Analysing Digital Interaction [1st Edition] 3030649210, 9783030649210, 9783030649227

This book investigates interaction-focused scholarship on online communication. It focuses on a broad range of online co

227 59 4MB

English Pages 229 Year 2021

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Contents......Page 6
Notes on Contributors......Page 8
List of Figures......Page 12
List of Images......Page 13
List of Tables......Page 14
1: Introduction: The Microanalysis of Digital Interaction......Page 15
Digital Interaction......Page 20
The Microanalysis of Interaction......Page 22
Development of ‘Digital CA’......Page 25
Overview of the Book......Page 29
References......Page 31
2: Ethical Challenges in Collecting and Analysing Online Interactions......Page 36
Public/Private Distinction......Page 38
Consent......Page 41
Anonymity and Searchability......Page 44
Visibility: Going Behind the Screen......Page 48
Conclusion......Page 50
References......Page 51
Introduction......Page 54
Digital Conversation and Context......Page 55
The Long Reach of the Digital Archive......Page 58
Case Study: Defending the Danny Baker Persona......Page 60
Unique Affordances of Twitter and Their Methodological Implications......Page 61
Racism Denials and Apologies......Page 63
Reactions to the First ‘Apology’......Page 66
Two ‘Conversations’ Triggered by Replies......Page 69
Conclusion......Page 72
References......Page 74
Introduction......Page 77
Identity and Transition......Page 79
Sampling Expatriate Blogs......Page 81
Analytic Framework: Category Fit and Category Change......Page 82
Local Categorial Negotiation......Page 85
Transcending the Local......Page 88
Discussion......Page 92
Conclusion......Page 95
References......Page 96
Introduction......Page 99
Methodology: Membership Categorization Analysis......Page 101
Data......Page 102
Person Identification and Characterization......Page 104
Linking Action Characterization and Perception......Page 107
Action Formulations and Establishment of Membership Categories......Page 109
Personal Experiences to Establish Identification and an Underlying Reaction Pattern Associated with ‘good people’......Page 111
Establishing Crying as an Unavoidable Reaction to Specific Stimulus......Page 112
Extreme Cases as Identification Devices......Page 113
Reaction Indications as Identification and Empathy Displays......Page 115
Discussion......Page 117
Conclusion......Page 118
References......Page 119
Introduction......Page 122
Closings in Spoken and Online Interaction......Page 124
Data Collection......Page 126
Data Analysis......Page 127
Analysis......Page 128
Directly Named Recipient Not Responding......Page 131
Idiomatic Endings......Page 134
Personal Comments and Meta Talk......Page 136
Discussion......Page 138
Conclusion......Page 140
References......Page 141
Introduction......Page 144
The Design of Message Extensions......Page 147
Interactional Implications of Messages that Look Incomplete......Page 151
Negotiating the End of the Instalment Series......Page 156
Conclusions......Page 159
Novelty of the Practice......Page 160
Peculiarities of Prospective Continuations in Mediated Settings......Page 162
References......Page 163
Introduction......Page 166
Data and Method......Page 170
Interweaving Adjacent Pairs in Multi-component ‘Turns’......Page 171
An Orientation Towards Providing Elaborate Answers......Page 174
The Open Question/Closed Question Device......Page 179
Accountability and Desertion......Page 180
Conclusion......Page 183
References......Page 185
Introduction......Page 188
Companions in Medical Consultations......Page 189
Medical Video-Consultations......Page 190
Participation Framework......Page 191
Data and Method......Page 192
Findings......Page 194
Companion as Bystander......Page 196
From Bystander to Ratified Participant......Page 197
Re-ratifying the Companion......Page 202
Navigating Companion Participation Statuses......Page 204
Discussion......Page 208
Conclusion......Page 210
References......Page 211
10: Conclusion: Future Directions in Analysing Digital Interaction......Page 215
References......Page 224
Name Index......Page 226
Subject Index......Page 228
Recommend Papers

Analysing Digital Interaction [1st Edition]
 3030649210, 9783030649210, 9783030649227

  • Commentary
  • TruePDF
  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

PALGRAVE STUDIES IN DISCURSIVE PSYCHOLOGY

Analysing Digital Interaction

Edited by Joanne Meredith · David Giles Wyke Stommel

Palgrave Studies in Discursive Psychology

Series Editors Cristian Tileaga˘ Social Sciences & Humanities Loughborough University Loughborough, UK Elizabeth Stokoe Department of Social Sciences Loughborough University Loughborough, UK

Palgrave Studies in Discursive Psychology publishes current research and theory in this established field of study. Discursive Psychology has, for the past 30 years, established an original and often critical understanding of the role of discourse practice for the study of psychological, social, and cultural issues. This book series will provide both introductions to discursive psychology for scholars new to the field, as well as more advanced original research for those who wish to understand discursive psychology in more depth. It is committed to the systematic representation of discursive psychology’s contemporary ethos into all things social – from everyday interactional encounters to institutional settings and the analysis of wider social issues and social problems. Palgrave Studies in Discursive Psychology will therefore publish ground-breaking contemporary contributions on the relevance of discursive psychology for key themes and debates across psychology and the social sciences: including communication, social influence, personal and social memory, emotions, prejudice, ideology, child development, health, gender, applied interventions, institutions. The series editors welcome contributions from ethnomethodology and conversation analysis, as well as contributions more closely aligned to post-structuralism, approaches to analysis combining attention to conversational detail with wider macro structures and cultural-­ historical contexts. We invite junior and senior scholars to submit proposals for monographs and edited volumes that address the significance of discursive psychology in psychology, communication, sociology, applied linguistics. Please contact the series editors (c.tileaga@lboro. ac.uk; [email protected]) or the commissioning editor ([email protected]) for more information. More information about this series at http://www.palgrave.com/gp/series/15475

Joanne Meredith  •  David Giles Wyke Stommel Editors

Analysing Digital Interaction

Editors Joanne Meredith Department of Psychology University of Wolverhampton Wolverhampton, UK

David Giles Department of Psychology University of Winchester Winchester, UK

Wyke Stommel Center for Language Studies Radboud University Nijmegen, Gelderland, The Netherlands

Palgrave Studies in Discursive Psychology ISBN 978-3-030-64921-0    ISBN 978-3-030-64922-7 (eBook) https://doi.org/10.1007/978-3-030-64922-7 © The Editor(s) (if applicable) and The Author(s), under exclusive licence to Springer Nature Switzerland AG 2021 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover credit: Panther Media GmbH / Alamy Stock Photo This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Contents

1 Introduction: The Microanalysis of Digital Interaction  1 Joanne Meredith, David Giles, and Wyke J. P. Stommel 2 Ethical Challenges in Collecting and Analysing Online Interactions 23 Hannah Ditchfield 3 Context, History, and Twitter Data: Some Methodological Reflections 41 David Giles 4 “It’s time to shift this blog a bit”: Categorial Negotiation as a Local and Cumulative Accomplishment 65 Linda Walz 5 The Radio Host Cried, the Facebook Users Identified: Crying as an Action Linked to ‘good people’ 87 Elisabeth Muth Andersen 6 “On that note I’m signing out”: Endings of Threads in Online Newspaper Comments111 Joanne Meredith v

vi Contents

7 Similarities and Differences Across Settings: The Case of Turn Continuations in Instant Messaging133 Anna Spagnolli, Sonia Genovese, and Mattia Mori 8 The Spectre of ‘Ghosting’ and the Sequential Organization of Post-match Tinder Chat Conversations155 Christian Licoppe 9 Participation of Companions in Video-­Mediated Medical Consultations: A Microanalysis177 Wyke J. P. Stommel and Martijn W. J. Stommel 10 Conclusion: Future Directions in Analysing Digital Interaction205 Janet Smithson Name Index217 Subject Index219

Notes on Contributors

David Giles  is Reader in Psychology at the University of Winchester. He

has promoted the use of qualitative methods in Psychology throughout the last two decades, co-founding the Taylor & Francis journal Qualitative Research in Psychology and the MOOD (Microanalysis of Online Data) international network of researchers developing and adapting methodologies for the study of digital data. He has published research in this tradition concerning online discussion forums, Twitter and YouTube, particularly in the field of media psychology where he has a particular interest in the dynamic interaction between celebrities and their audiences. Elisabeth  Muth  Andersen  is Associate Professor of Danish language and communication at the Department of Language and Communication, University of Southern Denmark, and is a member of the research Center for Social Practices and Cognition (SoPraCon). Mainly using ethnomethodological conversation analytic methodology, she has conducted research on online as well as face-to-face health communication practices. Approaching written and multimodally constructed mediated texts as interaction, her research includes work on reader constructions in online newspapers, membership categories used to do moral work in social media interaction, as well as constructions of viewer participation in a vii

viii 

Notes on Contributors

fictional net-series. Recent publications include papers in Journalism Studies and Discourse, Context & Media. Hannah  Ditchfield is a university teacher and researcher in the Department of Sociological Studies at the University of Sheffield. Her work focuses on digital media and society with specific interests in issues of online interaction, identity and online research ethics. Her research has drawn on methods of screen capture and conversation analysis and has recently been published in New Media & Society as well as the SAGE Handbook of Qualitative Data Collection. Christian  Licoppe is trained in history and sociology of science and technology, is interested in conversation analysis and multimodal interaction analysis, and more generally ethnographic studies of multi-participant interaction in mobile and institutional settings. He has developed an extensive research program on the use of mobile communication on the move, linking communication and mobility studies. In his research program, he has shown how mobile communication supports forms of “connected presence” and how location-aware mobile technologies and the way they promote fleeting encounters between “pseudonymous strangers”. It also led him to studies of the organization of message-based conversations in mobile dating, as a particular instance of such encounters, marked by a continuous tension between distance and intimacy. He has developed several research programs on video-mediated communication, showing how the “talking heads” configuration accounted for the observable order of video-mediated conversations and the ensuing importance of “camera actions” in video-mediated communication. Joanne Meredith  is Senior Lecturer in Psychology and a member of the CyberPsychology Research Unit at the University of Wolverhampton. Her research focuses on digital communication, particularly focusing on the similarities and differences between digital and offline communication. She is also interested in how political communication occurs in digital spaces, and is a founder of the Online Political Discourse Network which aims to advance research in this area. Her research uses interactional methods, including conversation analysis and discursive psychol-

  Notes on Contributors 

ix

ogy, to explore how interactions unfold in digital spaces. Her research articles have been published in journals such as Qualitative Research in Psychology, Journal of Pragmatics, Discourse and Communication and Public Understanding of Science. Janet Smithson  is a social psychology researcher and a senior lecturer in the School of Psychology at the University of Exeter, UK. My research focuses on discourse analysis of communication on online forums, blogs, social media and websites, especially concerning health, mental health and gender. Other research interests include cross-cultural and ­cross-­national research methodology, work-life practices and policies, life-­course transitions and mediation. Anna  Spagnolli  is an associate professor at the University of Padova, and a member of the Human Inspired Technology Research Centre at the same university, where she leads the Ethics Committee. She is also the teaching coordinator for the PhD course in Brain, Mind and Computer Science. She holds a master degree in Social Psychology and a PhD in Social and Personality Psychology, but her research interests and teaching activity are in the field of human-computer interaction and mediated conversation. She has participated so far in 13 EU-funded projects in the Information and Communication Technology (ICT) area. She was cofounder and editor-in-chief of PsychNology Journal, one of the first openaccess journals in the area of ICT, indexed by PsycInfo and SCOPUS. She is an associate editor of Springer’s journal Virtual Reality. Martijn W. J. Stommel  is a pancreas and hepatobiliary surgeon at the Radboud University Medical Center in Nijmegen, the Netherlands. His clinical work and research is focused on innovative surgical techniques and outcomes on the one hand, and on innovation in the delivery of care and implementation of eHealth on the other. He is involved in the development and implementation of smart remote care solutions, especially for patients whose care is centralized in centers of expertise. He has published over 60 peer reviewed research articles, both in surgical journals and in journals like The Lancet, Journal of Medical Internet Research and Journal of Computer-Mediated Communication.

x 

Notes on Contributors

Wyke  J.  P.  Stommel is Assistant Professor of Language and Communication at the Centre for Language Studies, Radboud University Nijmegen, and affiliated with the Radboud Interdisciplinary Hub for Security, Privacy and Data Governance. Her research interests are mainly in health interaction and digitally mediated interaction, centrally using conversation analytic methods. She is one of the initiators of the MOOD (Microanalysis of Online Data) network that aims to develop CA as a method for the analysis of “digital” interaction and has an interest in ethical issues in this area. Her research articles have been published in journals like Research on Language and Social Interaction, Social Science & Medicine, Discourse Studies and Discourse, Context & Media. Linda  Walz is Lecturer in Linguistics at Leeds Trinity University, UK.  Her research focuses on the analysis of social interaction and the linguistic construction of identity. She is particularly interested in personal experiences and narratives of transition, drawing on Discourse Analysis and Membership Categorisation Analysis. Her recent publications in Text & Talk (2020) and Discourse, Context & Media (2020, with Richard Fitzgerald) explore identity work in blogs about transnational relocation. She has co-edited the online proceedings of the 51st annual meeting of the British Association for Applied Linguistics.

