In providing a pragma-stylistic account of image macros, this chapter advances the analytical toolkit available to stylisticians dealing with digital texts and contexts.

82

CONTEMPORARY MEDIA STYLISTICS

1 INTRODUCTION Although the term ‘meme’ has an interesting and varied history, it is most recently applied to the proliferation of replicated and re-worked multimodal texts in online contexts. Until now, memes have not been subject to any stylistic analysis despite, I argue, their fit with the core concern of stylistics: linguistic creativity, as expressed through users’ grasp of complex pragmastylistic rules and ability to break those rules – in the ‘right’ ways – to serve communicative effects. This volume offers a unique opportunity to explore the pragma-stylistics of memes, focusing on the most successful ‘image macros’ and selecting their most liked and disliked iterations to better understand their core characteristics. In defining ‘pragmastylistics’, Hickey observed how, ‘utterances with the same, or virtually the same, meaning may differ in their linguistic form and situational appropriateness, and these differences may have either stylistic or pragmatic explanations’ (1993: 578). The different degrees of success of different versions of the same meme family can therefore be explained through a pragma-stylistic approach.

2 LITERATURE REVIEW 2.1 The meaning of ‘meme’ While the internet users of today might be familiar with ‘memes’ in reference to replicated online content, the term has an older and broader significance which is necessary to outline in order to understand its current status in academic research as well as in popular usage. The term ‘meme’ was coined by biologist Richard Dawkins in his influential book The Selfish Gene (1976), by shortening the Greek word mimema, meaning ‘something which is imitated’. The rhyme with ‘gene’ served to emphasize Dawkins’s proposal that ‘meme’ describes a unit of human cultural evolution analogous to the gene. In Dawkins’s view, like genes, memes spread through human culture by imitation and a process of Darwinian selection. In the extreme form of this metaphor (A UNIT OF CULTURE IS A GENE) ideas, tunes, poems, fashions and behaviours replicate, mutate and exist above and beyond the human individual, who is nothing more than an infected carrier. The last two decades of the twentieth century saw the emergence of memetics as a field dedicated to understand the spread of information and culture through memes, culminating in the relatively short-lived Journal of Memetics, 1997–2005. A concept coined by a biologist was developed and applied by psychologists, philosophers and anthropologists in order to explore the spread of culture and ideas. As the reader might expect, taking Dawkins’s metaphor too literally can have serious implications for understanding (or misunderstanding) the transference of information and human culture. For a start, evolutionary biology is not

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

83

analogous to culture and society. Viewing humans as ‘hosts’ (Blackmore, 1999) undermines the role of human agency in cultural and informational production, a point made by several critics of memetics, including J. R. Searle (1997: 105): [T]he spread of ideas though imitation requires the whole apparatus of human consciousness and intentionality. Ideas have to be understood and interpreted. And they have to be understood and judged as desirable or undesirable, in order to be treated as candidates for imitation or rejection. … The whole process normally involves language with all its variability and subtlety. In short, the transmission of ideas through imitation is totally unlike the transmission of genes through reproduction, so the analogy between genes and memes is misleading from the start. As well as providing a sound critique of the analogy of genes and memes, the philosopher and pragmaticist makes two important points pertinent to my thesis in this chapter: notwithstanding the misguiding origins of the term ‘meme’, the spread of ideas is performed through (a) the subtleties of language and (b) processes of interpretation and appraisal, which are vital for their transmission and further spread. Thus, it would appear that in order to fully understand memes (in their current online form), we must consider their textual features and pragmatic functions. Other critics of memetics included semioticians such as Deacon (1999) and Kull (2000), who pointed out that the concept of ‘sign’ already describes what ‘meme’ purports to signify; like memes, the sign can be copied, and has the added value of distinguishing between the signifier and the signified, as well as a long history of research into semiotic structure and use. Perhaps even more worryingly, what exactly a ‘meme’ signifies in memetic research – an idea, a unit of information, an observable behaviour – is unclear, as is the scope of its meaning, that is, the defining ‘unit’. I argue that with a century of academic endeavour towards describing, classifying and theorizing units of language, linguists have plenty to contribute towards a more rigorous and scrupulous study of memes, if indeed it can be ascertained what they are. The new popular usage of the term ‘meme’ specific to an internet text, which denotes a more concrete, text- or genre-bound (Wiggins and Bowers, 2014) definition, may be the impetus needed to inspire linguists to pay attention to these artefacts and their use in online discourse. 2.2 Internet memes While the controversies waged around the meme in the academic field of memetics, the term gained fresh currency in non-academic circles, where it was adopted by internet users to describe ‘“catchy” and widely propagated ideas

84

CONTEMPORARY MEDIA STYLISTICS

and phenomena’ (Knobel and Lankshear, 2007: 201). Of course, the internet is abound with texts of this sort, never more so than since the invention of Web 2.0 where users are not passive receivers of on-screen text but can like share, create and reformulate multimodal content. As Shifman notes, ‘What Internet users seem to have grasped – and Richard Dawkins couldn’t have imagined back in 1976 – is that the meme is the best concept to encapsulate some of the most fundamental aspects of the Internet in general, and of the so-called Web 2.0 culture in particular’ (2014: 18). Indeed, Dawkins has since remarked that the term he coined has been subject to memetic transformation in its application to online discourse, or as he puts it, has ‘mutated’ (2013). Internet memes can be distinguished from more general cultural memes by their propensity to user adaptation (a feature of the new media we use to communicate). This paradox between internet memes’ capacity for both cohesion across iterations and, at the same time, uniqueness has since been noted by several scholars (e.g. Davison, 2012; Segev et al., 2015). I argue that this tension between formal patterns and creative alterations makes internet memes prime candidates for stylistic analysis, where style is understood as a result of patterned choices in expression. This study focuses on ‘image macros’ in particular, which Davison defines as ‘a set of stylistic rules for adding text to images’ (2012: 123), adding that ‘what is replicated from instance to instance is the set of formal characteristics’ (2012: 130). Given that image macros are defined according to their stylistic properties, this study aims to ascertain what those are. Returning to the subject of internet memes in general, Knobel and Lankshear (2007) identified specific characteristics pertaining to internet memes: intertextuality, anomalous juxtapositions and humour. Their findings have been confirmed in subsequent studies. Supporting the presence of incongruity and resultant humour in internet memes, Lou (2017) analyses image macros which use the text ‘[That awkward moment] when …’ and complete the multimodal simile with an image. He finds that their ‘communicative and rhetorical function … is to make a humorous comparison between an everyday activity and a strange photo’ (2017: 121). Other scholars identify specific kinds of humour as prevalent in internet memes, namely absurdity and parody (Jenkins et al., 2009; Shifman, 2011), although without recourse to rigorous accounts of these forms of humour. Simpson (2013) advances one such account of parody, defining it as a form that involves replicating characteristics of an anterior discourse. It seems clear then that digital media is a platform on which parody can flourish through the adaptation and sharing of internet memes. Shifman remarks, ‘The singular attributes of the internet – on which copying and imitating texts have become workaday activities – turned these genres [parody and pastiche], previously produced by the dedicated few, into a cultural logic shared and employed by the many’ (2011: 190–1). This observation of the significance of parody in digital media is drawn

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

85

from Shifman’s investigation of successful YouTube memes, which she found to have several features in common: ordinary people, flawed masculinity, humour, simplicity, repetitiveness and whimsical content. Several of these characteristics will also bear relevance for this study of the most popular image macros. Despite their popularity among digital communicators and their potential to offer fascinating insights to linguists ‘of all stripes’ (Lou, 2017), internet memes have received limited attention in linguistic research (Shifman, 2013: 363), although that lacuna is being redressed more recently. Memes have been explored by linguists using a cognitive (Dancygier and Vandelanotte, 2017; Lou, 2017) or social lens (Blommaert, 2015; Blommaert, 2018; Varis and Blommaert, 2015). Varis and Blommaert (2015) emphasize the ‘phatic’ function of memes and viral texts, whose meaning may matter less than their expression of group membership. Dancygier and Vandelanotte (2017) and Lou (2017) analyse a range of image macros, describing the processes a reader goes through to create meaning from the multimodal texts. While all these studies recognize the significance of memes’ interactive contexts, little attention has been paid to the pragmatics of internet memes (i.e. how they are used), which the present study goes some way to addressing, by examining liked/disliked iterations of the top five image macros. 2.3 Analytical framework Unsurprisingly, internet memes have received most attention from researchers in computer-mediated communication, where the definition of meme is drawn from the vernacular of digital communicators. While she recognizes their status as a ‘conceptual troublemaker’ (2013), Limor Shifman, in particular, has contributed greatly to our understanding of internet memes’ form and function (2011, 2013, 2014). She advances a very useful framework for their analysis based on a comprehensive three-part definition of the internet meme: a) [A] group of digital items sharing common characteristics of content, form and/or stance, which b) were created with an awareness of each other, and c) were circulated, imitated and/or transformed via the internet by many users. (Shifman, 2014: 41) Unlike the slippery object of research in memetics, Shifman’s definition captures the distinction between an individual iteration of an internet meme text and the group of texts with which it coheres. In fact, she helpfully distinguishes between a viral text and a meme on the basis of iterations: ‘Whereas the viral comprises a singular cultural unit (such as a video, photo or joke) that propagates in many

86

CONTEMPORARY MEDIA STYLISTICS

copies, an Internet meme is always a collection of texts’ (2014: 57) which inspire ‘extensive creative user engagement in the form of parody, pastiche, mash-ups or other derivative work’ (Shifman, 2011: 190, author’s italics). In further contrast to most research in memetics, Shifman recognizes the role of user engagement in the (re)production of internet memes, underscoring the fact that the people make ‘choices’ in their production. Most significantly for a pragma-stylistic analysis, she categorizes the features which are shared across groups of memes into three ‘mimetic dimensions’ along which cultural items can be imitated: content, form and/or stance. Shifman describes how content refers to the ‘ideas and ideologies conveyed’, form is ‘the physical incarnation of the message, perceived through our senses’ and stance encapsulates ‘the ways in which addressers position themselves in relation the text, its linguistic codes, the addressees and other potential speakers’ (2014: 40). It is clear that these dimensions help delineate the ways in which memes carry ideas and assert positions, and they provide the framework for the pragma-stylistic analysis in Section 4. Recognizing that ‘stance’ is a very broad category, Shifman (2014: 40–1) divides it into three sub-dimensions, drawing on other frameworks: i. participation structures, indicating who is entitled to participate and how (Philips, 1972); ii. keying, which captures the tone and style of communication (Goffman, 1974/1986); and iii. communicative functions, according Jakobson’s (1960) typology. Again, Shifman indicates clear ways in which internet memes can be used to take a stance in interaction, and the frameworks listed here are already in use by scholars of pragmatics and linguistics in general. The toolkit Shifman advances has enormous potential for the close analysis of memes, yet she does not apply it as systematically as a stylistician might. While she advances an incredibly insightful and comprehensive account of internet memes, the framework has untapped potential for a close, pragma-stylistic account of memes and image macros in particular. Moreover, I contend that there are some additional concepts which can be borrowed from Goffman to help us consider memes in interaction. For example, Shifman’s brief description of (i) participation structures can be enhanced through Goffman’s account of the ‘production format’ of an utterance (1979/1981: 145). Goffman notes that the terms ‘speaker’ and ‘hearer’ are not nuanced enough to capture the complex roles that participants actually enact in discourse. As such, among the various types of speakers, he distinguishes between an ‘animator’ as ‘an individual active in the role of utterance production’ and an ‘author’ as ‘someone who has selected the sentiments that

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

87

are being expressed and the words in which they are encoded’ (1979/1981: 144). These will prove useful concepts in our analysis of image macros (Section 4), where I use ‘author’ to refer to the person who creates each instantiation of an image macro (including the overlaying meme phrase), and ‘animator’ to refer to the character who features in the image macro and is attributed with voicing the meme phrase. Shifman also lists (ii) keying as significant in the enactment of stance through memes, but the concept’s relevance to memes is not developed nor applied, which I do so more fully here. Within Goffman’s frame analysis (1974/1986), frames are used to organize experiences so that participants can define situations. Keying, then, is the process by which we understand one frame in terms of another. Goffman’s primary example of keying is ‘makebelieve’, at which ‘playfulness’ is at the heart – that is, ‘the relatively brief intrusion of unserious mimicry during interaction between one individual and others’ (1974/1986: 48). Forty years ago, he recognized ‘the growing use of replicative’ media (1974/1986: 68) and identified ‘technical redoings’, which include recordings, demonstrations and simulations, as additional examples of keying. Brooks (2007: 10) points out that Goffman’s frame analysis ‘afford[s] strong potential for integrating the study of virtual action and interaction with much of what is already known about social action and interaction more generally’, while keying in particular can explain how ‘people interact with simulated images and processes’. The concept of keying has clear significance for digital discourse and mimetic practice. Based on Goffman’s understanding of different performances of the same play, I suggest individual iterations of memes are, likewise, ‘keyings of a common model’ (1974/1986: 78). The last indicator of stance in Shifman’s model requires less elaboration: (iii) communicative functions. Jakobson proposed six fundamental functions of language and, as they will underscore the analysis of image macros, they are summarized here. The referential function is the role of language to denote something, that is, to carry referential meaning. Jakobson describes the expressive function as ‘a direct expression of the speaker’s attitude towards what he is speaking about. It tends to produce the impression of a certain emotion whether true or feigned’ (1960: 354). While the expressive function is more focused towards the addressee than the referential, the conative function is fully directed towards the addressee and usually finds its form in the vocative and imperative moods (indicating an ‘addressed recipient’ in Goffmanian terms). Less relevant to the present data are ‘messages intended to establish, prolong or discontinue communication’ (1960: 355), which serve a phatic function (although see Varis and Blommaert [2015] on the phatic function of internet memes and viral texts). ‘[S]peaking of language’ serves a metalingual function (1960: 356), whereby speakers discuss the code itself. Finally, and most importantly for stylisticians, it the poetic function which is fulfilled by

88

CONTEMPORARY MEDIA STYLISTICS

focusing ‘on the message for its own sake’ (1960: 356) and many decades of stylistic research has demonstrated how marked choices in expression serve to foreground the message itself and to prolong the attention paid to it by addressees. The poetic function, as Jakobson and scholars in stylistics argue, is particularly prominent in literary texts, but can also be found in other communicative contexts. I argue that highly competent authors of memes, including image macros, are skilled at using the genre to fulfil the poetic function and will demonstrate this in Section 4.2. Incorporating Shifman’s three dimensions in their quantitative study of internet ‘meme families’, Segev et al. identify ‘cohesive quiddities’, that is, ‘recurring features that are unique to each family and constitute its singular essence’ (Segev et al., 2015: 3). The analysis of image macros in Section 4 will use Shifman’s model to demonstrate precisely which features contribute towards a meme family’s cohesive quiddities and which features are open to meme authors’ choice.

3 DATA SELECTION: CREATING A MEME POOL In order to investigate the pragma-stylistics of popular internet memes, I have created a ‘meme pool’ (Knobel and Lankshear, 2007) of the all-time most popular image macros, selecting for each their highest and lowest scored iterations, which helps ascertain their acceptability to online users. As Segev et al. acknowledge, identifying the most popular meme families is an impossible task, ‘due to the dynamic nature of memes and the different methods for evaluating their popularity (view counts, likes/votes, number of derivations, number of mentions in search results, and so on)’ (2015: 7). Inevitably, the memes selected for analysis are indicative of their popularity during a certain time period, and by a certain measure. Unlike previous researchers, I do not believe in using offline (Knobel and Lankshear, 2007) or academic (Segev et al., 2015) sources as indicators of internet memes’ popularity, as this can lead to circularity in the research. Despite their accessibility, it is best to refrain from using popular search engines (e.g. Google) to select data, as they provide manipulated and unreplicable results (Lew, 2009). Instead, I use a combination of online user-generated indicators of meme popularity, both to select the meme families (Phase 1) and then to select the individual image macros (Phase 2). 3.1 Phase 1: Selecting meme families In order to identify the most popular ‘families’ of image macros, I used Memegenerator.net. Launched in March 2009, it was the first online meme generator (a website for users to create memes by selecting from established

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

89

TABLE 5.1 The top five most popular image macro families, according to Memegenerator.net Rating

Image Macro Family

No. of Likes

1

Y U No

4,109

2.5 million

2

Futurama Fry

8,246

1.8 million

3

The Most Interesting Man in the World

6,819

2.1 million

4

Philosoraptor

5

Condescending Wonka

12,104 4,084

No. of Posts

837k 2.1 million

images, or uploading their own, and adding the text of their choosing). This website has a constantly updated page dedicated to the ‘most popular’ image macro templates ‘of all time’, which at the time of research included the following top five (Table 5.1). I limited the study to the top five due to the spatial constraints governing an in-depth qualitative study of these meme families and their (un)successful iterations. In being one of the original websites for creating image macros, Memegenerator.net’s all-time popularity records go back as far as is possible, which ensures that the resultant popular meme families are those that have persisted in the years since image macro creation began. As you can see from the data Memegenerator.net provides, the popularity ranking is based on the number of ‘likes’ from users for the meme family, as well as the number of ‘posts’, meaning individual instantiations of that meme family. Unfortunately, Memegenerator.net does not make clear how these two factors are weighted in ranking the meme families. Although this long-standing website provides the researcher with longitudinal, observational and retrospective data on successful meme families, it is not without its drawbacks. This is only one of an eventual proliferation of meme generating platforms and the data only pertains to the image macros created on this website. In fact, this website’s popularity has waned since its peak usage in the years from 2009 to 2015 (Knowyourmeme.com, 2018) as meme creators have since turned to other websites or applications that serve the same purposes. The implication of this for the present study is that the meme pool may be skewed towards meme families that were more popular during the website’s peak period. Nonetheless, it is not the aim of the present study to comment on the most contemporary image macros. Instead, the present aim is to analyse the most popular image macro families and to select them in as rigorous a method as is possible. No other meme generator or database provides such a transparently sampled list of all-time popular image macro families.

90

CONTEMPORARY MEDIA STYLISTICS

3.2 Phase 2: Selecting the un/popular uses While Memegenerator.net helped me to identify the top five most popular meme families, the next phase of the data selection was to find individual instantiations of those image macros which are deemed successful and unsuccessful by users. This user-based evaluation of image macros would determine the core features of the image macros and, in their disliked iterations, the contested features. Knowyourmeme.com is an extensive online database of all kinds of internet memes, including image macros. Much like a wiki, any registered member can submit a meme to this independently owned website. The editorial staff and moderators evaluate each submission by further researching the online presence of the meme for confirmation or invalidation. Site users vote and comment on the content, feeding back into the site’s research. Because of its comprehensivity and the fact that its content is verified and voted on by internet users, Knowyourmeme.com has been used as a source in several other academic studies of memes (e.g. Ryan Vickery, 2014; Segev et al., 2015; Lou, 2017). For the present study, it offers the same benefits. I entered the name of each of the top five meme families of all time into Knowyourmeme.com’s search bar, repeating the search process for each of the five. From the results, I selected ‘images’, as the present study focuses on image macros solely. Once filtered to view image versions of this meme family only, I opted to order the results by ‘highest score’, then ‘lowest score’. These scores are based on the number of likes and dislikes that each image macro has received from the website’s users. Other researchers of digital communication have found that various online platforms’ ‘in-built popularity measurements’ (Shifman, 2011: 190) are useful indicators of user-informed approbation of the data (see also Blommaert, 2018). Therefore, user responses to individual image macros on the most comprehensive online meme database inform the selection of successful and unsuccessful image macros to be analysed in Section 4.2. Using a combination of user-driven popularity measurements, as provided by two prominent meme websites provides me with a list of the top five image macro families, and one of their most popular iterations, as well as one of the most unpopular. This has resulted in ten image macros (see Appendix) to be analysed in Section 4. I found that using likes and dislikes to select the specific image macros to be analysed meant that ‘middle-of-the-road’ instantiations of these meme families were overlooked, that is, image macros that played by the rules but did not break any rules in ways that pleased or displeased other users. This method resulted in the fact that the most popular image macros, while conforming to the meme families’ core characteristics, often demonstrate extreme creativity and intertextual references to other memes (i.e. ‘meta memes’ [Shifman, 2011]). Moreover, the most unpopular image macros often espouse social prejudices,

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

91

such as racism or misogyny, or alternatively, fail to play by the meme’s formal rules. In previous studies of memes, researchers have tended to self-select ‘typical’ examples of image macros for analysis. What this study’s method offers, in contrast, is a meme pool that – in its inclusion of approved rulebreakers and disapproved rule-breakers – gets closer to identifying the rules shared by online meme users.

4 PRAGMA-STYLISTIC ANALYSIS In this section, I apply Shifman’s framework for the analysis of memes, augmented with additional Goffmanian concepts. In Section 4.1, I focus on the ‘cohesive quiddities’ or essential characteristics that constitute each of the top five meme families. Then, in Section 4.2, I turn to analyse the most liked and disliked versions of these, considering the reasons for their (dis)approbation in relation to how they carry out the meme family’s rules. 4.1 The meme families In this section, I apply Shifman’s framework to the five most popular meme families, discussing their typical form, content and the kinds of stance they perform. Throughout, I refer to the meme families in the abstract, using my own ‘meme literacy’ (Shifman, 2014: 100) and Knowyourmeme.com as sources for their descriptions. However, where the individual examples from my meme pool (Appendix) serve to illustrate a point, I will refer to them specifically. The all-time most popular meme family in my meme pool is ‘Y U No’, which usually features a stick figure character with his hands in an open, pleading, gesture. The character, common to a series of other ‘rage comic’ memes, is accompanied by the eponymous phrase ‘Y U No’, an abbreviation for ‘Why do/are you not …’, which is completed with an action the meme author wishes the ‘addressed recipient’ (Goffman, 1979/1981: 134) would do. Therefore, in terms of stance, this meme family performs Jakobsonian communicative functions that are expressive (of the authors’ desires towards the addressee) and conative (engaging the addressee in a negated, but desired activity). The referential function, however, is variable as the meme generator can insert the desired activity. Usefully, the abbreviated nature of the meme phrase ‘Y U No’ means there are no auxiliary verbs that compel the author to use a specific tense or aspect. The cohesive quiddities of this meme family, then, are the stickman character, the adaptable phrase ‘Y U No’, and the expressive and conative functions it performs, leaving the referential function open to change. The second most popular meme family is Futurama Fry, which features a still from the cartoon series Futurama, in which the character Fry has semi-closed eyes. This image is overlaid with the phrase, ‘Not sure if X or Y’, a protasis

92

CONTEMPORARY MEDIA STYLISTICS

completed with the coordinated apodosis of the meme generator’s choice. The first half of a coordinated apodosis construction is usually placed at the top of the image and the second half at the bottom of the image (see Appendix, Figure 5.2, where both image macros meet the standard formal criteria for this meme family). The use of text on two sides of the image and two sides of a coordinated apodosis allows for a multimodal contrasting of two situations, for humorous or other effect. Thus while the form follows strict rules, the ideational content is subject to change. In terms of this meme’s communicative function, the epistemic uncertainty manifest in the conditional construction and replicated in the character’s squinting expression fulfils an expressive function (uncertainty), while the referential function is again up to the author to complete. The third most popular family in the present study’s meme pool is known as ‘The Most Interesting Man in the World’. The image is always the same still from Dos Equis beer advert featuring US actor Jonathan Goldsmith in a suit jacket, gazing directly at the camera, a beer within reach. In the original advert, he used a phrase that inspires the repeated phrase in this meme family, ‘I don’t always X, but when I do, I Y’. Both verb phrases are chosen by the meme generator, according to the referential function s/he wants to perform. Generally, the format dictates that the verb phrase Y is a very cool and interesting way of doing X. The implication is that the speaker ‘I’ is an alignment of the meme generator and ‘The Most Interesting Man in the World’s’ identity; in Goffman’s terms, the ‘author’ aligns him/herself with the ‘animator’ as being cool and interesting (1981/1979). Thus, the propositional content should reflect this, with Y being a cool way of doing X, performing the expressive function of proclaiming the subject’s status as cool. Once again the form has textual and visual cohesive quiddities (the image, the repeated clausal structure), and the function of the meme is expressive, with the referential variation open to change within certain parameters. ‘Philosoraptor’ is the fourth most popular family in the meme pool, and invariably features a green dinosaur head, whose talon rests on his upturned chin meditatively. His name is a portmanteau of his species, Velociraptor, and his philosophical nature. This character always co-occurs with a ‘philosophical’ conditional question of the meme author’s choosing, usually an apodosis at the top of the image followed by a protasis in the form of a question at the bottom of the image, often ‘If X, [then] Y?’. Philosoraptor’s musings are usually far from philosophical, however, and simply present a logical conundrum or an incongruity between two states of affairs, such as ‘If actions are stronger than words, [picture of Philosoraptor] why is the pen mightier than the sword?’. As this typical example demonstrates, the two states of affairs advanced in the proverbs on either side of the image are presented as incompatible and the Philosoraptor character is used as the animator to express this conundrum, which originates from the meme author. In relation to Jakobson’s communicative functions, the

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

93

referential function is very malleable, while the expression of puzzlement is a constant. As the example above demonstrates, Philosoraptor’s musings can be of the linguistic variety, and thus may also perform a metalingual function. The fifth and last meme family in this study is ‘Condescending Wonka’, which always features a still of Gene Wilder in the 1971 film Willie Wonka and the Chocolate Factory. With his head propped on his hand and his gaze directed at someone or something, half smiling, it appears the character is listening, bemused. The text of this meme is highly variable, but often begins with, ‘So [tell me more about how] …’ and is completed with a subordinate clause that contains an action that is deemed implausible by the author. In employing the imperative to ask the addressee to elaborate on a point, this meme performs the conative function. Once again, the character is an animator for the meme author’s attitude, which serves an expressive function, whereby the author expresses condescension or disalignment with Wonka’s addressee. Participant roles are particularly complex in this meme, as the author quotes the animator, who quotes the target of the condescension. In this way, Condescending Wonka offers a form for ‘rekeying’ (Goffman, 1974/1986: 81) a primary frame within the mimetic frame, in order to take a critical and condescending stance. Consideration of the interactive context wherein this image macro is used would be necessary to understand if the target of the sarcasm is a ‘ratified’ participant, or if the ‘addressed recipient’ (Goffman 1979/1981) is invited to share in the author’s condescending stance towards an ‘unaddressed recipient’. The preceding account of the top five image macro families demonstrates that there are clear patterns in how image macros are formed and function, summarized in Table 5.2. All five of these meme families use a character as an animator for the meme author’s utterance. By employing the comprehensive framework advanced by Shifman (2014), I have identified the core characteristics of the five most popular image macro families, including the ‘cohesive quiddities’ (Segev et al., 2015) that are shared across individual instantiations of those families, as well as the elements open to variation. The character represents a consistent stance, fulfilling a consistent expressive function, which varies according to the character and meme family it constitutes. Sometimes, the character is used as an animator for the author to fulfil the conative function by directly involving other discourse participants in the discourse. All of these popular meme families use a fixed linguistic structure which, in concert with the image, reflects the expressive function. The referential function is open to variation, as semantically rich elements (such as verb or noun phrases or clauses) are those which are left open to the meme author’s choice. As such, the cohesive quiddities of popular image macros are the character and the image, structural elements of the text and the expressive (or conative) function performed. While all of these cohesive quiddities contribute to the meme families’ coherence, the referential element

Content

The idea that the addressee is obliged to do something and they have not

The idea that a state of affairs may be perceived in two ways

The idea that the meme generator has a cooler way, Y, of doing X

Meme Family

1. Y U No

2. Futurama Fry

3. The Most Interesting Man in the World

Visual: Still of man expressive (of author’s in Dos Equis advert cool or interesting status) sitting at a table with a beer Verbal: ‘I don’t often X, but when I do, I Y’

The coordinated clauses or nominalizations, that is, referential function

expressive (of uncertainty, due to perceived similarity between two states of affairs)

Visual: Still of Fry from Futurama gazing off-camera squinting Verbal: ‘Not sure if X or Y’

Character: Dos Equis man The verb phrases, that is, referential Phrase: ‘I don’t often X, function but when I do, I Y’ Function: expressive of author’s cool status

Character: Fry from Futurama Phrase: ‘Not sure if X or Y’ Function: expressive of uncertainty and perceived similarity

The verb phrase, Character: stick man that is, referential Phrase: ‘Y U no …’ function Functions: expressive of obligation and conative of addressee’s obliged action

expressive (of the author’s perceived obligation of the addressee) and conative (engaging the addressee in a negated, but desired activity)

Visual: stick man with hands outspread. Verbal: ‘Y U no …’ completed with action

Variables

Cohesive Quiddities

Stance

Term derived from Segev et al. (2015)

Form

Categories derived from Shifman (2014)

TABLE 5.2 Analysis of the top five families of image macros based on Shifman (2014) and Segev et al. (2015)

Visual: Still of Willie Wonka, head resting on hand Verbal: ‘So [tell me more about how] …’

The juxtaposition of Visual: Cartoon of two ideas through green dinosaur the conditional Verbal: ‘If X, [then] Y?’