List of Figures

Fig. 7.1

The Skype window of one participant in the group discussion, Cippa, magnified from the recording of his screen activity 147 Fig. 7.2 An extension of Vorreiter’s (2003) turn continuation taxonomy151 Fig. 9.1 Physician facing patient and companion (#2) 183 Fig. 9.2 Pointing and gazing, line 1, Extract 9.1 186 Fig. 9.3 Line 4, Extract 9.1 186 Fig. 9.4a Physician’s screen preceding Extract 9.2 188 Fig. 9.4b Line 1, Extract 9.2 188 Fig. 9.5 Line 1, Extract 9.3 191 Fig. 9.6 Gazes at companion, line 6, Extract 9.3 191 Fig. 9.7 Line 1, Extract 9.5 194 Fig. 9.8 Physician’s screen, Extract 9.6 195

xi

List of Images

Image 1.1 Original tweet Image 1.2 https://images.app.goo.gl/1vg1dcsiLaPgFacPA

2 4

xiii

List of Tables

Table 3.1 Table 3.2 Table 3.3 Table 3.4

Danny Baker’s tweets following accusations of racism Initial responses to first apology Conversation 1 following reply to Baker’s 7.08 tweet Conversation 2 following reply to Baker’s 7.08 tweet

51 54 57 59

xv

1 Introduction: The Microanalysis of Digital Interaction Joanne Meredith, David Giles, and Wyke J. P. Stommel

In January 2010, a man in the UK was looking forward to flying abroad to meet his girlfriend, whom he had met on Twitter. He woke up about a week before his flight to see that snow has closed the airport he was due to fly from. He went on Twitter and posted the following tweet (Image 1.1): The question arises—is this a joke or is it a credible threat against the airport? This question took on real-world significance when the author of the tweet was charged with sending a message which was ‘grossly offensive or of an indecent, obscene or menacing character’. If this same utterance “It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

J. Meredith (*) Department of Psychology, University of Wolverhampton, Wolverhampton, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_1

1

2 

J. Meredith et al.

Image 1.1  Original tweet

were to be said in a spoken interaction, the meaning could be understood based on the responses to the initial utterance. Yet on Twitter there was no immediate response. This poses challenges for understanding the meaning of the message, both for other users of Twitter and for analysts such as ourselves. How can we understand the meaning of a message when there are no responses? What contextual information could be used to determine whether this is a joke, when things like body language or tone of voice are not available? These are all questions which we seek to explore in this book. A related issue relevant to digital interaction is the question who counts as a participant in the interaction. Although participation is not necessarily self-evident in co-presence either (e.g. is a person who is reading available as a participant?), the negotiation of who counts as a participant is probably even more tangible in digital interaction and inevitably related to the technology or setting in which the interaction takes place. On social media like Twitter, information pours into users’ devices in a continuous stream that, on the surface, bears little resemblance to offline interaction. The table below displays the ‘conversation’ that appeared below a tweet posted by a British comic actor during a live announcement by the UK’s Prime Minister Boris Johnson during the 2020 Covid-19 pandemic. (Bear in mind that five days had elapsed since the tweet was initially posted, and the ‘chain’ of responses stretches to well over 200). The actor’s tweet, a derisory observation of the government’s presentation skills, is time-stamped 7.04, but the tweet that immediately follows (R1) was posted over an hour later. D. Giles Department of Psychology, University of Winchester, Winchester, UK e-mail: [email protected] W. J. P. Stommel Center for Language Studies, Radboud University, Nijmegen, The Netherlands e-mail: [email protected]

1  Introduction: The Microanalysis of Digital Interaction 

3

Extract 1.1  Twitter Thread OP

7.04

R1 R2

8.15 8.23

R3

7.14

R4

7.22

R5

8.21

We face a pandemic and we’re in the hands of people who couldn’t even format PowerPoint slides for their presentation. #Pressconference Yes! Needed. Slides that would be clear to a simple layman. I didn’t see the press conference but unsure if slides would help given that if Johnson’s lips are moving, he is not to be trusted on a single thing he says ( same for Gove). It is always going to get better in two weeks, in a month, sometime in the future. Absolutely no confidence in this man. I really hope my MP[Twitter handle] is calling the PM to account. I hope ALL MPs call him to account. He's obviously got Steve Baker on side.... Typical Leftie ! Ha ha get used to it 4 more years [Hugging face emoticon]

Seemingly, R1 and R2 have jumped the queue, because R3 came in at 7.14, just ten minutes after the original (OP). R5 is another interloper. This chain of tweets is far from unusual. It is typically what you find when retrospectively viewing responses to a popular Twitter user’s account (he has nearly half a million followers). The analyst’s job might be to somehow sift through all this material to try and reconstruct ‘what happened’—who tweeted what, when. But did anyone actually see the ‘conversation’ unfold this way? What are we actually attempting to ‘reconstruct’ by neatly stacking these fired-off snippets of communication into some chronological sequence? Furthermore, some of these tweets are responses to other tweets in the chain. Do we count these as separate ‘conversations’? The only thing we can be sure about is that the original poster has triggered the interaction, while the communication event will be different for each participant—and for each member of the silent audience following it from the outside. In video-interaction, who counts as a participant plays out differently. Visibility is crucial to whether or not a person is treated as a conversational participant. Hence, appearance on screen is a cue to participation (Licoppe, 2017b). Due to the enormous increase in video-interaction as a result of the Covid 19-pandemic, video-visibility caught the attention of media in cases where it became particularly spicy or embarrassing. Image 1.2 shows a man who appears on screen in underpants while his partner is in the middle of a videoconference. Her hand gesture treats the

4 

J. Meredith et al.

Image 1.2  https://images.app.goo.gl/1vg1dcsiLaPgFacPA

visual availability of the man as unwanted. Hence, on-screen appearance may unwillingly invoke participation. However, in other cases participants may fail to be recognised as a conversational partner as they are invisible on screen for the person on the other side. To further complicate the issue, participants may move off screen and/or cameras may be re-adjusted and thus participants’ status with regard to the interaction may shift (repeatedly) during a video call. Appearance on and off screen may seem a futility of video-interaction, but such apparent details of digital interaction may have real-life implications, for instance in institutional settings such as medicine. Companions who sit next to a patient in the consultation room, who are gazed at and talked to by the physician, may be off screen in the case of a video consultation and thus merely be a bystander. Chapter 9 addresses the question how companions of patients participate in oncological video consultations related to their appearance on screen. These examples demonstrate some of the issues which we seek to address in this book. The ideas in this book are a culmination of many

1  Introduction: The Microanalysis of Digital Interaction 

5

years of consideration of how best to analyse digital communication using interactional methods which have most commonly been used for spoken interaction. The initial germination of the ideas came from a meeting at the 25th anniversary conference of the Discourse and Rhetoric Group (DARG) at Loughborough University in 2012. It became clear at this meeting that there were a number of researchers who were engaged in analysing digital interaction using ‘microanalytic’ methods, and yet there still appeared to be some scepticism about the applicability of such methods to digital interaction. This initial meeting developed into the ‘microanalysis of online data’ (MOOD) network, which has subsequently held a number of symposia in the Netherlands, UK and Switzerland. It has also led to an influential paper on ‘digital conversation analysis’ (Giles, Stommel, Paulus, Lester, & Reed, 2015) and a special section of the Journal of Pragmatics (Giles, Stommel, & Paulus, 2017). We are writing this introduction during the Covid-19 global pandemic of 2020, when many countries of the world instigated some form of ‘lockdown’ with individuals unable to see others in person. In this context, the use of digital platforms for keeping in contact has become not only the norm, but essential for maintaining relationships. In the earliest periods of lockdown, we saw social activities such as pub quizzes, birthday parties or watching films together move into the digital realm. Social media was used to maintain contact with others and to gain information about the pandemic. In a more professional context, education was conducted digitally, academics started attending ‘virtual conferences’, doctor-­ patient consultations were held via video-conferencing software, and legal hearings were held remotely. The increase in the use of online platforms for communication during lockdown highlighted some of the concerns which have been raised previously about digital interaction. One key concern related to the ways in which digital interaction differs from offline interaction, and whether this impacts upon the ‘quality’ of the interaction (e.g. McKenna & Bargh, 2000; Turkle, 2017; Wen, 2020). This has been seen as particularly relevant for written communication, where paralinguistic cues (facial expressions, body language, tone of voice) are absent. These concerns are not new: there has been research across a variety of fields which has explored the ways in which users have negotiated some of the apparent ‘deficits’ of

6 

J. Meredith et al.

digital (textual) communication, such as the use of smilies, emojis and other punctuation markers to replace the offline cues provided by nonverbal communication (e.g. Crystal, 2006; Derks et al., 2008; Markman & Oshima, 2017; Meredith, 2014; Petitjean & Morel, 2017). There is often a focus on the ways in which the technology itself may lead to negative digital behaviours, such as trolling, abuse, harassment or bullying (e.g. McVittie, Sambaraju, & Bain, 2020; Sambaraju & McVittie, 2020; Suler, 2004). There has also been some concern that you cannot express real empathy in both written and video-mediated digital communication, which may be particularly necessary now (Turkle, 2011, 2020). The apparent deficits in video communication are not related to the difference between writing and speaking but to issues like time lags in the transmission of the video and/or audio channel, the impossibility of mutual gaze and to on-screen visibility. Rather than viewing these issues as merely practical hurdles, scholars working on video-interaction have analysed how these technology-related features actually affect the interaction (Heath & Luff, 1993; Licoppe, 2017b; Licoppe & Morel, 2012). This book seeks to address the questions of how individuals interact on digital platforms and how the technology itself impacts upon the interaction. This book provides empirical evidence of digital interaction, and brings together research from authors across a range of disciplines who all have one thing in common: they take a microanalytic approach to investigating digital communication. All of them explore the ways in which digital interaction can be studied using microanalytic methods and its associated challenges. The focus of this introductory chapter is to define digital interaction, to provide an overview of what we mean by ‘microanalysis’ and the similarities and differences between the different methods. Finally, we provide an overview of the development of digital conversation analysis.

Digital Interaction For the purposes of this book when we talk about ‘digital’ interaction we mean any communication which takes place within a digital environment which is designed to facilitate interaction. This can include blogs,

1  Introduction: The Microanalysis of Digital Interaction 

7

below-the-line responses to news, YouTube comments, video-mediated interaction as well as other social media platforms such as Twitter, Facebook and so on. These are distinct from sites which are seen as purely informational (such as a university website) or those which have a more commercial function (such as eBay). The platforms focused on in this book vary from digital dating apps, to YouTube, to instant messaging, and each platform has different technological features which are discussed in the chapters. Interaction can be broadly defined as communication where there are at least two speakers taking alternating turns. To a great extent this is also the case with digital interaction: we are interested in the ways in which people have conversations. In spoken interaction we would examine any responses to explore how an  utterance has been understood. Non-­ responses in offline interaction can be inspected for whether they indicate potential trouble in the interaction. However, as the first example in this chapter showed, the nature of digital interaction is potentially different from offline interaction, as not every message online will get a response. As the second example further demonstrated, responses which are received may be difficult to ‘reconstruct’ as a coherent thread of interaction. This raises questions for those analysing digital interaction: do we simply ignore a status or blog entry because it is not ‘interaction’ as we might understand it from the perspective of conversation? How do we inspect responses which may be presented in a different way from how they appeared for later participants? We would argue that by examining digital communication from an interactional perspective we are able to uncover significant insights into how digital interaction occurs. In this sense, our perspective is that any message posted on a digital platform is ‘designedly interactional’ (Meredith & Potter, 2014). By this we mean that although there may not be a specific recipient, the posts are still recipient designed. Recipient design is a key concept in conversation analysis and refers to the ways in which any specific turn is designed for its recipients (Schegloff, Jefferson, & Sacks, 1974). Some types of digital interaction are designed for recipients who are effectively unknown, such as people who are ‘lurkers’ on large sites (Nonnecke & Preece, 1999). These ‘lurkers’ do not tend to engage in the interaction, but rather simply read posts or messages.

8 

J. Meredith et al.

Therefore, they are not just anonymous strangers, but they are also invisible. In other contexts, such as in instant messaging or video-­mediated chat, the recipients tend to be known, although there may still be multiple recipients. We can, therefore, reconsider the definition of recipient design for the digital context, where interactions are often multi-party and multi-recipient (Giles et al., 2015). In this sense it is important to consider the participation framework (Goffman, 1981) for any digital interaction. Goffman’s (1981) framework discusses the idea of ratified or non-ratified participants. Ratified participants may be direct recipients (e.g. the person to whom a message is aimed) or indirect recipients (e.g. others who may be part of the conversation but not the direct recipient of a message). Non-ratified participants may be bystanders or overhearers, those who are not directly engaged in an interaction but have heard parts of the interaction. Digital messages may be designed for a direct recipient, such as when replying to someone’s Tweet. They may also be designed for indirect recipients, that is those who have been engaged in the interaction to this point but the message is not directed to them. We can also consider whether messages might be designed for non-­ratified participants. Dynel (2014) argues that it is simply not possible to have overhearing participants in digital interaction, as this suggests that it is a non-sanctioned overhearing, whereas reading a post on Twitter which is not directed at you is not non-sanctioned. However, in terms of recipient design we can consider that these overhearing recipients may be important in terms of thinking about designing messages for more general audiences who may be simply reading these messages (Meredith & Potter, 2014). When analysing data which do not seem interactional, then, we can consider who these messages are designed for—who might the audience be?

The Microanalysis of Interaction This book focuses on the use of microanalytic approaches to studying digital interaction. Microanalysis is not only a method of analysis, but is also a theoretical approach which argues that the details of interaction are relevant to our understanding of social life (Bull, 2013). This approach to

1  Introduction: The Microanalysis of Digital Interaction 

9

the analysis of communication has been made possible by the widespread availability of video- and audio-recording equipment. Being able to record data and play it back allows for the detailed analysis of interaction. Microanalytic research is interested in interaction as the object of analysis. This is in contrast to the way in which interaction might be analysed to make claims about a user’s intentions or state of mind. Microanalysis demonstrates how interaction might orient to, or be shaped by, the context in which it is taking place. Another focus is on how the interaction itself develops in a sequential and rhetorical way. In this section, we will outline the microanalytic approaches used in this book, their underlying approach and some key concepts. The three key methods of analysis used in this book are conversation analysis (CA), membership categorisation analysis (MCA) and discursive psychology (DP). The three methods are similar in terms of their treatment of interaction, although there are some key differences in their analytic interests. All three methods take the perspective that the analysis of the fine-grained detail of talk allows us insight into the way in which we relate to one another and the ways in which we maintain understanding in interaction. CA was developed in the 1960s by Harvey Sacks, Emanuel Schegloff, and Gail Jefferson (see Maynard, 2012 for more detail on the development of CA). CA aims to ‘examine how, through talking, people live their lives, build and maintain relationships and establish who they are to one another’ (Stokoe, 2009, p. 81). CA treats talk as not just talking about things but also as doing things—requesting, inviting, denying, refuting and so on. As analysts we are interested in how any utterance is understood through its design and sequential placement as doing a particular action. CA finds that every action within a conversation is shaped by the previous interaction, and also provides the context for the rest of the interaction. Schegloff and Sacks (1973) found that sequences are organised as adjacency pairs. Adjacency pairs comprise two utterances which are (in spoken interaction) adjacent to one another and are produced by different speakers. Schegloff and Sacks note that when the first turn in this sequence is delivered (known as a ‘first pair part’) this makes a second pair part relevant, to be produced by a different speaker. CA research on sequence organisation explores how sequences are built to be coherent

10 

J. Meredith et al.

(see Schegloff, 2007). Another interest within CA is in how individuals know who should take the next turn and when they should do so. Sacks, Schegloff, and Jefferson (1974) explicated in detail how turn-taking functions in spoken interaction, noting that each turn-at-talk was made up of turn constructional units (TCUs). Each party in a conversation has the rights to a single TCU, which may be a word, a phrase or a sentence, and at the end of this TCU there is a transition relevance place (TRP) where another party in the interaction may take a turn. As such, there are specific positions in spoken interaction where speaker transition occurs. Like conversation analysis, MCA was first proposed during Harvey Sacks’s classic series of lectures in the early 1970s (Sacks, 1992). However, its focus is less on the structural features of talk and more on the use of category labels to frame substantive topics, typically social issues such as gender, race and ethnicity, and how these kinds of social identities are invoked in interaction (Stokoe, 2012). Therefore, whilst CA has a clear focus on the sequential aspects of interaction, MCA focuses on members’ methods for categorising the world or displaying their understanding of it. However, MCA still focuses on the ways in which categories are used in specific interactional contexts and so combines an interest in the sequential organisation of talk with the use of categories. A number of key concepts and ideas are relevant for MCA (see Stokoe, 2012 for an extensive list of these). Category-bound activities are activities which are linked to a particular category (e.g. ‘the baby is crying’, with the category of baby linked to the activity of crying). Category-tied predicates are a characteristic of a category (e.g. ‘the politician is a liar’, with the category of politician linked to the characteristic of lying). Category-bound activities and category-tied predicates are important as the activities and predicates can be inferred from the context without the characteristic being used (e.g. we may be able to infer the characteristic of ‘liar’ from the use of the phrase ‘typical politician’). At the same time, these tied and bound predicates and activities are not completely context-free, as their meaning is dependent on the context in which they are used. For this reason, it is important to analyse the interaction and discourse around the use of a category to fully understand the inferences which can be drawn from the use of that category in that context.

1  Introduction: The Microanalysis of Digital Interaction 

11

Discursive psychology (DP) has a different historical trajectory from CA and MCA, although it has been influenced by both. DP explores how speakers construct psychological categories such as personality, disposition, emotions, views and beliefs (see Edwards & Potter, 1992). It shares a similar approach to talk as CA. For example, it treats talk as action-­ oriented and as situated in a particular sequence or context. It also treats talk as situated rhetorically: for example that a description may be built to anticipate any potential counter-arguments (see Billig, 1996). Discourse is also viewed as both constructed and constructive. It is constructed in the sense that it uses a range of resources (e.g. grammar, words, categories, idioms) to build descriptions and accounts. It is constructive in the sense that it is used to build version of the world, such as particular events. Finally, in terms of its methods, DP is not a universal approach to discourse (Tileagă & Stokoe, 2015). Some kinds of DP are very strongly influenced by CA, in that they focus on how psychological concepts are used within the sequence of talk-in-interaction. However, some DP studies focus on written texts, including both traditional and digital media (e.g. newspapers) as well as formal and institutional talk (e.g. parliamentary exchanges, courtroom interaction). Using microanalytic approaches to analyse digital interaction offers two benefits: (1) it allows us to understand and explore the norms and conventions of digital interaction in depth; (2) through doing so we can offer concrete, empirical evidence of specific aspects of digital interaction and differences between digital and offline communication.

Development of ‘Digital CA’ While CA, MCA and DP were developing into the methods and approach we see now, a parallel development was occurring in computing. In the 1960s, the first message was sent from one computer to another. In the 1970s, the first e-mail was sent. And in the 1990s, the World Wide Web was developed, which allowed for the growth of digital communication platforms. These parallel developments of digital communication and microanalytic methods eventually led to digital CA. In this section we will provide an overview of digital CA and of the literature in this area.

12 

J. Meredith et al.

The first studies which examined digital interaction using CA and MCA were published in the late 1990s and early 2000s (e.g. Reed, 2001; Rintel, Mulholland, & Pittam, 2001; Werry, 1996). These studies focused on exploring the suitability of CA methods for analysing this kind of data (Reed, 2001), highlighting how the ‘naturalness’ of this data made it particularly suited for understanding interaction which has not been influenced by the researcher. This early research also investigated empirically how digital interaction is organised (Rintel et al., 2001; Cherny, 1999), with a strong focus on sequence organisation. It was noted that in much digital interaction, the two parts of an adjacency pair were often not actually adjacent in the interaction. This led to the finding that digital modes of communication often led to ‘disrupted turn adjacency’ (Herring, 1999). There was, then, a focus on how it was that users of these kinds of platforms were able to maintain coherence. It has been found that users make use of addressivity, lexical repetition and grammatical structures in order to maintain coherence in an otherwise seemingly disordered interactional context (Berglund, 2009). As a result, ‘disrupted turn adjacency’ does not seem to cause considerable communicative issues or breakdowns of understanding (Meredith, 2014). Instead, users appear to have adapted to the interactional constraints and maintain coherence. Another focus of early studies of digital communication was on turn-­ taking (e.g. Garcia & Jacobs, 1999). Significant differences between digital and spoken turn-taking have been identified. A key factor in this difference is the fact that it is not possible to mutually coordinate turn-­ taking in digital written interaction, as interlocutors do not have access to the turn-in-progress as they would with spoken interaction. Users can only post a final, fully completed turn, at which point the recipients are able to respond. As such there are questions about whether terms such as ‘turn constructional unit’ or ‘transition relevance place’ are relevant in digital interaction (Meredith, 2019). These challenges in adapting a method which has predominantly been used for spoken interaction and applying it to digital communication have been key in the development of digital CA (Giles et al., 2015). At this point, it might have been timely to start developing new methods that could better capture the phenomena of online interaction than CA, a method that had predominately evolved out of the study of

1  Introduction: The Microanalysis of Digital Interaction 

13

landline telephone calls. But then other things happened that dragged digital communication even further away from offline conversation. The emergence of social media in the 2000s (Facebook 2004/Twitter 2006/ Instagram 2010 and so on) brought about a new set of interactional interfaces that caused even more problems for academics trying to apply CA to computer-mediated interaction. By this time, many of those academics had moved on, as if the matter had been solved with a few caveats. Furthermore, these differences were not merely technological, but cultural. One of the most interesting aspects of early online interaction was that, in most environments, users were anonymous, at best represented by a choice username. Online interaction tended to remain online. But social media, driven by the possibilities of sharing visual information among intimate friends and relatives, shifted the focus on to the (offline) identity of the individual user. When someone tweets, we see their name, what they look like, what they ‘do’, how many followers they have and who (some of ) their friends are. This is social data, not just the content of a text. Each new medium evolved its own culture (just the phenomenon of a ‘tweet’ needs to be understood in the context of the medium, as a unique communicative act) (Giles, 2018). How do we address these contextual features in qualitative analysis? One important contribution has been Hutchby’s (2001) work on affordances (see also Arminen, Licoppe, & Spagnolli, 2016; Meredith, 2017). The concept of affordances puts forward the possibility that features of the environment (here the specific technology technological/cultural environment) can be perceived as having a number of potential actions for those who share the environment. However, the environment is also shaped by its users and which features those users choose to make use of. Regardless of designers’ intentions, features of technology can both afford and constrain the interactional potential (Hutchby, 2001). The concept of affordances moves away from the prospect of a technological deterministic approach, and rather allows for an analysis of digital interaction which also demonstrates how the interaction orients to particular technological features. On a practical level, the concept of affordances allows analysts to examine the interaction itself first and then explore if and how that interaction orients to the relevant technological features of the medium (Arminen et al., 2016; Meredith, 2017).