5. Condescending The idea that Wonka the embedded proposition is implausible

4. Philosoraptor

Character: Philosoraptor Phrase: ‘If X, [then] Y?’ Function: expressive of puzzlement Character: Willie Wonka Phrase: ‘So [tell me more about how] …’ Function: expressive (condescension or disalignment) towards the interlocutor, that is, conative

expressive (of puzzlement)

expressive (of condescension or disalignment) conative (asking the addressee to elaborate)

The proposition pertaining to an addressee’s stance, implausible to the Actor/Animator, that is, referential function

The contents of the apodosis and the protasis, that is, referential function

96

CONTEMPORARY MEDIA STYLISTICS

is open to the meme authors’ creative input. This confirms Jenkins et al.’s observation that essential to spreadable media is the presence of a ‘gap’, which encourages ‘producerly engagement’ (Jenkins et al., 2009, no pages). In the subsequent section, the most ‘liked’ and ‘disliked’ instantiations of these meme families are investigated in order to further understand, from a user viewpoint, what makes a good and bad version of these image macros. In Goffman’s terms, we now turn to look at the other side of the participation framework, that of the effect on meme readers. 4.2 The most liked/disliked image macros from each meme family Above, we saw how the top five meme families in the image macro genre work, using Shifman’s (2014) framework to describe the rules and cohesive quiddities that underscore them. In this section, we examine the most liked and most disliked instantiations of each of top five image macro families, as voted by users on Knowyourmeme.com and compiled in Appendix. Beginning with the most liked iteration of ‘Y U No’ (Appendix, Figure 5.1a), this image macro maintains all of the cohesive quiddities pertaining to the meme family: the rage face character, the eponymous text and the expression of an unfulfilled action (see Table 5.2). The elaborated meme phrase ‘You say you like it/ Why U no put ring on it?’ refers to a popular song by Beyoncé called ‘Single Ladies’. This cultural text has pre-established the expectation ‘If you like it, then you shoulda put a ring on it’ through its chorus’ repeated refrain, which serves as the schema from which this meme author draws on to create the expected action key to ‘Y U No’. While referring to the song lyrics, the meme phrase also manages to retain the ungrammatical, abbreviated style typical of the meme family. The intertextual reference made in the meme phrase is strengthened through mimetic adaptations made to the image, where the stick-man character is crudely adorned with Beyoncé’s hair, make up and a ring. Moreover, the one-shouldered black bodysuit and triplicate characters replicate the song’s equally famous music video. As outlined in Section 2.1, Knobel and Lankshear (2007) found that successful early internet memes tended to include three characteristics: rich intertextuality, anomalous juxtapositions and some element of humour. It seems then that this image macro is particularly successful because of its intertextual references to shared knowledge of popular culture, of which meme readers may feel a part. The juxtaposition of a glamorous celebrity with an ugly, pleading stick man is certainly anomalous and, as is widely accepted in humour studies (Dynel, 2013), incongruous juxtapositions are fundamental to humorous discourse. In contrast, the most disliked iteration of ‘Y U No’ does not uphold the meme family’s necessary cohesive quiddities. Although it uses the necessary linguistic formulation in the meme phrases ‘Girls / Y U No date me. Forever

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

97

alone’, gone is the requisite character and, thus, the visual force behind the expressive and conative stance. In his place is an image featuring a character from another meme family, ‘Scumbag Steve’ who, as his name suggests, is used to represent actions that the meme author deems reprehensible (Dancygier and Vandelanotte, 2017). While memes can – and often successfully do – refer to other memes (Shifman, 2011; Lou, 2017), using Scumbag Steve as an animator for the author’s utterance produces no additional meaning, other than perhaps to suggest that Scumbag Steve is in need of a date. While breaking the rules of the meme family may be permitted if the result is to perform additional referential or poetic functions, this text manages none of those feats and only manages to lose expressive and conative power. I would argue that this violation of mimetic practice is the main reason for its dislike within the Knowyourmeme.com community. A good example of a ‘meta meme’ (Shifman, 2011) or a ‘meme about a meme’ can be found in Figure 5.2a (Appendix). Not strictly an image macro, this animated image (only partially captured in print) is the result, in Phase 2 of the data selection, as the most liked ‘Futurama Fry’ image and requires some additional description here. The meme phrase ‘Not sure if cop tailgating me wants me to go faster / or is just testing me’ preserves the formal cohesive quiddity of this family by using ‘Not sure if ’ followed by two alternative states of affairs that could be potentially confusing, capturing the meme’s typical expressive function. As usual for this meme family, the confusion is reflected in the character’s expression and in the division of the coordinated construction between top and bottom of the image. However, seconds into reading the meme, a character from another meme family, ‘Socially Awkward Penguin’, waddles across the image accompanied by the text ‘Wanders into the wrong meme’. By doing so, ‘Socially Awkward Penguin’ upholds his/her typical characteristics, but in interaction with this meme. As before, the most successful iteration of a meme family is that which fits Knobel and Lankshear’s (2007) account of intertextuality, anomalous juxtaposition and, as a result, humour. In addition, this animated image macro visually and verbally calls attention to the status of ‘Futurama Fry’ as a meme, and as distinct from the ‘Socially Awkward Penguin’ meme which interrupts it. As such, this ‘meta meme’ could also be described as fulfilling a metalingual function and, in the process of foregrounding the mimetic form, as poetic. Once again, the most successful iteration of a meme is that which maintains all the cohesive quiddities, while providing additional communicative functions. Yet the poetic and the metalingual functions are only fulfilled if the reader is ‘meme literate’, once again highlighting the significance of in-group knowledge and identity as key factors in a meme’s approval rating. Unlike the unpopular ‘Y U No’ meme, the most disliked iteration of the ‘Futurama Fry’ meme (Appendix, Figure 5.2b) does uphold all the cohesive quiddities that constitute the meme family. Formally, the Futurama character

98

CONTEMPORARY MEDIA STYLISTICS

is presented in his usual squinting expression, and the accompanying text ‘Not sure if gay / or just European’ presents two alternative states of affairs. Because all the cohesive quiddities are met, it can be assumed that the Knowyourmeme. com community dislike this image macro because of its particular referential content, and the idea it expresses. By coordinating the attribute ‘gay’ with ‘European’ as a potential alternative judgement, the meme author is using the characters’ gaze to express their opinion that these two personal categories are similar, in terms of their homosexuality and being European. While there is the potential for humour in the juxtaposition of these incongruous categories, the online community’s disapprobation of the referential content demonstrates that this attempt at humour has misfired. In this image macro, we can see how mimetic forms can be used to advance social prejudices (in this case on continental and homosexual identities) which are, in turn, subject to community policing. The dislike towards this image macro can be explained through consideration of sharing memes as ‘keying’, specifically of the playful variety (see Section 2.2). Goffman (1974/1986: 49) observed: [A]lthough individuals can playfully engage in an extremely broad range of activity, limits on playfulness are established in various groups. … Among familiars, for example, there will be appeals to ‘taste’; it is not nice to make light of certain aspects of the lives of friends. If we understand sharing an image macro as playful keying, this particular instance makes light of an idea that the community does not find acceptable, that of stereotyping and comparing the identities of these two distinct social groups. In the most liked version of ‘The Most Interesting Man in the World’ (Appendix, Figure 5.3a), the meme author maintains the meme family’s visual and verbal formal cohesive quiddities. However, they do not perform the expressive function (coolness) in the usual way (i.e. by taking claim to a particular way of doing an action). Instead, in using the meme phrase ‘I don’t always see, but when I do, it’s what you did there’, this meme author has used a phrase popular in internet communication, ‘I see what you did there’, and blended it with the formal features of this particular meme. The phrase, which originates from 1990s sitcom Friends, is used when the speaker overtly recognizes a ‘move’ made by their interlocutor, an utterance serving a particular communicative function (Swales, 1981, 1990). In this way, the phrase inherently serves a metalingual function. By integrating the phrase into the formal structure of this meme, the meme author upholds the formal quiddities of ‘The Most Interesting Man in the World’ and draws on shared knowledge of another popular internet phrase to fulfil the referential function. In doing so, he or she makes reference to the addressee’s discourse, as well as intertextual reference to wider internet

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

99

discourse, performing a complex metalingual function. Moreover, by violating one rule of this meme (that verb Y be a cool way of doing verb X), the author foregrounds the linguistic trickery, fulfilling the poetic function. I would argue this image macro is deemed successful by the online community because it fulfils so many additional communicative functions, while remaining largely faithful to the meme family’s cohesive quiddities. While the expressive function is not performed directly (through reference to doing X in a Y fashion), the meme author’s status as cool is affirmed indirectly through their skill at mimetic practice, in terms of imitating and adapting content, complete with the intertextuality, incongruity and resultant humour that are found in the most successful memes. In contrast, the least popular iteration of this meme family (Appendix, Figure 5.3b) seems to fulfil all the necessary cohesive quiddities. The lengthy meme phrase ‘I don’t always march/ but when I do, I march my ass to work because I’m a responsible adult with bills to pay and a family to support’ uses the generic structure of this meme to suggest that marching one’s ass to work is a superior way of marching. The image macro was uploaded to a series ‘March for our Lives’ which was based on the anti-gun violence protests, following a Florida high school shooting in February 2018. Thus, there are real-world referents for the two senses of ‘marching’, which form the basis of the meme author’s pun. Accordingly, in addition to expressing the meme author’s coolness or superiority, this iteration also performs a somewhat poetic communicative function, foregrounding the dual sense of the word ‘march’ and the contrasting ways in which the action can be carried out. Nonetheless, despite the image macro’s adherence to the meme family’s cohesive quiddities and use of additional resources for humorous purposes (which elsewhere contributed to the success of the image macro), this iteration proves unpopular. This effect is unsurprising, given that the meme author employs the superior stance afforded by this meme to criticize the forum and its members’ political action. The most popular ‘Philosoraptor’ image macro sees the character ponder the proposition ‘If an illegal immigrant fought a child molester / would it be Alien vs. Predator?’ (Appendix, Figure 5.4a). Preserving the meme family’s cohesive quiddities, including the visual and verbal formal features, the meme author uses the character as animator to create a pun. Referring to the science fiction movie Alien vs. Predator, the meme author plays on the dual senses of these words to also refer to the people introduced in the apodosis. Indeed, the conditional structure used in the Philosoraptor meme family provides the ideal vehicle for a formulaic linguistic ‘joke’ such as this one. In their influential General Theory of Verbal Humor, largely based on canned and written jokes, Attardo and Raskin (1991) propose that jokes have three essential stages: (i) setup, (ii) incongruity and (iii) resolution. In this image macro, the apodosis serves as the

100

CONTEMPORARY MEDIA STYLISTICS

setup, introducing two different kinds of people according to their societal roles (both, incidentally, contentious ones). The apodosis also introduces an incongruity, in the unlikely pairing of these two kinds of people in a fight. Lastly, the protasis serves to provide a fictional frame in which this incongruity would make sense, calling on the meme reader’s pop culture knowledge of the film and linguistic knowledge of the polysemes for the joke’s resolution. Thus, through the feigned puzzlement of the Philosoraptor as animator for the author’s joke, this image macro foregrounds the word meanings, fulfilling the metalingual and, arguably, the poetic function. Like the other popular image macros, adherence to the meme family’s cohesive quiddities and the fulfilment of additional functions, including humour, give rise to its success. This success is despite the fact that the joke’s content is contingent on dehumanizing both categories of people, including a derogatory term for a migrant (alien) and an insouciant term for a paedophile (predator). While gays and Europeans were deemed unfair game (Appendix, Figure 5.2b), migrants and paedophiles seem to be acceptable targets of meme humour in this domain. The most unpopular of the ‘Philosoraptor’ memes asks, ‘If there are no girls on the internet / what am I doing here?’. While the author has fulfilled the linguistic criteria for this meme, the conditional question does not require any additional resources for its resolution, lacking the logical conundrum necessary for this to cohere fully with the meme family. With the opposition of the ‘girls’ in the apodosis and ‘I’ in the protasis, the implication is that the meme author is a male who wants to interact with females. However, this suggestion at a linguistic level clashes with the visual level, where the Philosoraptor, usually the animator for the author’s words, is painted female. Thus, the meme author has failed to coherently align discourse participation roles and provide the logical complexity necessary to this meme family. I would suggest that these failings lead to this image macro’s unpopularity. The most liked iteration of the ‘Willy Wonka’ meme dates back to 2016, shortly after the death of Gene Wilder, the actor who plays the meme’s character (Appendix, Figure 5.5a). Willy Wonka is attributed with the words ‘Oh, I died? / Please suddenly tell the internet how much of a fan you’ve always been’. As outlined in Section 4.1 this meme family does not demand a strict linguistic form, but, rather, the character is used as an animator for the expression of condescension or sarcasm, usually achieved through a feigned expression of interest and a subsequent request for further information, both evident here. In this case, the meme author’s disparagement is directed at those internet users who, in reaction to news of the death of a public figure, claim to have been long-standing fans, jumping on a bandwagon of mourners. Through embedding the words of the ‘fan’ in the utterance, the meme author is echoing their discourse and parodying their behaviour. Moreover, by doing so through the medium of a Gene Wilder meme, the meme author uses the fictional

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

101

animator to comment on non-fictional behaviour around the animator/actor’s death, appealing to his/her peers who share the same disparaging view of such behaviour. Thus, as well as upholding the meme’s cohesive quiddities, it makes intratextual links to directly relevant current affairs and to objectionable internet behaviour. By drawing attention to the real-world relevance and consequently the fictionality of the text, the meme author imbues it with new meaning and fulfils the poetic function. As with the other popular image macros, scrupulous attention to the meme’s rules along with performing additional communicative strategies merits high levels of approbation among the online community. The most unpopular ‘Willy Wonka’ image macro depicts the character as asking, ‘You’re in university and still believe in God? / That’s adorable’ (Appendix, Figure 5.5b). Through the use of a question with the fictional interlocutor’s words embedded in the proposition, the format allows the meme author to take a stance on the proposition through the condescending attitude of Willy Wonka. The use of the adjective ‘adorable’ to describe the dual state of both being in university and believing in God suggests that these are incompatible concepts, that acquiring knowledge is at odds with believing in a religious deity. While ‘adorable’ is a positive evaluation, its use in a meme family used to express condescension frames it as insincere. Despite the meme author’s adherence to the necessary cohesive quiddities across the levels of form and stance, this meme was negatively evaluated by the community. It seems that associating religion with lack of education is not a position this community endorses. As with the unpopular iterations of ‘Futurama Fry’ (Appendix, Figure 5.2b) and ‘The Most Interesting Man in the World’ (Appendix, Figure 5.3b), the meme readers seem to have taken issue not with the author’s formal rendering of the meme but with its referential content and ideological stance, where the playful keying is deemed inappropriate.

5 CONCLUSION The pragma-stylistic analysis of the most liked and disliked iterations of the top five image macro families has yielded several findings which add to contemporary understanding of memes. In Section 4.1, an augmented version of Shifman’s (2014) framework was applied to the top five image macro families: Y U No, Futurama Fry, The Most Interesting Man in the World, Philosoraptor and Condescending Wonka. Although Shifman’s three mimetic dimensions – content, form and stance – were designed to be dimensions along which cultural items can be imitated (Section 2.2), I have demonstrated that certain elements of each dimension are cohesive quiddities, while certain elements are open to user adaptation. In these image macro families, the cohesive quiddities spanned the formal dimension (the character, linguistic structures), as well as the specific stance communicated through the expressive and/or conative functions.

102

CONTEMPORARY MEDIA STYLISTICS

The referential function was consistently subject to change, meaning users could employ an image macro’s formal properties and consistent expressive or conative functions to take a particular stance in an interactive context. These referential ‘gaps’ and the endless playful iterations they afford are what make internet memes so engaging and spreadable (Jenkins et al., 2009). Elaborating on the significance of Goffman’s work for Shifman’s model aided in understanding the participation structures inherent in each meme, delineating the author (meme creator) of the meme phrase from the animator (character). Moreover, understanding each iteration of a meme as a ‘keying of a common model’ helps to explain why some image macros were disliked, either (a) because their playful keyings made light of content that was deemed unsuitable or (b) because their technical redoings failed to cohere with the meme family’s quiddities. I believe there is further work to be done on the significance of keying in digital discourse in general, as digital platforms afford imitations, (re)iterations and extensive sharing. While I have considered the pragmatic level in relation to participant roles and user approbation, there is a need for more research on the use of memes, including image macros, in interactive contexts. It would be helpful to understand why, for example, some social groups were deemed acceptable targets of meme humour and why some were not. Such a study requires consideration of the particular discourse and interactive contexts, which would only be possible through further pragmatic or critical linguistic research. It is worth noting, however, that other scholars’ findings on the social biases of memes tend to represent white people and men more frequently (Shifman, 2011; Segev et al., 2015). All of the top five meme families feature male characters as animators. The androgynous ‘Y U No’ character (Appendix, Figure 5.5.2a) and Philosoraptor (Appendix, Figure 5.5.4b) are doctored for feminine characteristics, suggesting the unmarked version is male. This pattern indicates a gender bias in mimetic representation. Furthermore, although Shifman (2011) found ‘no social commentary’ in her corpus of popular YouTube memes, my meme pool contained a high proportion of image macros taking political or ideological stances, or identifying social groups as the target for humour. However, this may have been a result of the method of data collection, whereby the most disliked memes were also included, resulting in unpopular political stances being represented. Section 4.2 also confirmed the findings of previous research which indicates that memes often make use of incongruous juxtapositions, and this was demonstrated to add to their potential for humour. Moreover, the meme family was sometimes incongruously juxtaposed with other memes or aspects of pop culture, which confirms the idea that intertextuality plays a key role in mimetic practice. I would add that this serves to solidify in-group membership in terms of shared cultural knowledge and plays a part in creating humour, which we know to be founded on incongruity. Moreover, the imitative nature of internet memes makes them ideal vehicles for parody. Yet, the image macros studied here emphasized that ‘it is

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

103

not only the parody form, but the quality and subtlety of execution that matters’ (Jenkins et al., 2009). In some of the most successful iterations, that ‘quality and subtlety’ was achieved through mastery over the mimetic form, including adherence to the cohesive quiddities and skill at using the freedoms afforded through the referential content to make additional communicative functions, namely the poetic. Through careful application of Shifman’s model (expounded with Goffmanian concepts) to the most salient image macros, I hope to have demonstrated their extreme rule-governed nature, as well as the spaces within which meme authors make their own creative choices to achieve pragma-stylistic effects.

APPENDIX: MEME POOL OF MOST LIKED/DISLIKED IMAGE MACROS

FIGURE 5.1: (a) Y U No – high score; (b) Y U No – low score.

FIGURE 5.2: (a) Futurama Fry – high score; (b) Futurama Fry – low score.

104

CONTEMPORARY MEDIA STYLISTICS

FIGURE 5.3: (a) The Most Interesting Man in the World – high score; (b) The Most Interesting Man in the World – low score.

FIGURE 5.4: (a) Philosoraptor – high score; (b) Philosoraptor – low score.

THE PRAGMA-STYLISTICS OF ‘IMAGE MACRO’ INTERNET MEMES

105

FIGURE 5.5: (a) Condescending Wonka – high score; (b) Condescending Wonka – low score.

REFERENCES Attardo, S. and Raskin, V. (1991). ‘Script theory revis(it)ed: Joke similarity and joke representation model’. Humor: International Journal of Humor Research, 4(3), 293–347. Blommaert, J. (2015). ‘Meaning as a nonlinear effect: The birth of cool’. In T. Lillis (Ed.), Theory in Applied Linguistics Research: Critical Approaches to Production, Performance and Participation (pp. 7–27). Amsterdam and Philadelphia: John Benjamins. Blackmore, S. (1999). The Meme Machine. Oxford: Oxford University Press. Blommaert, J. (2018). ‘#HableWasErbij: The political and social impact of memes’. Online lecture. Diggit Magazine, 1 March. URL: https://www.facebook.com/diggi tmagazine/videos/2024456734494841/. Accessed 11 April 2018. Brooks, J. (2007). ‘Understanding virtuality: Contributions from Goffman’s frame analysis’. School of Information Studies, Faculty Scholarship, 87. Syracuse University. URL: https://surface.syr.edu/istpub/87. Accessed 7 June 2018. Dancygier, B. and Vandelanotte, L. (2017). ‘Internet memes as multimodal constructions’. Cognitive Linguistics, 28(3), 565–98. Davison, P. (2012). ‘The language of internet memes’. In M. Mandiberg (Ed.), Social Media Reader (pp. 120–34). New York: New York University Press. Dawkins, R. (Performer) and Marshmallow Laser Feast (Director). (2013). Just for Hits – Richard Dawkins. https://youtu.be/GFn-ixX9edg. Accessed 4 April 2018. Deacon, T. (1999). ‘The trouble with memes (and what to do about it)’. The Semiotic Review of Books, 10(3). http://projects.chass.utoronto.ca/semiotics/srb/10-3edit.html Dynel, M. (2013). ‘A view on humour theory’. In M. Dynel (Ed.), Developments in Linguistic Humour Theory (pp. vii–xiv). Amsterdam and Philadelphia: John Benjamins. Goffman, E. (1974/1986). Frame Analysis: An Essay on the Organization of Experience. Boston: Northeastern University Press.

106

CONTEMPORARY MEDIA STYLISTICS

Goffman, E. (1979/1981). ‘Footing’. In E. Goffman (Ed.), Forms of Talk (pp. 124– 159). Oxford: Basil Blackwell. First published in Semiotica, 25(1979), 1–29. Jakobson, R. (1960). ‘Closing statement: Linguistics and poetics’. In T. A. Sebeok (Ed.), Style in Language (pp. 350–449). Cambridge, MA: MIT Press. Jenkins, H., Li, X., Domb Krauskopf, A. and Green, J. (2009). ‘If it doesn’t spread, it’s dead (Part 7): Aesthetic and Structural Strategies’. http://henryjenkins.org/blog/2 009/02/if_it_doesnt_spread_its_dead_p_6.html. Accessed 15 June 2018. Knobel, M. and Lankshear, C. (2007). ‘Online memes, affinities and cultural production’. In M. Knobel and C. Lankshear (Eds), A New Literacies Sampler (pp. 199–227). New York, NY: Peter Lang. Knowyourmeme.com. (2018). ‘Memegenerator’. http://knowyourmeme.com/memes/ sites/meme-generator. Accessed 27 April 2018. Kull, K. (2000). ‘Copy versus translate, meme versus sign: Development of biological textuality’. European Journal for Semiotic Studies, 12(1), 101–20. Hickey, L. (1993). ‘Stylistics, pragmatics and pragmastylistics’. Revue belge de philologie et d’histoire, 71(3), 573–86. Lew, R. (2009). ‘The web as corpus versus traditional corpora: Their relative utility for linguists and language learners’. In P. Baker (Ed.), Contemporary Corpus Linguistics (pp. 289–300). London and New York: Continuum. Lou, A. (2017). ‘Multimodal simile: The “when” meme in social media’. English Text Construction, 10(1), 106–31. Philips, S. (1972). ‘Participant structure and communicative competence: Warm Springs children in community and classroom’. In C. Cazden, V. John and D. Hymes (Eds), Functions of Language in the Classroom (pp. 370–94). New York: Teachers College Press. Ryan Vickery, J. (2014). ‘The curious case of Confession Bear: The reappropriation of online macro-image memes’. Information, Communication and Society, 17(3), 301–25. Searle, J. R. (1997). The Mystery of Consciousness. London: Granta Books. Segev, E., Nissenbaum, A., Stoldero, N. and Shifman, L. (2015). ‘Families and networks of internet memes: The relationship between cohesiveness, uniqueness and quiddity concreteness’. Journal of Computer-Mediated Communication, 20(4), 417–33. Shifman, L. (2011). ‘An anatomy of a YouTube meme’. New Media and Society, 14(2), 187–203. Shifman, L. (2013). ‘Memes in a digital world: Reconciling with a conceptual troublemaker’. Journal of Computer-Mediated Communication, 18(3), 362–77. Shifman, L. (2014). Memes in Digital Culture. Cambridge, MA: MIT Press. Simpson, P. (2013). On the Discourse of Satire: Towards a Stylistics Model of Satirical Humour. Amsterdam and Philadelphia: John Benjamins. Swales, J. (1981). Aspects of Article Introductions. Birmingham: University of Aston. Swales, J. (1990). Genre Analysis: English for Academic and Research Settings. Cambridge: Cambridge University Press. Varis, P. and Blommaert, J. (2015). ‘Conviviality and collectives on social media: Virality, memes and new social structures’. Multilingual Margins, 2(1), 31–45. Wiggins, B. E. and Bowers, G. B. (2014). ‘Memes as genre: A structurational analysis of the memescape’. New Media and Society, 17(11), 1886–906.

CHAPTER SIX

The stylistics of emoji: An interactional approach DWI NOVERINI DJENAR AND MICHAEL C. EWING

Emoji have been studied in different disciplines using various approaches. From historical descriptions and experiment-focused studies to pragmatic and interactionally based accounts, emoji have been highlighted as a resource people use to convey affect and a sense of connectedness. Although these studies have helped us understand the social functions of emoji, they have mainly focused on analysing the various communicative functions of emoji and users’ perceptions of the affective value of these icons. Examining emoji as a stylistic resource is useful for gaining insight into how people convey feelings of connectedness and how usage patterns arise as they do this within an online community. We draw on the concept of foregrounding from stylistics and insights from sociocultural linguistics and conversation analysis (CA) to demonstrate how emoji form part of the intersubjective and dialogic enactment of stance that contributes to style construction. Our data are taken from Kaskus, the largest online community in Indonesia. We conduct a qualitative analysis of emoji use in a collection of texts comprising 1,000 posts from two discussion forums (approximately 107,000 words in total) and analyse four dialogic uses of emoji that give rise to stylistic effects: emoji for indexing convergent alignment, divergent alignment, intensity of emotion and to report one’s sentiments towards a third party. We demonstrate that the social meanings of emoji are not merely created by individual users but rather emerge through interaction between users. This can be seen in the

108

CONTEMPORARY MEDIA STYLISTICS

dialogic resonances produced by use of emoji in posts that repeat or contrast with the emoji of others through processes of intersubjective alignment. Finally, while others have noted that emoji can disambiguate online messages with respect to affect, we show that emoji can also be ambiguous with regard to other aspects of a message and argue that ambiguity can be useful for mitigating uncomfortable situations.

1 INTRODUCTION Emoji, the ‘colourful symbols – the winks, smileys, lovehearts and so on – embedded as single character images, or glyphs, in our digital keyboards’ (Evans, 2017: 7), can be found within different mobile phone platforms, email programs and popular social media sites such as Facebook, Twitter and WhatsApp, where they are often supplemented by additional site- or app-specific images. The word ‘emoji’ combines two Japanese words – e ‘picture’ and moji ‘letter, character’ – and these images are said to form a universal language comprising ‘prefabricated signs’ and carrying a standard set of nuances, which users imbue with individual meanings (Danesi, 2016; Evans, 2017). In this study we show that emoji are not only signs people employ to index individual meanings and stances but symbolic resources they deploy intersubjectively and dialogically across different posts and messages. Style, as we use the term here, is understood as ‘the style of texts’ (Leech and Short, 2007: 11), that is, style as examined through the choices of elements that are characteristic of the texts (Leech and Short, 2007: 11; Jeffries and McIntyre, 2010: 15), including non-linguistic elements such as emoji. Combining verbal expressions,1 emoji and specific layout on the screen, the texts we examine are inherently multimodal. As Nørgaard (2019: 17) points out, ‘[M]ultimodal meaning extends beyond the simple addition of discrete modes and the meanings they realise individually.’ This accords with Dresner and Herring’s (2010) approach to emoji, which treats these icons as part of text rather than as a separate, non-verbal addition to text. Working with linguistic elements, emoji can playfully reinforce, augment or frame the content of a verbal message in a particular way, and can ‘instruct’ the addressee on how to interpret the message. Various studies (e.g. Stark and Crawford, 2015; Danesi, 2016; Evans, 2017; Gibson, Huang and Yu, 2018) have examined the social and cultural dimensions of emoji, underlining their main function in enabling users to inject humour and light-heartedness into online communication. Emoji have also been claimed to be signs with universal import. Evans (2017), for example, states that emoji represent a universal language that enables people to convey an ‘emotional voice’ and to add nuance they cannot create with verbal expressions alone. Similarly, Stark and Crawford (2015) describe emoji

THE STYLISTICS OF EMOJI

109

as ‘signifiers of affective meaning’ and cultural objects that ‘enable sociality in the digital networks’, providing users with a ‘quick and efficient way to bring some color and personality into otherwise monochrome networked spaces of text’ (p. 1). Danesi (2016: 51), citing a study by Novak et al. (2015), points out that emoji are primarily employed to ‘add positive nuances to the content of texts, no matter what their intent or purpose’. A similar point is made in other work, which ranges from experimental studies designed to determine the level of affect conveyed by different emoji (e.g. Riordan, 2017) to studies testing market preferences by measuring the association between commercial products and emotion (e.g. Swaney-Stueve, Jepsen and Deubler, 2018). Much of this previous work has focused on the affective function of emoji in relation to the linguistic expressions with which they are produced. In this study we move beyond the direct linguistic expression–emoji connection to examine the dialogic resonances that emoji can produce across texts and between interactants. We examine emoji use within the Indonesian language online community, Kaskus, in order to show how emoji have emerged as an important stylistic resource within intersubjectively informed language use that promotes feelings of interconnectedness through convergent and divergent alignment. By looking at computer-mediated communication in Indonesian – a language rarely examined in studies of CMC – our study contributes to understanding the evolution of affective markers in digital communication more generally.

2 BACKGROUND Emoji, like emoticons before them, are a recent innovation. On early online communication platforms in the 1980s, punctuation began to be used to duplicate the popular smiley face as a sign that a message should not be taken seriously. This gave rise to emoticons, a series of conventionalized graphic representations of emotional states, utilizing the limited resources of the ASCII keyboard (Danesi, 2016: 2). Once graphical interfaces become more developed and widely available, Shigetaka Kurita developed the first set of emoji in 1999 for a Japanese mobile phone platform. These provided a more visually satisfying alternative to the graphic imagery of emoticons (Danesi, 2016: 2–3). Over the following decade, a standardized set of emoji has evolved and is available across platforms, although the actual visual implantation of each emoji will vary with different operating systems and service providers. At the same time, different organizations have also introduced their own specialized emoji, both mirroring and augmenting the standard forms. The emoji we examine are one such set. While non-verbal signs such as capitalization and the exclamation mark are commonly used to convey affect by indicating increased volume, stronger tone or greater intensity of emotion (Dresner and Herring, 2010), emoji, like emoticons, have become important tools for augmenting the communication

110

CONTEMPORARY MEDIA STYLISTICS

of affect. Users achieve this by drawing on various communicative resources, including structural positioning, combining different emoji and repeating the same one(s), and as we will show, by the size of emoji. While emoji carry a conventionalized set of meanings, users also imbue them with individual nuances (Danesi, 2016: 53). Emoji are not simply pictograms, representing what each is a picture of, but are also ideograms, serving as symbolic representation of concepts (Stark and Crawford, 2015). Riordan (2017) shows that both emoticons and emoji are flexible in the sense that the meaning they convey will be dependent on user and context. They can thus be effective in doing emotion work (Hochschild, 1979), that is, work that helps formulate and convey one’s own emotions in relation to others’ emotions for the purpose of preserving and enhancing social relationships. In sum, emoji are effective because they do not simply convey emotion but are fundamentally relational and social as well. Emoji currently consist of face and non-face icons. Like emoticons, face emoji convey a range of facial expressions associated with different emotions. Non-face emoji include pictures of hand gestures, everyday objects (e.g. shoes, handbags, sunglasses, hats and animals), places, objects associated with certain activities (e.g. tent for camping) and so forth. According to Riordan (2017: 563), non-face emoji fulfil a similar function to face emoji in enabling users to ‘act out emotion work that preserves and enhances social relationships’. Riordan also argues that, although emoji can communicate positive affect, they cannot change ‘the valence’ of the verbal message itself. Thus, a negative message can be softened by adding an emoji, but the emoji cannot alter it into a positive message (2017: 562). Structurally, emoji can be placed at the end of a clause or sentence where punctuation is usually placed, or as a free-standing element in place of verbal forms. Danesi (2016: 52) refers to these as ‘adjunctive’ and ‘substitutive’ uses respectively. Adjunctive use is by far the most common. For example, a smiley face is often placed at the end of a sentence where one would expect a full stop, in order to ‘instruct’ the recipient on how to read the affective tone of the message. Substitutive use is when emoji substitute for lexical items. For example, linearly sequenced emoji showing a hand grenade, a seashell and a bikini are understood as standing for ‘bombshell bikini’ (Danesi, 2016: 78). Emoji are also often ‘pluralized’ (Danesi, 2016: 82), that is, the same icon is repeated to indicate intensified emotion. Users have augmented static emoji by placing them alongside animated GIFs, and more recently, moving emoji (animated emoji GIFs) have become widely available through various internet platforms. Like GIFs in general, animated emoji can be used to attract attention. Another important aspect of the relationship between text and emoji is the relative size of the inventory. Generally, language users have a wide variety of lexical and grammatical resources to draw from when composing text. In contrast, only a limited number of emoji are available. Yet, possibly because

THE STYLISTICS OF EMOJI

111

of this limitation, emoji are employed to convey a wide range of meanings and provide various emotional nuances. For example, a user may employ the grinning-face emoji to communicate sarcasm, while another user may deploy the same emoji to index a self-deprecating stance. The Kaskus emoji, which form the focus of our study, are created in-house and include emoji that are similar to many of the standard internationally available emoji as well as many that are particular to this online site. Kaskus administrators refer to their emoji as smilies and a page of the Kaskus Help section lists the available emoji and provides short one-word explanations of their meaning (Kaskus, 2018b). The inventory of emoji currently available to Kaskus users includes eighty-six large, unique emoji that are characterized as cuma di KASKUS ‘only on Kaskus’, twenty-seven small, unique emoji that are also labelled cuma di KASKUS and seventy-seven small emoji which are described as ‘standard’. The large emoji are very popular among users. They all involve human-like cartoon figures in a variety of bright colours (e.g. examples (2), (4) and others). Most include heads and shoulders, cut off at the chest. A few are full body and involve two characters interacting. The large unique emoji are all animated GIFs showing the characters variously smiling, blinking, waving or otherwise gesturing. The Kaskus emoji represent a wide range of emotions, events, actions and objects. As noted, all the large emoji involve human-like characters and when objects are included (e.g. bedug ‘large drum kept in mosque and used to call people to prayer’, ketupat ‘rice cake wrapped in coconut leaves’ or bajaj ‘auto rickshaw’), characters are shown utilizing these objects. The small, unique emoji are also animated GIFs. They include only a round head and no body, although occasionally the character is holding something in their hand. Again, they appear in a variety of colours and show emotions, events, actions and utilized items. The seventy-seven non-Kaskus-specific emoji are almost all static images (six are animated). About half of them are round heads, twenty-four are objects (e.g. a television or telephone) that do not include a human-like character and ten are animals, including cats, dogs and the Linux penguin. Most of the heads are yellow, which is reminiscent of the iconic smiley face graphic popularized in the 1960s (See example (1)). As noted, they are described as ‘standard’ and many are similar to different smileys, emoticons and emoji seen elsewhere, but the images are still Kaskus specific. From this inventory, the various face emoji are by far the most commonly used in the Kaskus forums, including the buying and selling forums and the relationship forums from which we draw our data. The object emoji are underutilized on Kaskus, possibly due to the fact that they are restricted to the less interesting ‘standard’ set. The face emoji are more popular, probably because they make up the majority of Kaskus emoji, including all of the most colourful large emoji. It also possibly reflects the fact that, by and large, users in

112

CONTEMPORARY MEDIA STYLISTICS

this online community, like other online communities, do not know each other in real life and project their online identities only through their online name and avatar, and the ways they construct their posts. In this respect, face emoji provide them with a convenient and effective resource for styling themselves online. Having an online name and using face emoji enable users to present themselves as real people with emotions rather than obscure voices that could be interpreted as the voices of intruders or people with questionable intent.