14 

J. Meredith et al.

A second issue is more methodological, and concerns the challenges faced when using terminology that was originally developed for spoken interaction. Previous research has varied to the extent to which it uses such terminology. As noted above, some authors have chosen to talk about turn constructional units (TCUs) as relevant in digital interaction (e.g. Garcia & Jacobs, 1999; Tudini, 2015), whilst others have argued against this (Meredith, 2019). Similarly, there is an argument about the extent to which terms such as ‘repair’ or ‘summons’ might be relevant for digital interaction (see Meredith, 2019). These types of issues will be addressed in the empirical chapters of the book. Studies in the digital CA field have focused on a range of interactional phenomena. A recent review of the literature by Paulus, Warren, and Lester (2016) found a total of 89 articles which used CA to analyse online interaction between 1994 and 2016. They found that as well as sequence organisation and turn-taking, authors have been interested in turn design (Petitjean & Morel, 2017), with a particular interest in how turns are recognisable as doing certain actions when there is a lack of visual, nonverbal and paralinguistic cues. There is some interest in how emoticons, emojis or even gifs are used in interaction to make turns recognisable as doing certain actions (Markman & Oshima, 2017; Tolins & Samermit, 2016). Digital CA research has also focused on how trouble in interaction is managed on digital platform, often referred to using the CA term ‘repair’. Studies have found how repair often occurs visibly for participants in digital interaction, such as through correcting misspellings or typos (Collister, 2011; Schönfeldt & Golato, 2003). It has also been demonstrated, through the use of screen-capture software, that writers do repair when constructing their messages (Meredith & Stokoe, 2014). Writers tend to be oriented to similar issues as in spoken interaction, but their repairs are ‘hidden’ when performed in message construction. However, participants are still oriented to the sequential implications of doing certain actions. As digital platforms have developed, so digital CA has branched out into analysing these platforms. While multi-party chat rooms such as ICQ or IRC were the focus of much of the early research (e.g. Rintel et al., 2001; Schönfeldt & Golato, 2003), other studies focused on other platforms such as instant messaging (Meredith, 2014) and online forums

1  Introduction: The Microanalysis of Digital Interaction 

15

(Stommel & Koole, 2010). It was noted by Paulus et al. (2016) that DP tends to be used more commonly for asynchronous interaction, such as online forums, while CA is more typically used with synchronous interaction such as instant messaging. This may be because the latter is less obviously suited to the analysis of written text (Potter & Edwards, 2012). It may also be because interactional issues are often more obvious in instant messaging, and CA explicitly focuses on interaction. Research which is broadly located within digital CA has also started to investigate social media sites. Housley, Webb, Edwards, Procter, and Jirotka (2017) analysed Twitter data, showing how it is possible to merge microanalytic approaches with big data collection. Other more recent research has explored the organisation of sequences on Periscope (Licoppe & Morel, 2018) and the music sharing service Soundcloud (Reed, 2017). Yet another string of research focuses on video-mediated interaction. With the increasing use of video technology in various domains, interaction scholars examine how the technological features interfere with the social and/or institutional setting and goals (Oittinen, 2018; Seuren et al., 2020; Stommel, Van Goor, & Stommel, 2019). More specifically, such studies tend to focus on both verbal and embodied practices in video-interaction afforded by the technology, such as showing (Licoppe, 2017a; Stommel, Licoppe, & Stommel, 2020). Hence, digital interaction covers a wide range of different types of interaction occurring across a multitude of different platforms. Digital CA and associated microanalytic methods have developed as ways to explore digital interaction in depth. The aim of this book is to demonstrate how these methods can be used, on a variety of data, and to address the opportunities and potential challenges of applying such methods to digital data.

Overview of the Book This book comprises seven empirical chapters in which the authors all apply microanalytic methods to digital interactions. The focus is on providing an in-depth analysis of some phenomenon of digital interaction on a specific platform. In other words, the authors do not intend to make generalisable claims about how digital interaction functions, but rather

16 

J. Meredith et al.

focus on exploring specific aspects of digital interaction. All of the empirical chapters focus on the digital nature of the interaction, and how the interaction may be impacted by being online. The authors also address the challenges of using microanalytic methods for digital interaction, and the considerations that need to be made when applying these methods to such interaction. Although there are similarities in the approaches taken in each chapter, there are also differences. The chapters are, out of necessity, very varied. The empirical chapters are organised so that the platforms studied move from asynchronous platforms in the earlier chapters, through to more synchronous platforms such as instant messaging, with the final empirical chapter examining video-conferencing. Similarly, the methods used in the book move from a more discursive approach, through to MCA approaches, and then finally towards digital CA. The book does not, though, just focus on the analysis of digital data. It is acknowledged that the collection of digital data also raises a number of questions, particularly around ethics. Therefore, prior to the empirical chapters we include a chapter on ethical concerns and issues around collecting such data. This chapter introduces some of the key issues which may need to be addressed in order to collect data in an ethical way. A final chapter brings together the findings from all of the chapters and addresses what we know so far, and where we go from here. This book offers an overview of areas of interest for researchers of digital interaction. At the same time it highlights the potential of the microanalytic approach for understanding contemporary social life. And as the number of platforms and formats increase, so do the possibilities for research investigating the detail of this kind of interaction. Through the meticulous disentangling of the many facets of digital interactions, we can see that communication in these environments is rich, challenging and social in new ways compared to face-to-face communication. The community of researchers in this area is broader than this book is able to capture and is still growing. We hope that this book not only demonstrates the range of research which is being conducted, but also provides inspiration for further exploration in this area.

1  Introduction: The Microanalysis of Digital Interaction 

17

References Arminen, I., Licoppe, C., & Spagnolli, A. (2016). Respecifying mediated interaction. Research on Language and Social Interaction, 49(4), 290–309. https:// doi.org/10.1080/08351813.2016.1234614 Berglund, T. Ö. (2009). Disrupted turn adjacency and coherence maintenance in instant messaging conversations. Language@Internet, 6(2). Retrieved from http://www.languageatinternet.org/articles/2009/2106 Billig, M. (1996). Arguing and thinking: A rhetorical approach to social psychology. Cambridge: Cambridge University Press. Bull, P. (2013). Communication under the microscope: The theory and practice of microanalysis. London: Routledge. Cherny, L. (1999). Conversation and community: Chat in a virtual world. Stanford, CA: Center for the Study of Language and Information. Collister, L. B. (2011). *-repair in online discourse. Journal of Pragmatics, 43(3), 918–921. https://doi.org/10.1016/j.pragma.2010.09.025 Crystal, D. (2006). Language and the internet (2nd ed.). London: Cambridge University Press. https://doi.org/10.1017/CBO9780511487002 Derks, D., Bos, A. E., & Von Grumbkow, J. (2008). Emoticons and online message interpretation. Social Science Computer Review, 26(3), 379–388. Dynel, M. (2014). Participation framework underlying YouTube interaction. Journal of Pragmatics, 73, 37–52. Edwards, D., & Potter, J. (1992). Discursive psychology. London: Sage. Garcia, A. C., & Jacobs, J. B. (1999). The eyes of the beholder: Understanding the turn-taking system in quasi-synchronous computer-mediated communication. Research on Language and Social Interaction, 32(4), 337–367. Giles, D., Stommel, W., Paulus, T., Lester, J., & Reed, D. (2015). Microanalysis of online data: The methodological development of “digital CA”. Discourse, Context & Media, 7, 45–51. https://doi.org/10.1016/j.dcm.2014.12.002 Giles, D., Stommel, W., & Paulus, T. M. (2017). The microanalysis of online data: The next stage. Journal of Pragmatics, 115, 37–41. https://doi. org/10.1016/j.pragma.2017.02.007 Giles, D.  C. (2018). Twenty-first century celebrity: Fame in digital culture. Bingley: Emerald. Goffman, E. (1981). Forms of talk. Philadelphia: University of Pennsylvania Press. Heath, C., & Luff, P. (1993). Disembodied conduct: Interactional asymmetries in video-mediated communication. In G. Button (Ed.), Technology in work-

18 

J. Meredith et al.

ing order: Studies of work, interaction, and technology (pp. 35–54). London: Routledge. Herring, S. C. (1999). Interactional coherence in CMC. Journal of Computer-­ Mediated Communication, 4(4). Retrieved from http://www.ascusc.org/jcmc/ vol4/issue4/herring.html. Housley, W., Webb, H., Edwards, A., Procter, R., & Jirotka, M. (2017). Digitizing Sacks? Approaching social media as data. Qualitative Research, 17(6), 627–644. https://doi.org/10.1177/1468794117715063 Hutchby, I. (2001). Technologies, texts and affordances. Sociology, 35(2), 441–456. https://doi.org/10.1177/S0038038501000219 Licoppe, C. (2017a). Showing objects in Skype video-mediated conversations. From showing gestures to showing sequences. Journal of Pragmatics, 110, 63–82. https://doi.org/10.1016/j.pragma.2017.01.007 Licoppe, C. (2017b). Skype appearances, multiple greetings and ‘coucou’. The sequential organization of video-mediated conversation openings. Journal of Pragmatics, 27(3), 351–386. Licoppe, C., & Morel, J. (2012). Video-in-interaction: “talking heads” and the multimodal organization of mobile and skype video calls. Research on Language and Social Interaction, 45(4), 399–429. https://doi.org/10.108 0/08351813.2012.724996 Licoppe, C., & Morel, J. (2018). Visuality, text and talk, and the systematic organization of interaction in periscope live video streams. Discourse Studies, 20(5), 637–665. https://doi.org/10.1177/1461445618760606 Markman, K. M., & Oshima, S. (2017). Pragmatic play? Some possible functions of English emoticons and Japanese Kaomoji in computer-mediated discourse [Preprint]. SocArXiv. https://doi.org/10.31235/osf.io/qa764 Maynard, D.  W. (2012). Everyone and no one to turn to: Intellectual roots and contexts for conversation analysis. In The handbook of conversation analysis (pp.  9–31). John Wiley & Sons, Ltd. https://doi. org/10.1002/9781118325001.ch2 McKenna, K. Y., & Bargh, J. A. (2000). Plan 9 from cyberspace: The implications of the internet for personality and social psychology. Personality and Social Psychology Review, 4(1), 57–75. McVittie, C., Sambaraju, R., & Bain, F. (2020). ‘I love James blunt as much as I love herpes’–‘I love that you’re not ashamed to admit you have both’: Attempted insults and responses on twitter. Language & Communication, 76, 23–34.

1  Introduction: The Microanalysis of Digital Interaction 

19

Meredith, J. (2014). Chatting online: Comparing spoken and online written interaction between friends. Thesis, Loughborough University. Retrieved from https://repository.lboro.ac.uk/articles/thesis/Chatting_online_comparing_ spoken_and_online_written_interaction_between_friends/9479531 Meredith, J. (2017). Analysing technological affordances of online interactions using conversation analysis. Journal of Pragmatics, 115, 42–55. https://doi. org/10.1016/j.pragma.2017.03.001 Meredith, J. (2019). Conversation analysis and online interaction. Research on Language and Social Interaction, 52(3), 241–256. https://doi.org/10.108 0/08351813.2019.1631040 Meredith, J., & Potter, J. (2014). Conversation analysis and electronic interactions: Methodological, analytic and technical considerations. In Innovative methods and technologies for electronic discourse analysis (pp.  370–393). Hershey: IGI Global. Meredith, J., & Stokoe, E. (2014). Repair: Comparing Facebook ‘chat’ with spoken interaction. Discourse & Communication, 8(2), 181–207. Nonnecke, B., & Preece, J. (1999). Shedding light on lurkers in online communities. Ethnographic Studies in Real and Virtual Environments: Inhabited Information Spaces and Connected Communities, Edinburgh, 123128. Oittinen, T. (2018). Multimodal accomplishment of alignment and affiliation in the local space of distant meetings. Culture and Organization, 24(1), 31–53. https://doi.org/10.1080/14759551.2017.1386189 Paulus, T., Warren, A., & Lester, J. N. (2016). Applying conversation analysis methods to online talk: A literature review. Discourse, Context & Media, 12, 1–10. Petitjean, C., & Morel, E. (2017). “Hahaha”: Laughter as a resource to manage WhatsApp conversations. Journal of Pragmatics, 110, 1–19. https://doi. org/10.1016/j.pragma.2017.01.001 Potter, J. (2012). Discourse analysis and discursive psychology. In H. Cooper (Ed.), Quantitative, qualitative, neuropsychological and biological (Vol. 2, pp. 111–130). Washington, DC: American Psychological Association. Potter, J., & Edwards, D. (2012). Conversation analysis and psychology. In J.  Sidnell & T.  Stivers (Eds.), The handbook of conversation analysis (pp.  701–725). John Wiley & Sons, Ltd. https://doi. org/10.1002/9781118325001.ch35 Reed, D. (2001). ‘Making conversation’: Sequential integrity and the local management of interaction on Internet newsgroups (p 10).