3 THEORETICAL ORIENTATION People use emoji to convey emotion to others, and therefore a stylistic study of emoji would appropriately take as its departure point this dialogic dimension. Although the affective function of emoji has been much discussed, and many studies include screen capture of online messages as an illustration of how people employ them in online interaction, how emoji are responded to in extended interaction and how individual emotional voices are accommodated in multiparty interaction are only recently being explored (Al Rashdi, 2018; Gibson, Huang and Yu, 2018; Kelly and Watts, 2015; Sampietro, 2019). Moreover, the question of how emoji serve as a stylistic resource is currently under-explored. In what way can we understand the use of emoji as part of style building? As interactants use emoji to convey feelings of connectedness, what kinds of stylistic effects emerge? The systematic study of texts is the business of stylistics as a discipline (Jeffries and McIntyre, 2010: 21). As Jeffries and McIntyre (2010: 24) point out, stylistics draws its strength from being open to different theoretical approaches to and methodologies for analysing language and literature, but the central concern of the discipline is ‘with the style of particular texts, whether they are representative of a genre, an author, or themselves alone’ (2010: 15). Although meta-textual factors (e.g. the context of a text’s production, the author’s personal background, etc.) may come into consideration in the analysis of meaning, ‘the unavoidable basis of all stylistics remains the text itself, and the linguistic choices that have been made (albeit unconsciously) to arrive at a particular form of words’ (Jeffries and McIntyre, 2010: 15). In terms of our study, and indeed, stylistics studies whose object of analysis is not limited to words alone, such as multimodal stylistics (e.g. McIntyre, 2008; Nørgaard, 2019), these ‘choices’ include linguistic and non-linguistic signs. Our study draws on the concept of foregrounding (e.g. Douthwaite, 2000; Emmott and Alexander, 2014; Leech, 2008; Leech and Short, 2007) and we combine this with insights from sociocultural linguistics (Bucholtz, 2008; Du Bois, 2007; Du Bois and Kärkkäinen, 2012), particularly with regard to intersubjectivity, stance and alignment, to examine how style emerges from practices of using emoji in online interaction. Foregrounding, a foundational

THE STYLISTICS OF EMOJI

113

notion in stylistics, is concerned with the way certain parts of texts are highlighted or made prominent through techniques such as repetition and parallelism, giving rise to unusual or defamiliarizing effects. Stockwell (2002: 14) includes unusual naming, innovative descriptions, creative syntactic ordering and the use of creative metaphor in the range of devices employed for foregrounding. Foregrounding is understood as a relational notion; certain parts of a text stand out – they are foregrounded – relative to other parts. As Leech and Short (2007: 40) point out, a prominent feature of style stands in a relationship of coherence with other features of style. In cognitive poetics, the relationship between foregrounded and backgrounded features is understood as corresponding to the figure/ground relation in Gestalt psychology (Stockwell, 2002: 14). In digital communication, emoji are a resource that can be exploited to achieve similarly unusual effects. Generally speaking, a single emoji in a mainly verbal text stands out visually relative to the rest of the text. When multiple emoji are used, other foregrounding effects can come into play. In this study, we take a dialogic approach and begin from the basic assumption that people employ emoji to convey feelings to others and proceed to identify what kinds of practices stand out as they do so in an online community. As we pay careful attention to how interactants use emoji in constructing their posts in response to others, insights from CA, particularly the notions of adjacency pair and sequentiality, help elucidate how the dialogic use of emoji is structurally organized. By taking this ‘dialogic-sequential’ approach (Du Bois and Kärkkäinen, 2012: 441) we support the broad understanding in stylistics that style is not randomly built but emerges through empirically observable patterns of use. 3.1 Intersubjectivity, alignment and stance Intersubjective understanding is both prerequisite for social interaction and constantly being shaped by social interaction. It emerges through shared activity (Mead, 1934), joint attention and experience of others’ embodied actions (Gallagher and Hutto, 2008), rather than through some form of mentalist analogy between one’s own and others’ minds (Zlatev, 2008). According to Verhagen (2008: 307), patterns of language use emerge as ‘mutually shared solutions to coordination problems, rules that are followed because of the expectation that others will follow them and because one knows that others expect one to follow them’. The dynamic nature of intersubjectivity can be seen through two important related concepts, common ground and alignment. Common ground is the shared knowledge, beliefs and suppositions that allow people to communicate (Clark, 1996). This does not mean that common ground is identically or exhaustively shared but rather that it is sufficiently

114

CONTEMPORARY MEDIA STYLISTICS

shared for communication to proceed successfully for the immediate purposes of interlocutors. Common ground is also not a fixed commodity but amenable to negotiation, augmentation and modification throughout the course of interaction. The concept of alignment helps us understand the dynamics of how common ground is (re)negotiated. Alignment is not simply a matter of agreement but is rather about degrees of ‘commensurability’ (Du Bois, 2007), which can be viewed along a cline of alignment and disalignment (Du Bois and Kärkkäinen, 2012). In relation to conversation, Kiesling (2011) states that alignment is ‘achieved when interactants are cooperative in the project of creating an interaction’, pointing out that this is ‘not the same as saying interactants agree about denotational content, but only that they are on some level engaged in moving the conversation forward’ (p. 4). Intersubjective alignment is often taken for granted by language users; it simply allows interaction to progress unproblematically, and so is often not explicitly noted by interactants. Nonetheless, from time to time ‘intersubjectivity rises to focal prominence’ (Du Bois, 2007: 159) and is explicitly attended to by participants. The way language users attend to intersubjectivity is not just a matter of individual actions but involves responses to others’ actions and the anticipation of actions to come. Intersubjective alignment is thus a kind of ‘distributed action’ (Du Bois and Kärkkäinen, 2012), dependent on the constant (re)alignment and (re)negotiation of common ground. Intersubjectivity and alignment are also crucial to the enactment of stances and thus the emergence of styles. In sociocultural linguistics, style is understood broadly as socially meaningful ways of doing and being that contrast with other ways of doing and being (Hymes, 1974). In particular, Bucholtz (2015) has shown that ‘style is a system of sociocultural positioning through modes of semiotic action’ (p. 32), that is, it involves multimodal and multidimensional clusters of linguistic and other semiotic practices that speakers use to display and construct identities in interaction. Although this conception of style is broadly concordant with the understanding of style in the discipline of stylistics, the main concern of stylistics, as noted, is with the style of texts. This means that the primary aim of analysing semiotic and multimodal practices in style building is to understand meaning in texts. Stance is important to our understanding of how styles emerge and are enacted by social agents. Du Bois (2007) usefully defines stance as ‘a public act by a social actor, achieved dialogically through overt communicative means, of simultaneously evaluating objects, positions, subjects (self and others), and aligning with other subjects, with respect to any salient dimensions of the sociocultural field’ (p. 163). Crucial to this is the notion that people engage with, react to and attempt to influence each other in relation to an object of shared attention, the object of discourse. Stances are enacted locally in

THE STYLISTICS OF EMOJI

115

discourse through the use of linguistic and non-linguistic resources drawn from the wider community. Stances themselves will be responded to as participants engage in ongoing processes of (re)alignment, and Du Bois and Kärkkäinen (2012) use the term ‘dialogic resonance’ to describe this ‘process of alignment between subsequent stances’ (p. 445). Furthermore, it is through the repetition of stances, and their association with the semiotic resources (both linguistic and otherwise) through which stances are enacted, that styles emerge as stabilized repertoires of ways of doing things linked to situations and social identities (Bauman, 2004; Bucholtz, 2015; Eckert, 2012; Johnstone, 2009; Ochs, 1990). Style, as it contributes to the meaning of text, emerges from this ongoing (re) association of semiotic resources and stances. 3.2 Emoji and social action The basic goal of CA is to identify particular conversational practices within contexts of use and to extrapolate the larger orders of organization in which they fit. This is done to develop an ‘understanding of how basic social actions are produced and recognized, and how their production and recognition are located and shaped within the institution of interaction’ (Heritage and Clayman, 2010: 16). By social action is meant semiotic acts that are publicly recognizable, culturally contextualized, as a response to something, and which call for a response (Enfield and Sidnell, 2017: 3). Examples include complaining, questioning, requesting and advising. In conversational interaction, social actions are regularly followed in sequence by responses, in a structure referred to in the CA literature as ‘adjacency pairs’ (e.g. question followed by answer or request for advice followed by provision of advice). The asynchronous nature of online interaction and the way users quote and respond to each other means that the two parts of such pairs are often not physically adjacent in the Kaskus data we analyse. Nonetheless, users regularly respond to others’ contributions, and the links between initiating actions and responsive actions are usually clear. Insights from CA have been increasingly applied to online interaction through analytical practices that have come to be known as ‘digital CA’ (Giles et al., 2015; Giles, Stommel and Paulus, 2017). This approach is valuable for providing tools to analyse turn construction in interaction as it pays careful attention to recipient design, that is, the ‘multitude of respects in which the talk by a party in a conversation is constructed or designed in ways that display an orientation and sensitivity to the particular other(s) who are the co-participants’ (Sacks, Schegloff and Jefferson, 1974: 727). At the same time, digital CA recognizes CMC-specific issues such as the fact that the linearity of turn constructions is disrupted, with users able to create their own linearity by choosing which past posts they want to respond to and in what order. In addition, online interaction, even when ostensibly between two users, is ‘polylogal, with

116

CONTEMPORARY MEDIA STYLISTICS

an unlimited number of interactants dropping in and out of the interaction’ (Giles et al., 2015: 49). It also involves not only explicit interactants but also observers such as the moderator, other members and the general online public.

4 METHODOLOGY In this study we use data from Kaskus, the largest online community in Indonesia, where emoji have become important markers of stance, not unlike the ubiquitous interactional discourse markers of conversational Indonesian (Djenar, Ewing and Manns, 2018). Kaskus is a social networking site originally set up by three Indonesian university students in 1999. In 2005 it was named the best and largest online community by PC Magazine Indonesia, and by 2014 it boasted more than 7.8 million members (Kaskus, 2018a). Kaskus offers a range of services, including a buying and selling platform, reviews of hangout places, books, comics (particularly manga) and games, advertisements (including job vacancies), and online news, but by far its most distinctive feature is its chat forums, from which our data are taken. We conduct a qualitative analysis of emoji use from two Kaskus discussion forums: Ask Da Girls, a forum intended for (mostly) female users to discuss men and relationships, and Ask Da Boys, a companion forum intended for (mostly) male users to discuss women and relationships. These forums are public and open. While one must be a registered member to post, all posts are publicly available and can be read by anyone with internet access. Having publicly available data is a way of ensuring replicability; that is, other researchers can test the validity of our analysis by replicating our study using the same data and similar or different analytical approaches. Our data set consists of 1,000 posts, 500 from each forum, all published in 2011. The same female moderator, whom we call XP, oversees the two forums, often raising issues for discussion and answering members’ questions. Forum members also often raise questions and respond to both the moderator and other members. Kaskus owns copyright to all material on their website – something users acquiesce to in agreeing to Kaskus terms and conditions – and the material analysed here, including text and emoji, is used with the permission of Kaskus. In order to respect the privacy of individual users, and in line with the approved ethics protocol, user names, avatar images, metadata relating to date and time of posting and all other kinds of identifiers that appear on Kaskus are not reproduced here, and users are only identified by code letters.2 The 1,000 posts from Ask Da Girls and Ask Da Boys were produced by 282 users. Most users only produced one or two posts and just twenty-four (9 per cent) of the contributors (including the moderator) account for half of the posts. The data consist of approximately 107,000 words of published material (excluding metadata), of which 55,200 is original text and 51,800 is quoted

THE STYLISTICS OF EMOJI

117

material. On average then, posts across both forums contain fifty-five words of original text and fifty-two words of quoted text. Of the total number of posts, only 200 (20 per cent) do not contain emoji at all. The rest contain emoji either in the quoted material or in the original text, or both. In this study we are particularly interested in how emoji are used as resources for displaying dialogic resonance across posts and thus contribute to the intersubjective enactment of stances through interaction, which in turn give rise to the emergence of style. To this end, we analyse only posts that are relevant to the dialogic use of emoji. That is, we only look at posts containing emoji that speak to a prior post or posts that also contain emoji. We examine these posts in their interactional context to determine emergent social meanings of emoji use. Not all of these posts can be said to contain emoji use that speaks to prior emoji use. Example (1) illustrates the occurrence of emoji in both a quote and an original text that responds to the quote, but the emoji themselves do not contribute to the dialogic resonance between these two parts. (1) Post by RG: Non-dialogic use of emoji Quote: Original Posted by CP ➤ cewe yang enak didekati itu yang friendly, cheerful & bisa diajak susah. jadi sekalinya kalo si cowo ngajak jalan dengan budget minimum ga akan ada masalah.Ok ane rasa ini jawaban yang ane butuhin.. thx ya om.. mau ane taro pejwan..

eit bentar dulu ... definisi enak itu kan beda2 loh buat tiap cowo. tapi kalo emang menurut lo cewe kek gitu enak buat didekati, then go ahead

CP Quote: girls that are great to approach are those who are friendly, cheerful & don’t mind if you’re poor. so when a guy on low budget takes her out there won’t be a problem. Ok I think this is the answer I need..thanks om..3 I will put it on page one …

118

CONTEMPORARY MEDIA STYLISTICS

RG:

whoa wait a minute … the definition of great is different for every guy. but if you think it’s that kind of girl that’s great to approach, then go ahead.

CP ends their post with a handshake emoji. However, this emoji does not contribute affect to the specific points made by CP; rather, it closes the post by indicating connectedness between fellow Kaskus users. RG’s response comments on the specific point CP had made about the kind of girl who is enak, ‘nice, great’, to approach and CP ends this point with a smiley emoji. Thus CP’s emoji use is not picked up on by RG. Similarly, RG uses an emoji in response to a point by CP that does not use an emoji. In this way the emoji used by CP and RG are operating independently of each other and do not display the kind of dialogic resonance we focus on, nor do they contribute to the intersubjective construction of style. In contrast, dialogic resonance between emoji occurs when both an initiating action and a responsive action contain emoji, and the emoji contribute to and highlight the intersubjective negotiation of stance that occurs between the contributors. Example (2) provides a simple illustration of this process. (2) Post by XP: Dialogic resonance between emoji Quote: Original Posted by GW ➤ chat sm dia pun susah amat bkin dia bls, kdg pertanyaan” ga djawab

karena dia emg ngaku pendiem sih so, what can I do ? hmm pendiem mah relative pendiem karna belom nemu yang klop aja emang ga ada cara laen ya selaen di FB? deketan ga?

THE STYLISTICS OF EMOJI

119

GW Quote: even when I chat with her it’s difficult to get her to respond, sometimes she doesn’t respond to my questions

because she has even admitted she’s a quiet person so, what can I do? XP:

hmm quietness is relative (someone may be) quiet because she hasn’t found the right person yet isn’t there another way other than through FB? are you close to her?

GW comments that a girl he likes is a pendiem, ‘quiet person’, and he indicates humorous affect with a smiley face. In her response, XP comments twice about quiet people and uses the same smiley emoji each time. The repetition of emoji is indexical of XP and GW’s shared attention to the object of discourse (i.e. pendiem ‘quiet person’) and the dialogicality that obtains between their posts. This is a straightforward example of the kind of dialogic resonance between emoji that we are looking at. Of the total 1,000 posts that make up our data, we identified 482 posts that include emoji both in quoted material and in the original text of the post. Of these, 284 posts (58 per cent) contain emoji used for dialogic resonance, that is, the emoji occur in both the initiating action and the corresponding responsive action. While many examples of dialogic emoji use are as straightforward as that illustrated in example (2), many are much more nuanced and interactionally complex. The analysis presented in Section 4 explores this complexity by demonstrating ways that emoji can be deployed, dialogically and intersubjectively, to link social actions and responses to create convergent and divergent alignment between Kaskus participants. Following the basic tenets of CA methodology (Seedhouse, 2005), we aim to take the participants’ perspective in the context of the online interactional environment. Reliability is achieved by making the data transparent and taking the reader through a detailed understanding of the analysis in order to determine how convincing it is (Seedhouse, 2005). To this end, we present several extended examples to illustrate how emoji can be deployed for the intersubjective and dialogic enactment of stance. Although ethics considerations prevent us from presenting screenshots from the forums, we lay out the original Indonesian in a format mirroring how it appears online, together with free English translations. Rather than simply claiming a given case of emoji use has a certain meaning, we lead the reader through a detailed micro-analysis of each example, showing how we have come to this conclusion.

120

CONTEMPORARY MEDIA STYLISTICS

Due to the sequencing complexities that arise as users quote previous posts and respond to them asynchronously, it can sometimes be difficult to work through the interactions. A post may contain responses to several previous posts, and these posts do not always occur immediately prior to the responses. Moreover, sometimes participants do not clearly indicate which of their responses attend to which previous post(s). Following CA practice, we attempt to ensure internal validity of the analysis by paying close attention to the interaction (Seedhouse, 2005). That is, it is through user responses to the way emoji are deployed that we glean evidence for their social meaning at any particular point in interaction. Because this is essentially a case study of particular moments of online interaction, external validity is more difficult to achieve. Nonetheless, we follow the CA understanding of ‘individual instances as products of a machinery, [therefore] the analysis is not limited to one particular [case], so some generalizable findings emerge’ (Seedhouse, 2005: 257). By ‘machinery’ we mean context-free norms of organization, that is, rules or regularity that informs use. Thus, the micro-analyses presented in the following section contribute to building up a larger picture of regularity in online interaction by Kaskus users.

4 EMOJI USE AS STYLISTIC PRACTICE The practices of emoji use we explore involve the dialogic resonances of Kaskus users’ emoji deployed in response to the emoji of other users. In the following subsections we present an analysis of four ways of using emoji, showing how emoji contribute to stance taking in online interaction, and, in turn, how this contributes to the emergence of Kaskus-specific styles of interaction: use of emoji to index convergent alignment, divergent alignment, intensity of emotion and as reported speech, conveying to a co-user the sentiment expressed to a third party. 4.1 Indexing convergent alignment through emoji repetition Kaskus users often repeat another user’s emoji to signal convergence of stance intersubjectively, for example, when tacitly agreeing with a point another user has made. Al Rashdi (2018) also notes that Arabic WhatsApp users similarly repeat another’s emoji to ratify and approve of what has been posted. In the Indonesian data we examine, the repetition of another’s emoji is often accompanied by the repetition of another’s wording and grammatical construction. As noted previously, the number of extant emoji is limited relative to the inventory of lexical and grammatical resources that are available to users. Thus a participant may repeat the emoji employed

THE STYLISTICS OF EMOJI

121

by another participant, but the nuances conveyed may differ as the verbal expressions that accompany the emoji differ. Examples of this dialogic use of emoji are given in (3) and (4), showing posts involving four participants: DK, XP (moderator), ND and OK. Example (3) shows the original post by DK (note that the elements in brackets in the English translation indicate material that is grammatically required for well-formedness and comprehensibility in English but are not explicitly expressed in the Indonesian original). In this post, DK asks why guys are not good at waiting. DK mentions that she4 loved a guy but does not want to commit to a relationship, so she told him to wait for five years before she and he should commit to a relationship. DK asserts that, if they wait long enough, by the end of that waiting period, and if everything goes well, they could proceed straight to marriage (implying this is better than dating for a number of years with no certainty of marriage). DK then argues that, if the person really loves her, he would surely be willing to wait. DK inserted a grinning-face emoji at the end of her assertion to convey a light-hearted stance towards the object of discourse and to mitigate potential criticisms. The structural position of the emoji at the end of DK’s assertion is important in indexing these. Without the emoji, the assertion could be interpreted as a stance of arrogance or a misled assumption about relationships. (3) DK’s original post knp cowo gag mau bgt nunggu ‘dlm artian’ nunggu jwban gitu.. masa gw suruh nunggu 5th gag mau,biar langsung nikah gitu kalo dia cinta mah pasti mau. ini malah gw di tinggal.menurut agan2 gmn?? DK:

why can’t guys wait ‘meaning’ waiting for an answer.. I asked (him) to wait for 5yrs (but he) refused, so (we if he loved could) go straight to marriage from there (me) (he’d) definitely want (to wait). but instead he left me. what do you think agan??5

DK’s post was subsequently responded to by three co-users: moderator XP, ND and OK, as shown in (4), with the quotes from the co-users given in shaded boxes and the elements discussed, in bold.

122

CONTEMPORARY MEDIA STYLISTICS

(4) Post by DK: Convergent alignment Quote: Original Posted by XP ➤ nyahahah 5 taon sih lo yang keterlaluan nunggu j glom ada kepastian buat apa? mending lo terima, pacaran lima taon, itu lebih manusiawi oo gitu gan, gw kan mikirnya kalo cinta ya lug a bakalan berpaling. tp ternyata salah gan :”( Quote: Original Posted by ND ➤ buset 5 taon.. cuma nunggu jawaban lage.. klo 5 taon sudah a a hubung n kemungkinan m si bisa. klo jawabannya iya, klo n ga.. (jamuran busuk) dia tau gan gw jg sayang sm dia, tp gw ga mau pacaran, ane jg bingung ama diri ane gan … sekarang udah telat kayaknya. Quote: Original Posted by OK ➤ Edan, 5 tahun

XP Quote: nyahahah you’ve gone too far asking (him to wait) 5 years

why (would anyone) wait that long with no certainty better if you accept him, date for five years, that’s more humane DK:

oh I see gan, I thought if (you) really love someone you won’t go with someone else. but (I was) wrong gan :”(

ND Quote: That’s ridiculous 5 years.. just to get an answer.. if (you’re waiting for) 5 years and you two are already dating that might be ok.. if the answer is yes (that’s great), but if not.. ((he’ll) grow mouldy + rot)

THE STYLISTICS OF EMOJI

DK:

123

he knows gan that I love him, but I don’t want to date him, I’m also confused about myself … looks like it’s too late now.

OK Quote: That’s crazy, 5 years (of waiting) This thread highlights several ways dialogic resonance is produced through lexical, grammatical and emoji repetition. First, in terms of lexical repetition, all contributors make a reference to ‘five years’: DK initially mentions 5th, ‘5yrs’, and this is repeated by XP, ND and OK but with different spellings. XP and ND both wrote 5 taon, with the word tahun, ‘year’, being rendered here in the colloquial spelling, taon, while OK’s version, 5 tahun, is given in standard spelling. By using different spellings for the same temporal reference, these participants convey convergent alignment while also showing individual stylistic preference. Second, in response to DK’s post, XP, ND and OK all use exclamatory constructions. XP begins her post with the exclamatory, sarcastic laugh nyahahah directed at DK, and this is echoed by ND through use of the exclamatory expression, buset, ‘that’s ridiculous’. OK follows with a similar exclamatory expression, edan, ‘that’s crazy’. Though not lexically identical, these expressions are grammatically resonant and exemplify a dialogically produced parallel structure indexical of these responders’ convergent attitude towards each other. Third, dialogic resonance is further shown in the repetition of the large laughing face, animated emoji (as mentioned, all large-size Kaskus emoji are animated). Among the three participants who reply to DK’s post, XP, the moderator, is the first to respond. Instead of employing the same standard-size grinning-face emoji used by DK in her response, XP upgrades the level of affect by opting for a large-size laughing-face emoji to highlight a divergent stance (further discussion on divergence is given in the next subsection). OK joins in to use the same large-size emoji, placing it at a structurally similar position, namely, at the end of the exclamatory sentence, thereby indexing alignment with XP’s stance intersubjectively. The foregoing examples additionally demonstrate how users create ‘individual nuances’ of meaning (Danesi, 2016) through a multimodal combination of emoji and language resources, including spelling and verbal expressions. They express tacit agreement by repeating another’s use of emoji and also grammatical and lexical resources, though with differing amount of informational detail in the verbal expressions. For example, OK’s verbal response, edan, ‘(that’s) crazy’, indexes a similar evaluative stance to that of XP’s, but XP’s post is more elaborate. In this sense, the individual nuances of emoji that Danesi (2016) alludes to can be interpreted as nuances that emerge

124

CONTEMPORARY MEDIA STYLISTICS

as participants use emoji to intersubjectively construct and align their affective stances. Repetition of emoji as an index of convergent alignment is more dramatically shown in instances in which a participant repeats another participant’s multiple emoji in a structurally similar manner. One context in which this occurs is explicit agreement, that is, where a participant clearly indicates they hold the same opinion as another, as in the following example, taken from the same thread as (3) and (4). CG is responding to DK by quoting the latter part of XP’s post – the same post quoted by DK in (4). Foregrounding effects emerge in this example through the parallelism created by CG’s repetition of XP’s emoji and their structural position. (5) Post by CG: Explicit agreement Quote: Original Posted by XP ➤ cowo tuh disuru nunggu 10 menit aja ngomel apalagi 5 taon bener cinta juga bakal pudar kalo ga dihangatin terus bener banget dah makanya buat ngelupain cinta tuh gampang cari aja yang lebih oke

XP Quote:

guys complain when you ask them to wait 10 minutes let alone 5 years even serious love will fade if you don’t keep it warm

CG:

so right that’s why it’s so easy to forget the one you love just go and find a better one

In the quote, XP makes two assertions: first, guys are not good at waiting; second, relationships require constant maintenance, or they will quickly break down. XP attaches an animated laughing-face emoji and a shy face with eyes down emoji to these assertions, respectively. CG begins her message to DK by explicitly agreeing with XP’s first assertion (bener banget dah, ‘so right’) and attaches the same laughing-face emoji, thereby creating the first instance of dialogic parallelism of emoji in this example. In the next line CG tacitly agrees with XP’s second point (i.e. that relationships require constant maintenance)

THE STYLISTICS OF EMOJI

125

by repeating XP’s point but using alternative wording and the adverb makanya, ‘therefore; that’s why’, as a linker to the previous line. CG attaches the same ‘shy’ face emoji as that employed by XP at the end of her sentence to index her convergent alignment with XP’s stance, thus creating the second instance of parallelism of emoji dialogically. In addition to the two emoji, CG attaches a different non-face emoji in the form of an animated penguin wearing a red scarf (the colour not shown here) pacing back and forth to mark a humorous new point that twists the point made by XP. XP suggests that people need to put in continuous efforts if they want to maintain a relationship. In the final line of his message, CG writes that, if you fail to maintain your relationship and forget the person you love, you can just go and find another love. The pacing penguin indexes this differential, joking stance. 4.2 Indexing divergent alignment through emoji size and type In addition to using different emoji, choosing emoji of a different size is another strategy for indexing divergent alignment. As mentioned, the Kaskus emoji come in standard as well as large sizes, and the large emoji are all animated, while only a few of the standard-size ones are. The contrast in size and animation is often exploited by participants to index divergent alignment. Contrast is a relational notion; therefore, to interpret the meaning of someone’s large emoji, it is crucial to consider what the emoji is used in response to. In line with the dialogic approach we adopt in this study, the meaning of a participant’s use of a large emoji is determined relative to another participant’s emoji to which the emoji responds, and the linguistic expressions to which they are attached. Examples (3) and (4) were discussed earlier in terms of convergence between the three respondents – XP, ND and OK. These examples also illustrate divergence, in this case between the three respondents and the original contributor, DK. In her initiating post, DK uses standard-size laughing-face emoji. In her response, moderator XP deploys a large, animated face emoji. The emoji foregrounds her post in the following respects. First, although XP’s large-size emoji aligns with DK’s standard-size emoji in terms of conveying light-heartedness, the contrast in size and the fact that her emoji is animated while DK’s is not accentuates XP’s differential stance. Second, the choice of this emoji by someone in moderator role, who is normatively required to exercise neutral judgement in her response, produces an unexpected stylistic effect. Instead of giving neutral advice, the moderator ridicules DK, indirectly accusing her of being inhumane. Use of the large emoji in conjunction with the verbal expressions contributes to this playful but mocking stance. Third, the contrast between DK’s emoji in the initiating action and XP’s upgraded emoji in the responsive action suggests that DK’s mitigating use of emoji appears to

126

CONTEMPORARY MEDIA STYLISTICS

backfire. Instead of receiving a similar emoji, DK’s standard-size static emoji are ‘outdone’ by XP’s large, animated emoji. A similar example showing contrast in the choice of emoji between the initiating action and the responsive action is given in (6). MP is admonishing RE for engaging in name-calling with his girlfriend. RE’s post (as quoted by MP) is itself a response to another user (not shown here) who responded to the original post by yet another user (also not shown) who had asked for advice about how guys should behave in front of a girl they like. In his post, RE mentions that he and his girlfriend used to joke around by calling each other names, and he justifies this by suggesting that he and his girlfriend were very comfortable with each other. But the name-calling play went too far and they ended up quarrelling. (6) Post by MP: Indicating admonishment Quote: Original Posted By RE ➤ gw setuju nih … jgn jaim … pokokny jadi diri sendiri aj … gw dulu am ce gw sk kata2 an ampe parah emang paling enak pdkt di bawa becanda gan.. paling asik tuh ngatain cewe.. tp kudu liat2 cewe nya lo ya … ada batesan nya ngatain cewe gan RE Quote: I agree … don’t worry about your image … you should just be yourself … me and my girlfriend used to call each other names till it got really bad MP:

the best thing to do when you approach a girl is not to be too serious gan.. it’s really fun calling a girl names.. but depends on the girl right.. there is a limit (to what’s okay) when you call a girl names gan

THE STYLISTICS OF EMOJI

127

MP begins his message by seemingly aligning with RE’s stance, using no emoji. Then, he prefaces his admonishment with a large animated laughing-face emoji, which contrasts in size with RE’s standard-size laughing-face emoji. The emoji creates an impression of a light-hearted and friendly stance and is presumably chosen to downplay the admonishment. However, its sheer size and the linguistic component of the post make it clear that it is not just a playful gesture MP is signalling. MP subsequently attaches the same emoji adjunctive to his final admonishing sentence to reinforce his point. This ‘wrap around’ structural position of the two emoji is significant in foregrounding the admonishment. The emoji demarcate the admonishing part of the message from the opening line in which MP intersubjectively aligns with RE, and by having the emoji placed at each end of the message, a kind of parallel effect is created. As with XP’s large emoji, MP also uses the large emoji in a responsive action to express divergent alignment. Crucially, this indexical meaning obtains when the emoji is considered relative to the emoji in RE’s initiating action. In other words, we can interpret the social meaning of MP’s emoji by applying the notion of ‘backwards’ orientation in adjacency relationship (Schegloff, 2007: 16). Like a turn at talk that orients to a prior turn, MP’s emoji is a display of his understanding of a prior emoji employed by another. As mentioned, the laughing-face emoji creates an impression of lightheartedness and helps to downplay the disapproving tone of MP’s message. In this sense, the effectiveness of emoji in communicating affect can be attributed to the fact that people can deploy them to signal multiple stances. By choosing a laughing emoji, MP creates light-heartedness through mirroring the sentiment indexed by the laughing face standard-size emoji employed by RE, and by choosing a larger size emoji, MP is able to signal divergence. The sizes of the emoji thus iconically index the co-participants’ affective positions: the smaller emoji conveys RE’s sense of amusement, while the larger emoji indexes MP’s disapproving stance. 4.3 Pluralization of emoji and intensity of emotion The previous examples show repetition of emoji by participants across posts as a practice of conveying feelings of connectedness through convergent and divergent alignment. In the following, we explore repetition of emoji by the same participant – what Danesi (2016) refers to as ‘pluralization’ of emoji – to show how repetition also provides a resource for conveying intensity of emotion. We use the term ‘pluralization’ here to refer to repetition of the same emoji or the use of multiple different emoji by the same participant. However, pluralization is not a monologic practice. As will be shown, as one participant