20 

J. Meredith et al.

Reed, D. (2017). Performance and interaction on Soundcloud: Social remix and the fundamental techniques of conversation. Journal of Pragmatics, 115, 82–98. https://doi.org/10.1016/j.pragma.2017.01.012 Rintel, E.  S., Mulholland, J., & Pittam, J. (2001). First things first: Internet relay chat openings. Journal of Computer-Mediated Communication, 6(3), JCMC634. Sacks, H. (1992). Harvey sacks: Lectures on conversation (G.  Jefferson, Ed.). Blackwell Publishers. Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the organization of turn-taking in conversation. Language, 50(4), 696–735. Sambaraju, R., & McVittie, C. (2020). Examining abuse in online media. Social and Personality Psychology Compass, 14(3), e12521. Schegloff, E. A. (2007). Sequence organization in interaction: A primer in conversation analysis I (Vol. 1). Cambridge: Cambridge University Press. Schegloff, E. A., Jefferson, G., & Sacks, H. (1974). A simplest systematics for the organization of turn-taking for conversation. Language, 50(4), 696–735. Schegloff, E.  A., & Sacks, H. (1973). Opening up closings. Semiotica, 8(4). https://doi.org/10.1515/semi.1973.8.4.289 Schönfeldt, J., & Golato, A. (2003). Repair in chats: A conversation analytic approach. Research on Language and Social Interaction, 36(3), 241–284. https://doi.org/10.1207/S15327973RLSI3603_02 Seuren, L., Wherton, J., Greenhalgh, T., Cameron, D., A’Court, C., & Shaw, S. (2020). Physical examinations via video? Qualitative study of video examinations in heart failure. using conversation analysis. Journal of Medical Internet Research, 22(2), e16694. https://doi.org/10.2196/16694 Stokoe, E. (2009). Doing actions with identity categories: Complaints and denials in neighbor disputes. Text & Talk, 29(1), 75–97. Stokoe, E. (2012). Moving forward with membership categorization analysis: Methods for systematic analysis. Discourse Studies, 14(3), 277–303. Stommel, W., & Koole, T. (2010). The online support group as a community: A micro-analysis of the interaction with a new member. Discourse Studies, 12(3), 357–378. https://doi.org/10.1177/1461445609358518 Stommel, W., Licoppe, C., & Stommel, M. (2020). “Difficult to assess in this manner”: An “ineffective” showing sequence in post-surgery video consultation. Social Interaction. Video-Based Studies of Human Sociality, 3(3), DOI: 10.7146/si.v3i3.122581.  Stommel, W., Van Goor, H., & Stommel, M. (2019). Other-attentiveness in video consultation openings: A conversation analysis of video-mediated ver-

1  Introduction: The Microanalysis of Digital Interaction 

21

sus face-to-face consultations. Journal of Computer-Mediated Communication, zmz015. https://doi.org/10.1093/jcmc/zmz015 Suler, J. (2004). The online disinhibition effect. Cyberpsychology & Behavior, 7(3), 321–326. Tileagă, C., & Stokoe, E. (2015). Discursive psychology: Classic and contemporary issues. Abingdon: Routledge. Tolins, J., & Samermit, P. (2016). GIFs as embodied enactments in text-­ mediated conversation. Research on Language and Social Interaction, 49(2), 75–91. https://doi.org/10.1080/08351813.2016.1164391 Tudini, V. (2015). Extending prior posts in dyadic online text chat. Discourse Processes, 52(8), 642–669. https://doi.org/10.1080/0163853X.2014.969138 Turkle, S. (2011). Life on the screen. New York: Simon and Schuster. Turkle, S. (2017). Alone together: Why we expect more from technology and less from each other. London: Hachette. Turkle, S. (2020, July 9). Technology and empathy after COVID-19. American Academy of Arts & Sciences. Retrieved from https://www.amacad.org/ events/tech-­empathy-­covid-­19 Wen, T. (2020, April 9). How coronavirus has transformed the way we communicate. Retrieved from https://www.bbc.com/worklife/article/20200408-­ coronavirus-­how-­lockdown-­helps-­those-­who-­fear-­the-­phone Werry, C. C. (1996). Internet relay chat. In S. Herring (Ed.), Computer-mediated communication: Linguistic, social and cross-cultural perspectives (pp.  47–63). Amsterdam and Philadelphia: John Benjamins.

2 Ethical Challenges in Collecting and Analysing Online Interactions Hannah Ditchfield

This chapter explores the key ethical issues facing research that collects and analyses online interaction. Specifically, it focuses on the issues that impact research drawing on the approaches of conversation analysis (CA) and discursive psychology (DP). CA and DP are interested in understanding talk in interaction and aim to analyse the kinds of talk produced in everyday naturally occurring situations (Hutchby, 2001). With online spaces providing ample opportunity for naturally occurring talk to occur, such approaches have increasingly been adopted in online contexts (see Giles, Stommel, Paulus, Lester, & Reed, 2015). With the rise of research into and on internet technologies, ethical practices that researchers have built up within the “It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

H. Ditchfield (*) Department of Sociological Studies, University of Sheffield, Sheffield, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_2

23

24 

H. Ditchfield

study of offline activities have been unsettled. Traditional ethical debates around privacy, consent and anonymity have been reignited by the need to respond to the characteristics of these new technologies and practices (Whiteman, 2012). Such interest has resulted in a rapid growth of debate regarding the specific challenges and questions of online research ethics (see Baym & Markham, 2009; Eynon, Fry, & Schroeder, 2017; Tiidenberg, 2018; Whiteman, 2012) with the development of specific ethical guidelines for digital researchers to follow (now in its third edition: Franzke, Bechmann, Zimmer, Ess, and the Association of Internet Researchers, 2020). Although CA and DP work into online interaction faces many of the same ethical dilemmas of any research collecting online data (e.g. public/ private distinctions, processes of informed consent and the challenge of anonymisation) there are also specific challenges that are raised due to the central aims and premises of CA/DP research. Notable scholars in the field of digital research ethics, for example, have spoken of the risks involved in tracing direct quotations through simple Google searches, posing threats to the identities of users and participants (Markham, 2012). With CA relying on the analysis of direct quotations and the close examination of the very organisation of talk, common approaches designed to prevent the searchability of online data, such as stripping or fabrication (Markham, 2012), become problematic. Alongside this, technological developments have led to methodological innovations in data collection. CA work is now drawing on technologies to capture fundamental CA structures, such as repair, in unique ways. Such innovations, for example use of screen capture technology, are working to reveal previously invisible online spaces and raise concerns around how to access such data in an ethically sensitive way (Ditchfield, 2020; Meredith & Stokoe, 2014). Due to the specific ethical challenges faced by online CA/DP, there is a growing need for a focused discussion on the issues and approaches impacting this field. This chapter attends to this by exploring four of the key ethical challenges facing digital CA/DP researchers: (1) public/private distinction, (2) informed consent, (3) anonymity and searchability, and (4) the visibility of going behind the screen. In discussing these challenges, I present cases of good ethical practice from CA/DP studies with the aim of providing researchers examples of how these ethical challenges have been approached in practice. In doing this, this chapter reveals how CA and DP work has found ethical solutions to researching a variety of online

2  Ethical Challenges in Collecting and Analysing Online… 

25

interactive spaces. It also reveals where there are currently gaps in ethical discussion within the field and argues for the importance of consistent and transparent reflection on ethical decision making within digital CA/DP publications.

Public/Private Distinction Defining data as either public or private is something all CA and DP research needs to consider. This is because a researcher’s understanding of whether the data they are collecting is public or private has implications for other fundamental ethical concerns such as informed consent. Within literature on digital ethics, there is much debate surrounding the public/ private distinction with many arguing the task of defining online research sites in this way is increasingly difficult (Eynon et al., 2017; Markham & Buchanan, 2012; Nissenbaum, 2010; Tiidenberg, 2018; Whiteman, 2012). danah boyd is one scholar who has discussed these complications claiming that, offline, people are accustomed to having “architecturally defined boundaries” (e.g. walls, limited audio ranges) that help give a sense of how public actions are (boyd, 2008, p. 14). The digital world, however, lacks these clear boundaries making it difficult to define spaces, and therefore online interactions, as either private or public. In CA and DP research, scholars have commonly understood the public/private nature of online data based on what can be referred to as access barriers (Whiteman, 2012). These barriers refer to elements that restrict access to certain online content. Account membership, for example, is one such barrier. Without a Facebook, Twitter or Instagram account individuals are unable to gain access to certain interactions. To overcome such a barrier, an account simply needs to be made. A harder barrier to overcome, though, is that of privacy settings. Here, users have the agency to set restrictions on who can, and who cannot, see and gain access to the content they post online. Online CA research such as Giles (2017a) and Housley, Webb, Edwards, Procter, and Jirotka (2017) have approached data with minimal access barriers as public. In the case of Housley et al., for example, API technology was used to collect ‘public’ tweets where no privacy restrictions were in place. In studies that have drawn on interactions more heavily guarded by access restrictions, such as Meredith and

26 

H. Ditchfield

Stokoe (2014) and Stommel and te Molder (2015), data has been understood and approached as private. Although using access barriers as a guide to defining public/private spaces is drawn upon within online interaction research it is essential to consider the complex nature of online environments. As Markham and Buchanan (2012) highlight: Individual and cultural definitions and expectations of privacy are ambiguous, contested, and changing. People may operate in public spaces but maintain strong perceptions or expectations of privacy. (p. 6)

This means that although researchers may be able to technically gain access to interactions online and thus define it as public data, users themselves do not perceive these interactions to be public and subsequently do not expect this data to be accessed and used for research. boyd eloquently described users’ expectations as the difference between “being public” and “being in public” (2014), the difference being that when being public individuals act in the knowledge that their performance, or interactions, will be widely used and accessible. While just being in public, for instance talking with a friend in a park, expectations are different; individuals do not expect to be listened to, recorded or followed. In response to acknowledging user’s expectations, scholars such as Helen Nissenbaum (2010) have advocated for “contextual integrity”, that is, respecting users’ expectations of what will happen to their data in a given context. Discussion and reflection around what online user’s public/private expectations are is largely absent from CA and DP work. This is not to say it is not considered, it is rather to say that it is not often explicitly addressed within empirical publications. One approach that is helpful when it comes to considering the expectations of online users is through approaching “ethics in context” (Whiteman, 2012). This refers to an approach in which “ethical decision making is embedded in the local details of research rather than involving the application of general ethical principles” (Whiteman, 2012, p.  9). This means that whether data is understood as public or private, and the eventual ethical decisions that are then made, should be based on the specific local details of the study rather than just access barriers. This involves considering the platform,