128

CONTEMPORARY MEDIA STYLISTICS

uses pluralized emoji, another may respond in like manner in accommodation. Foregrounding effects emerge through the parallelism of emoji use by the same participant and also by different participants across posts. While pluralization of emoji, whether the same or different emoji, generally indexes intensity of emotion, pluralization of different emoji produces what we call a composite affective structure, by which we mean a structure of affective meaning that is composed of different emotions linked together by means of different emoji positioned consecutively. For example, a laughing-face emoji followed by a walking penguin wearing a red scarf may be used as a whole to indicate a mocking stance. The stance is composed of a mocking act (indicated by the laughing) and a distancing act from the object of mocking (a metaphorical walking away from the object). The two emoji in this case form a composite meaning that indexes divergence from a particular stance. Pluralization of the same emoji is exemplified in (7). This example draws attention due to the size of the emoji, its pluralization at the beginning of the post and the parallel structure produced through repetition of the emoji throughout the post. GB is asking advice about what to do regarding a girl he loves. He attaches four large ‘no hope’ animated face emoji in his address to female Kaskusers (as indicated by the address term sista, ‘sister(s)’) to indicate his state of desperation and uses the same emoji singularly throughout the post. The consistency of emoji choice in this post and its spread throughout the message convey the overall tone of desperation, with details about what causes this feeling being laid out in the linguistic portion of the post. (7) Post by GB: Long-distance relationship

Sista aku lagi ni aku bener bener sayang sama cew yang baru PDKT sama aku, aku kgk mau ninggalin dia, aku udah anggap tuh cew kayak pacar aku, tapi gara-gara aku pengen nglanjutin kuliah ke luar kota ku, dia kgk pengen LDR and udah mau ahkirin hubungan kita agar dia tidak sakit hati

. ni malem

malem saya aku sangat sayang sama tuh cew kasih solusi sista.. dia udah kgk mau buat LDR, usah aku paksa ia tetep tidak mau

THE STYLISTICS OF EMOJI

GB:

129

sista I’m feeling all I really love this girl who has approached me, I don’t want to leave her, I consider her my girlfriend, but because I want to continue my study in another city, she refuses to have a long distance relationship and wants to end our relationship so she doesn’t feel hurt

.

and at this time of night I am I really love that girl give me a solution sista .. she really doesn’t want to be in a long distance relationship, I really tried to persuade her but she refused This example is particularly striking in the way the emoji are organized in terms of number and structure, and the foregrounding effects this gives rise to. First of all, the four initial emoji in the address are repeated as four single emoji distributed throughout the body of the text. Second, these four single emoji are used in both ‘adjunctive’ and ‘substitutive’ functions (Danesi, 2016: 52) and these functions are distributed in an alternate structure, reminiscent of A-B-A-B rhyme scheme. The adjunctive function is shown in the second single emoji that follows the salutation (at the end of the clause ‘and wants to end our relationship so she doesn’t feel hurt’) and the fourth (final) single emoji in the post. Substitutive use is illustrated in the first and third single emoji in the post. Notice the sentence below the salutation, which is linguistically incomplete and is rendered complete with the ‘no hope’ face emoji inserted in place of a lexical adjective. Similarly, the second last emoji (ni malem malem saya + emoji, ‘and at this time of night I am + emoji’) is used in place of a linguistic expression describing emotion. The parallelism in this text is schematized as follows: emoji – emoji – emoji – emoji substitutive emoji adjunctive emoji substitutive emoji adjunctive emoji Unlike the previous examples in which stylistic effects emerge as users repeat another’s emoji to index convergent or divergent alignment and produce

130

CONTEMPORARY MEDIA STYLISTICS

dialogic resonance, this example demonstrates how stylistic effects can also emerge as a single user deploys emoji in a dialogic context to enact stance and invite alignment intersubjectively. Pluralization of different emoji and the construction of composite meaning are illustrated in (8). GB was devastated when his girlfriend told him, out of the blue, that she wanted to split up. HZ had previously responded to GB’s request for advice. In (8) HZ quotes GB’s response to her advice before providing a further comment. In the quoted message, GB includes two grinning-face emoji and a non-face emoji showing an animated penguin in a red scarf pacing back and forth to convey his joy at being able to take his girlfriend out despite her having expressed that she wanted to end the relationship. In response to GB, HZ also pluralizes the emoji (and appropriately excludes the walking penguin as HZ is not the one involved in the activity of going out). However, unlike GB who employs crying-face emoji and the grinning-face emoji to index mixed emotions, HZ uses only the grinning-face emoji throughout to convey empathy and encouragement. (8) Post by HZ: Pluralization and composite meaning Quote: Original Posted by GB ➤ sakit hati rasanya kalo berpisah sama tuh cewek tapi kok tadi waktu aku ajak maen maen dia malahan seneng bgt sist kok jadi aneh ya sista,kemaren malem dia pengen berpisah, tapi tadi dia malahan care banget sama saya ya tuh kan saya bilang apa sebenernya itu cewe juga ga pengen pisah sama situ, cuma dia liat kedepannya, dia ga mau nyakitin hati dia dengan LDR sama situ. dan mungkin dia memanfaatkan waktu2 terkahir dgn situ bersama hal yang baik2 ya sperti dia care sama situ mungkin dia ingin membuat kenangan akhir yang indah dan mungkin belum waktunya aja itu

GB Quote: (I) feel so heartbroken if (I have to) split up with that girl

THE STYLISTICS OF EMOJI

131

but when I took her out earlier today she was really happy sis It seems strange now sista, last night she wanted to split up, but earlier today she was so caring towards me HZ:

yes see what did I tell you that girl actually doesn’t want to split up, she was just thinking about the future, she doesn’t want to hurt herself by having a long distance relationship with you and maybe she wanted to spend the last moments with you doing nice things like showing she cares for you maybe she wanted to have a good memory of your final moments together and maybe she’s just not ready (to split)

Rather than indicating intensified emotion, in this example the pluralization of emoji represents accommodation to another’s style. In HZ’s previous posts, she uses very few emoji. This suggests that the use of emoji in (8) is prompted by GB’s style and thus accommodates to it. HZ’s emoji do not mirror GB’s exactly in terms of number or type but her choice of rhetorical strategy does. HZ uses pluralized emoji twice, mirroring GB’s use of pluralization. The first pluralization occurs at the end of her response to GB’s statement about his girlfriend being pleased when GB took her out, to which he attaches the mixed pluralized emoji. HZ’s pluralized emoji is thus dialogically linked to it. The second pluralization occurs in the subsequent line (at the end of her post), in which HZ expands on her point about why the girl was looking happy on being asked out. The pluralization thus functions to visually foreground that point, and in both instances, also to index convergent alignment. Example (8) also illustrates lamination of functions. While pluralization of emoji can signal intensified emotion, and this can occur in any post (either dialogic or non-dialogic), it can also be used as a strategy for accommodation, therefore, examining the pluralized emoji in both the initiating and responsive posts is useful for determining which social action by another a user is responding to and hence which function the emoji are being used to fulfil. 4.4 Using emoji as ‘reported speech’ The emoji use discussed so far is mainly concerned with conveying to co-users how one feels about someone, some situation, or what someone wrote, and it is relatively clear that the sentiments indexed by the emoji are displayed to or directed at co-users. But emoji use is not limited to this; they are also deployed to ‘report’ to co-users a sentiment expressed towards a third party. However,

132

CONTEMPORARY MEDIA STYLISTICS

unlike reported speech that can show clearly what one purportedly said to someone else, either through direct or indirect reporting strategies, emoji, being pictorial, can be ambiguous between conveying one’s feeling to a co-user and ‘reporting’ to a co-user how one conveyed one’s feeling towards a third party. For example, the head-hit-by-hammer emoji can be used to convey one’s own feeling of frustration, or to ‘report’ annoyance shown to a third party. The difference is subtle, but we argue that this capacity of emoji to convey sentiments in an imprecise way is precisely what makes them so effective for mitigating tension and potential conflict in online communication. Examples (9)–(11) are taken from a thread showing moderator XP and user SN discussing men and their smoking habits. Similar to the previous examples, dialogic resonance emerges in these examples through repetition of wording and convergent and divergent choices of emoji. Example (9) shows SN’s original post and XP’s response to it. SN is asking for advice about what to do to make her boyfriend stop smoking. She mentions that she has given her boyfriend various threats to make him stop and attaches a red angry-face emoji facing left (colour not shown) to convey to XP the sentiment she expressed towards her boyfriend. The direction of the emoji’s gaze is a clue that the anger is directed at a person external to the interaction. SN then attaches a yellow head-hitby-hammer emoji following her complaint that the situation is giving her a headache. This emoji can be interpreted as either directed at SN herself, saying she feels frustrated, feeling like her head has been hit by a hammer, or directed at the boyfriend, indicating that she feels like hitting his head with a hammer in annoyance for giving her a headache. In response, XP uses three different emoji, all adjunctively: a static grinning-face emoji, an animated emoji showing a blue hammer-on-head with tongue out emoji and an animated laughing-face emoji. (9) Post by XP: ‘Reported speech’ Quote: Original Posted by SN ➤ mo tny ni girls … knp cowok ane kgk bs brenti ngrokok sihhh … pdhl udah diancem ini ituu … puusiing … susah! kecuali kesadaran diri sendiri untung co gw ga ngerokok mau diancem ini itu juga didepan ente doank ga ngerokoknya

THE STYLISTICS OF EMOJI

133

SN Quote: (I) want to ask a question girls … why can’t my boyfriend stop smoking huuuhh … despite the fact that I’ve threatened him with this and that … (gives me a) headache … XP:

(it’s) hard! except if he (stops) at his own will luckily my boyfriend doesn’t smoke even if you threaten your boyfriend with this and that he’ll only stop smoking while you’re looking

Moderator XP attaches the grinning emoji in the first line of her response to inject a sense of light-heartedness and to ‘cool down’ SN’s anger. In her role as moderator, XP is expected to provide advice, and in undertaking this role, she often uses grinning or laughing-face emoji in her responses to mitigate facethreatening acts. These emoji create an impression of friendliness and increase the likelihood of XP’s advice being taken in good faith (though as we have seen, XP also uses a larger size emoji to mock her interlocutors). In her second line, XP attaches a blue hammer-on-head with tongue out emoji adjunctive to her statement saying she is relieved her boyfriend does not smoke. The emoji functions here like a conditional clause, with its propositional content aimed at the absent boyfriend: ‘If you smoke, I’ll hammer your head (and give you a headache).’ This blue emoji mirrors SN’s yellow head-hit-byhammer emoji; hence one might say this creates parallelism. However, the dialogic relation between the two emoji in the posts is more nuanced than this. First, in SN’s post, the sentiment of threat is conveyed with the red angry-face emoji, while XP’s hypothetical threat is expressed with the blue hammer-onhead with tongue out emoji. Second, while SN’s yellow head-hit-by-hammer emoji visually aligns with XP’s blue hammer-on-head with tongue out emoji, its ambiguity does not, as the latter emoji only makes sense if interpreted as a sentiment directed at XP’s boyfriend, not XP herself. The adjacency relationship between the different emoji in the two posts is schematized below (with the arrow indicating ‘responded to by XP with’). SN

XP

anger conveyed to third party

➔ light-hearted advice

experience of headache

➔ experience of headache reinforcement of light-hearted advice

134

CONTEMPORARY MEDIA STYLISTICS

SN responds to XP by first mirroring XP’s grinning face with the worded laugh huahahaha, showing dialogic resonance, before explicitly agreeing with XP’s point. The laugh, positioned at the beginning of SN’s response, additionally serves as a cohesive device that links her post to XP’s. After pondering that perhaps she should split up with her boyfriend, SN then repeats her complaint, upgrading it from having ‘a headache’ to feeling ‘tired of being angry all the time’, and reattaches the yellow head-hit-by-hammer emoji she used previously, this time to convey self-blame or regret. (10) SN’s response: ‘Reported speech’ Quote: Original Posted by XP ➤ susah! kecuali kesadaran diri sendiri untung co gw ga ngerokok mau diancem ini itu juga didepan ente doank ga ngerokoknya huahahaha.. bener bgt.. putus aje kali yee … capek juga ngamuk2 trus

XP Quote: (it’s) hard! except if he (stops) at his own will luckily my boyfriend doesn’t smoke even if he is threatened with this and that he’ll only stop smoking while you’re looking SN:

huahahaha.. (you’re) so right.. maybe we should just split up right … (I’m) sick and tired of getting annoyed all the time

In the quote box in (11), XP responds to SN’s second instance of complaint by admonishing her for having put herself in such a situation, attaching the ‘eek’ face emoji at the end of her admonishment. In reply, SN defends herself by saying her boyfriend did not tell the truth when they first got together, and she attaches the same yellow head-hit-by-hammer emoji adjunctively. As with the ambiguity in (9), the emoji here is ambiguous between conveying SN’s self-blame and SN ‘reporting’ that she expressed anger at her boyfriend.

THE STYLISTICS OF EMOJI

135

In her final clause, to which she inserts three exclamation marks, SN attaches the same red angry-face emoji that she used in her initial post, thus ‘concluding’ her complaint. As with her initial use of this emoji, the red face is being used here to convey to XP the angry sentiment SN expressed to her boyfriend. (11) Post by SN: ‘Reported speech’ – final response Quote: Original Posted by XP ➤ putus gara2 dia ngerokok? kalo gitu berat kenapa bisa jadian dl? dulu blg ny kgk ngrokok.. ternyataa!!!

XP Quote: SN:

splitting up just because he smokes? if it’s that bad why did you get together?

:

he said he didn’t smoke but in fact!!!

The above examples illustrate the affordances that the basic semantics of emoji can provide their users. Although the head-hit-by-hammer emoji essentially conveys a negative sentiment, participants can use it to index a range of feelings, from annoyance, self-blame, regret and self-deprecation. What is most interesting in these examples is the way the emoji is used in conjunction with linguistic expressions to create a kind of double-voice construction reminiscent of free indirect discourse. For example, in (11), the reported speech dulu blg ny kgk ngrokok, ‘he said he didn’t smoke’, is grammatically constructed as a report about what a third party said. The emoji, on the other hand, conveys a sentiment directed at that same third party but conveyed as if direct reported speech, that is, as if addressed to the boyfriend as second person. Riordan (2017) argues that affect communicated through verbal expressions is often ambiguous and that emoji can disambiguate this (also see Danesi, 2016). The foregoing examples show that emoji can themselves be ambiguous with regard to the kind of affect being conveyed and at whom it is directed. This should not be surprising, given that signs of alignment can be subtle and

136

CONTEMPORARY MEDIA STYLISTICS

elusive, producing what Du Bois and Kärkkäinen (2012: 440) characterize as ‘participants’ careful wrought achievement of strategic ambiguity’. Although the use of the red angry-face emoji and the yellow hammer-on-head emoji makes it clear that SN is expressing annoyance and frustration, her verbal expressions and the non-verbal sign (i.e. the multiple exclamation marks) render her post in (11) ambiguous with respect to who the affect is directed at – XP who admonishes her, herself for believing her boyfriend’s words or her boyfriend who supposedly lied to her. Whichever the interpretation, drawing on ambiguity can be a strategic move in an interactional context such as Kaskus where making assertions can be responded to with admonishment, and giving advice to individuals a participant does not personally know comes with the risk that those individuals may respond with hostility. Ambiguity can minimize this risk.

5 CONCLUSION This study has examined texts dynamically produced by multiple members of an online community in order to demonstrate the role of emoji, verbal forms and other non-verbal signs within people’s stylistic practices. Many past studies have analysed how people use emoji to augment and highlight aspects of their own posts. We have shown that in addition to this, it is important to look at the interaction across posts showing people responding to each other. Much of the emoji use we have identified on Kaksus is about responding to others’ ideas and feelings and therefore the role of emoji in texts can be better understood by taking intersubjective alignment into account. Emoji are not just expressions of the users’ own feelings but, crucially, also expressions of the intersubjective connection with others manifest through these dynamically produced texts. Several key points have emerged from our study. First, emoji occur in the majority of the posts we have examined and were often used dialogically, that is, deployed in response to others’ emoji use. In this sense, emoji can be said to form part of one’s style of relating to others and also relating to the styles of others. Users relate to others’ styles by taking up their styles of emoji use and in many cases, their verbal styles or by showing differential styles. Thus, style never speaks to itself but always exists in dialogic relationship with other styles (Djenar, Ewing and Manns, 2018: 1). Repeating others’ wording and grammatical constructions, as well as their use of emoji, gives rise to dialogic resonance between posts. Second, affective nuances interpretable from use of emoji not only are part of the idiolect of individual users but are subject to specific interactional contexts. Danesi (2016: 61) states that ‘emoji become communicative signs not in and of

THE STYLISTICS OF EMOJI

137

themselves but because we imbue them with meaning through symbolism and for communicative purposes’. Our study has shown that meaning emerges as users deploy emoji to engage with others in purposive activities (including to establish intimacy and foster a sense of belonging to a community) and index convergent or divergent alignment through dialogic actions. Third, through a close study of the dialogic use of emoji we can gain insight into the kinds of resources people draw on to convey different degrees of affect and to indicate a link between one’s message and another person’s message within the context of multiparty interaction. In the data we examined, the size of emoji, animation and pluralization are among the resources used to serve these purposes. It is a common practice within this community to begin messages with a cohesive device in the form of a smiley-type emoji (e.g. grinning-face emoji and animated laughing-face emoji) or a ‘worded’ laugh, to link what one is about to say to a prior post. Emoji are therefore valued both as markers of affect and devices for indicating cohesion between different elements within a message and across messages. Fourth, other studies (e.g. Danesi, 2016; Riordan, 2017) have pointed out that verbal messages can be ambiguous with respect to affect and that emoji can serve to disambiguate them. Emoji can help the sender ‘instruct’ the recipient on how to interpret a message. We have seen, for example, that a laughing-face emoji attached to an admonishing verbal message can be interpreted as the sender wanting the recipient to know the message is a friendly one. Our study has shown that, while it may be true that the addition of emoji to a verbal message can provide a clue to the kinds of emotions that the user intends to convey, emoji can also give rise to ambiguity. We have seen how an emoji can be ambiguous with regard to whom the feeling is directed at (e.g. whether to an interlocutor or a third party), or with regard to the feeling conveyed (e.g. frustration or annoyance). We argue that such ambiguity actually makes emoji an effective tool for mitigating potential conflicts, and that the potential for such ambiguity is an important additional characteristic of emoji, complementing the disambiguating function already noted by Danesi (2016) and Riordan (2017). Finally, understanding the functions of emoji in enhancing online communication requires looking beyond individual messages. A textual analysis of interaction between people who use verbal forms and emoji to communicate with each other online has provided a fruitful avenue for promoting this understanding. Our study contributes to contemporary media studies by taking a dialogic approach to analysing emoji. It also contributes to stylistics by adopting an interactional approach to analysing multimodal texts, an approach that can be usefully applied to other types of texts, produced online or offline, and adding to the existing interaction-focused stylistic studies.

138

CONTEMPORARY MEDIA STYLISTICS

NOTES 1. Following Danesi (2016), we use the terms ‘verbal expression’, ‘verbal message’ and ‘verbal form’ to refer to the linguistic elements in a message, in contrast to nonverbal elements, such as emoji. 2. Despite being anonymized in this chapter, at the time of publication these posts are still available online, allowing for the replicability of our analysis. 3. Om (from Dutch oom ‘uncle’) is used to address adult males of comparable age to one’s father but is here playfully used to address a member of one’s peer group. 4. Real-world sex or gender identification of users is, of course, unknown. These forums present discussions of heterosexual relationships in line with heteronormative expectations prevalent in wider Indonesian society. We base our choice of English gendered pronouns in translations and discussion on how users present themselves in this gendered social context. Note that choice of gendered pronominal reference only becomes an issue in English translation and discussion, as Indonesian pronouns are not marked for gender. 5. DK addresses fellow Kaskusers with agan. This term, and its short form gan (from juragan, ‘boss’), are historically associated with Arab traders in Indonesia (Djenar, Ewing and Manns, 2018: 57). Kaskus users employ this term and also Arabicderived pronouns ente, ‘2SG’, and ane, ‘1SG’, to address each other, including the moderator. These address terms entered Kaskus initially through the site’s buying and selling forum and are now used across its different forums.

REFERENCES Al Rashdi, F. (2018). ‘Functions of emojis in WhatsApp interaction among Omanis’. Discourse, Context & Media, 26, 117–26. Bauman, R. (2004). A World of Others’ Words: Cross-Cultural Perspectives on Intertextuality. Oxford: Blackwell. Bucholtz, M. (2008). ‘All of the above: New coalitions in sociocultural linguistics’. Journal of Sociolinguistics, 12(4), 401–31. Bucholtz, M. (2015). ‘The elements of style’. In D. N. Djenar, A. Mahboob and K. Cruickshank (Eds), Language and Identity Across Modes of Communication (pp. 27–60). Berlin: De Gruyter Mouton. Clark, H. H. (1996). Using Language. Cambridge: Cambridge University Press. Danesi, M. (2016). The Semiotics of Emoji: The Rise of Visual Language in the Age of the Internet. London: Bloomsbury. Djenar, D. N., Ewing, M. C. and Manns, H. (2018). Style and Intersubjectivity in Youth Interaction. Berlin: De Gruyter Mouton. Douthwaite, J. (2000). Towards a Linguistic Theory of Foregrounding. Alessandria: Edizioni dell’Orso. Dresner, E. and Herring, S. C. (2010). ‘Functions of the nonverbal in CMC: Emoticons and illocutionary force’. Communication Theory, 20, 249–68. Du Bois, J. W. (2007). ‘The stance triangle’. In R. Englebretson (Ed.), Stancetaking in Discourse: Subjectivity, Evaluation, Interaction (pp. 139–82). Amsterdam: John Benjamins. Du Bois, J. W. and Kärkkäinen, E. (2012). ‘Taking a stance on emotion: Affect, sequence, and intersubjectivity in dialogic interaction’. Text & Talk, 32(4), 433–51.

THE STYLISTICS OF EMOJI

139

Enfield, N. J. and Sidnell, J. (2017). The Concept of Action. Cambridge, UK: Cambridge University Press. Eckert, P. (2012). ‘Three waves of variation study: The emergence of meaning in the study of sociolinguistic variation’. Annual Review of Anthropology, 41, 87–100. Emmott, C. and Alexander, M. (2014). ‘Foregrounding, burying and plot construction’. In P. Stockwell and S. Whiteley (Eds), The Cambridge Handbook of Stylistics (pp. 329–43). Cambridge, UK: Cambridge University Press. Evans, V. (2017). The Emoji Code: The Linguistics Behind Smiley Faces and Scaredy Cats. New York: Picador. Gallagher, S. and Hutto, D. (2008). ‘Primary interaction and narrative practice’. In J. Zlatev, T. P. Racine, C. Sinha and E. Itkonen (Eds), The Shared Mind: Perspectives on Intersubjectivity (pp. 17–38). Amsterdam: John Benjamins. Gibson, W., Huang, P. and Yu, Q. 2018. ‘Emoji and communicative action: The semiotics, sequence and gestural actions of “face covering hand”’. Discourse, Context & Media, https://doi.org/10.1016/j.dcm.2018.05.005 Giles, D., Stommel, W. and Paulus, T. (2017). ‘Introduction: The microanalysis of online data: The next stage’. Journal of Pragmatics, 115, 37–41. Giles, D, Stommel, W., Paulus, T., Lester, J. and Reed, D. (2015). ‘Microanalysis of online data: The methodological development of “digital CA”’. Discourse, Context and Media, 7, 45–51. Heritage, J. and Clayman, S. (2010). Talk in Action: Interactions, Identities and Institutions. Malden, MA: Wiley-Blackwell. Hochschild, A. R. (1979). ‘Emotion work, feeling rules, and social structure’. American Journal of Sociology, 85, 551–75. Hymes, D. (1974). Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia: University of Pennsylvania Press. Jeffries, L. and McIntyre, D. (2010). Stylistics. Cambridge: Cambridge University Press. Johnstone, B. (2009). ‘Stance, style and the linguistic individual’. In A. Jaffe (Ed.), Stance: Sociolinguistic Perspectives (pp. 29–52). New York: Oxford University Press. Kaskus. (2018a). Sejarah KASKUS. Retrieved 28 May 2018. https://bantuan.kaskus.co. id/hc/id/articles/214603738-Sejarah-KASKUS Kaskus. (2018b). Smilies KASKUS. Retrieved 23 May 2018. https://bantuan.kaskus.co. id/hc/id/articles/224997327-Smilies-KASKUS Kelly, R. and Watts, L. (2015). ‘Characterising the inventive appropriation of emoji as relationally meaningful in mediated close personal relationships’. Paper presented at Experiences of Technology Appropriation: Unanticipated Users, Usage, Circumstances, and Design, Oslo, Norway, 20 September 2015. Kiesling, S. (2011). ‘Stance in context: Affect, alignment and investment in the analysis of stancetaking’. Paper presented at the iMean conference, The University of the West of England, 15 April. Leech, G. (2008). Language in Literature: Style and Foregrounding. London: Pearson Education. Leech, G. and Short, M. (2007). Style: A Linguistic Introduction to English Fictional Prose (2nd edn). London: Pearson Longman. McIntyre, Dan. (2008). ‘Integrating multimodal analysis and the stylistics of drama: A multimodal perspective of Ian McKellen’s Richard III’. Language and Literature, 17(4), 309–34.

140

CONTEMPORARY MEDIA STYLISTICS

Mead, G. H. (1934). Mind, Self, and Society from the Standpoint of a Social Behaviorist. Chicago: University of Chicago Press. Nørgaard, N. (2019). Multimodal Stylistics of the Novel: More than Words. New York: Routledge. Ochs, E. (1990). ‘Indexicality and socialization’. In J. Stigler, G. Herdt and R. Schweder (Eds), Cultural Psychology: The Chicago Symposia (pp. 287–308). Cambridge: Cambridge University Press. Riordan, M. A. (2017). ‘Emojis as tools for emotion work: Communicating affect in text messages’. Journal of Language and Social Psychology, 36(5), 549–67. Sacks, H., Schegloff, E. A. and Jefferson. G. (1974). ‘A simplest systematics for the organization of turn-taking for conversation’. Language, 50, 696–735. Sampietro, A. (2019). ‘Emoji and rapport management in Spanish WhatsApp chats’. Journal of Pragmatics, 143, 109–20. Schegloff, E. A. (2007). Sequence Organization in Interaction: A Primer in Conversation Analysis I. New York: Cambridge University Press. Seedhouse, P. (2005). ‘Conversation analysis as research methodology’. In K. Richards and P. Seedhouse (Eds), Applying Conversation Analysis (pp. 251–66). Hampshire: Palgrave Macmillan. Stark, L. and Crawford, K. (2015). ‘The conservatism of emoji: Work, affect, and communication’. Social Media + Society, 1(2), 1–11. Stockwell, Peter. (2002). Cognitive Poetics: An Introduction. London: Routledge. Swaney-Stueve, M., Jepsen, T. and Deubler, G. (2018). ‘The emoji scale: A facial scale for the 21st century’. Food Quality and Preference, 68, 183–90. Verhagen, A. (2008). ‘Intersubjectivity and the Architecture of the Language System’. In J. Zlatev, T. P. Racine, C. Sinha and E. Itkonen (Eds), The Shared Mind: Perspectives on Intersubjectivity (pp. 307–32). Amsterdam: John Benjamins. Zlatev, J. (2008). ‘The co-evolution of intersubjectivity and bodily mimesis’. In J. Zlatev, T. P. Racine, C. Sinha and E. Itkonen (Eds), The Shared Mind: Perspectives on Intersubjectivity (pp. 215–44). Amsterdam: John Benjamins.

CHAPTER SEVEN

Rape victims and the law: Victim blaming and victimization in reports of rape in the British press ALESSIA TRANCHESE

The discourse surrounding rape in the media has been the object of previous studies (e.g. Waterhouse-Watson, 2013); however, these have not been characterized by a systematic linguistic investigation integrating quantitative and qualitative methods. This chapter aims to fill this gap by using critical discourse analysis (Fairclough, 1995) and corpus analysis tools (Baker et al., 2008) to study the representation of rape in the British quality press. In particular, it focuses on a relatively under-investigated comparison between the representation of rape complainants (when referred to as victim(s)) and lawand-order institutions. The analysis is carried out by subjecting collocational patterns of each search term to a transitivity process analysis (Halliday, 1994) – based on categorizing the collocates of social actors in terms of material, mental, relational, behavioural, verbal and existential processes – as well as to a social actor analysis (van Leeuwen, 2008) – to explore the socio-semantic inventory and naming strategies used by the press to refer to different social actors. In order to contextualize the results of the quantitative and qualitative analyses, such

142

CONTEMPORARY MEDIA STYLISTICS

findings are compared with corpus-external data (e.g. statistics on rape in the UK), since rape is conceptualized as a social practice (Fairclough, 1995) that is articulated by both discoursal representations and non-discoursal factors such as social relations and structures. Findings show that the representation of rape in the British quality press reflects dominant understandings of rape rooted in patriarchal conceptualizations of gender roles and sexualities, and perpetuated by established professional practices in mainstream journalism, such as the reliance on authoritative ‘voices’ as news sources. This leads to subtle forms of victimization and victim blaming, including the deprivation of women’s voices of the authority and truth value that are instead accorded to established institutions.

1 INTRODUCTION Linguistic studies of representation of rape have traditionally emphasized discrepancies between agency attributed to men and women, often concluding that rape tends to be portrayed as an agentless crime in which complainants1 are the ‘done-to’, while perpetrators are mostly obscured or backgrounded (Benedict, 1992; Clark, 1992; Ehrlich, 1999; 2001; Meyers, 1997; WaterhouseWatson, 2013). In contrast, a comparison between the representation of complainants and established law-and-order institutions, with a focus on both agency and naming strategies, is a comparatively under-investigated subject. This chapter aims to address this gap and, in doing so, it also examines hierarchical structures of truth making, and the links between these and the reinforcement of stereotypical beliefs about gender roles in reports of rape in the British quality press. Through the analysis of syntactical, lexical and contextual elements, this study argues that representations of rape in the corpus under analysis deny women’s reality while, at the same time, legitimizing institutional ‘truths’ (e.g. law’s understanding of rape) instead of challenging them to promote a broader comprehension of rape, beyond its archetypal, phallocentric, interpretation (Smart, 1989). In this sense, this study expands the conceptualization of victimization and victim blaming in rape cases to include absence of semiotic mobility (Blommaert, 2005), that is, the inability to make oneself understood because of established gatekeeping practices in society that allow only a restricted group of people to create (and disseminate) meaning.

2 BACKGROUND Rape, like other forms of violence against women, is included by Walby (1990) among the six structures of patriarchy, that is, the system of social structures and practices in which men dominate, oppress and exploit women (1990: 20). This understanding of rape as part of a system of behaviours that are not

RAPE VICTIMS AND THE LAW

143

biologically determined but socially established negates the notion that the subordination of women and the dominance of men in society are natural and inevitable. That is to say that women are not inevitably helpless victims of rape and, by extension, of patriarchy. Admittedly, each woman who experiences rape makes sense and reacts to it in a deeply personal way – one that is subject to change, even within the same individual. While societal norms may suggest that the ‘ideal’ – or normative – reaction to rape is one of helplessness and despair, this should not be assumed, in order to not deny legitimacy to those reactions that deviate from the norm. For instance, women who have been raped by men are likely to describe their experiences in terms that could appear paradoxical, contradictory and possibly illogical because they challenge the archetypal understanding of male and female sexuality, in which men are natural predators, incapable of resisting sexual desires, and women are their inevitable preys (Smart, 1989: 43). On the surface, these women’s accounts may lack coherence or even consistency; yet, they should not be discarded as untrue or non-genuine simply because of their appearance of irrationality and deviance. On the contrary, they should be heard with the awareness that there simply does not exist a single, ‘correct’, standpoint to use as a compass for and from which to judge all other accounts and descriptions of this experience which exists on a continuum and is formally neat (i.e. it has one name, rape), but semantically confused (i.e. it has many meanings) for the women who experience it (van Leeuwen, 2008: 24), and may not call it rape. Even the legal definition of rape as forced penetration2 – which is held as ‘Truth’ and ‘sets and resets the parameters within which rape is dealt with … in society’ (Smart, 1989: 26) – is too narrow to incorporate such diverse meanings. In fact, the definition adopted by law-and-order institutions disqualifies women’s experience of rape, while celebrating a series of male-centric perspectives and myths which assume – with various degrees of mitigation – ‘natural male sexual need and female sexual capriciousness’ (Smart, 1989: 35). The pervasiveness of rape myths – that is, ‘prejudicial, stereotyped and false beliefs about rape, rape victims and rapists’ which create ‘a climate hostile to rape victims’ (Burt, 1980: 217) – is not always, or necessarily, expressed through overt mistrust and victim blaming. In fact, it often manifests itself at the level of discourse as victim blaming and linguistic victimization. Such forms are generally subtler or covert, in that they preserve an appearance of rationality which often makes it difficult to recognize, question, or refute the actual assumptions that shape and uphold them. In this scenario, the language used by the media to represent rape becomes particularly relevant. While it must be acknowledged that there is a difference between social practices and their representation(s) in the media, it cannot be denied that such representation(s) evaluate practices, shape perceptions and understandings within their social and ideological context and can contribute to

144

CONTEMPORARY MEDIA STYLISTICS

their legitimation (van Leeuwen, 2008: 6). In particular, media representations possess the ability to reinforce the perception of ‘naturalness’ of certain codes of conduct, and hierarchical social relations – including the assumption that there exist pre-defined sexual roles for men and women on which many rape myths are based – which originate from and at the same time uphold hegemonic power (Fairclough, 1995). Similarly, through their discourses and representations, the media can foster the sense of need for and acceptance of other hierarchical social relations, like a set of institutional sources (e.g. law-and-order institutions) seen as authoritative, legitimate and reliable truth holders (Richardson, 2007).

3 VICTIM BLAMING AND VICTIMIZATION IN AND THROUGH LANGUAGE Media representations of rape complainants have often appeared to differentiate women as ‘illegitimate victims’ and ‘legitimate victims’ (Benedict, 1992; Meyers, 1997). While the former refers to women who are blamed for ‘provoking’ the assault, the latter indicates women who are seen as innocent and victimized, often because they were killed by the attacker. The word victim applies to both groups and, as such, it encapsulated one of the contradictions of rape discourse: while, on the one hand, rape complainants are seen as recipients of violence, they can also be undermined and seen as helpless and void of agency (Kelly, 1988) and, therefore, to blame for not ‘doing the right thing’ or for not ‘doing enough’ to stop or prevent the rape. The sense of passivity associated with the term victim to refer to rape complainants is rooted in the more general assumption that this is a core trait of female personality (Scully, 1994: 42). Admittedly, women’s supposed passivity has also been found by several linguistic studies which have demonstrated how this is often produced and reproduced in and through language, especially in the representation of who acts and who is acted upon in literature and fiction (Burton, 1982; Wareing, 1990; Ehrlich, 1999; Cameron and Kulick, 2003). A number of studies (Simpson, 1993; Mills, 2008; Talbot, 1998; CaldasCoulthard, 1995) have also highlighted asymmetries concerning attribution of agency to men and women in the naming strategies used by the media to represent them. However, Mills (2012) challenges the existence of a direct connection between the use of linguistic structures that deny women’s agency, and the implication that these can only indicate women’s passivity. In contrast, she argues that language must be analysed in context, in order to account for its complexities and multiple ambiguities. This non-deterministic reading of linguistic structures also implies that passivity and agency cannot be assumed to be always associated with blamelessness or culpability, respectively. It follows that the word victim

RAPE VICTIMS AND THE LAW

145

should not be seen as inevitably suggesting empathy, since its use does not always translate into ascribing lack of culpability or responsibility; similarly, it should not be presumed that the active representation of women is always positive simply because this is normally ascribed to participants in dominant positions in social structures (Fairclough, 1995). Instead, together with the wider context of textual production, the syntactical and lexical co-text of the elements under analysis must be considered to understand the sociopolitical significance of any pattern (Jeffries and Walker, 2018). Given the inability of the traditional active/passive dichotomy to account for the many facets of the discourse of both victim blaming and victimization, this chapter sees them as two connected phenomena and conceptualizes these as (consequence of) disempowered voices. Crucially, they are understood not only in terms of material agency ascribed to complainants – for example, actions that, one way or another, ‘provoked’ the attack (Scully, 1994: 42) – or abuses committed against them, but also in terms of verbal agency (or lack thereof). The downgrading of women’s words from truth to opinion has been highlighted by Caldas-Coulthard (1995) who has argued that news reports tend to represent the voices of female people in indirect speech, often introducing their words with verbs like claim or argue, which cast doubts on their reliability. Mills (1995) highlights how women are also often talked about in the media rather than doing the talk. These institutionalized practices of covert sexism tend to be particularly common when women are represented in the public sphere, a domain which appears to be predominantly populated by male voices; in the private sphere (e.g. as mothers or wives), on the other hand, women have been found to often ‘regain’ their voices. The speech of women in the media has also been found to be generally obscured when compared with institutional voices. According to WaterhouseWatson (2013), rape complainants’ stories, in particular, are often marginalized and given much less space when compared with comments by public officials or established law-and-order institutions, like the police and judicial courts. These, as main sources of news (van Dijk, 1988; Fowler, 1991; Richardson, 2007; Bednarek and Caple, 2012), are in a favoured position to have their voices heard, while women, whose statements are often marginalized through value judgements, are silenced, deprived of authority and of the right to define their own experiences, and blamed. Admittedly, victim blaming also entails the possibility of speech. For example, failure to take ‘preventive’ measures and resist ‘strongly enough’ or ‘clearly enough’ to sexual advances can be seen as ‘lack of speech’. Moreover, if a subject is erased, or silenced, or if their competence to have a say in the process is questioned, they are denied the possibility of rebuttal. As suggested by Waterhouse-Watson (2013: 74), when ‘women are left without a position from which to speak, [they] cannot dispute the actions and

146

CONTEMPORARY MEDIA STYLISTICS

motivations which have been attributed to them’ and, therefore, they can be blamed because they are denied the possibility of defending themselves. Victim blaming and victimization can also be realized by lexical or ‘labelling’ choices, including lexical specificity that highlights only one side of an individual’s identity (Cruse, 1977). Especially when systematically repeated within a certain discourse, lexical choices can provide insight into the bias and ideology that surround certain actors. According to Talbot (1998: 217), naming is one way of classifying people and such categorization is a powerful normative force as it ‘reflects and sustains existing social relations and identities’. Therefore, these are an important side of the discursive practices which the press draws on to represent rape crimes and can contribute to further victimization and/or victim blaming, since labels can influence the ways we come to see not only others but also ourselves and our actions Fairclough (2003: 22). For example, whether a woman who experiences rape is referred to as a victim, a survivor or a complainant will (or may) have an effect on how she is (or thinks she will be) perceived in society, and on how people (including complainants) react to rape. This study focuses on deconstructing subtler forms of victimization and victim blaming in the press using transitivity analysis (Halliday, 2004) and social actor analysis (van Leeuwen, 2008). As key linguistic tools, naming strategies and transitivity patterns applied to specific social actors can be used to produce and reproduce said subtler forms of victim blaming and victimization of rape complainants, as well as a sense of legitimation of the authorities that mediate their stories (media, law-and-order institutions, etc.). The following sections will investigate how the system of authority and institutions that regulates the mainstream press plays a role in the propagation of biased discourses about rape, using language as a vehicle to establish and cultivate the belief in its legitimacy (van Leeuwen, 2008), often to the detriment of the legitimacy of non-established voices. While, for ease of explanation, these linguistic categories are discussed separately in the following, it should be noted that there are various degrees of overlap between them and they should not be seen as compartmentalized and disconnected from each other.

4 METHODOLOGY Although a growing body of research (Mills, 1995; Talbot, 1995; Lischinsky, 2017; Tranchese, 2019) has examined (gendered) patterns of agency and labelling/naming in various media forms, few studies have taken advantage of the systematicity offered by the use of corpora to carry out an integrated transitivity and social actor analysis. Here, corpus analysis tools are used to explore the way(s) in which social actors in press reports of rape are referred

RAPE VICTIMS AND THE LAW

147

to and the way(s) in which their participation in a given action is encoded grammatically. 4.1 Transitivity analysis The concept of transitivity is associated with Halliday’s work in systemicfunctional linguistics (2004). It is mostly concerned with how individuals express ideologies, account for their world views, and attribute agency and responsibility through a system of grammatical choices, particularly with regard to who is represented as participant in which process type. Halliday distinguished between three components of the clause that contains a process type: the process itself, the participants involved in the process, and the circumstances of the process (Simpson, 1993). The process type that appears in a clause and the roles taken by different participants (i.e. who does what to whom) are two crucial elements in the exploration of agency and are, therefore, the focus of this study. The process types (including their sub-categories and possible participant roles) that are relevant to this analysis are material, verbal, mental and relational processes. For most process type, participants can be identified in terms of whether they are performing the action or they are having it done to them (Simpson, 1993; Thompson, 1996; van Leeuwen, 2008; for a categorization of verbs into process types see Neale, 2019). Material processes (or actions which can be observed in the real world – for example, he was murdered by a stranger) normally have two participant roles: an Actor (or doer) and a Goal (or done to). However, some material processes (i.e. non-transactive processes – e.g. she tripped over the step) can also involve only one participant (the Actor). Verbal processes are processes of saying (e.g. she told them the truth) and normally involve a Sayer (the individual who speaks), a Receiver (the participant addressed in the process) and, in some cases, Verbiage (the thing that is said). Mental processes refer to states of mind; unlike verbal and material processes, these are ‘internal’ processes which are not visible or hearable externally (e.g. I did not understand the text). They can be subdivided into perception (e.g. seeing), emotion/reaction (e.g. feeling), desideration (e.g. wanting) and cognition processes (e.g. understanding). Participants normally involved in these processes are Senser (the individual who perceives, reacts, wants or thinks) and Phenomenon (the thing that is perceived, reacted to, wanted or thought about). Finally, relational processes refer to states of being. In this study, two relational processes will be considered: attributive (intensive) relational processes, where an entity has a quality ascribed to it (e.g. the woman was tall), and identifying relational processes, where one entity is identified in terms of another (e.g. he was the victim of the attack). Participant roles in attributive (intensive) relational processes are Carrier (or what the clause is

148

CONTEMPORARY MEDIA STYLISTICS

about) and Attribute (a comment about the Carrier), while Value (or the more general category) and Token (the more specific category) are participant roles in identifying relational processes. 4.2 Social actor analysis Studying the socio-semantic categories attributed to certain social actors – individuals or groups of individuals – allows us to examine which aspects of their identity are legitimized, suppressed, backgrounded and/or foregrounded. Thus, it is a useful tool to explore the ideological underpinning of such choices and to identify power and status in social practices, as well as the dynamics between participants (van Leeuwen, 2008). Social actor analysis was carried out using van Leeuwen’s (2008) system network to categorize references to social actors within discourse. The naming types used in van Leeuwen’s network and relevant for this study are assimilation (i.e. when social actors are referred to as groups rather than individuals) and, particularly, aggregation (i.e. when individuals are not only represented as groups but also quantified and/or treated as statistics). Identification – referring to social actors by what they unavoidably are – is also considered in this analysis and, in particular, its sub-categories of classification (i.e. references to societal labels such as age, religion, gender or class) and physical identification (i.e. references to social actors by means of mention of their physical attributes). Finally, nomination – references to social actors by their name – and, particularly, aspects of titulation, that is, the use of titles (or lack thereof – detitulation) to refer to certain social actors, are also used in this study. 4.3 Corpus building and analysis The corpus used for this analysis was collected through the online newspaper database LexisNexis. The following daily (paper) versions of four national British quality papers were considered: The Guardian, The Independent, The Times and The Daily Telegraph. The search terms used to query the newspapers were rape* OR rapist* OR raping. The time span taken into consideration for this study is from 1 January 2008 to 31 December 2008. The corpus was built by selecting only articles concerning incidents of rape perpetrated by men against women. The corpus includes feature articles, comments and editorials for a total of 484 articles and 250,841 words. In order to investigate and compare the aspects of the representation of victim(s) and established law-and-order institutions that were highlighted in Section 3, police and court, in addition to victim(s), were selected from the corpus for analysis. These were chosen to represent the institutions that normally mediate complainants’ stories to the press and are involved in the judicial procedures that deal with rape, from the investigation to the trial stage.

RAPE VICTIMS AND THE LAW

149

These words were subject to collocation and concordance analysis, in order to – as suggested by Baker and McEnery (2015: 250) – understand the ways in which they contribute towards certain discourses. Collocation analysis was carried out using Word Sketch, a tool that provides a ‘summary of the word’s grammatical and collocational behaviour’ (sketchengine.eu) and can point towards certain strong collocation tendencies. In order to focus on repetitions of transitivity patterns and naming strategies, victim(s), police and court were studied through the extraction and in-context examination of their most frequent3 pre-modifiers and verbal4 collocates. Concordance analysis was used to infer contextual elements that would not be noticeable through collocation analysis alone (Baker et al., 2008), as this only provides a de-contextualized snapshot of collocation patterns. For example, concordance analysis allowed to distinguish between occurrences of be and have as auxiliaries and as full verbs. Subsequently, each verb phrase was categorized into a process type and presented visually in a chart. This quantification of process types allowed to observe systematic repetitions of certain grammatical structures which were assumed to represent traces of opinions and ideas that shaped certain discourses and could, therefore, be revealing in terms of the bias and ideology that surrounds certain social actors (Mills, 2004). For ease of presentation, only processes with more than twenty-five (normalized per 100) occurrences were presented visually; similarly, sub-categories of process types were not included in the charts but will be discussed in the analysis.

5 ANALYSIS The focus of this section is to present the key findings of the transitivity and social actor analysis of the three search terms (victim(s), police and court). In order to examine constant features of the discourse surrounding these words in the corpus, the analyses of their respective WordSketches is presented in the following. While these are mostly shown separately, they nonetheless intersect, since both naming strategies and process types contributes to the creation of certain discourses. 5.1 Victim(s) As shown in Figure 7.1, the type of role in which victim(s) appeared the most was as Goal in material processes. Victim(s) were often also represented as being at the receiving end of some adverse circumstances (as a Goal or Phenomenon in material and mental processes respectively). The representation of victim(s) as subject to negative circumstances is in line with the prototypical meaning of the term, but closer observation showed that

150

CONTEMPORARY MEDIA STYLISTICS

victim/s 300 250 200 150 100 50 0 sayer in verbal process

possessor in actor in possessive material relational process process

goal phenomenon carrier in attribute in value in material in mental attributive attributive identifying relational relational relational process process process process process

FIGURE 7.1: Victim(s) – process types and participant roles.

the verbs often indicated negative circumstances in a bureaucratic context (e.g. victims are routinely disbelieved or rape victims are being let down by the law), rather than physical aggression that would highlight their vulnerability and passivity, like, for example, murder, knock down, strangle or attack. While these did appear, they were less dominant than verbs that expressed victimization without physical assault. In some cases, the word victims(s) was used on its own, as a stand-alone identity, without any reference to either the crime or the person(s) that committed it (e.g. the victim was a schoolgirl). Victim could be considered an ontological metaphor since it does not refer to an ‘entity bounded by a surface’ (Lakoff and Johnson, 1980: 25) but is created through naming, and its characteristics are not permanent and natural but are collectively created by those who create and sanction its meaning. Thus, victimhood can become a fixed status, an indelible, overriding, identity, independent of time and/or space – more similar to sex, ethnicity or other categories that one is born with – rather than the temporary result of another person’s action. As such, it cannot be escaped and is a permanent ‘stain’ or ‘damage’ caused by an action imposed by one person upon another. Thus, naming a complainant victim can be seen as a form of disempowerment, since victim(s) is a term in which the done-to status is embedded. A victim’s powerlessness is an accepted fact, a reality that singles her out as an outsider whose status is irreversibly and permanently changed and can never be reinstated. Survivor(s) has often been proposed by activists and feminists as a more appropriate term to refer to complainants, in that it stresses their reaction rather than their victimhood (Kelly, 1988). In order to see whether there existed any differences and/or similarities in the ways in which women were presented when referred to as victim(s) or survivor(s), a search was carried out in the corpus for the term survivor(s).

RAPE VICTIMS AND THE LAW

151

However, this was found to be mostly used in reference to people who survived war or genocide, not to rape survivors. The predominant use of victim in the press may be a trace of the language used by the judicial system to officially refer to rape complainants, rather than the consequence of a direct choice of news reporters. However, its prevalence shows how the views of powerful law-and-order institutions can be reinforced through replication in the press because of the intersection between the two and of the strong influence that the former has in shaping the narratives of the latter. 5.1.1 Agency Perhaps surprisingly, given the topic of the articles, victim(s) did not regularly co-occur with rape in expressions such as the victim was raped or victims had been raped; on the contrary, there seemed to be a tendency to obscure agency and background the crime through nominalization (Fairclough, 1992: 95). Admittedly, rape was the most frequent pre-modifier of victim(s), and other variants, such as abuse or sex attack, were also among the most common modifiers. Moreover, victim(s) of rape was the strongest prepositional phrase containing victim(s). The use of the expressions rape victim(s) and/or victim(s) of rape, combined with the use of the term on its own, as shown in Section 5.1, has consequences in terms of obscuration of the (responsibility of the) perpetrator, in addition to presenting victimhood as the main identity of complainants. Admittedly, it could be argued that, on a continuum of visibility – which would virtually go from he raped to victim(s), with expressions such as he raped her, she was raped and rape victim(s) in between these two extremes – victim(s) is the form that expresses the least amount of visibility of the perpetrator, by totally shifting the attention onto the complainant and completely obliterating the perpetrator and the crime he committed. Perpetrator and crime become facts that simply happened and exist a priori; as such, they do not require further discussion or questioning and, consequently, they become invisible. The only element that remains visible is the woman who was affected by it, while the abuser disappears. As one possible choice among countless options, the selection of a specific way to represent action and participants, especially when done systematically and repetitively as in this case, can signal a specific world view or idea about certain individuals (Mills, 1995). For instance, while the examples on the visibility continuum refer to the same participants, each version conveys a different message in terms of agency (from the most to the least transparent) and, arguably, responsibility. In turn, these linguistic structures can be interpreted as traces of the ideology that discounts and justifies men who commit rape (Benedict, 1992). Further evidence of the erasure of the perpetrator was found when victim(s) was followed by of; only in one of those instances, it was followed by a noun referring to a human being (violent men). In all other cases of

152

CONTEMPORARY MEDIA STYLISTICS

the prepositional phrase victim(s) of rape, the action – or crime – itself, not human agency, appeared to turn women into victims. This was also evident in another expression highlighted by the concordance analysis – fall victim to sexual attacks – where the emphasis on the fact that women were victimized by circumstances rather than by a human being was even stronger. Indeed, as shown by the analysis of the expression to fall victim to in the British National Corpus (BNC – bncweb.lancs.ac.uk), this tends to co-occur with events or nonhuman entities, such as flu-epidemic, violence, hunger, virus and so on. Arguably, it is the sense of vulnerability and victimhood embedded in this term that makes it possible to avoid any specification of what exactly happened to her or who turned her into a victim, because victim is ‘what she is’. In other words, the term itself functions as a ‘package-word’ including both Actor, process and Goal combined into one identity (Section 5.1). This way, it is the victim status that becomes central and, it could be argued that, as long as victimhood remains an unquestioned part of her identity, the crime can be nominalized and ascribed to her as an ornament – or attribute – of that very identity (rape victim), or as the maker of that identity (victim of rape), rather than as an act of violence perpetrated by a human agent against her. The sense that the use of victim(s) allowed to discuss rape complainants in isolation was reinforced by the recurrent use of this term in relational processes. Having only one animate participant (about whom something is being said), these do not require in fact the explicit mention that a participant affected the other. For instance, victim(s) appeared as Carrier in attributive (intensive) relational processes, in which attributes were ascribed to her. These focused mostly on her powerlessness (e.g. the victim was frightened/trembling with fear/vulnerable) or on the debate on women’s responsibility in rape cases (e.g. the victim was responsible/drunk/intoxicated). In one case, what was said about them was an expression of their strength (the victim was determined not to let the attack destroy her), although here the negation may also be seen as presupposing that the attack was indeed capable of destroying her. In other cases, victim(s) appeared as Attribute (the quality ascribed to an entity) in attributive (intensive) relational processes (e.g. thousands of women have been victims of sexual attacks) or Value (the entity in terms of which another entity was defined) in identifying relational processes (e.g. disabled people were often the victims of sexual assault). As suggested by Thompson (1996: 98), who is represented as Value/Token in identifying relational processes can reveal the writer’s beliefs about the Token by looking at the Value (in this case, victims(s)) used to categorize them. In the example above, it could be argued that the characteristics that are usually associated with victim were carried over to disabled people, without the necessity to question the validity of such characteristics, nor the circumstances under which the process of victimization took place.

RAPE VICTIMS AND THE LAW

153

5.1.2 Disbelieving victim(s) With their temporary vulnerable status transformed into fixed identity, victim(s) can also become the Actor and/or Goal of actions that are not directly related to the crime. Their status as affected parties can also be challenged, thus showing that victim(s) is not necessarily linked to demonstrations of empathy. For example, the second most frequent pre-modifier of victim was alleged, occurring mostly in the expression the alleged victim. Baker (2006) showed how the lemma ALLEGE is associated with a pattern of denial, while Tranchese (2019) illustrated how it is used when journalists try to distance themselves from their reporting in rape cases. The semantic prosody (Louw, 1993: 157) of disbelief was also visible in the BNC, where victim(s) was the first lexical collocate that referred to human beings in the list of collocates of the adjectival form alleged. It was also the only one that referred to people at the receiving end of criminal actions, since the other terms in the list referred to criminals (criminals, killer, accomplice, perpetrators, murderer, terrorists, etc.) or to a series of illegal activities like fraud, offences, abuse, assault, killing, and torture. Arguably, while in the latter cases the emphasis is on the presumption of innocence or questioning of the crime, in the case of alleged co-occurring with victim(s), it is their innocence, and, therefore, their blamelessness, that is under examination. This use of alleged seems to reverse the empathy normally attributed to vulnerable people, since it casts doubt on their version of the story, shifting the burden of proof onto them. If one assumed that the absence of alleged would automatically imply the assumption that a crime did happen, then the choice to use alleged would signal that the press casts doubt not only on the innocence of the perpetrator but also on the fact that a crime did happen. This reflects broader discourses surrounding rape, in particular the way in which the law treats rape cases. As suggested by Smart (1989: 34), every time the law finds the accused innocent, it is also, indirectly, placing the blame to some degree onto the woman, if not for having lied, at least for her sexual complicity. The victim(s) culpability was also implied when her agency was foregrounded and she was presented as Actor in material processes (the second most frequent role attributed to victim(s), after Goal of material processes – Figure 7.1) that emphasized the complainant’s behaviours before the attack (e.g. drinking, going to nightclubs or taking drugs). These processes were references to circumstantial information that reflect rape myths which see complainants as actively provoking rape, thus blaming women for the crimes that men commit against them (Burt, 1980). While some of these examples appeared in a context of resistance to this stereotypical understanding of rape (e.g. to challenge the idea that the way a woman is dressed matters in rape cases), they demonstrate how the responsibility of women is still very much an open debate. These contrasting discourses are evidence of the tension that still exists between blaming complainants and

154

CONTEMPORARY MEDIA STYLISTICS

representing them empathetically, a tension that is exemplified by the expression alleged victim. 5.1.3 Victim(s) as Sayers The third most frequent role in which victim(s) appeared (see Figure 7.1) was as Sayer in verbal processes. In most cases, it appeared that women were being used as sources of news material to tell readers about their own ordeal and victimization, by giving detailed accounts of the violence they had been subjected to (e.g. the alleged victim told detectives she had suffered a ‘harrowing’ ordeal; some victims have described being subjected to extreme violence; or one of his victims described how he would hold her head near to a gas flame). However, rather than being a way of letting women recount their own stories, through their own narratives, and express their feelings and emotions about the rape, these accounts could be seen as a way to draw attention to their violated (and sexualized) bodies. Such accounts are also a way to re-enact the violence, both for complainants and for those who listen to or read them. As suggested by Smart (1989, pp. 38– 43), asking women to recount details of the attack is a common process during rape trials. Given that the words of rape complainants reported by newspapers are, in most cases, those heard during rape trials and that these are expression of the ‘maleness’ of the law (Smart, 1989: 26), it could be argued that this structure of truth making reproduces and reinforces male-centred understandings of rape. It appears that women’s narratives could not develop freely, in a womancentred way, but they had to be filtered through and channelled into the (heavily sexualized and phallocentric) definition of rape provided by the law (and propagated by the media) and could only come to light through the questions (and answers) that the law allows (and that the press chose to replicate). 5.1.4 Youth In addition to being labelled in terms of the violence they underwent (sex abuse victim(s), rape victim(s), etc. – Section 5.1.1) victim(s) were also identified, or, in van Leeuwen’s (2008: 42) words, referred to in terms of what ‘they, more or less permanently, or unavoidably, are’. More precisely, victim(s) was identified through two modifiers (child and young: for example, child victims of rape) that referred to their (young) age, with an emphasis on very young age, as shown by the concordance analysis (e.g. The youngest victim was Elisabeth Brichet, 12). The emphasis on youth can be seen as a way to accentuate the sense of vulnerability evoked by victim(s) thus, possibly, accentuating the gravity of the crime. However, stressing the young age of these women can also reinforce another rape trope, that is, that only young women are raped or that the ‘ideal (or real) rape victim’ is a woman who is ‘innocent and unspoiled’, violated by a violent, stranger-rapist’ (Hirsch, 1994: 1029). In reality, women of all ages are raped (Office for National Statistics, 2017).

RAPE VICTIMS AND THE LAW

155

As with the violent details provided by rape complainants or the association between violence and victim(s) (Sections 5.1.1 and 5.1.3), it could be argued that youth also highlights what could be considered a brutal aspect of rape. In fact, youth and violence, in addition to sex (Talbot, 1998: 175), can be seen as details that the press offers to readers for their consumption in order to sell newspapers. Admittedly, the stress on youth suggests that the rape of young women (or children) – if perceived as innocent and blameless – is endowed with high news value, since these cases can be presented as being more ‘serious’ or ‘negative’ and, therefore, more newsworthy (Bednarek and Caple, 2012). While the emphasis on youth does not necessarily translate in the complete absence of reporting on older women, rape crimes that are considered journalistically unimportant do tend to receive the least amount of coverage (Benedict, 1992), thus skewing the representation towards those rapes that are more sensational but not necessarily more frequent (and certainly not ‘more serious’) than those that have less journalistic value. 5.1.5 Quantifying victim(s) Victim(s) was also preceded by quantifiers such as many, more or other (e.g. other victims). In van Leeuwen’s terms (2008: 37), this is a form of quantification and, more precisely, aggregation, that is, references to social actors as groups rather than individuals. This portrayal of rape complainants stressed the epidemic nature of violence against women and, arguably, encouraged the perception that sexualized violence is a constant threat for women. It could also be argued that the quantification had the effect of presenting these women as a homogenous group of people whose experiences were unified into the same identity defined by victimhood. In this sense, the press played a role similar to that of the rape trial, which, according to Smart (1989: 42), ‘constructs a category of Woman as if it was a unity’. In the press, this undiscussed and unquestioned identity was carried over from one entity to the whole unit, which was then discussed, framed and presented by other voices. Indeed, as shown in the following sections, women, like minorities or less powerful groups that are aggregated – for example, migrants (van Leeuwen, 2008: 38) – tended to be talked about rather than doing the talking. 5.2 Police and Court Both police and court refer to institutions, not to specific individuals, despite the fact that, in reality, they are an aggregation of individuals. Normally, it is an (unidentifiable) individual or small group of individuals in the organization that is being referred to by the press when journalists quote them and it is these individuals’ statements that are being reported to represent the wider organization.

156

CONTEMPORARY MEDIA STYLISTICS

Police and court are, therefore, labels that function like aggregators and their presence is meaningful given that they have the effect of distancing a specific individual from an action or opinion by presenting it as the words of a larger – and, therefore, arguably perceived as less biased and more neutral – social actor. Furthermore, presenting the subjective thoughts of a single individual as the voice of an authoritative institution may help legitimize the content of their statements or actions, thus making the subjective seem objective. Like in the case of victim(s) (Section 5.1), the use of these terms is an established and common professional practice in journalism and refers to individuals as a group. Unlike victim(s), though, police and court are not the group that is being talked about and defined externally, but it is the group whose values are being presented as an authoritative unified voice. 5.2.1 Official denominations There was a predominance of nomination, particularly official denominations, among the most frequent lexical choices used to modify police and court (Crown Court, High Court, Supreme Court and Metropolitan Police, Manchester Police). While, in van Leeuwen’s (2008) framework, nomination generally applies to specific individuals, it could be argued that using the official denomination of a given institution serves a similar function in that it ascribes a formal and official ‘identity’ to said institution. The use of formal names can transmit a sense of factuality and authority which highlights the high status of these organizations and, consequently, supports their legitimation, and the sense that they are objective seekers of justice, without agenda or bias. 5.2.2 Police agency As summarized in Figure 7.2, the most frequent process types with police were material, verbal and mental processes (in all of which it appeared in the active role or ‘doer’ – as Actor, Sayer or Senser). In quantitative terms, the appearance of police predominantly as Actor in material processes was in contrast with the portrayal of victim(s), often represented in the passive role of Goal – or ‘done to’ (Figure 7.1). However, there also appeared to be a qualitative difference when police and victim(s) co-occurred as Actors in material processes (the second most frequent role attributed to victim(s) – Figure 7.1). Victim(s) appeared as the Actor of material processes primarily in non-transactive processes (i.e. actions that involved only one participant) or in transactive, instrumental processes (i.e. actions that involved two participants, but in which the affected participant was a thing, rather than a person) like, for example, come forward, dress, drink, take (drugs) (van Leeuwen, 2008: 60–2). These verbs often indicated what women did and how they behaved before the attack (cf. Section 5.1.2). By contrast, police was often the Actor in material, transactive processes (i.e. actions that involved two participants and in which the done-to was another

RAPE VICTIMS AND THE LAW

157

1400 court

1200

police

1000 800 600 400 200 0 sayer in verbal process

senser in mental process

possessor in possessive relational process

actor in material process

goal material process

phenomenon in mental process

receiver in verbal process

carrier in attributive relational process

FIGURE 7.2: Police and court – process types and participant roles.

person or a group of people) (van Leeuwen, 2008: 60–2), like, for instance, arrest, question, interview or release. Such verbs often referred to actions that presented the police as actively participating and investigating rape cases and exercising their power. The representation of such actions stressed their role as agents who routinely tackle crime, thus arguably further reinforcing their authoritative role. 5.2.3 Authority and legitimation: Police and court in mental and verbal processes suggests that the association between institutions and verbal and mental processes is a typical feature of various legitimation processes. Here, police and court appeared frequently in these process types (especially police as Sayer in verbal processes and court as Senser in mental processes – Figure 7.2). The Verbiage in verbal processes attributed to police, in particular, was mainly introduced by the two speech-act or glossing verbs, tell and say (e.g. Police initially said she had accidentally drowned), which Caldas-Coulthard (1994: 305) defines as ‘neutral and structuring “glossing” verbs’, or verbs which simply signal the illocutionary act of saying and introduce ‘saying’ without explicitly evaluating. Bell (1991) and Richardson (2007) also showed that these verbs are the canonical neutral speech verbs used in news reporting; they convey a sense of impartiality and authenticity, since what has been said by someone else is presented as being reported verbatim and without filters. Moreover, other studies (Glasgow University Media Group, 1980; Geis, 1987; van Dijk, 1991) found that there is a correlation between the status of the speaker and the speech-act verb used by the press. For example, there appear to be systematic differences between the use of say with ‘credible’ sources (such as politicians or the police) and claim with sources that are considered less authoritative or ‘biased’ (such as ordinary people or minority groups).