2  Ethical Challenges in Collecting and Analysing Online… 

27

the topic, the type of users and what this might mean for user’s public/ private expectations. Although not explicitly addressed within publications, there are examples of CA and DP work where the local details of the research can be seen to impact the ethical approaches adopted. Here, I take Burke and Goodman (2012) and Giles (2017a) as examples. Burke and Goodman (2012) explored discussions about asylum seeking on Facebook. They collected and analysed interactions from Facebook groups: an online space on the platform where you can connect with people with shared interests (Facebook, 2020) and were specifically interested in exploring how talk about “Nazis” was used in discussion on asylum seekers. In terms of access barriers, the data collected by Burke and Goodman was publicly available with groups being found through the search function on Facebook. However, despite this, consent was obtained from the group’s administrator. In this way, the administrator acted as consensual gatekeeper to interactions within this particular space. Giles’ (2017a) research explored the relationships between fans and celebrities in the digital era by examining a group of emerging celebrities (namely crime authors) and their followers on Twitter. Similarly to the research of Burke and Goodman the tweets analysed by Giles were not restricted by any access barriers and were therefore technically public in nature. Giles, however, did not obtain consent from users or gatekeepers stating that all users were “unaware” of their participation (p. 449). Despite the data sets for Burke and Goodman (2012) and Giles (2017a) having the same access barriers in place different ethical strategies were adopted. Although not explicitly addressed in either article, I argue that such approaches could be explained through the consideration of the expectations and local details of the research. In the context of Burke and Goodman, data was collected from ‘Facebook’: a platform that is more commonly associated with interpersonal interaction with people who are known to one another. The topic of the interactions within the chosen Facebook groups were also of a more sensitive nature in that taking about Nazis and asylum seekers link to issues of prejudice and race. In the case of Giles, data was collected from the profiles of emerging celebrities. Although boundaries between who and who can’t be defined as a public figure are indeed blurred, emerging book authors’ interactions

28 

H. Ditchfield

with fans are arguably more public in nature than the interactions of everyday Facebook users. It is therefore reasonable to assume that fans interacting with authors on their public profiles will know (and expect) that their posts can be viewed by a wide range of Twitter users outside of their network. The interactions collected by Giles were also not of a sensitive nature and were more focused on the everyday communications that occur on authors’ Twitter profiles including happy birthday and networking exchanges. These examples reveal how defining public and private data on access barriers is a common strategy employed in CA and DP research; however, it is not the only aspect that is considered. Through taking a closer look at CA and DP work, it can be seen how the potential expectations of users and the specific local details of the research (e.g. platform, type of interaction and topic) can also impact where on the public/private spectrum data can fall and the implications this then has for ethical decisions such as consent. Although this is highlighted through this discussion, the complexities of making public/private distinctions of online data are not readily, or explicitly, discussed in many CA and DP studies revealing a need for more transparent reflection on this ethical issue in empirical work.

Consent Informed consent is something that every researcher collecting and analysing online interactions needs to consider. Consent refers to gaining permission from research participants for the use of their information, interactions or participation in research. The process allows potential participants to learn about the research aims, the benefits and harms of the research as well as data security and proposed outputs. It is through this process that participants ultimately agree, or disagree, to participate in the study. As highlighted by the examples discussed in the previous section (Burke & Goodman, 2012; Giles, 2017a) whether or not informed consent is obtained, or in what capacity it is obtained, is very much linked to the local details of the study and whether researchers understand the research site as public or private. Data defined as private, for

2  Ethical Challenges in Collecting and Analysing Online… 

29

example, is often understood as requiring consent and data defined as public is not. This can be seen across research into online interaction with Giles (2017a) and Housley et al. (2017) not gaining consent to use public Tweets and Meredith and Stokoe (2014) and Ditchfield (2020) seeking consent for the use of private Facebook messages, status updates and comments. However, the decision of whether to obtain informed consent is not as simple as just thinking about access barriers. As highlighted by Burke and Goodman’s research (2012), the local details of a project, such as the sensitivity of the topic or the specific platform from which data is collected, are also important factors when deciding on a consensual approach. For research that needs to obtain consent, there are further complications to consider when collecting and analysing online interaction. Collecting and analysing online interactions, in many cases, involves using interactions from multiple participants as well as data from a variety of communicative forms that coexist in the same online space. Some online content does involve just one participant, for example a single tweet, an Instagram caption or a Facebook status. However, much online interaction, and particularly the interactions conversation analysts are interested in, is dyadic. Therefore, if researchers intend to also collect and analyse the comments to tweets or Instagram posts the number of users involved in the research begins to increase. This also applies to more private interactions that occur from inbox to inbox (Facebook Messenger, WeChat, WhatsApp). Here, there are always at least two individuals involved in the exchange with the potential for a lot more in group chat settings. This means that consideration not only needs to be given to the consent of the central participant (the individual being recorded or who produces the initial content) but also the other individuals involved in the interactions and whose content will also be collected, observed and analysed. One approach to handling this in online interaction research has been to categorise participants as “primary” and “secondary” (Meredith & Potter, 2013). Primary participants are those initially approached for the research and whose content is central to the project (they could be the user recording their screen or a profile that is a case study). Secondary participants are those with whom the primary participant interacts and

30 

H. Ditchfield

whose interactions, by default, also become part of the sample. When collecting private interactions (such as Facebook messages) scholars have argued that it is ethical to gain informed consent from both primary and secondary participants (see Ditchfield & Meredith, 2018 for further discussion). This issue of secondary participants is not new or unique to online interaction. Offline interaction is also dyadic and interactionist research in offline contexts also needs to consider consent in relation to all interlocutors (primary and secondary). What makes this process different online, and arguably more challenging, is that due to the networked nature of many research sites researchers can collect data far beyond that for which informed consent was gained (Livingstone & Locatelli, 2014) meaning the potential number of secondary participants can be significantly larger. This was a challenge faced in my own work on Facebook (Ditchfield, 2018). In my research, not only were Facebook messages collected and analysed but so were comments and status updates. Such interactions occur on Facebook profiles and appear on newsfeeds rather than chats or messages which are sent from inbox to inbox. In this way, comments and status updates are more public in nature (as they are usually visible to user’s entire networks). Yet, they remain private through the use of certain access barriers such as privacy settings. With these forms of interaction being visible to entire networks complexity is added to the consent process by the larger number of secondaries involved within the research. With an individual’s Facebook network commonly made up of hundreds of users, obtaining informed consent from every secondary participant within the network is an unfeasible task (Ditchfield & Meredith, 2018). Considering these factors, the approach taken was to inform potential secondaries of the research and offer the option to opt out (see Ditchfield & Meredith, 2018 for details). This positioned myself, the researcher, in the position of an “overhearer” rather than an “eavesdropper” (see D’Arcy & Young, 2012). As an overhearer, researchers are in a position that allows the analyst to sit outside of the action but be a known intruder rather than an unknown ‘eavesdropper’. As part of being informed, participants can be offered the option to opt out of the research and have their interactions excluded. If this is not explicitly taken, though, consent is assumed. Adopting the position of an overhearer can be a useful ethical

2  Ethical Challenges in Collecting and Analysing Online… 

31

approach when obtaining informed consent is unfeasible, particularly in contexts that are semi-private (such as Facebook comments and updates). The example of Meredith and Potter (2013) and that of my own research (Ditchfield, 2018) highlight both some of the complexities, but also potential solutions, to the ethical conundrum of obtaining consent. Although detailed reflections exist around obtaining consent in CA and DP research, there is little reflection or attention paid to the issue of obtaining consent from potentially vulnerable participants, including that of under 18s. Livingstone and Locatelli (2014) highlight that due to the possibilities for identity play online there are increased opportunities to play and experiment with gender, age, race and sexuality. Performing different ages online (e.g. minors performing being older) poses specific challenges for informed consent. This is because there are different procedures for gaining informed consent for participants under the age of 18, usually involving also gaining the consent of a parent and guardian. This is problematic in two ways: first, due to affordances of online technology it is difficult to even know if your participant is above 18 and therefore what consent procedures are appropriate. Secondly, even if researchers are confident participants are underage, traditional processes of gaining parental consent may also be ethically difficult. This is because online spaces are often recognised as private spaces and revealing the usage of such spaces to parents may be damaging to the identity expression and community forming practices of the individual (Spriggs, 2010). It is therefore important for CA/DP researchers to consider these risks in their ethical decision making, especially when researching platforms that have large followings of younger users.

Anonymity and Searchability Anonymising data in online CA/DP research is a particularly pertinent challenge due to the reliance of using direct quotations for analysis and the risk that such quotations can be traced. Anonymity is often a “classic promise made to research participants” within social science research (Tiidenberg, 2018, p. 427) and refers to the protection of participant’s personal identity information such as names, workplaces and

32 

H. Ditchfield

locations—the idea being that it is not possible to trace the data back to the original source. Common approaches adopted when anonymising data in CA/DP research involve changing identity markers such as names, locations and institutions through blanking out or substituting for pseudonyms. Anonymising interactional data is important to consider in both research where consent is gained (such as Ditchfield, 2020; Meredith & Stokoe, 2014) but also in cases where it is not (see Giles, 2017a). In research where consent is not sought, anonymity becomes especially important as users are not aware that their interactions are being collected and therefore may find the publication of their online talk in academic forums “intrusive” (Giles, 2017a, p. 449). For much CA/DP work it is feasible for researchers to anonymise their participants. As Giles (2017a) highlights, in research where individuals are simply representatives of a group or category (e.g. author, blogger, fan), there is no need to make such individuals traceable as doing so brings no or little analytical value. In such contexts, there is more to gain in protecting individual’s identities, particularly when they are unaware of their participation. Within research on online interaction, there are two specific challenges that researchers face in relation to anonymity: (1) accessing data sets where the anonymity of participants needs to be kept from researchers themselves and (2) the risk of publicly available data being traced back to the original authors. I will begin by discussing challenge one. Researchers needing to access particularly sensitive contexts, such as medical settings and interactions, can face enhanced ethical challenges due to patient data needing to remain confidential even from the researchers themselves. Such a challenge was present in the online CA work of Stommel and te Molder (2015) in which they collected and analysed online counselling chats. The data collected within this study was both private (e.g. chats between counsellors and patients not available on public platforms) and also sensitive in nature with discussions centring on problematic drug and alcohol use and treatment options. Consent to use these interactions could not be gained as the chats were collected from stored conversations rather than recorded live. In order to protect the identity of these vulnerable participants, then, counsellors of the service were asked to select the calls in which none of the participants “mentioned anything that could lead to their person” (2015, p.  285). The

2  Ethical Challenges in Collecting and Analysing Online… 

33

counsellors also replaced the names with pseudonyms and deleted all other person-related information. Although this approach was effective in protecting the identity of chat participants from both researchers and academic audiences, it required significant work on the part of the counsellors. In this case, labour was balanced out due to counsellors maintaining a practical interest in the research with training workshops being developed based on the analysis. Stommel and te Molder’s work provides a good example of how anonymity can be used to access hard-to-reach data sets and highlights how the labour involved in such an approach can be justified when the research has wider societal impacts. Working to change identifying information or protecting participant identities from researchers themselves are not the only ethical challenges faced in the quest for anonymity in online interaction research. Markham (2012), for example, has argued that even when anonymised, online materials are still potentially traceable through search engines. Livingstone and Locatelli (2014) expand on this by stating that a “simple google search may link data taken from online source back to its original author” (p. 68) making ensuring the anonymity of sources difficult. This is a particularly prevalent issue when collecting online interactions as researchers use direct quotations within their published work making the risk of traceability high. This risk is further enhanced in research that collects publicly accessible data as users have not consented to, and are unaware themselves of, the use of their data in the research context. Outside of interactional research, one commonly applied approach to deal with this risk of searchability is to ‘strip’ the data. This could include altering certain words within a direct post, missing certain words or phrases out or avoiding using exact quotations at all instead paraphrasing the data in ways that still communicate key analytical themes. Annette Markham (2012) has taken such debates further by proposing “fabrication” as a methodological approach that works to avoid searchability of sensitive research materials. Markham talks of using composite accounts, interactions and events in order to represent the data collected within research whilst maintaining ethical sensitivity. Stripping data, or indeed methods of fabrication, may be appropriate in certain research contexts, for example a thematic or narrative analysis of social media posts; however, they are problematic for approaches such as CA/DP. CA’s primary