158

CONTEMPORARY MEDIA STYLISTICS

It follows that the little glossing which characterized the embedding of police statements in the articles in the corpus, achieved through the selection of neutral reporting verbs such as say and tell, served to validate the authority of the police as news sources by stressing their impartiality and the factuality and reliability of their words. At the same time, it could be argued that this was a way of providing journalistic legitimacy and making newspaper reports more authoritative. It should be noted that victim(s) too, co-occurred as Sayer in verbal processes introduced by the same verbs (Section 5.1.3); however, in addition to being a quantitatively less dominant pattern when compared to other process types and participant roles, the role of Sayer in verbal processes for victim(s) appeared predominantly in relation to personal, subjective experiences, rather than in relation to broader discussions of rape. On the other hand, the status of authority given to the police allowed them to tell other people’s (including complainants’) stories and have their opinion heard as universal truths that became part of the dominant rape discourse in the corpus. Further legitimation was also granted by the predominant use of police as the Receiver in verbal processes, such as tell, or the Goal in material processes, such as alert, contact, phone or call, which emphasized their authority as the institution that is entrusted with other people’s narratives and is allowed by the press – which will then amplify its voice and messages – to process and recite these through their own lens. Finally, police received additional validation as Senser in cognitive mental processes (e.g. review, plan, discuss or hear) which – unlike emotional mental processes that emphasize the subjective nature of certain practices – suggest objectivity, neutrality and factuality. Like police, court was often associated with verbal and mental processes (mostly as Senser of hear – for example, Portsmouth Crown Court heard – or as either Sayer and Receiver in verbal processes – for example, the court was also told how the couple were caught) (Figure 7.2). Court was also legitimized through the association with processes that focused on cognition rather than emotion or desideration (van Leeuwen, 2008: 106) and its authority was validated with the emphasis on actions that represented it as routinely engaged in tackling crime (as suggested, for example, by the process of ruling). The truth-holder status attributed to this law-and-order institution in society allowed the press to represent it as having the unquestioned authority to be entrusted with or to receive information and, subsequently, make an impartial and fair evaluation of such information followed by decisions with binding effect. As in the case of police, it may be argued that the use of courts of law as a source – presented as unemotional, detached and impartial – by another authoritative, established organization (i.e. the press), reinforced the sense of its reliability and neutrality and, ultimately, authority.

RAPE VICTIMS AND THE LAW

159

5.2.4 Reported speech The use of reported speech in conjunction with police and court – in many cases introduced by implicit or explicit that-embedded clauses (e.g. Police said Fritzl had got rid of the evidence or the court was told that he had difficulty understanding) further reinforced the sense of neutrality of the press, already established through the use of authoritative sources, presented as factual and reliable entities. With reported speech, the action of telling or hearing can be seen as being even more neutral and factual and, therefore, harder to refute than the content of what is being heard or said. Unlike the content, the actions of telling or hearing do not need verifying or questioning, thus allowing reporters to put enough ‘distance’ between themselves and what is being reported, while communicating a sense of detachment and impartiality. Furthermore, by representing actions as stories which are told or heard (rather than actions that are done), journalists can avoid taking responsibility for the truthfulness of what they report. As Tuchman (1972) pointed out, news makers constantly operate under the threat of criticism of bias and of libel suits; as long as they can prove that what they are writing has been said/written by a set of authorized knowers, they have provided themselves with a defence for whatever is printed. Thanks to the authoritative position of the criminal justice system, what is reported as statements made in court or by the police needs no justification or verification, since it is all part of the (socially accepted and legitimated) ‘judicial career’ (Fishman, 1980: 94) of a criminal case.

6 CONCLUSION This chapter has sought to fill a gap in the literature on the representation of rape in the media by comparing agency and naming strategies attributed to complainants (when referred to as victims) and law-and-order institutions in reports of rape crimes. This approach goes beyond the traditional comparison between complainants and perpetrators and adopts a broader understanding of victimization and victim blaming. The analysis highlighted that the representations of victims(s), police and court in the corpus shared some patterns of transitivity and lexical choices. For example, all social actors were found to occur in active roles in both material and verbal processes. This would seem to contradict previous studies which suggested that women’s voices are often absent from news stories. It would also appear to negate the connotation of lack of agency associated with the term victim. However, closer observation of these patterns in context showed that formal similarities between the representation of women and institutions did not always correspond to similar meanings. Victim(s), that is, women who experienced rape and would, therefore, be in a better position to define it, were presented as bystanders, with no valuable knowledge beyond the account of their own

160

CONTEMPORARY MEDIA STYLISTICS

subjective experience and testimony. While complainants did occur as Sayers in verbal processes, they mostly did so to narrate their suffering and recount the violent details of their experiences. Women were only allowed to express opinions and emotions, rather than facts; on the other hand, police and court were allowed to ‘neutrally’ and ‘factually’ speak about other people’s experiences. In the corpus, this was achieved by associating law-and-order institutions with formal denominations, neutral speech-act verbs, and collective nouns that hid the one-sided nature of their statements. Combined, these structures promoted a sense that their words and actions were part of a standardized, objective process which, in reality is a deeply individual and haphazard way of dealing with cases. By presenting the law as an impartial and independent entity and by using the words of official sources as objective and trustworthy ways to convey accounts of rape complainants, the press provided a justification for the regular and frequent reliance on them as normative suppliers of news (about rape). Thus, it made the legal discourse of rape its own and, in doing so, it granted law-and-order institutions the discoursive, rather than legal, authority to determine which notion of rape is acceptable, while also legitimizing these institutions role as truth holders entitled to make decisions on the basis of their own claim to truth. It could be argued that the validation provided by the press, itself a powerful and authoritative institution, functions as a ‘seal of validity’ for this very sense of authority and allows the law to continue presenting its own version of events as the only possible and reliable truth. In fact, one of the ways in which the law gains and preserves its authority is by asserting its supposed impartiality, by setting itself ‘outside the social order, as if through the application of legal method and rigour, it becomes a thing apart which can in turn reflect upon the world from which it is divorced’ (Smart, 1989: 12). Since the use of established institutions as sources could also be seen as a way for the press to gain its own validation, it could further be speculated that the press and the judicial system mutually legitimize each other and constantly reinforce each other’s status as unquestioned truth makers. In other words, by legitimizing law-and-order institutions, newspapers also legitimize their own reporting. The analysis showed that the lack of representation of women as authoritative voices, combined with the over-reliance on law-and-order institutions as sources, has serious implications for the representation of rape complainants. For example, the fact that the voices of women did not appear to have the same weight as that of the law led to a series of misrepresentations and reinforcements of myths, since rape appeared to be presented mostly from the institutions’ perspective. As a consequence, the press paid particular attention to rape cases which were reported and had come to the attention of the police. This law-andorder angle, however, obscures the overwhelming majority of rape crimes that are not reported to or pursued by the police (Smart, 1989: 38).

RAPE VICTIMS AND THE LAW

161

Similarly, given that women could not see their own definition of rape validated and their stories did not seem to hold truth value, the stereotypical beliefs about gender roles that shape the authoritative legal ‘truth’ could not be challenged easily. In fact, these were legitimized and maintained, consequently opening up the possibility of (indirect) victim blaming. Finally, despite the fact that only a very small minority of women have been found to have lied5 about rape, the analysis of alleged victim – which openly questions their credibility or innocence – together with the insinuations of blame and responsibility when women were represented as having agency (especially in material processes), showed that the debate (and doubts) about women’s responsibility and reliability in rape cases is still a dominant feature of the discourse that surrounds this crime. This creates a paradoxical situation: because of the possibility that the woman may be culpable of lying, her testimony is considered insufficient evidence to ascertain that a crime took place (Smart, 1989: 34); at the same time, the invalidation of her voice leads to a reliance on other sources which reinforce the very idea that she may be lying. The analysis of this corpus brought to attention the inner contradiction that characterizes the perception and representation of rape complainants: their victimization followed or preceded by a lack of action does not translate into absolution from blame or lack of responsibility (as it is the case with perpetrators) but is condemned. Yet, so is their agency, with their actions before and after the attack constantly judged and scrutinized, even when they are linguistically labelled as victim(s). While the (verbal and material) passivity of rape complainants has often been constructed as negative and culpable (Smart, 1989), agency for rape complainants is not necessarily positive either. Agency for rape complainants means responsibility. In the corpus, women’s victimization and victim blaming took the form and were a consequence of silencing, not through omission, but through deprivation of the authority, truth value and universality that were instead accorded to established institutions, which, instead were considered factual and reliable sources. The analysis showed that, while official sources were used by the press to help create the credibility of a news report, the use of victim(s) as sources served the purpose of providing newsworthy information for the reader. While a few examples from the corpus suggested that a different representation is possible (i.e. by stressing the strength of rape complainants or their resistance to the attack), these were mostly exceptions. It may be argued that this paradoxical representation of victim(s), as ‘guilty if they do and guilty if they don’t’, incapable of constructing their own identities, and not sufficiently trustworthy to contribute to debates about their own experiences, is a result of the maleness of the structure within which such representations are created. Through the institutions that constitute it, this structure infantilizes and blames women and does not grant them the freedom to create meaning and truth, which remain

162

CONTEMPORARY MEDIA STYLISTICS

a prerogative of the institutions and are limited to what the structure allows and understands, even when it shapes and confirms sexist attitudes and beliefs towards rape and rape complainants.

NOTES 1. Terminological note: The word complainant is chosen in this chapter to refer to women who have been raped by men. While terms such as victim or survivor also have their own merits, the term complainant emphasizes the act of speaking about rape and is, therefore, particularly relevant in a discussion of silencing as a form of victimization (for further discussion, see Waterhouse-Watson, 2013, A Note on Terminology). 2. According to British law, ‘a person (A) commits an offence if he intentionally penetrates the vagina, anus or mouth of another person (B) with his penis’ and if ‘B does not consent to the penetration’. http://www.legislation.gov.uk/ukpga/2003/ 42/section/1 (27 July 2018). 3. In order to limit the number of occurrences to a manageable sample, only collocates that had at least five occurrences were considered. 4. Both ‘object of ’ and ‘subject of ’. 5. https://www.cps.gov.uk/legal-guidance/false-allegations-rape-andor-domestic-abus e-see-guidance-charging-perverting-course (28 July 2018).

REFERENCES Baker, P. (2006). Using Corpora in Discourse Analysis. London: Continuum. Baker, P., and McEnery, T. (2015). ‘Who benefits when discourse gets democratised?: Analysing a Twitter corpus around the British Benefits Street debate’. In P. Baker and T. McEnery (Eds), Corpora and Discourse Studies: Integrating Discourse and Corpora (pp. 244–65). London: Palgrave Macmillan. Baker, P., Gabrielatos, C., Khosravinik, M., Krzyzanowski, M., McEnery, T. and Wodak, R. (2008). ‘A useful methodological synergy? Combining critical discourse analysis and corpus linguistics to examine discourses of refugees and asylum seekers in the UK press’. Discourse and Society, 19(3), 273–306. Bednarek, M. and Caple, H. (2012). News Discourse. London: Continuum. Bell, A. (1991). The Language of News Media. Oxford: Blackwell. Benedict, H. (1992). Virgin or Vamp: How the Press Covers Sex Crimes. New York: Oxford University Press. Blommaert, J. (2005). Discourse: A Critical Introduction. Cambridge: Cambridge University Press. Burt, M. R. (1980). ‘Cultural myths and supports for rape’. Journal of Personality and Social Psychology, 38(2), 217–30. Burton, D. (1982). ‘Through dark glasses through glass darkly’. In R. Carter (Ed.), Language and Literature (pp. 195–214). London: Allen and Unwin. Caldas-Coulthard, C. R. (1994). ‘On reporting reporting: The representation of speech in factual and fictional narratives’. In M. Coulthard (Ed.), Advances in Written Text Analysis (pp. 295–308). London: Routledge.

RAPE VICTIMS AND THE LAW

163

Caldas-Coulthard, C. R. (1995). ‘Men in the news: The misrepresentation of women speaking in news-as-narrative-discourse’. In S. Mills (Ed.), Language and Gender: Interdisciplinary Perspectives (pp. 226–39). London: Longman. Cameron, D. and Kulick, D. (2003). Language and Sexuality. Cambridge: Cambridge University Press. Clark, K. (1992). ‘The linguistics of blame: Representations of women in The Sun’s reporting of crimes of sexual violence’. In M. J. Toolan (Ed.), Language, Text and Context (pp. 183–97). London: Routledge. Cruse, D. A. (1977). ‘The pragmatics of lexical specificity’. Journal of Linguistics, 13, 153–64. Ehrlich, S. (1999). ‘Communities of practice, gender and the representation of sexual assault’. Language in Society, 28(2), 239–57. Ehrlich, S. (2001). Representing Rape: Language and Sexual Consent. London: Routledge. Fairclough, N. (1992). Discourse and Social Change. Cambridge: Polity Press. Fairclough, N. (1995). Media Discourse. London: Edward Arnold. Fishman, M. (1980). Manufacturing the News. Austin: University of Texas Press. Fairclough, N. (2003). ‘“Political correctness”: The politics of culture and language’. Discourse and Society, 14(1), 17–28. Fowler, R. (1991). Language in the News. London: Routledge. Geis, M. L. (1987). The Language of Politics. New York: Springer-Verlag. Glasgow University Media Group. (1980). More Bad News. London: Routledge and Kegan Paul. Halliday, M. A. K. (1994). An Introduction to Functional Grammar (2nd edn). London: Edward Arnold. Halliday, M. A. K. (2004). An Introduction to Functional Grammar (3rd edn). London: Arnold. Hirsch, S. F. (1994). ‘Interpreting media representations of a “Night of Madness”: Law and culture in the construction of rape identities’. Law and Social Inquiry, 19(4), 1023–58. Jeffries, L. and Walker, B. (2018). Keywords in the Press: The New Labour Years. London: Bloomsbury. Kelly, L. (1988). Surviving Sexual Violence. Cambridge: Polity Press. Lakoff, G. and Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press. Lischinsky, A. (2017). ‘Doing the naughty or having it done to you: Agent roles in erotic writing’. Porn Studies, 5(2), 156–74. Louw, B. (1993). ‘“Irony in the text or insincerity in the writer?” – The diagnostic potential of semantic prosodies’. In M. Baker, G. Francis and E. Tognini-Bonelli (Eds), Text and Technology: In Honour of John Sinclair (pp. 157–76). Amsterdam: Benjamins. Meyers, M. (1997). News Coverage of Violence against Women: Engendering Blame. London: Sage. Mills, S. (1995). Feminist Stylistics. London: Routledge. Mills, S. (2004). Discourse. London: Routledge. Mills, S. (2008). Language and Sexism. Cambridge: Cambridge University Press. Mills, S. (2012). Gender Matters: Feminist Linguistic Analysis. London: Equinox. Neale, A. (2019). Process Types and Participant Roles in the English Clause. London: Equinox.

164

CONTEMPORARY MEDIA STYLISTICS

Office for National Statistics. (2017). Sexual Offences in England and Wales. Retrieved from https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/ar ticles/sexualoffencesinenglandandwales/yearendingmarch2017 Richardson, J. (2007). Analysing Newspapers. An Approach from Critical Discourse Analysis. New York: Palgrave Macmillan. Scully, D. (1994). Understanding Sexual Violence: A Study of Convicted Rapists. New York: Routledge. Simpson, P. (1993). Language, Ideology and Point of View. London: Routledge. Smart, C. (1989). Feminism and the Power of Law. London: Routledge. Talbot, M. (1995) Fictions at Work: Language and Social Practice in Fiction. London: Longman. Talbot, M. (1998). Language and Gender: An Introduction. Cambridge: Polity Press. Thompson, G. (1996). Introducing Functional Grammar. London: Arnold. Tranchese, A. (2019). ‘Covering rape: How the media determine how we understand violence against women’. Gender and Language (forthcoming). Tuchman, G. (1972). ‘Objectivity as strategic ritual’. American Journal of Sociology, 77, 660–79. van Dijk, T. A. (1988). News as Discourse. Hillsdale: Lawrence Erlbaum Associates Publishers. van Dijk, T. A. (1991). Racism and the Press. London and New York: Routledge. van Leeuwen, T. (2008). Discourse and Practice: New Tools for Critical Discourse Analysis. Oxford: Oxford University Press. Walby, S. (1990). Theorizing Patriarchy. Oxford: Blackwell. Wareing, S. (1990). ‘Women in fiction: Stylistic modes of reclamation’. Parlance, 2(2), 72–85. Waterhouse-Watson, D. (2013). Athletes, Sexual Assault and ‘Trials by Media’. New York: Routledge.

CHAPTER EIGHT

Changing media representation of GinaLisa Lohfink as the icon of the ‘Nein heißt nein’ (no means no) movement in Germany ULRIKE TABBERT

This chapter will use tools from critical stylistics (Jeffries, 2010) to examine two newspaper articles from Spiegel Online (Der Spiegel is a weekly German news magazine; www.spiegel.de is the accompanying website), both reporting on the criminal court case against Gina-Lisa Lohfink, a German reality TV star. Lohfink had initially filed a rape case against two men she had had sex with in 2012 and claimed she had been sedated and then raped. The rape charges against the men were dismissed by the prosecution office after an expert witness stated that he found no evidence of her being drugged. The prosecution office subsequently filed charges against Lohfink for making false rape accusations arguing that her saying ‘Hör auf ’ (stop it) in a video related to the filming but not to the intercourse. Before the first verdict against her

166

CONTEMPORARY MEDIA STYLISTICS

in 2016, Lohfink became the short-lived icon of the German ‘Nein heißt nein Bewegung’ (no means no campaign) calling for tougher rape laws, specifically that sexual persistence post verbal refusal (like the aforementioned ‘Hör auf ’) should constitute an offence. The newspaper articles analysed in this chapter were published in summer 2016 (before the verdict against Lohfink) and February 2017 (after the appeal court ruling). The aim is to show by means of identifying stylistic choices made in these texts how the reportage on the case and thus Lohfink’s linguistic construction changed from before to after her conviction, in particular when the judge stated that she had not been raped. Before the verdict, transitivity patterns as well as direct speech presentation reveal the illocutionary force behind the then published article aiming at a change of the relevant paragraph 177 of the German Criminal Code. Lohfink as a person is of less importance (which can be seen, for example, in naming strategies) and almost appears to be mentioned en passant and for argument’s sake as the main goal of this article is arguably to create moral outrage and panic (Cohen, 2002) and to impose pressure on political decision makers. In the second article, however, the focus is on Lohfink herself, her lifestyle and appearance, as her public perception has meanwhile changed from her being a rape victim to being a sentenced offender. The ruling appeal court judge even stated that Lohfink has done a disservice (Bärendienst) to other rape victims. Due to Lohfink’s risk-prone lifestyle (O’Malley, 2010) she has never been regarded as an ‘ideal victim’ (Christie, 1986) in terms of victimhood status. However, her conviction is reported on almost unrelated to the meanwhile implemented change of the German Criminal Code which came into effect on 10 November 2016.

1 INTRODUCTION In summer 2016, a German reality TV star named Gina-Lisa Lohfink stood trial in Berlin, Germany, for having made false rape allegations against two men: footballer Pardis Fardjad-Azad and party promoter Sebastian Pinto. Her trial was widely reported on, even outside Germany, as it coincided with political efforts to toughen governing rape laws in Germany. The mainstream media, in support of these efforts, turned the trial into what it actually was not, namely, a rape trial, and Lohfink into a victim of rape whose story was believed neither by prosecutor Corinna Gögge nor by the ruling judge Antje Ebner. Only after Lohfink had been sentenced and her appeal against the verdict had been dismissed did the media reportage change, putting forward a distancing stance on Lohfink’s story. The primary interest of this chapter is to examine how these two ways of reportage happened linguistically and what ideological meaning this has for Lohfink’s adverse construction as an alleged victim of rape versus

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

167

a convicted offender of false rape accusations. The findings are continually linked to criminological frameworks born out of research in victimology and in particular on victims of rape. For the purpose of examining this change in reportage I chose two articles from Spiegel Online, both published on the magazine’s website. Der Spiegel is a weekly German quality news magazine, and www.spiegel.de is the accompanying website. The magazine takes pride in the quality of its journalistic work and attracts a higher educated readership than, for example, tabloid journalism. I therefore expected a balanced reportage on the case complemented by wellresearched facts. The first article (Article 1 hereafter, (Meiritz, 2016)) dates from 11 June 2016, a time when the initial trial against Lohfink was ongoing. It consists of 845 words. The second article (Article 2, (Siemens, 2017)) was published on 10 February 2017, the date when Berlin’s Court of Appeal (Kammergericht) dismissed Lohfink’s appeal against the verdict. Article 2 is 1,053 words in length. This chapter presents tools from critical stylistics (Jeffries, 2010), a framework particularly suited to detect ideological meaning in texts. Critical stylistics is able to expose values attached to ‘the construction of a particular description of the world through language’ (Jeffries, 2015b: 384; Tabbert, 2016: 27ff). Based on a relativist understanding of ideologies as a ‘coherent and relatively stable set of beliefs and values’ (Wodak and Meyer, 2009: 8), critical stylistics takes this notion further by stating that ‘ideation’ (i.e. the embodiment of our experiences of the external and internal world in language) and ideology are in fact ‘delivered by the same set of textual features’ (Jeffries, 2015b: 384). Critical stylistics, in providing a framework of ten textual–conceptual functions of texts, covers a broad range of analytical tools to examine ideological meaning as well as a route into the text that can ensure replicability of the analysis. Before I start with the linguistic analysis, I provide some background information on the trial against Lohfink, the sociopolitical situation at that time in Germany and, to a limited extent, also on Lohfink herself because this information proves to be relevant to fully appreciate the manipulation that went on in the reportage of the trial.

2 BACKGROUND In June 2012, Lohfink spent one night with footballer Pardis Fardjad-Azad and party promoter Sebastian Pinto in Berlin. She had sexual intercourse with both men, which the men filmed. Afterwards, the men offered these clips for sale to tabloid media outlets. The outlets declined purchase of the material and informed Lohfink’s management. After the clips proved to be unsellable, the men published them online. Lohfink subsequently filed charges for unpermitted publication of those films and shortly afterwards also for rape, claiming she had

168

CONTEMPORARY MEDIA STYLISTICS

been sedated and raped. The rape charges against the men were dismissed by the prosecution office after an expert witness, on grounds of an examination of the video material, stated that he had found no evidence for her being drugged. The men nevertheless were sentenced to a fine in May 2014 for publishing those videos without Lohfink’s permission. The prosecution office subsequently charged Lohfink for having made false rape accusations. The initial charge was made by means of a summary sentence (Strafbefehl), issued by Berlin’s District Court (Amtsgericht Tiergarten). Key arguments were that her saying ‘Hör auf ’ (stop it) and ‘nein, nein, nein’ (no, no, no) in the videos related to the filming but not to the intercourse and that there was no evidence of her being in a state of lost willpower during the event. Owing to Lohfink’s objection, a trial was held at the District Court where Lohfink, through provisions in the German Code of Criminal Procedure, was not obliged to attend. It would have sufficed for her lawyer to represent her. Lohfink nevertheless attended in person, which brought about a press rush. After a six-day trial, the District Court sentenced Lohfink to a monetary penalty (€20,000 in total). Lohfink appealed against the decision, choosing a type of appeal (Sprungrevision) which only allows for review of the correct application of the law based on the written judgement but does not provide opportunity for a re-examination of the evidence. On 10 February 2017, Berlin’s Court of Appeal confirmed the guilty verdict and thus dismissed Lohfink’s appeal against the verdict itself. The court further ruled that the sentence needed to be reviewed concerning her income and the amount of money she was sentenced to pay. In reasoning his decision, the ruling judge Fischer stated, among other things, that Lohfink had done a disservice to other rape victims (Article 2). After recommitment to the District Court and further investigation into how much money Lohfink earns per month, the court eventually confirmed the monetary penalty on 16 October 2017. Lohfink’s conviction is thus binding since February 2017 concerning the verdict itself and since 24 October 2017 concerning the sum of monetary penalty. It is not automatic that in the case of a dismissal of criminal charges following a report made by an alleged victim this person faces charges themselves for having made false accusations. If the prosecution office charges the complainant, however, this is a first indication of there being sufficient evidence not just for a wrongly made accusation but also for criminal intent. The court, following a preliminary examination of the evidence presented by the prosecution office, then issues either a summary sentence (Strafbefehl, as in Lohfink’s case) or a committal for trial as a necessary prerequisite for a trial to actually take place. All these legal barriers had successfully been passed at the time of the trial against Lohfink. Spiegel Online, however, appeared to be almost in denial of these facts in its reportage on the case in Article 1. Before the first verdict against Lohfink, she was made an icon of the German ‘Nein heißt nein Bewegung’ (no means no campaign) demanding tougher rape

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

169

laws. In Article 2, she was referred to as ‘Lichtgestalt der Frauenbewegung’ (shining light of women’s movement). This movement becomes meaningful for the trial against Lohfink when seen in its historical and social context. Originally, the notion ‘No means no’ dates back to second-wave feminism and Susan Brownmiller’s book Against Our Will (1975). Key demands put forward in those days were, among others, a right to self-determination over one’s own body, including refusal of (marital) sexual intercourse, as well as a woman’s right to determine a pregnancy. A long-standing patriarchal world view saw women as being inferior to men if not even in a property-like relationship to them. This, according to Brownmiller, stems from the early days of mankind when at some point men discovered their penis as a weapon. Rape thus became a driving force in history and secured patriarchy. Such ideological stance is traceable in the German penal code (Strafgesetzbuch) up until 1997 when rape could only be committed outside marriage, only through penetration and only against women. In 1997, the law was reformed and other sexual acts than (vaginal) penetration and if particularly humiliating became criminal offences. Further, rape within marital relationships or committed against men also became criminal offences. This mirrored societal change and, in line with Brownmiller’s demands, was a declared belief in gender equality and against patriarchy. However, the offender had to either use force or threaten the victim (or the victim had to be defenceless) in order for the act to constitute rape. Seen from an evidentiary perspective, it thus had to be proven in court that the victim either had traces of said violence on his/her body or had complied out of fear. Physical opposition (Gegenwehr) on the part of the victim is not required. However, the more forceful the resistance and the more the traces of violence on the victim’s body, the more conclusive becomes the evidence and the more likely a conviction. Apart from leaving the prosecution office with a sometimes challenging situation of proving (or disproving) the accusation, the then governing law had some other serious ‘protection gaps’ (Schutzlücken). For example, actions like touching another person’s private parts above clothes were difficult to subsume under assault, not to mention rape norms. Also, if the attack caught the victim by surprise and provided not enough time to form an ‘adverse will’ (entgegenstehender Wille), such actions were not considered as rape. Already before the Lohfink trial, a reform of the governing law was under way which aimed at closing said protection gaps. A first draft put forward to parliament (Bundestag) by the Ministry of Justice, dating from 2015, was nevertheless not as substantial and far-reaching as the one that eventually passed and was implemented in November 2016. During ongoing debate about how to best reform the relevant sections of the German penal code, one event in Germany in particular accelerated

170

CONTEMPORARY MEDIA STYLISTICS

the reform. On New Year’s Eve 2015, several hundred women claimed to have been sexually molested and mugged, and some even raped, on the streets of Cologne and other major German cities while attending public celebrations. In some cases, these women had their private parts touched in the middle of a crowd of people and too quickly for these women to form the said ‘adverse will’. In addition, most of the attackers were described as supposedly being immigrants. Germany at that time had just offered refuge to large numbers of people, among them many fleeing from the war in Syria in 2015. A controversial discussion had been going on in German society about this ‘welcome culture’ so that, unsurprisingly, the events in Cologne led to a state of ‘moral panic’ (Cohen, 1980, 2002) and lent a sense of urgency to the legislative process. The eventual changes of the penal code implemented, for example, that verbal refusal already suffices for the act to be classified as rape, as long as it was perceptible by the attacker. This was a clear commitment to the notion behind the slogan ‘No means no’ from Brownmiller’s times, also stated in the rationale accompanying the new piece of legislation. At the time of the first trial against Lohfink, however, this change of legislation was still pending and the public at large was still shaken by the events in Cologne. This set the scene for the Lohfink trial at the District Court. Activists as well as politicians sensed a chance to instrumentalize the trial to further awareness of gender inequality and to close the aforementioned ‘protection gaps’ in the governing law. Thus, during the trial, supporting crowds gathered, showed placards stating ‘Nein heißt nein’ (no means no) and formed #TeamGinaLisa on Twitter. They maintained that Lohfink was a victim of rape and thus wrongly accused because she was neither believed by the prosecution nor the court. Politicians openly sided with Lohfink during her ongoing trial, blurring the separation of powers anchored in the German constitution. The trial also showed how politicians in unison with the media followed Lohfink’s narrative lead and neglected that Lohfink was not a victim of rape according to the prosecution office. This fact (and not just whether Lohfink’s false rape allegations were made with criminal intent) was now called into question by Lohfink, her defence team, activists, politicians and the media alike. Manuela Schwesig, Prime Minister of Mecklenburg-West Pomerania and at that time Federal Minister for Family Affairs, Senior Citizens, Women and Youth, joined #TeamGina Lisa. She stated, ‘“Nein heißt nein” muss gelten. Ein “Hör auf ” ist deutlich’ (Article 1) (‘No means no’ must count. ‘Stop it’ is clear), referring to the aforementioned videos used as evidence. In aligning both Brownmiller‘s notion of ‘no means no’ with Lohfink‘s utterance in the video, ‘Hör auf ’, Schwesig mixed a long-standing feminist claim with contested evidence in a court trial. In doing so she imposed moral pressure on a criminal trial which, if

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

171

followed through, would theoretically mean that a conviction on legal grounds can be overturned on moral grounds. Schwesig was not the only politician using the trial for political purposes. Heiko Maas, now Federal Minister of Foreign Affairs and then Federal Minister of Justice and Consumer Protection, stated in June 2016, ‘Die Verschärfung des Sexualstrafrechts darf nicht länger blockiert werden. Die Reform ist dringend notwendig, um eklatante Schutzlücken zu schließen’ (SpiegelOnline, 2016). (The tightening of the law governing sexual offences must not be blocked any longer. The reform is urgently needed to close blatant protection gaps.) Later criticized for interference with an ongoing criminal trial in his capacity as Minister of Justice, he rejected any relation between his timely utterance and the trial against Lohfink. In March 2016, Minister Maas published a highly contested figure on his ministry’s webpage, namely, that in only 8 per cent of all reported rapes a conviction is reached. This figure is problematic as it neglects several facts. First, it turns a blind eye to the particularly high number of wrong rape accusations compared to other forms of delinquency in Germany. Second, the required level of proof (beyond reasonable doubt) often provides problems as mentioned above not least because such crimes are usually not witnessed by many people. Third, an acquittal does not mean that there is sufficient proof to the contrary, namely, that a rape did not happen but instead merely means that reasonable doubt remains. In those cases, the legal rule ‘in dubio pro reo’ (when in doubt, for the accused) needs to be applied. Fourth, in 20 per cent of all rape allegations the suspect is unknown and a fifth point is the already mentioned reform of rape law in 1997. The latter forbids direct comparison of rape figures from before and after this point. What can be taken from this is, in summary, that rape cases are generally difficult to prove and that in 92 per cent of all rape accusations a conviction could not be reached, the reasons for which are manifold (Rückert, 2017). Bearing this in mind, Lohfink’s narrative frame of being an unbelieved rape victim touched a nerve, particularly given the historical background of her trial. Frames, according to Lakoff (2004: XV), are ‘mental structures that shape the way we see the world’. Such frames are powerful as ‘they shape the goals we seek, the plans we make, the way we act, and what counts as a good or bad outcome of our actions’ (Lakoff, 2004: XV). Framing does not just happen through metaphorical structures but on various linguistic levels as outlined in the analysis. As ‘framing’ refers to the initial frame given to a story, I better talk about ‘re-framing’ due to the fact that the initial frame with her being the accused had already been set by the prosecution office and the court with the indictment and committal proceedings. At the time Article 1 was published, the frame of Lohfink being the accused had already been successfully undermined. Historically, rape victims have faced the dilemma of being perceived as ‘to be blamed’ for the crime, at least partially. One of the first examples of victim

172

CONTEMPORARY MEDIA STYLISTICS

blaming (Moore, 2014; Van Dijk, 2009) is the ancient mythological story of Lucretia, a Roman wife of high moral standards who got raped by a king’s son, Tarquinius. Lucretia, although being assured by her husband and her father that she was not to be blamed for it, felt she could only restore her ‘honour’ by committing suicide. Whereas Lucretia would nowadays be easily given the status of an ‘ideal victim’ according to Christie’s (1986) criteria, her suicide following her perceived share of blame for the rape (even if there was nothing she could have done to prevent it) was a means to restore her moral purity. Centuries later, the character Sarah Tobias in the Hollywood movie The Accused1 (Kaplan, 1988), starring Jodie Foster, was gang raped in a bar and blamed for it afterwards because she wore a miniskirt and led a ‘risk-prone lifestyle’ (O’Malley, 2010). In other words, she was a ‘non-ideal’ victim in terms of her victimhood status. Other than morally flawless Lucretia, Sarah Tobias’s miniskirt was perceived as ‘issuing an invitation’ and thus led to credibility problems and victim blaming.

3 LINGUISTIC ANALYSIS The aim of this chapter is to show by means of identifying stylistic choices made in both articles how the reportage on the case and thus Lohfink’s linguistic construction changed from before to after her conviction. Lohfink’s construction changed from being a ‘Lichtgestalt der Frauenbewegung’ (shining light of the women’s movement) to being a convicted offender, although now even wider known in public than before her trial. The framework applied in this chapter is critical stylistics (Jeffries, 2010) which serves ‘to investigate the construction of ideological meaning in all texts’ (Jeffries, 2015b: 381). Critical stylistics is ‘a text-based methodology for CDA’ (critical discourse analysis) (Jeffries, 2014b: 476) as well as a ‘mainstream textbased stylistics with a particular (critical) purpose’ (Jeffries, 2015a: 159). Using the ideational metafunction of language (Halliday, 1985) as a starting point, Jeffries framework lists ten textual–conceptual functions of texts. They are ‘a combination of textual features (triggers) and ideational function’ (Jeffries, 2014a: 412). These ten textual–conceptual functions carry the following labels: naming and describing, representing actions/events/states, equating and contrasting, exemplifying and enumerating, prioritizing, implying and assuming, hypothesizing, negating, presenting others’ speech and thoughts and, finally, representing time, space and society (see Jeffries, 2010 and Tabbert 2016 for an introduction). Through these functions, a text represents a reality. Studying these functions, therefore, provides a way to discover said reality in terms of how the text means what it means. However, I am interested not merely in discovering what the newspaper articles mean and what kind of reality they construct but, first and foremost, in

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

173

the ideological meaning they project. The detection of said ideological meaning is possible because ideational processes ‘produce worlds which have values attached to them’ (Jeffries, 2015b: 384). These values are of pivotal importance as they are socially constitutive and ideological in meaning. No text is free from values and thus ideological meaning as texts represents (social) reality in a particular way. Van Leeuwen (2018: 2) states that ‘representations of social reality select, interpret and evaluate social reality’. This selection and evaluation of social reality has ‘a profound impact on people’s knowledge of most aspects of social reality’ (Van Leeuwen, 2018). In the case of the Lohfink trial, a selective representation of reality allowed for a one-sided perception of events which became social reality at the time of the first trial. Analysing the newspaper reportage provides insight into how manipulation worked which became evident, for example, in the protests accompanying the trial. In the analysis, I move along the ten textual– conceptual functions as signposts. The presentation of results echoes five out of ten textual–conceptual functions; the findings from the other five are mentioned and discussed under those headlines which enables to see joined effect. 3.1 Naming strategies for Lohfink Table 8.1 lists all naming choices for Lohfink in both articles under scrutiny. The table shows how references to Lohfink’s profession are made in Article 1 only, whereas a focus on her body parts (her fingers) and an explicit mention of her short-term role in the feminist movement (‘Lichtgestalt der Frauenbewegung’) is limited to Article 2. In Article 1, the noun ‘model’ is a key naming choice, something Gregoriou (2011: 34) has already pointed out to be an important victim-naming strategy in cases where it applies. All three mentions of ‘model’ in Article 1 are not part of quotations but are journalistic naming strategies. In two instances, they pre-modify the proper noun ‘(Gina-Lisa) Lohfink’ in an extended noun phrase functioning as object. In the third instance, ‘model’ is used in a subject position. The noun ‘model’ itself is pre-modified twice by the adjective ‘former’, negating the fact by means of a presupposition that Lohfink is still a model at the time of the trial. This needs to be viewed against the background of some biographical facts about Lohfink. As a routine beauty pageant contestant, Lohfink had risen to fame following her appearances in various television reality shows, starting with Germany’s Next Topmodel, through (amateur) pornography, erotic photographs in magazines like Playboy, as well as through being the face of Venus, an annual erotic fair in Berlin. Modelling has thus played a key part in her gaining celebrity status. Using the pre-modifying adjective ‘former’ contradicts information Lohfink publishes herself on Facebook at the time of writing this chapter. On her Facebook profile page (https://www.facebook.com/pg/GinaLisa2309/about/?r ef=page_internal), she lists ‘Fotoshootings’ as one booking option. Negating

174

CONTEMPORARY MEDIA STYLISTICS

TABLE 8.1 Naming strategies for lohfink Category

Article 1 from 11 June 2016

in relation to legal proceedings

●

●

●

●

Fall Gina-Lisa Lohfink/ Der Fall von Gina-Lisa Lohfink/der Fall Lohfink/ der prominente Fall the (prominent) case GinaLisa Lohfink Der Umgang mit Gina-Lisa Lohfink the dealings with Gina-Lisa Lohfink

Article 2 from 10 February 2017 ● ● ●

● ● ● ● ● ● ●

in relation to crime

●

●

●

●

● ● ● ● ● ●

die mögliche Vergewaltigung von Gina-Lisa Lohfink the possible rape of Gina-Lisa Lohfink die mutmaßliche Vergewaltigung des Models Gina-Lisa Lohfink the alleged rape of model Gina-Lisa Lohfink nicht Täter sondern Opfer not perpetrator but victim ein Opfer (2) a victim Täterin Perpetrator

press reportage

● ● ● ● ● ●

● ●

proper noun

● ●

profession

●

●

●

●

● ●

Gina-Lisa Lohfink (5) Lohfink (5) die mutmaßliche Vergewaltigung des Models Gina-Lisa Lohfink the alleged rape of model Gina-Lisa Lohfink die frühere ‘Germany’s Next Topmodel’Kandidatin Lohfink the former ‘Germany’s Next Topmodel’ contestant Lohfink das frühere Model the former model

● ●

im Fall Lohfink (2) in the Lohfink case seiner Mandantin/seine Mandantin/mit seiner Mandantin (with) his client die Angeklagte the defendant das Verfahren gegen Lohfink the trial against Lohfink ihre Interessen her interests ein Opfer a victim Täterin perpetrator Vergewaltigungsopfer rape victim

das Interesse an ihrer Person the interest in her person Gina-Lisa Lohfink (2) Lohfink (21)

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

175

TABLE 8.1 (Continued) Category

Article 1 from 11 June 2016

age

● ●

die 29-jährige the 29-year-old

Article 2 from 10 February 2017 ● ● ●

(personal) pronoun

● ● ● ●

sie (3) she ihr she (3rd case)

● ● ● ● ● ● ●

determiner

● ●

gender/feminism

● ●

body parts

● ● ● ● ●

●

der 30-Jährigen die 30-jährige (2x) the 30-year-old sie (13x) she ihr (3x) she (3rd case) wer who jemand somebody die the Lichtgestalt der Frauenbewegung shining light of women’s movement hinter ihrem Rücken behind her back ihre eigene Unverwundbarkeit her own invulnerability die Finger mit den langen bunten Nägeln the fingers with the long, coloured nails

her ongoing modelling career in Article 1 therefore suppresses a fact that is perceived as potentially jeopardizing ideal victimhood. Idealness, according to Christie (1986), is key for being easily granted victimhood status and a means to distract from credibility issues (see further below). Naming strategies for Lohfink in relation to the crime never mention false accusations but only ‘rape’ (Vergewaltigung). Nominalization of a process (‘to rape’) existentially presupposes that a rape had occurred by ‘packaging up’ (Jeffries, 2010: 19) such ideas and by turning a process into an entity. The noun is pre-modified by epistemic modal adjectives mutmaßlich (alleged) and möglich (possible), indicating doubt. The use of such pre-modifiers shields the magazine from possible libel lawsuits given the already dismissed rape allegations at the time of publication. However, Article 1 elsewhere states: Im Zuge dieses neuen Prozesses sind Zweifel an der früheren Entscheidung der Staatsanwaltschaft laut geworden – auch weil das frühere Model mehrfach ‘Hör auf ’ in dem Video sagt.

176

CONTEMPORARY MEDIA STYLISTICS

[In the course of this new trial doubts about the prosecution office’s earlier decision have become loud – also because the former model repeatedly says ‘Stop it’ in the video.] Example 1 from Article 1 In Example 1 from Article 1, Spiegel Online favours the frame that this is a rape trial with Lohfink being a victim both of rape and of a miscarriage of justice. In quoting unnamed sources who supposedly uttered doubt, Spiegel Online achieves to plant the idea of doubt into the readers’ minds without being held accountable for it. This narrator’s representation of speech act allows for presentation of the illocutionary force of an utterance with only little indication of its content (Semino and Short, 2004: 52). I will return to speech presentation in Sections 3.2 and 3.3. As stated above, Lohfink’s construction as a victim of rape was not without credibility problems as she shared few characteristics with Christie’s (1986) ‘ideal victim’ in terms of victimhood status. An ideal victim, according to Christie, is weak, sick, old or very young, was carrying out a respectable task at the time of the offence and could not possibly be blamed for what happened (p. 19). Lohfink, in contrast, met the men in a club in Berlin, Maxxim, and was under the influence of alcohol at that time. With one of them, Pardis FardjadAzad, she had had consensual sex the night before. Lohfink’s defence lawyer Burkhard Benecken (Benecken, 2018; Benecken and Wöhrle, 2016) was quoted as saying elsewhere, Kurzer Rock, große Brüste – die taugt nicht als Vergewaltigungsopfer. [Short skirt, large breasts – she isn’t good for a rape victim.] Example 2 from Spiegel Online (Siemens, 2016) He anticipated the public’s hesitance in lending her victimhood status (Van Dijk, 2009; Van Wijk, 2013) (see also Example 4 below). It is, however, pivotal that a victim of crime convinces ‘potential givers of the victim status’, that is the audience, by successfully framing her/himself as a victim (Van Wijk: 160f). Benecken, in anticipating the audience’s noncompliance in lending Lohfink victimhood status, implicitly argues in Example 2 that Lohfink’s version of events is true, namely that she is a genuine victim of rape. Seen from Benecken’s point of view, not lending Lohfink victimhood status happens because of her appearance (and not because the rape was invented). The Lohfink case therefore adds a further dimension to Van Wijk’s argument about how to best convince the audience of lending victimhood status. In fact, it emphasizes two aspects: first, the victim has to convince the audience that s/he is a true victim of crime and second; s/he needs to present her/himself as a victim worthy of

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

177

being granted victimhood status. The latter brings in credibility issues and thus a moral dimension and builds on the grounds of the first. In focusing exclusively on the moral dimension as in Example 2, any doubt about whether the rape actually happened is successfully suppressed. The morality aspect is seized by politicians (see Example 3 below) who follow the lead and rightfully demand that victim blaming because of appearance and lifestyle should belong in the past. As a result, the first aspect, namely, whether Lohfink was actually a victim of rape, is shielded away, although this had already been negated according to the authorities. What links the question of actually being a rape victim with morality issues is credibility. Lohfink’s story about the events of the night in question is considered to be credible in Article 1 despite authorities’ judgement. Discussion thus focuses on idealness, lending victimhood status and morality issues, which might explain why the court’s press officer Jani, as mentioned to me in a personal conversation, noticed no particular interest of the media in the facts of the trial, that is, whether Lohfink actually got raped and whether she made accusations with criminal intent. The following example from Article 1 serves to illustrate how Benecken’s discursive lead on the second aspect is seized by politicians, who partake in the complete neglect of the first aspect and the question of who is the victim. Katja Dörner, then vice parliamentary party leader of the Green Party, is quoted in Article 1 using direct speech, saying, Diskussionen, ob ein Opfer aufgrund von Klamotten oder Lebensstil selbst verantwortlich für eine Vergewaltigung ist, müssen endlich der Vergangenheit angehören. [Discussions whether a victim is responsible for being raped due to clothes or lifestyle must finally belong to the past.] Example 3 from Article 1 Joint efforts to dispel doubts about Lohfink’s alleged victimhood status bring to mind the notion of ‘impure’ victims (D. T. Meyers, 2016: 26ff). Impure victims are those who neither fit a ‘pathetic victim paradigm’ nor a ‘heroic victim paradigm’ and thus face credibility problems as well as difficulties in being granted victimhood status. An example of an impure victim Meyers mentions are trafficked sex workers. Lohfink’s construction bears traits of all three categories. She fits the pathetic victim paradigm because she allegedly was unconscious (knockout drops) and thus in no way complicit in her alleged suffering (Meyers, 2016: 35). She also fits the heroic victim paradigm as she is supposedly brave enough to file charges against the men against authorities’ better judgement and to speak up for other girls and women. Lastly, Lohfink fits the impure victim paradigm. In publicly

178

CONTEMPORARY MEDIA STYLISTICS

discussing Lohfink’s appearance and lifestyle the question of her not being ‘morally pure’ is successfully argued against by Lohfink’s lawyer Benecken as well as by politicians (as demonstrated in Examples 2 and 3). Their arguments were heard not least because during her trial, Lohfink and Benecken occupied centre stage with numerous media appearances. There, Lohfink effectively complained about not being believed and repeated her narrative. This was not surprising as back in 2013 she had stated in an interview for Die Tageszeitung taz (Smechowski and Gernert, 2013) that in case she were to be accused, ‘Dann laden wir alle Zeitungen und Sender ein, das wird wie eine Bombe’ [we then invite all newspapers and broadcasters, that will be like a bomb]. At the time she was actually facing trial, the audience was ‘able and willing to hear’ (Brison, 2002: 51) and Lohfink successfully competed for media attention against all other news items (Van Wijk, 2013: 167). In Article 2, naming strategies indicate distance from Lohfink by using legal terms like Mandantin (client) and Angeklagte (defendant) and a higher number of the anaphoric personal pronouns, sie (she) and ihr (she, third case), which are deprived of any ideological meaning. Lohfink is referred to by her surname alone in twenty-one instances and thus far more often than in Article 1, another means to maintain distance. Her being named Opfer (victim) and Vergewaltigungsopfer (victim of rape) are part of speech presentations, quoting earlier press reportage on the case as well as Lohfink’s lawyer Benecken. The context in which these two nouns are presented in Article 2 is of particular ideological importance for the magazine’s revised stance on the case: (1)Die Justiz mache ein Opfer zur Täterin, so die verbreitete Lesart damals. (2)Es gelte offenbar das Motto: ‘Kurzer Rock, große Brüste – die taugt nicht als Vergewaltigungsopfer ’, schimpfte Anwalt Benecken. [(1)Judiciary allegedly turned a victim into an offender, such was the common interpretation back then. (2)The motto apparently is, ‘Short skirt, big breasts – she isn’t good for a rape victim’, lawyer Benecken complained.] Example 4 from Article 2 (numbers added for ease of reference) A discussion of this example (see also Example 2) brings in other textual– conceptual functions besides naming and describing. Although the linguistic phenomena can be clearly distinguished from one another, the effect for the construction of social reality is a combined one and thus needs to be discussed as such. In sentence 1 from the above Example 4, temporal distance is created by the adverb damals (back then). This time deixis anchors a text on a time scale. Here, it indicates that those times when Lohfink was constructed as a victim and judiciary got it all wrong belong in the past. In addition, a distancing stance

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

179

towards the quoted assertion is achieved by means of presenting it in indirect speech, as we have seen in Example 1. Another marker of stance is modality, belonging to the textual–conceptual function of hypothesizing. In sentence 1, an epistemic modal verb form (mache, best translated by adding an epistemic modal adverb ‘allegedly’) is used to indicate doubt. In sentence 2 (seen in context with sentence 1), the use of an epistemic modal verb form (gelte), presenting an amalgam of direct and indirect speech quoting Lohfink’s lawyer, together with an epistemic modal adverb offenbar (apparently) create doubt and thus contribute to a distancing effect. Additionally, the use of past tense schimpfte (complained) is a temporal pointer (time deixis) which adds temporal distance to the already mentioned distance of stance. The repetition of the lexeme Opfer (victim) is rhetorical by nature but here ceases to convince the reader for reasons given before and, additionally, because it is negated (as was also the case in Example 2) and therefore constructs an ‘unrealized world’ (Nahajec, 2009: 109). Note the syntactic structure of the verbiage presented in direct speech: the noun phrase before the hyphen is metonymy as Rock (skirt) and Brüste (breasts) stand for the entire human being. By giving preference to these particular details instead of, for example, Lohfink’s hands (see Example 10), her sexually driven attractiveness is emphasized which in an apposition-like structure further describes the subject die (she/the). This distancing (and even derogatory form) of anaphoric reference, where the definite determiner is used as a pronoun, bears ideological meaning as it indicates social deixis. Further description of the referent is to be found in a subject complement position, Vergewaltigungsopfer (rape victim). Here, negation invites the reader to imagine an alternative scenario, namely when somebody who is looked down upon (due to appearance) can actually be a victim of rape and worthy of being given victimhood status. This effect worked in Example 2 as discussed above but no longer because of the context in which this quote is now presented. 3.2 Representing actions/events/states and presenting others’ speech In Article 1, naming choices for Lohfink rarely appear as the head noun in a subject position and even less often as an Actor of a material process (Simpson, 1993). Here are the two sole occurrences (underlined segments) from a total of sixty sentences: Lohfink erstattete Anzeige: Sie sei gegen ihren Willen sediert und zum Geschlechtsverkehr gezwungen worden, sagt sie. [Lohfink filed charges: She had been sedated and forced to have sexual intercourse against her will, she says.] Example 5 from Article 1

180

CONTEMPORARY MEDIA STYLISTICS

Jetzt soll die 29-Jährige wegen mutmaßlicher Falschaussage 24.000 Euro zahlen, sie hat Einspruch eingelegt. [Now the 29-year-old ought to pay 24,000 Euros for alleged false accusations, she has lodged an appeal.] Example 6 from Article 1 These two material actions with Lohfink as Actor are her filing charges and her lodging an appeal. These two instances refer to the alleged rape and not the crime she is accused of. Both relate to legal procedure, although the first is in fact the action she had been accused of doing falsely and with criminal intent. Nevertheless, both actions are also those commonly done by rape victims (if they come forward at all). Mentioning these two and no other material actions done by Lohfink is another rather subtle means to align Lohfink with other rape victims and also to construct her as someone who is not being believed and has to (legally) fight for her rights. As such she is not merely constructed as a victim of rape but her construction covers ‘another important element in being an ideal victim’ (Christie, 1986: 21). This other aspect Christie mentions is her being ‘strong enough to be listened to, or dare to talk. But she (he) must at the very same time be weak enough not to become a threat to other important interests’ (Christie, 1986). I have introduced this notion under Section 3.1 and mentioned the timeliness of the Lohfink trial. Lohfink’s voice is amplified by politicians as it serves their political purpose. Lohfink is placed ‘at the centre of criminal justice policy’ (Meyers, 2016; Poletta, 2006; Walklate et al., 2018: 4). To achieve this means that her story needs to be a ‘canonical’ narrative (Meyers, 2016; Poletta, 2006; Walklate et al., 2018). Her not being believed actually aligns her story to those of other rape victims in triggering existing schemas, namely, that victims of rape are sometimes not believed by the police and other authorities (Meyers, 1997). Raising doubts about the prosecution office’s decision to dismiss rape charges aligns with an investigative journalistic approach to news reportage, claiming a public interest in investigating the truth (Pressekodex, published by Deutscher Presserat, http: //www.presserat.de/pressekodex/pressekodex). These idealistic professional ethics are undoubtedly worth pursuing. However, it appears that no thorough journalistic investigation had been made (as is often the case in relation to crime news (Rowbotham, Stevenson and Pegg, 2013: 200; Soothill and Walby, 1991: 35)). Instead, one-sided preference is given to Lohfink’s narrative seized by politicians. Here are two examples from Article 1 expressing doubt about the prosecution office’s earlier decision. Im Zuge dieses neuen Prozesses sind Zweifel an der früheren Entscheidung der Staatsanwaltschaft laut geworden – auch weil das frühere Model mehrfach ‘Hör auf ’ in dem Video sagt.

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

181

[In the course of this new trial doubts grew loud about the prosecution office’s earlier decision – also because the former model says ‘Stop it’ in the video multiple times.] Example 7 from Article 1 (same as Example 1) Das Berliner Amtsgericht schätzte die Videosequenzen mit Lohfink und den beiden Männern nicht als Vergewaltigung ein. Doch die Entscheidung von damals wird öffentlich angezweifelt. [Berlin local court did not judge the video sequences with Lohfink and the two men as rape. But the earlier decision is being publicly contested.] Example 8 from Article 1 It is noteworthy that in both examples passive verb voice is used to hide agency, namely, who doubts the prosecution office’s earlier decision. The reason behind the said doubt is an out-of-context quote from the video, namely, Lohfink saying ‘Stop it’. This and Example 5 are the only two occasions where Lohfink is quoted, in direct (Example 7) and indirect (Example 5) speech. Lohfink is thus equally rarely the Sayer in a verbalization process as she is the Actor in a material process. She is denied an active role in Article 1 in contrast to Article 2 and is thus aligned with passivity and voicelessness, as Meyers (2016: 34) points out concerning the pathetic victim paradigm when stating, ‘Pathetic victims … are people whose capacities for choice and action have been so completely neutralized that there can be no doubt that they are innocent.’ Transitivity processes in relation to Lohfink in Article 2 provide a strong contrast to findings from Article 1. Although material processes with Lohfink as Actor are equally rare, a difference can nevertheless be noticed even there: Um zehn vor zehn betritt Gina-Lisa Lohfink das Berliner Kammergericht und rauscht wortlos an den Journalisten vorbei – begleitet von ihrem Freund Florian Wess, von Manager Helmut Werner und Verteidiger Burkhard Benecken. [At ten to ten, Gina-Lisa Lohfink enters Berlin Court of Appeal and sweeps wordlessly past journalists – accompanied by her friend Florian Wess, by manager Helmut Werner and defence lawyer Burkhard Benecken.] Example 9 from Article 2 These material processes, the second being metaphorical, describe Lohfink’s movement when entering court. These material actions are different from the ones in Article 1 where they describe Lohfink’s legal actions. A particular focus on her body in Article 2 (see also Example 4) can also be witnessed in the

182

CONTEMPORARY MEDIA STYLISTICS

next example with an additional zooming in on her body parts by means of a meronymic relationship between Lohfink and her fingers: Lohfink blickt vor sich ins Leere, die Finger mit den langen bunten Nägeln aber kann sie nicht stillhalten, sie müssen sich ständig berühren. [Lohfink stares dead ahead, but she cannot rest her fingers with long, colourful nails, they must constantly touch each other.] Example 10 from Article 2 The described movement of her fingers serves as a glimpse into her inner world by indicating that despite her outwardly shown indifference (‘stares dead ahead’), she is in fact nervous. This is achieved by means of a concessive opposition (Davies, 2013: 71ff), indicated by the contrastive conjunction aber (but). Deontic modal auxiliaries (kann/can; müssen/must) underline a split appearance, constructing her fingers as being almost separated from her body and having a life of their own. Apart from these material actions, the majority of sentences in Article 2 with Lohfink in a subject position are mental and verbalization processes, providing further access to Lohfink‘s inner world. Beside her being quoted with ‘Hör auf ’ (Stop it) from the video as in Article 1, her verbiages, sometimes mixed with her lawyer’s, are presented in indirect speech and other less faithful categories of speech presentation. The following example, in addition to speech presentation, further illustrates how glimpses of Lohfink’s mental state are constructed: Lohfink und ihr Anwalt schäumten öffentlich, kündigten an, im Rahmen einer Sprungrevision sofort die höchste Instanz anzurufen. [Lohfink and her lawyer fumed publicly, they announced an immediate leapfrog appeal to supreme authority.] Example 11 from Article 2 After Berlin District Court’s ruling, Lohfink (and her lawyer) ‘fumed’ and announced their intended next legal step. ‘Fuming’ is narrator’s report of speech act with the exact verbiage not being delivered. Presented like this, it is also a mental reaction to the verdict, namely one of anger and fury. This verb allows the reader to watch Lohfink’s reaction from the sidelines rather than sharing her disappointment over the verdict. The viewpoint is that of a camera angle filming Lohfink and Benecken and differs greatly from that presented in Example 3, where a politician created immediacy by turning Lohfink’s experience into one that affects at least other victims of rape if not society at large.

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

183

A distancing stance is to be found on several other occasions in Article 2, such as in the following example. Here, the ruling judge is quoted arguing his decision: Warum denn, fragt Fischer, habe die Angeklagte an allen Sitzungen im Amtsgericht teilgenommen, wenn das Verfahren sie so sehr belastet habe? Sie sei dazu nicht verpflichtet gewesen, weil es vor dem Amtsgericht um einen Strafbefehl gegangen sei. [Why then, Fischer (the ruling appeal court judge) asks, had the accused taken part in all court sessions if the trial had burdened her so much? She had not been obliged to do so because at the District Court it had been about an order of summary punishment (opposed to an accusation script which binds the accused party to appear in court).] Example 12 from Article 2 The judge’s words are presented in indirect speech and contain three different process types, all present Lohfink in a subject position (sie/she): teilgenommen (take part), belastet habe (burdened) and sei nicht verpflichtet gewesen (had not been obliged to do so). The first is material action intention, the second mental reaction and the third relational intensive in combination with deontic modality and negation. In following the judge’s argument that Lohfink was not obliged to attend in person but did so despite the alleged effect it had on her, an ulterior motive for Lohfink’s personal appearances in court is implied. The argument, that Lohfink’s purpose was to enhance her media presence and thus attract lucrative contracts, is put forward by means of transitivity choice: what Lohfink did, the mental effect it had on her, and what she hypothetically could have done instead. Distance is achieved by presenting the judge’s verbiage and thus his viewpoint instead of Lohfink’s or her lawyer’s. The scene is watched through the eyes of a legal authority other than in Article 1 where quotes from politicians dominate. 3.3 Presenting others’ speech and thought Although I have talked about speech (but not thought) presentation in connection with naming strategies and transitivity patterns on several occasions throughout this chapter, I wish to emphasize its importance for Lohfink’s construction by dedicating a separate section to it, starting with some facts. In Article 1, thirty-one out of sixty sentences include a quote, fourteen of which are a presentation in direct speech. In Article 2, in contrast, we find speech presentation in forty-eight out of sixty-eight sentences and thought presentation in three more. In nine sentences the verbiage is presented verbatim

184

CONTEMPORARY MEDIA STYLISTICS

(direct speech); in five more sentences the quote is an amalgam of direct and indirect speech. The person most quoted in Article 2 is the ruling appeal court judge Fischer (in thirty-two sentences) whereas in Article 1, politicians are quoted in twenty-five sentences. These figures underline the newspaper’s changing approach to the Lohfink trial, where a clear preference for the political agenda of the trial has given way in favour of what the court said. Instances of thought presentation, namely Lohfink’s alleged thoughts on the day of the appeal court’s ruling, also fit into the already stated focus on inner world in Article 2: Ob die 30-jährige schon ahnt, dass man heute kurzen Prozess mit ihr macht? Dass ihr Anwalt so zerrupft und zerzaust dastehen wird, wie es Verteidigern wohl nur sehr selten passiert? [Does the 30-year-old already sense that one makes short work of her today? That her lawyer will be torn to shreds and look rumpled as it very seldom happens to defence lawyers?] Example 13 from Article 2 Es sind Sätze, die Lohfink und Benecken wie Ohrfeigen schmerzen müssen. [These are sentences that must hurt Lohfink and Benecken like slaps.] Example 14 from Article 2 In German, the idiom ‘kurzen Prozess machen’ (to make short work of somebody/something) is best literally translated with ‘to make short trial with somebody’. The time aspect relates to the duration of the trial before the District Court (six days) versus less than one hour at the appeal court. Beside its literal meaning the use of this expression brings in a metaphorical dimension (Lakoff and Johnson, 2003), namely TIME IS VALUABLE GOODS. The judge, in keeping the procedure short, shows that the legal facts on which to base his decision are straightforward, not least because of the type of appeal Lohfink chose. The anticipation of the outcome of the appeal, presented as Lohfink’s thoughts, was written with hindsight, of course, but constructs doubts in Lohfink’s mind about her chance of success (in contrast to constructed doubts about the prosecution office’s decision, see Examples 1 and 7). This is achieved by the sentence type (question) as well as the temporal adverb schon (already). Such doubts before the appeal court’s ruling contrast Lohfink’s construction in Article 1 where doubts about her being an actual rape victim are minimized. In Example 14, Lohfink and her lawyer, in listening to the judge, are constructed as if suffering pain. Here, a simile between ‘sentences’ and ‘slaps’ is

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

185

used to construct the court scene in a metaphorical sense, namely A COURT TRAIL IS A FIGHT. The defeated party, that is Lohfink and Benecken, are ‘slapped’ by the judge, which constructs the criminal justice system as prevailing party when in fact justice was being served. This together with the large number of sentences quoting the judge provide a stark contrast to Article 1 where legal authorities were constructed as supposedly getting the facts all wrong. The following Example from Article 1 underlines this contrastive construction: ‘(1)Der Umgang mit Gina-Lisa Lohfink ist erschreckend. (2)Ein Opfer wird zur Täterin gemacht, öffentlich bloßgestellt, es wird ihr nicht geglaubt. (3) Das nimmt anderen Frauen den Mut, eine Vergewaltigung anzuzeigen’, sagt die Vizefraktionschefin der Grünen, Katja Dörner, SPIEGEL ONLINE. [(1)’The social dealings with Gina-Lisa Lohfink are alarming. (2)A victim is turned into an offender, publicly disposed, she is not being believed. (3)This disheartens other women to report a rape’, says vice parliamentary party leader of the Green Party Katja Dörner to SPIEGEL ONLINE.] Example 15 from Article 1 (numbers added for ease of reference) Presenting Dörner’s words as direct speech indicates the magazine’s earlier stance on the matter and is symptomatic for the public perception of the trial against Lohfink back then. Spiegel Online in quoting Dörner in direct speech allows not only for immediacy but also for her viewpoint to become the magazine’s, which van Krieken, Sanders and Hoeken (2016: 147, 150) call ‘viewpoint embedding’. As we have seen in relation to Example 12, viewpoint embedding is crucial as it allows for the magazine to back up their own viewpoint by choosing quotes which support it. In Example 12 it was the judge whereas here attention is given to a politician and thus a non-participant in the Lohfink trial. Lohfink being a victim is existentially presupposed by means of a change-ofstage verb (wird gemacht/is turned, see underlined segment) (Levinson, 1983: 181). Additionally, this phrase presents the original verbiage for verbreitete Lesart (common interpretation), as later quoted in Article 2 (see Example 4). Third, this phrase is part of a three-part list symbolizing completeness (Jeffries, 2010: 70). Enumeration is another textual–conceptual function of texts and contributes to Lohfink’s construction. Dörner in Example 15 lists three processes (turned, disposed and not being believed) Lohfink allegedly had to passively endure which are two material and one mental process (Simpson, 1993: 89). However, none of the Actors/Sensers to any of these processes are mentioned. This omission is highly manipulative as from context the reader is able to deduce that the Actors/Sensers Dörner is referring to are the prosecution office and court. This quote is thus a criticism of judiciary without Dörner

186

CONTEMPORARY MEDIA STYLISTICS

having to point a finger as the readers can do so themselves. Examples 13 through 15 thus vividly illustrate the difference between Lohfink’s construction in Articles 1 and 2 which can be witnessed across several linguistic devices like noun phrases, transitivity patterns, speech presentation, deixis, implicature and presupposition. It serves best to end this chapter with a look at oppositional meaning as another textual–conceptual function. 3.4 Opposition and negation The last example demonstrates embedded viewpoint but also takes the argument further. What we can see in the next example is how Spiegel Online under the disguise of explaining legal matters to their readers, in fact reverses (if arguably not even ridicules) them: (1)Der Fall ist politisch relevant, weil Bundesregierung und Bundestag gerade um eine Reform des deutschen Sexualstrafrechts ringen. … (2)Die Justiz ist in Deutschland unabhängig. (3)Allerdings ist es nicht ungewöhnlich, dass öffentliche Debatten politische Prozesse beeinflussen. [(1)The case is politically relevant because Federal Government and German Parliament struggle for a reform of the German law governing sexual offences. … (2)The German justice system is independent. (3)However, it is not unusual that public debates influence political trials.] Example 16 from Article 1 (numbers added for ease of reference) This extract not only describes the trial against Lohfink as being political by means of a subject – subject complement structure but explicitly mentions limitations to the principle of an independent justice system, namely, in case of political trials. In a democratic society where this independence is termed a constitutional principle, any interference should be viewed critically. Instead, it is presented as a commonly accepted fact. This is achieved in Example 16, among others, by a contradictory adverb allerdings (however) in combination with double negation (nicht – not; ungewöhnlich – unusual). Sentence 3 begins with a contrastive opposition, triggered by allerdings (however), and indicates an exception to the rule laid out in sentence 2 (Davies, 2013; Jeffries, 2010: 55). Further, a pragmatic presupposition is triggered in sentence 3 by the negating adverb nicht (not), namely, the addressee expects public debates not to influence court trials. In case they do, this is considered unusual. However, notice that two and not just one negation take scope over gewöhnlich (usual).