34 

H. Ditchfield

interest is in the very structure and organisation of talk with approaches such as DP being specifically concerned with how things like accountability and identity are constructed and put together within interaction itself. This means that any changes to exact quotations—be it word swaps, missing turns of interaction or composite accounts—would dramatically distort the data and make the unique contribution CA and DP make to knowledge impossible. A further issue presented by data stripping or fabrication is related to how CA and DP ground their analysis. One of the central challenges in any discursive analysis is the “interpretative gap” (Edwards, 2012). The interpretative gap refers to the process that exists between “phenomena, data, analysis, and conclusions” (Edwards, 2012, p.  1) and is “the distance between the object under scrutiny and, via method, data processing and inferences, what you eventually want to say about it” (p. 4). This gap exists within all types of research, both of quantitative and qualitative forms, but it is a particularly contested issue within discursive approaches to analysis (Edwards, 2012). The central concern in relation to this gap is with how researchers arrive at the interpretations made. One way to deal with the challenge of the interpretative gap is not necessarily to narrow it, but, instead, map out the journey of analysis through the process of grounding analytical interpretations (see Ditchfield, 2018). According to Wood and Kroger (2000), grounding analytical interpretations is “not about how you come up with patterns, interpretations and so forth, but how you justify your identification of patterns” through drawing on evidence in the discourse itself (p. 95). By staying loyal to, and accurately displaying, the original empirical material you can purposively show other academics how you reached the interpretations you have leaving the data open to criticism and other analytic interpretation. With CA and DP analysis focusing on the very structures and discursive details of text, it would become extremely difficult to ground interpretation using significantly altered extracts. Interactional studies that have collected public online interactions for analysis have considered these ethical challenges and adopted approaches that work to protect both the identities of users and the original structures of the data. Giles’ (2017a) work on authors’ interactions on Twitter, for example, took measures by converting interactions into readable data

2  Ethical Challenges in Collecting and Analysing Online… 

35

tables rather than displaying screen shots. Additional care was also given to remove as much detail as possible in order to lower the risk of searchability. Cino and Vandini (2020) collected public posts from the parenting forum “BabyCenter” and worked to further ensure anonymity by checking reported quotes in the Google search engine to make sure they could not be traced. Beyond these techniques, alternative suggestions have been made by scholars that may be feasible for CA and DP work. In the AoIR 3.0 ethical guidelines, Franzke et al. (2020) propose that researchers could ask for consent to use direct quotations within final reports and publications. Due to the potential size of interactional data sets collected from platforms such as Twitter, as well as the increasing number of secondary participants involved in research, asking for informed consent before data collection can often be unfeasible. However, authors of the 3.0 guidelines suggest that consent could be asked at the dissemination stage of research. Here, consent is gained to use a quote, rather than the more generic consent to be including in the data set. In this way, researchers could outline the risks of searchability to quoted participants and gain permission for is direct use. This would not solve the problem of searchability but it would mean participants were informed of the risk and ultimately happy for this information to still be published. This approach, though, is not without its challenges. In David Giles’ (2017b) chapter on online discussion forums, for example, it became difficult to obtain this form of consent from public figures with messages being passed on and ignored by celebrity agents. Such an approach, therefore, may only be feasible for certain research contexts.

Visibility: Going Behind the Screen The technologies researchers use to collect and analyse online data present ethical challenges in relation to what becomes visible to researchers. In addition to technologies developing in ways that give users more opportunities for interaction, technologies are also developing that give researches alternative approaches to data collection. This is apparent in research such as Housley et  al. (2017) where API technologies were

36 

H. Ditchfield

adopted to gather large data sets of tweets. However, there are also technologies being utilised that go beyond collecting published content (e.g. content that has been posted on profiles, forums or in chats) and actually work to capture the construction of such content. One such example is research using screen capture software (Ditchfield, 2020; Meredith & Stokoe, 2014). Screen capture software can be downloaded onto a laptop and run in the background of any computer activity in order to record whatever is on the screen. In other words, what the computer user sees on their screen, the software captures and saves as video files. Recordings display interactive detail ranging from keystrokes and deletions to cursor movements and click-throughs revealing, moment by moment, how users navigate through a social networking platform. Through the use of data collection approaches such as screen capture it is not only published interactions that are being collected and analysed but also the construction process of these interactions including the edits made to messages before they are even shared with their intended audience. Through such technologies, spaces of online environments are now being accessed that are usually only ever visible to the user themselves, such as the edits they make before posting, the way they navigate platforms and where attention is placed. This opens up an alternate layer of private data accessible due to technological advancements in researchers’ methodologies. In ways this space is more private than a private, direct, message on sites such as Facebook. It is a space where there is usually no audience whatsoever. Yet, through methods such as screen capture, researchers have the opportunity to become spectators to these backstage online worlds. Gaining access to these usually invisible spaces has implications for the ethical decisions and approaches adopted in research. In the case of screen capture technology and my own work (Ditchfield, 2018) the sensitivity of accessing this space was handled in two ways: (1) through enhanced participant control and (2) through enhanced researcher-researched rapport. Participants used the screen capture technology on their own devices and decided for themselves when they would record themselves interacting on the social media platform. They also had control over the data they eventually sent over for analysis meaning that participants could still keep elements of their online life ‘invisible’. Participants were also recruited

2  Ethical Challenges in Collecting and Analysing Online… 

37

from my own network to ensure a foundation of trust and rapport was present before the research began (for discussion on using networks for research see Taylor, 2011). Of course, adopting such approaches posed methodological challenges around representativeness and research relationship dynamics (for full discussion, see Ditchfield, 2018) but ultimately provided an ethical route into a unique and revealing interactive space.

Conclusion This chapter has explored four ethical challenges that face CA and DP research into online interaction: (1) the public/private distinction, (2) informed consent, (3) anonymity and searchability and (4) the visibility of going behind the screen. Through discussing examples of good ethical practice in relation to these challenges, this chapter has brought together some approaches and strategies for CA/DP researchers to draw upon in the process of collecting and analysing online interaction. These examples have revealed how complex ethical conundrums can be sensitively approached with even hard-to-access data sites becoming available for investigation. This demonstrates how ethical procedures are not necessarily restrictive to the research agenda of approaches like CA and DP. In considering local details when defining private data (e.g. Burke & Goodman, 2012), adopting primary and secondary levels to consent (e.g. Meredith & Potter, 2013), thinking creatively about anonymity (e.g. Stommel & te Molder, 2015) and considering searchability and the visibility of our methods (e.g. Ditchfield, 2018; Giles, 2017a;) CA/DP work is able to make significant, and ethical grounded, contributions to the field of online interaction. Despite these successes, there remains space for further reflection and discussion. Issues such as the privacy expectations of online users or the potential age (and therefore vulnerability) of potential posters is largely absent from CA/DP reflections. There are also alternative approaches to challenges that have not yet been utilised, but may be feasible, in the context of online interactionist work (e.g. gaining consent for quoting, and not analysing, online talk). Most importantly, however, this chapter

38 

H. Ditchfield

calls for consistency in ethical reflection. Although examples of good practice were indeed present, the topic of ethics was occasionally completely missed in CA/DP empirical publications. This is not to say that reflections did not happen, it is only to note the lack of transparency that can occur when addressing ethical conundrums. As stated in the latest “Association of Internet Researchers” guidelines (Franzke et al., 2020), it is through processes of deliberation and discussion that best ethical practice can be achieved. It is therefore essential for online interaction researchers not only to conduct ethical research, but also to talk (and write) about it.

References Baym, N., & Markham, A. (2009). Introduction: Making smart choices on shifting ground. In N.  Baym & A.  Markham (Eds.), Internet inquiry: Conversations about method (pp. vii–xix). Thousand Oaks, CA: Sage Publications. boyd, d. (2008). Facebook’s privacy trainwreck. Convergence, 14(1), 13–20. boyd, d. (2014). It’s complicated: The social lives of networked teens. New Haven: Yale University Press. Burke, S., & Goodman, S. (2012). Bring back Hitler’s gas chambers’: Asylum seekers, Nazis and Facebook—A discursive analysis. Discourse & Society, 213(1), 19–33. Cino, D., & Vandini, C. (2020). “Why does a teacher feel the need to post my kid?”: Parents and teachers constructing morally acceptable boundaries of children’s social media presence. International Journal of Communication, 14, 1153–1172. D’Arcy, A., & Young, T. M. (2012). Ethics and social media: Implications for sociolinguistics in the networked public. Journal of SocioLinguistics, 16(4), 532–546. Ditchfield, H. (2018). Behind the screen of Facebook: An interactional study of pre-­post editing and multicommunication in online social interaction. PhD Thesis, University of Leicester, Leicester. Ditchfield, H. (2020). Behind the screen of Facebook: Identity construction in the rehearsal stage of online interaction. New Media & Society, 22(6), 927–943.

2  Ethical Challenges in Collecting and Analysing Online… 

39

Ditchfield, H., & Meredith, J. (2018). Collecting qualitative data from Facebook: Approaches and methods. In U. Flick (Ed.), The SAGE handbook of qualitative data collection (pp. 707–727). London: Sage Publications. Edwards, D. (2012). Discursive and scientific psychology. British Journal of Social Psychology, 51(3), 425–435. Eynon, R., Fry, J., & Schroeder, R. (2017). The ethics of online research. In N. Fielding, R. Lee, & G. Blank (Eds.), The SAGE handbook of online research methods (pp. 19–37). London: Sage Publications. Facebook. (2020). Retrieved from https://www.facebook.com/ help/1629740080681586 Franzke, A.  S., Bechmann, A.  Zimmer, M., Ess, C., & the Association of Internet Researchers. (2020). Internet research: Ethical guidelines 3.0. Retrieved from https://aoir.org/reports/ethics3.pdf Giles, D. (2017a). How do fan and celebrity identities become established on twitter? A study of ‘social media natives’ and their followers. Celebrity Studies, 8(3), 445–460. Giles, D. (2017b). Online discussion forums: A rich and vibrant source of data. In V. Braun, V. Clarke, & D. Gray (Eds.), Collecting qualitative data: A practical guide to textual, media and virtual techniques (pp. 189–210). Cambridge: Cambridge University Press. Giles, D., Stommel, W., Paulus, T., Lester, J., & Reed, D. (2015). Microanalysis of online data: The methodological development of “digital CA”. Discourse, Context & Media, 7, 45–51. Housley, W., Webb, H., Edwards, A., Procter, R., & Jirotka, M. (2017). Digitizing Sacks? Approaching social media as data, Qualitative Research, 17(6), 627–644. Hutchby, I. (2001). Conversation and technology: From the telephone to the internet. Cambridge: Polity Press. Livingstone, S., & Locatelli, E. (2014). Ethical dilemmas in qualitative research with youth on/offline. International Journal of Learning and Media, 4(2), 67–75. Markham, A. (2012). Fabrication as ethical practice: Qualitative inquiry in ambiguous internet contexts. Information, communication, and society, 5(3), 334–353. Markham, A., & Buchanan, E. (2012). Ethical decision-making and internet research: Recommendations from the AoIR Ethics Working Committee (version 2.0). Retrieved from https://aoir.org/reports/ethics2.pdf

40 

H. Ditchfield

Meredith, J., & Potter, J. (2013). Conversation analysis and electronic interactions: Methodological, analytic and technical considerations. In H. L. Lim & F. Sudweeks (Eds.), Innovative methods and technologies for electronic discourse analysis (pp. 370–393). Hershey: IGI Global. Meredith, J., & Stokoe, E. (2014). Repair: Comparing Facebook ‘chat’ with spoken interaction. Discourse & Communication, 8(2), 181–207. Nissenbaum, H. (2010). Privacy in context: Technology, policy, and the integrity of social life. Stanford, CA: Stanford University Press. Spriggs, M. (2010). Ethical difficulties with consent in research involving children: Findings from key informant interviews. AJOB Primary Research, 1(1), 34–43. Stommel, W., & te Molder, H. (2015). Counselling online and over the phone: When pre-closing questions fail as a closing device. Research on Language and Social Interaction, 48(3), 281–300. Taylor, J. (2011). The intimate insider: Negotiating the ethics of friendship when doing insider research. Qualitative Research, 11(3), 3–22. Tiidenberg, K. (2018). Ethics in digital research. In U. Flick (Ed.), The SAGE handbook of qualitative data collection (pp.  466–481). London, Thousand Oaks: Sage Publications Ltd. Whiteman, N. (2012). Undoing ethics: Rethinking practice in online research. New York: Springer. Wood, L.  A., & Kroger, R.  O. (2000). Doing discourse analysis: Methods for studying action in talk and text. Thousand Oaks: Sage.