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

187

The second negation by means of a prefix (‘un-’) reduces the aforementioned expectation to absurdity by cancelling out the first negation (Nahajec, 2012: 170). If something is not unusual, it is usual and thus common. According to sentence 3, it is thus common for political debates to influence political court trials. This stands in diametrical opposition to the rule laid out in sentence 2, which in context with sentence 3 states that the German justice system is only independent in case of non-political trials. Note that double negation here has the effect that sentences 2 and 3 do not provide a rule and its exception but in fact two equally valid rules – one applies to non-political and the other to political trials. Double negation allows for hedging such disturbing facts and making them sound more nuanced. Further, sentence 1 does not appear next to sentences 2 and 3 in Article 1; they are in fact separated by several paragraphs. Nevertheless, they are linked to each other by means of lexical coherence (Fall, politisch, politische Prozesse – case, politically, political trials). In sentence 1, it is established that the trial against Lohfink is ‘politisch relevant’ (politically relevant) in a subject – subject complement sentence structure. Sentence 3 goes beyond mere relevance by naming the trial against Lohfink a ‘political trial’. There is a change of syntax as the adjective ‘political’ moves from a subject complement to a subject position by being integrated as a pre-modifier to the head noun ‘trials’ in the noun phrase. The trial against Lohfink is assigned to a group of ‘political trials’ by means of an existential presupposition. This illustrates what Jeffries (2010: 19) refers to as ‘packaging up’ ideas: it is less likely to be questioned whether the Lohfink trial is actually a political trial (e.g. over politically motivated crimes like hate crimes) or whether the Lohfink trial is merely politicized by politicians and the media.

4 CONCLUSION This chapter has shown how Lohfink has been framed as a victim of rape during her first trial on charges of false rape allegations and how this happened linguistically in news reporting about her case. Spiegel Online, probably in an attempt to support politicians in their efforts to bring about a change in the governing law concerning sexual crimes, apparently uncritically copied Lohfink’s narrative which went against facts established by the prosecution office, violated constitutional principles and contradicted journalistic standards (Pressekodex). Lohfink strived to be portrayed as the ‘good girl’ in a dichotomous construction of good and bad girls, the latter being blamed for their victimization (Meyers, 2016: 9). This aspired ‘goodness’, not least based on her declared ‘fight for all women’ [‘Ich kämpfe für alle Frauen’ (Bruckner, 2017)], indicates that Lohfink took notice of her trial being politicized and she herself contributed to this discourse.

188

CONTEMPORARY MEDIA STYLISTICS

Difficulties in defining her as a ‘good girl’ resulted in pushing boundaries of what might be seen as socially appropriate female behaviour (Meyers: 9, 63) and have thus almost succeeded in finally making the ‘impure victim’ socially acceptable had not her rape narrative been found untrue by both deciding judges. Thus, attempts to construct Lohfink’s story as a canonical rape victim story and a ‘morality tale’ (p. 23) resulted in an effort to put things right (Article 2). Lohfink’s construction in Article 2 reverses Mendelsohn’s victim typology where a victim starts as a perpetrator and ends as victim. In Lohfink’s case it is exactly the other way round as arguably she never was a victim of rape but only was successfully constructed as such (Mendelsohn, 1963; Walklate, 2007: 31).

NOTE 1. This film was synchronized and screened in Germany under the title ‘Angeklagt’. A connection between Lohfink and the film was made by Sanyal, Mithu M. (2016). Vergewaltigung: Aspekte eines Verbrechens. 2. Auflage. Hamburg, Edition Nautilus GmbH, p. 163.

REFERENCES Benecken, B. (2018). Stars zwischen Macht und Ohnmacht: Wie das Promi-Phänomen uns alle beeinflusst. Berlin: Goldegg Verlag. Benecken, B. and Wöhrle, C. (2016). Tatort Unterwelt: Ein Strafverteidiger gibt unzensierte Einblicke in kriminelle Parallelgesellschaften. München: riva Verlag. Brison, S. (2002). Aftermath: Violence and the Remaking of the Self. Princeton, NJ: Princeton University Press. Brownmiller, S. (1975). Against Our Will: Men, Women and Rape. New York: Somon and Schuster. Bruckner, J. (2017). Gina-Lisa darf machen, was sie will. Available: https://www.sue ddeutsche.de/medien/rtl-dschungelcamp-gina-lisa-darf-machen-was-sie-will-1.33314 54. Süddeutsche Zeitung. Christie, N. (1986). ‘The ideal victim’. In E. A. Fattah (Ed.), From Crime Policy to Victim Policy (pp. 17–30). Basingstoke: The Macmillan Press. Cohen, S. (1980). Folk Devils and Moral Panics: The Creation of the Mods and Rockers. New York: St. Martin’s Press. Cohen, S. (2002). Folk Devils and Moral Panics: The Creation of the Mods and Rockers (3rd edn). London: Routledge Taylor & Francis Group. Davies, M. (2013). Oppositions and Ideology in News Discourse. London: Bloomsbury. Gregoriou, C. (2011). Language, Ideology and Identity in Serial Killer Narratives. London: Routledge. Halliday, M. A. K. (1985). An Introduction to Functional Grammar. London: Edward Arnold. Jeffries, L. (2010). Critical Stylistics: The Power of English. Basingstoke: Palgrave Macmillan.

CHANGING MEDIA REPRESENTATION OF GINA-LISA LOHFINK

189

Jeffries, L. (2014a). ‘Critical stylistics’. In M. Burke (Ed.), The Routledge Handbook of Stylistics (pp. 408–20). Milton Park and New York: Routledge. Jeffries, L. (2014b). ‘Interpretation’. In P. Stockwell and S. Whiteley (Eds), The Handbook of Stylistics (pp. 469–86). Cambridge: Cambridge University Press. Jeffries, L. (2015a). ‘Critical Stylistics’. In V. Sotirova (Ed.), A Companion to Stylistics (pp. 157–76). London and New York: Bloomsbury. Jeffries, L. (2015b). ‘Language and ideology’. In L. Cummings and N. Braber (Eds), Introducing Language and Linguistics (pp. 379–405). Cambridge: Cambridge University Press. Kaplan, J. (Writer). (1988). The accused: Paramount Pictures. Krieken, K. v., Sanders, J. and Hoeken, H. (2016). ‘Blended viewpoints, mediated witnesses: A cognitive linguistic approach to news narratives’. In B. Dancygier, W.-l. Lu and A. Verhagen (Eds), Viewpoint of the Fabric of Meaning (pp. 145–68). Berlin and Boston: Walter de Gruyter. Lakoff, G. (2004). Don’t Think of an Elephant! Know Your Values and Frame the Debate. White River Junction: Chelsea Green Publishing Company. Lakoff, G. and Johnson, M. (2003). Metaphors We Live By. Chicago: The University of Chicago Press. Leeuwen, T. v. (2018). ‘Moral evaluation in critical discourse analysis’. Critical Discourse Studies, 15(2), 1–14. Levinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press. Meiritz, A. (2016). Reform des Sexualstrafrechts. Schwesig schaltt sich in Fall GinaLisa Lohfink ein. Retrieved from http://www.spiegel.de/politik/deutschland/ginalisa-lohfink-gruene-fordern-schnelle-reform-des-sexualstrafrechts-a-1096892.html Mendelsohn, B. (1963). ‘The origin of the doctrine of victimology’. Excerpta Criminologica, 3, 239–45. Meyers, D. T. (2016). Victims’ Stories and the Advancement of Human Rights. New York: Oxford University Press. Meyers, M. (1997). News Coverage of Violence against Women: Engendering Blame. Thousand Oaks, London, and New Delhi: Sage. Moore, S. E. H. (2014). Crime and the Media. Basingstoke: Palgrave Macmillan. Nahajec, L. (2009). ‘Negation and the creation of implicit meaning in poetry’. Language and Literature, 18(2), 109–27. Nahajec, L. (2012). ‘Evoking the possibility of presence: Textual and ideological effects of linguistic negation in written discourse’. PhD Doctoral thesis, University of Huddersfield, Huddersfield. O’Malley, P. (2010). Crime and Risk. London: Sage. Poletta, F. (2006). It Was Like a Fever: Storytelling in Protest and Politics. Chicago and London: The University of Chicago Press. Rowbotham, J., Stevenson, K. and Pegg, S. (2013). Crime News in Modern Britain: Press Reporting and Responsibility, 1820–2010. Basingstoke: Palgrave Macmillan. Rückert, S. (2017). Sexualstrafrecht: Keine Strafe für Vergewaltiger? Retrieved from https://www.zeit.de/2017/18/strafrecht-vergewaltigung-strafe-statistik Sanyal, M. M. (2016). Vergewaltigung: Aspekte eines Verbrechens. Hamburg: Edition Nautilus GmbH. Semino, E. and Short, M. (2004). Corpus Stylistics: Speech, Writing and Thought Presentation in a Corpus of English Writing. London: Routledge. Siemens, A. (2016). Interview mit Gina-Lisa Lohfink ‘Muss ich erst umgebracht werden?’. Available: https://www.spiegel.de/panorama/justiz/fall-gina-lisa-lohfin k-muss-ich-erst-umgebracht-werden-a-1097049-druck.html

190

CONTEMPORARY MEDIA STYLISTICS

Siemens, A. (2017). Gescheiterte Revision im Fall Lohfink. Ein Richter teilt aus. Retrieved from http://www.spiegel.de/panorama/justiz/gescheiterte-revision-imfall-lohfink-ein-richter-teilt-aus-a-1134069.html Simpson, P. (1993). Language, Ideology and Point of View. London: Routledge. Smechowski, E. and Gernert, J. (2013). Die Gina-Lisa-Story. Retrieved from retrieved through ProQuest: https://search.proquest.com/docview/1268731501?accounti d=11526 Soothill, K. and Walby, S. (1991). Sex Crime in the News. London and New York: Routledge. SpiegelOnline. (2016). Verschärfung des Sexualstrafrechts. Maas und Oppermann erhöhen Druck auf Union. Retrieved from http://www.spiegel.de/politik/deutsch land/sexualstrafrecht-sozialdemokraten-erhoehen-druck-auf-union-a-1097133.html Tabbert, U. (2016). Language and Crime: Constructing Offenders and Victims in Newspaper Reports. London: Palgrave Macmillan. Van Dijk, J. (2009). ‘Free the victim: A critique of the western conception of victimhood’. International Review of Victimology, 16, 1–33. Van Wijk, J. (2013). ‘Who is the “little old lady” of international crimes? Nils Christie’s concept of the ideal victim reinterpreted’. International Review of Victimology, 19(2), 159–79. Walklate, S. (2007). Imagining the Victim of Crime. Maidenhead: Open University Press. Walklate, S., Maher, J., McCulloch, J., Fitz-Gibbon, K. and Beavis, K. (2018). ‘Victim stories and victim policy: Is there a case for a narrative victimology?’ Crime Media Culture, 15(2): 199–215. Wodak, R. and Meyer, M. (Eds), (2009). Methods of Critical Discourse Analysis (2nd edn). London: Sage.

CHAPTER NINE

Child victims of human trafficking and modern slavery in British newspapers ILSE A. RAS

The portrayal of crime, and those affected by, it shapes public perceptions and policy; the representation of human trafficking is in this regard no different (Small, 2012; Wilson and O’Brien, 2016; Wylie, 2016). Problematically, the (mis)representation of victims of this crime creates a victim hierarchy, resulting in non-ideal, but nonetheless real, victims being denied essential care and support (Wilson and O’Brien, 2016). This chapter focuses on how child victims of human trafficking have been presented by British newspapers since 2000, focusing on the stereotype of child victims of trafficking being girls and young women enslaved in the sex industry, and the problematic consequences of the perpetuation of these stereotypes. Using a critical stylistic (Jeffries, 2014) approach and drawing on corpus linguistics to facilitate the examination of a 62.5-million-word corpus of British news articles on the topic of human trafficking, I analyse nouns and noun phrases identifying children and young people (e.g. child*, boy*). Victims of human trafficking are often metaphorically described as COMMODITIES (Gregoriou and Ras, 2018a) and my analysis shows that not only

192

CONTEMPORARY MEDIA STYLISTICS

are child victims associated with commodifying processes, but the experiences of these children also become commodities, to be used in the aim of selling newspapers. My findings show that the stories of child victims of human trafficking and modern slavery are used to legitimate, among other things, governmental responses to migration-related matters. The data also show, however, some progress towards a more inclusive notion of who can be a victim of sexual abuse and exploitation, suggesting that at least part of the communicated understanding of human trafficking/modern slavery is beginning to respond to criticisms of the (previously) gendered assumption of who can be the victim of sexual abuse.

1 CHAPTER OVERVIEW This chapter focuses on how child victims of human trafficking have been described by British newspapers since 2000, as these representations will affect which children receive support, and what type of support these children receive (see also Gregoriou and Ras, 2018b; Hua and Nigorizawa, 2010). It answers the questions of how these children are described, and which types of exploitation are focused on. It is important to examine newspaper representations of societal issues, as newspapers continue to be a force to be reckoned with in influencing and reinforcing widely held opinions (Bednarek and Caple, 2012), and in legitimating certain responses over others. This is not necessarily, or at all, intentional; newspapers, and other media, are simply responding to (commercial) pressures, and journalists will have a pre-existing understanding of the issue on which they are reporting.

2 HUMAN TRAFFICKING AND MODERN SLAVERY Human trafficking and modern slavery are legally defined in the UK through the Modern Slavery Act 2015 (MSA 2015), as an umbrella term covering a variety of exploitative practices, including forced and bonded labour and domestic servitude (Schedule 1). The MSA 2015 also adopts the definition of human trafficking of the 2000 Protocol to the UN Convention against Transnational Organised Crime (‘Palermo Protocol’, UNODC, 2016), as the movement of people for the purpose of exploitation. There is a disagreement between the Palermo Protocol and the MSA 2015 on the question of consent, however, as the Palermo Protocol indicates that the consent of adult trafficked persons must explicitly have been violated, whereas the MSA 2015 notes that trafficking is movement for exploitation regardless of consent (to travel), as it is exploitation that is the problem. Presumably, the irregular movement of a person for purposes other than exploitation are different crimes; if their consent is given, this movement is illegal migration when transnational, and if

CHILD VICTIMS OF HUMAN TRAFFICKING

193

their consent is not given, this movement is kidnapping regardless of whether borders are crossed. This question, of whether a lack of consent should be a necessary prerequisite for human trafficking, has been debated internationally for over a century through the 1904 and 1910 International Agreements for the Suppression of the ‘White Slave Traffic’; the 1921 International Convention for the Suppression of the Traffic in Women and Children, the 1933 International Convention for the Suppression of the Trafficking of Women of Full Age, and the 1949 UN Convention for the Suppression of the Traffic in Persons and of the Exploitation of the Prostitution of Others. However, what all these international agreements and conventions have in common is a stipulation that even if the explicit lack of consent is a prerequisite for a person to be recognized as a victim of trafficking, this is only a prerequisite in the case of the trafficking of adults. The movement of children for exploitative purposes has never required an explicit violation of consent, given the legal presumption that children are unable to give informed consent. There are several types of exploitation outlined in the MSA 2015 (Schedule 1), which can be categorized as sexual, labour, domestic exploitation and organ donation; the latter is forbidden to adults and children alike under the Human Tissue Act 2004. Article 4 of the European Convention on Human Rights defines labour exploitation through the exclusion of those types of labour that are not exploitative: ‘work required to be done in the ordinary course of detention’, ‘service of a military character’, ‘service exacted in case of an emergency or calamity’ and ‘any work or service which forms part of normal civic obligations’. It is this latter exclusion which is problematic: what distinguishes ‘normal’ labour from ‘exploitative’ labour? The clearest available distinction is provided through the 1926 Slavery Convention and the 1930 Forced Labour Convention, which indicate that forced labour, that is, labour extracted against a person’s consent, can develop into conditions ‘analogous to slavery’ (Slavery Convention 1926). Clearer are the boundaries between ‘normal’ labour and ‘problematic’ labour: labour legislation is generally rather clear about the legal and illegal conditions of a job, and children are, in this case, more protected than adults, with age points to guide the transition from child to adult. In the UK, children are allowed to work part-time from the age of thirteen (with exceptions made for the entertainment industry), and full-time from age sixteen. Children are also legally protected by law to only work between 7.00 am and 7.00 pm, with limited hours and substantial breaks, and are prohibited from performing work in factories, industrial sites, pubs, betting shops, and other ‘work that may be harmful to their health, well-being or education’. This suggests that there is a continuum of ‘child-ness’, whereby adolescents over a certain age, often sixteen, are considered capable of taking on at least a few adult responsibilities and are less protected by law from certain types of work and working hours. The same laws protect children in domestic service.

194

CONTEMPORARY MEDIA STYLISTICS

As the age of sexual consent also lies at sixteen, children below that age also cannot consent to sex work. Over the age of sixteen, sex work is legal, but the prostituting of another person is not and can thus be considered exploitative also.

3 POWER OF LANGUAGE Chibnall (1977) notes that newspapers reflect and can change the opinions held by their readerships, thereby (re)establishing norms and values. Indeed, Sutherland (1955) highlights the importance of communication in the legitimation of behaviours and attitudes. It must be noted here that Sutherland (1955) focused on delinquency, and hypothesized that through communication, a potential delinquent learns (a) the means of committing a crime and (b) the attitude that justifies or neutralizes this crime. Sykes and Matza (1957) examine the communication used by delinquents and note that this communication tends to provide arguments that reduce or ‘neutralise’ the delinquent acts. Shoenberger, Heckert and Heckert (2012) further point out that these ‘techniques of neutralisation’ can also be applied to ‘positive’ behaviour. I here use the phrase ‘legitimation’, as ‘neutralisation’ implies a moral direction (from ‘negative’ or ‘positive’ to ‘neutral’), whereas I focus on the use of language to make a particular act be an ‘accepted’, ‘expected’ or ‘correct’ response, among perhaps a range of accepted and expected behaviours to a particular situation. Consider, for instance, the range of possible, versus ‘legitimate’, responses of female survivors to sexual assault, which in turn also have ‘legitimate’ responses. As Dewan (2017) points out, a ‘victim [who] doesn’t act like one’, that is, who does not respond to their assault in a manner that has become accepted and expected of survivors of sexual assault, is less likely to be accepted as a survivor of sexual assault. The understanding of what constitutes ‘legitimate’ must continuously be reinforced, for instance, against the threat of competing understandings of what is, or is not, legitimate (in the sense of ‘a person can be a survivor of sexual assault if they did not fight back’ vs. ‘a person was not sexually assaulted if they did not fight back’). Reinforcing the legitimacy of a particular behaviour may also pre-empt anticipated criticisms and confusion (Van Dijk, 1998), and serve to ‘fine-tune’ legitimated behaviours and attitudes. Sutherland (in Cressey and Sutherland, 1955) highlights intimate communication as the main avenue through which responses are legitimated. However, Bandura (1990), as well as Van Dijk (1998), note that print media and television play an equal, if not greater, role in communicating expected responses. Despite the spread of new and social media, newspapers and other traditional media remain important in creating and maintaining legitimations. Bednarek and Caple (2012), for instance, write that traditional media are

CHILD VICTIMS OF HUMAN TRAFFICKING

195

extremely powerful in influencing governments and other major institutions, and shaping the ideas and behaviours of the wider public. They do so through the systematic translating, ‘distorting’ (Chermak, 1994) or framing of events (Louw, 2005; Hartley, 1982; Kuhn, 2007), by foregrounding particular aspects of the event and obscuring others, or indeed by choosing to report some events over others. Chermak (1994), for example, shows that 10.6 per cent of all news in his study of British newspapers is crime-related news. However, if all events that happened at any given moment could be quantified, it would be unlikely that a full 10 per cent of them would be crime related. Similarly, minor and volume crimes are less likely to be reported than unusual and serious crimes (Howitt, 1998; Gray, 2009). This creates the perception that crime is more prevalent and more serious than it is, which affects people’s perception of the safety of an area and thereby influences their behaviour, but can also affect the prioritization of crime over other social issues. The legitimation of behaviours and attitudes, as a consequence of newspaper reporting, is not always intentional. It may simply be an effect related to more pressing journalistic concerns, such as writing a story that will be printed and sell newspapers, especially as traditional media are under ever-increasing (commercial) pressure. The journalistic ethic to present a story as objectively as possible can also lead to a preference to including the perspective of authorities, who are often part of powerful and privileged groups (Kuhn, 2007; Cottle, 2003; Machin and Niblock, 2006), thereby continuing the legitimation of the perspectives of these privileged groups. Views opposed to prevalent norms may, similarly, be de-legitimated (Louw, 2005; Cottle, 2003; Fowler, 1991), for instance through marginalization, criminalization and demonization.

4 REPRESENTATIONS OF HUMAN TRAFFICKING AND MODERN SLAVERY Gregoriou and Ras (2018a, 2018b) outline the representations of human trafficking and its effects in general and analyse specifically how British newspapers have portrayed human trafficking since 2000. The type of exploitation that tends to be foregrounded in fiction and news media is sexual (Alvarez and Alessi, 2012; Marchionni, 2012; Papadouka, Evangelopoulos and Ignatow, 2016), despite the reality that labour trafficking is much more prevalent (O’Brien, 2016: 210; Feingold, 2005). Stereotypical victims are female, young and coerced (Gregoriou and Ras, 2018b; Andrijasevic and Mai, 2016; Wilson and O’Brien, 2016). The problem is that male victims and victims of other types of exploitation, such as labour, are more likely to be overlooked as victims (Sharma, 2005; Duong, 2014; Lobasz, 2009). Through a gendered perception of who can be a victim and give consent, a gendered interpretation of consent and victimhood continues to be legitimated, and through focusing

196

CONTEMPORARY MEDIA STYLISTICS

on sexual exploitation rather than labour exploitation, questions about what constitutes abusive and exploitative labour practices are avoided. The structural factors underlying trafficking, such as poverty, inequality, conflict and the global demand for cheap labour and sex, are often ignored (Sharma, 2005). However, these factors increase a person’s vulnerability, leaving them more likely to be trafficked. At this point, the question of consent becomes a difficult one: even if a person has consented to their movement and exploitation, but did not have much of a choice due to their circumstances, is this consent at all valid? On the other hand, the assumption that a person cannot have given consent due to their circumstances denies their agency and is similarly problematic. Concerns about human trafficking have been linked to various other issues over the years, such as women’s rights (Farrell and Fahy, 2009) and, more recently, policing and border control (Marchionni, 2012). Gregoriou and Ras (2018a) show that British newspapers use metaphors that characterize human trafficking as a SPREADING, UNWANTED SUBSTANCE, that can be BROKEN through ‘crackdowns’, ‘breaking up’, ‘smashing’, ‘crushing’ and ‘stamping out’, with the legitimate response to trafficking being like a WAR, as evidenced through the use of fight-related verbs, including ‘tackle’, ‘combat’, ‘target’, ‘attack’ and ‘spearhead’. The effects of these frames on the victims of human trafficking are not negligible. By shifting the frame to one of policing and border control, attention is also shifted to identifying and capturing perpetrators, away from a potential focus on the needs of victims and the structural factors that underlie trafficking and the (gendered perception of) often blurry consent of those being moved and exploited. In fact, victims of trafficking or of the structural factors that facilitate and encourage trafficking are often prosecuted as illegal migrants instead (Coghlan and Wylie, 2011; Cunningham and DeMarni Cromer, 2016; Farrell and Fahy, 2009). The police raid, a preferred method of response as a result of the policing frame, can be incredibly traumatizing (Hill, 2016), while the use of these trafficked people as witnesses in the process of prosecuting perpetrators, legitimated also as a ‘correct’ response through the policing frame, can have similarly traumatizing effects on these witnesses (Meshkovska et al., 2016; Wijers, 2015). In many ways, victims are reduced to a measure of their ‘usefulness’ in the fight against trafficking and slavery.

5 METHODOLOGY AND METHODS: CRITICAL STYLISTICS AND CRITICAL DISCOURSE ANALYSIS To explore whether the reporting of child victims particularly differs from the reporting of victims of human trafficking (as examined in Gregoriou and Ras, 2018a), in general, I use the critical stylistic approach advocated by Jeffries

CHILD VICTIMS OF HUMAN TRAFFICKING

197

(2007, 2010, 2014). As I examine this reporting in the corpus of national British newspaper reporting on human trafficking and modern slavery that was collected by Gregoriou and Ras (2018a), I make use of corpus linguistic methods to facilitate this critical stylistic analysis. Critical stylistics analyses literary as well as non-literary texts (Jeffries, 2014: 408). It evolved from critical linguistics, which assumes that language shapes and is shaped by social structures (Fairclough, 2015; Wodak, 2001). The goal of critical linguistics is to ‘demystif[y]’ authors’ and speakers’ aims (Fowler and Kress, 1979; Jeffries, 2014). Like critical linguistics, critical stylistics draws on Halliday’s metafunctions of language and his systemic-functional linguistics (SFL). Halliday’s (1973) metafunctions are three: the ideational, or the use of language to communicate experiences and views; the interpersonal, which negotiates the relations between language users, and the textual, which is the use of language to create a coherent and cohesive text. Jeffries (2014) specifically focuses on the ideational, noting that it is not the language itself but its specific use within a specific context that has a specific (potentially problematic) meaning. Critical stylistics is, as Jeffries (2014: 408) indicates, a response to developments in critical discourse analysis (CDA), by returning the text to its central place in the analysis and to move away from the explicitly (Marxist) political nature of critical discourse analysis. CDA was explicitly conceived as a method of (Marxist and/or left-wing) academic activism (Fairclough, 2015: 52), although Poole (2010) shows that the approach of CDA can also be used to examine texts from non-Marxist and/or non-left-wing perspectives. Critical stylistics steps away from the movement-like aspects of CDA, although not abandoning them entirely: ‘I would offer critical stylistics as a method of finding the ideology in any text, whether or not you agree with it. The rest is personal choice’ (Jeffries, 2014: 410). Finnemore and Sikkink (1998: 892) suggest that norms are never considered bad by the person or institution holding them, which means that it is important to use a method of exploring the discursive strategies that negotiate and influence our understanding of issues and how we see them, that does not, from the outset, mark a specific author’s or speaker’s aim as ‘bad’. As such, critical stylistics should, be the preferred approach. The explicit politics of CDA have, over time, invited accusations of cherrypicking, with the most-suggested remedy to the problems with CDA, to use corpus linguistic techniques and software to conduct much of the analysis (Widdowson, 2004; Jeffries, 2010; 2014; Baker et al., 2008). Baker et al. (2008) coined the label ‘corpus-assisted critical discourse analysis’ (CACDA) for this particular strand of CDA, thereby also sidestepping the question of Tognini-Bonelli’s (2001) distinction between corpus-based and corpusdriven techniques, as CACDA may use either (or both) over the course of

198

CONTEMPORARY MEDIA STYLISTICS

the analysis. Baker et al.’s (2008) study of the representations of refugees, asylum seekers, and (im)migrants is a good example of the power of corpus linguistics as a tool in a CDA, as they explore UK newspaper reporting on these people over a decade. In doing so, they avoid the issue of cherrypicking: they do not, for instance, solely focus on reporting by a notoriously migrant-phobic newspaper (e.g. The Sun or the Daily Mail) nor do they focus on reporting on a contentious event. The methods they developed to collect this corpus and perform their analyses can also be applied to different corpora (e.g. Gregoriou and Ras, 2018a). Fairclough (2015) does, however, warn that corpus linguistics is not an analytical approach as such; it offers descriptive statistics of text features and can generate displays of how a word is used, but cannot make sense of these statistics and displays (see also McIntyre, 2012; 2015). However, in simplifying, accelerating and systematizing parts of the analytical process, corpus linguistics can certainly enhance the replicability of this process, and by offering the possibility of processing multiple texts rapidly, and offering up all instances of a text feature, it also protects (to an extent) against cherrypicking. The uptake of these methods in critical stylistics would prevent critical stylistics from encountering these same problems. Jeffries (2014) certainly advocates the use of corpus linguistics as part of a critical stylistic analysis, having used corpus linguistics in a paper (Jeffries and Walker, 2012) exploring sociopolitical keywords in UK broadsheet reporting between 1998 and 2007. The corpus used in this study was originally collected by Gregoriou and Ras (2018a). We did so using the method set out by Gabrielatos (2007) of first selecting a set of core search terms, then mathematically evaluating the relevance of potential additional search terms. These search terms were entered into Lexis Nexis, with search results limited to UK national newspapers and a publication date between 1 January 2000 and 30 September 2016. For this specific study, the corpus was extended to 31 December 2016. Table 9.1 details how many articles, types and tokens there are in each year of the corpus. The next step in Gregoriou and Ras (2018a) was to identify pertinent actors and acts in this corpus. We used the British National Corpus (Burnard, 2007) as a reference corpus to calculate which words were sufficiently key, that is, that occurred more often in the primary corpus than the reference corpus over a certain log-likelihood (LL) threshold. This LL threshold was set to 15.13 (Gregoriou and Ras, 2018a), which means that the statistical significance of the frequency of these nouns, compared with the reference corpus, is p