3 Context, History, and Twitter Data: Some Methodological Reflections David Giles

Introduction One of the main themes of this book is the suitability of existing qualitative methods for the analysis of communication in online (digital) environments. My colleagues and I in the MOOD network have discussed elsewhere the various challenges of adapting techniques from the study of talk to the study of written/typed interaction (Giles, Stommel, & Paulus, 2017; Giles, Stommel, Paulus, Lester, & Reed, 2015). Among these is the important issue of digital archives and the ease with which this information “It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

D. Giles (*) Department of Psychology, University of Winchester, Winchester, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_3

41

42 

D. Giles

can be used in online communication. In this chapter I consider some additional information that digital communicators use, particularly in social media, and I do this by addressing the unique affordances and communication culture of Twitter. I will make certain assumptions in doing so, not least that the reader is largely familiar with the way Twitter works: what it is, what it looks like, what sort of people use it and the sort of interaction that takes place in the medium. As a long-standing student of media psychology and audience-celebrity relations, my particular interest is in the way that Twitter (and, to a lesser extent, other social media) appears to level out the hierarchy between celebrities (and other public figures) and what used to be known, in the days of broadcast media, as the audience, or general public (Giles, 2017a, 2018). What relevance does the identity of Twitter users have for discourse and conversation analysts? I will argue in this chapter that some of the contextual details that conversation analysts (in particular) attempt to ‘bracket off’ from their analysis are in fact deeply ingrained in the interaction that takes place on Twitter, especially when one of the interactants is a celebrity. While conversation and discourse analysis offer the microanalyst of online data many useful methodological tools, the cultural and technological context of Twitter cannot be bracketed off in quite the same way that we can extract phone calls and (arguably) focus group discussions in order to produce transcripts in which the structure of talk can be explained by the verbal exchanges alone. Of course, there is certainly no consensus within the discursive psychology field as to the significance of contextual details even for studying talk itself, and the first part of the chapter is devoted to a short discussion of the debate in the literature. I then go on to discuss some of the unique affordances of social media, notably the preservation of archive content, before presenting a case study of a controversy surrounding a well-known British celebrity and his Twitter communication.

Digital Conversation and Context What do I mean by context when it comes to discourse and conversation analysis? Perhaps the best place to begin is the running debate between Emmanuel Schegloff and other authors in the pages of Discourse and

3  Context, History, and Twitter Data: Some Methodological… 

43

Society in the late 1990s/early 2000s. The debate was triggered by Schegloff’s (1997) broad critique of critical discourse analysis and the way researchers in this field import their own interests (around social inequality, largely) into their analyses. Much of the debate—perhaps because Schegloff used it to illustrate his own argument—has proceeded around the relevance of gender in talk. At what point should the analyst orientate towards speaker gender? Do we, as Schegloff argues, only bring gender into our analysis if explicitly referenced by speakers? Subsequent participants in the debate have argued that we cannot avoid importing such observations into analyses when they are so relevant to the social worlds that both analysts and speakers share: to do otherwise is naïve empiricism (Billig, 1999; Stokoe & Smithson, 2001; Wetherell, 1998). With online, or digital, interaction, the issues around context are somewhat different. For example, Schegloff (1988/1989) argues that conversation analysis can be applied to mediated data (in this case, television footage) so long as one brackets off all the contextual baggage in order to focus solely on the interactional dynamics of the talk itself. But when dealing with printed, or typed, interaction of the kind we encounter every day in social media and elsewhere, it is not such an easy matter: the talk always arrives embedded in other visual/textual information that neither the analyst nor the interactants can ignore. A TV interview follows the rules, up to a point, of any other kind of interview talk, such as a job interview. What we see in media such as Twitter is something quite different. In the first article to directly critique Schegloff’s (1997) original argument, Wetherell (1998) suggests that an important contextual issue concerns the identity of the speakers and their relationship to one another. She argues that it is important to identify subject positions that speakers occupy which both enable and constrain their contribution to conversation. In her own data, one speaker takes up the position of “I was drunk”, which enables certain allowances to be made, but also limits the claims he can make (see also Giles, 1999). Of course, these positions can be contested (by any of the interactants involved), but not without provoking disagreement. With traditional, or broadcast, forms of media, subject positions are more explicit still, since the media productions themselves are structured by the performers’ identities. The host or presenter of a show, like a teacher in a classroom, has certain rights to speak and expectations around turn design, including the ability (within reason) to silence other

44 

D. Giles

contributions if they are disrupting the preferred flow of events. Guests, particularly those without expert or celebrity status, have much more restricted subject positions (Giles, 2002a, 2003). Despite this, what goes on in media such as radio and television contains enough “natural” talk for Schegloff (1988/1989) to be able to bracket off even political identities in order to focus on the verbal interaction alone within a TV interview. In this particular case, he was discussing a famous encounter between American TV presenter Dan Rather and then Vice President1 George Bush, in which Bush’s supporters claimed their man had been “ambushed” by over-zealous interrogation around the Iran-Contra scandal. For Schegloff, as a scholar of conversation per se, these details are merely of ‘vernacular’ interest. The detail that one interactant is a (famous) TV anchor and the other a (Republican) politician can be separated from the ‘technical’ aspects of the interview, which are (or perhaps should be) of more interest to the analyst. These can be read off the interview transcript alone, through “mundane coloured glasses” (p. 218). Schegloff goes on to argue that the trajectory of the nine-minute episode, and the controversy it generated, could be attributed, in the main, to the violation of conversational norms produced by overlapping speech—a violation that, arguably, could threaten the satisfactory progress of any such interview, regardless of the participants’ identities. Is it possible though, never mind desirable, to simply bracket off such ‘vernacular’ detail when trying to understand how interaction proceeds in social media? In social media generally, subject positions are even more important than in broadcast media. They are predetermined by all kinds of identity detail that surrounds interactants—names, pictures, biographies, various metrics alluding to popularity, other status indicators, and indeed all the social demographic inferences (gender being one) that can be inferred from these details. What can we feasibly bracket off here? Following Schegloff to the letter, we could claim that interactants have to refer explicitly to any of these details in order to make them analytically relevant, but the relative status of the interactants may be an important determinant of the structure of the interaction itself.  He was actually running for the Republican presidential nomination at the time, hence the particular sensitivity towards perceived bias in his media representation. 1

3  Context, History, and Twitter Data: Some Methodological… 

45

On Twitter we cannot ignore imbalances of power whereby some members of the interactive community, such as celebrities, have visibly greater amounts of social capital than others. As analysts, we can never really know what social information members have access to on any given platform: specific messages arrive in different formats depending on the device the receiver is using, the preferences they have set up within the programme for accessing the messages, and so on. But messages are always tagged with the identity of the sender, since this is the raison d’être of social media.

The Long Reach of the Digital Archive We are all accustomed to reading news stories in which a public figure’s retrospective tweets, blog posts, and other historic archives have been unearthed at a point in time when they are of particular public interest. These digital materials enter public discourse around their author and constitute important evidence for that individual’s beliefs, or character, or even their competence to perform a public role. Historic digital content haunts all those who leave its traces on the Web. Most of it, most of the time, sits there unnoticed. But when its author becomes newsworthy, it tumbles out of the closet and the author is required to speak to it.2 Imagine you are the author of a fashion blog stretching back ten years, started at high school. Let’s call you Letitia. In the interim your YouTube channel (LettyLips) has taken off and you now have a million subscribers and you are receiving multiple media engagements. You have just been chosen to head up a task force tackling racism in the fashion industry. But wait: someone you once fell out with has recalled a blog post from when you were about 18 that made some mildly disparaging remark about a particular ethnic group. The post is still there in the archive: it is copied to a safe location, linked to by that someone, and circulated around the Twitterati, who are firing off polemic after polemic. You  I am deliberately using twentieth-century language for ironic effect here. Unlike traditional archives, which were typically physical artefacts deriving from an earlier time period, digital archives are dynamic, fluid data that force us to rethink the whole notion of historical primary sources (Featherstone, 2006). 2

46 

D. Giles

vaguely recall writing the post, but it reflects a teenage you, still learning about the world and the people in it, and you are horrified that this embarrassing comment is still online with your name attached. Or is it your name? You are now LettyLips… Pushed to defend the 18-year-old Letitia’s regrettable lapse of judgement, you come up with a watertight argument. That was the teenage me, another person. I am now someone else, so don’t stick that mud on me—while the 18-year-old Letitia might have had one or two dubious views (and I challenge you to find another, which you won’t, because I’ve suddenly deleted all my archive posts from the first eight years of the blog), LettyLips is not, in any way, a racist. It is surely no coincidence that writing—of any kind, handwritten, typed, or digital—seems to provide the most fertile conditions for this creative kind of self-presentation. We can see its effects even in the interaction that takes place in supposedly anonymous online contexts like discussion forums, where community members are identified by their usernames (which, as Stommel, 2007 points out, are often tied to personal characteristics of the individuals concerned). Because most (public) forums preserve their discussions in the form of archives, previous exchanges are easily available for present interactants to cut and paste excerpts from, or provide links to. In this way, interaction spreads across the forum’s history, and makes it harder for any specific thread to stand entirely alone (Giles, 2017b). In a data set that I have recently been working on, I and my co-author (Jones & Giles, under review) found a very interesting instance of this phenomenon in a thread on Mumsnet, a popular UK parenting website that features very lively and often fractious discussion (Giles, 2016; Pedersen & Smithson, 2013). The OP (opening post) is written by a woman who had ‘just discovered’ that her partner is on the sex offenders’ register for an inappropriate relationship with a student in a school where he had been teaching. After several minutes’ advice from fellow Mumsnet members (in each case, to get rid of this man as soon as possible), a member refers to a previous thread started by the OP which added new information to the story: the OP’s current relationship is only two months old, and the earlier thread contained “red flags” about him.

3  Context, History, and Twitter Data: Some Methodological… 

47

Subsequent posts continue to reference various previous threads, disclosing further that the OP is herself a teacher, and the tenor of the thread quickly becomes more critical, at times confrontational and even hostile (‘it should make you get shot’ says one poster). One post consists simply of a hyperlink to one of the OP’s earlier threads. A little over half an hour after her first post, the OP makes her final contribution to the thread in which she replies to a query about her partner’s suspended sentence. However, the thread remains very active for another 12 hours or so, with much speculation about whether or not the OP is a ‘troll’ (a provocateur), which includes a formal (inconclusive) investigation by a community moderator. After 5 days and 300 posts, the thread grinds to a halt with the final few posters wondering, not unsympathetically, what the OP has decided to do. So even with anonymous usernames, the archival nature of digital communication ties regular contributors to their own histories of activity, and this history is capable of structuring the present interaction. When the online persona is clearly representing an offline individual whose (offline and online) history is well known to the community or audience, this continuity of persona is surely impossible for the analyst to ignore.

 ase Study: Defending the Danny C Baker Persona The case study I report here is intended to illustrate some of the issues I have been discussing so far. It is, I believe, a fairly typical example of contemporary Twitter interaction involving a celebrity (with half a million followers on the medium) and other Twitter users, who all share the same entitlement to contribute to discussion, although, as I will argue, their relative status is significant to some extent in the way it structures the interaction. It concerns Danny Baker, a white broadcaster from London, who was once a hero of mine for his work on the weekly rock publication New Musical Express (NME) in the late 1970s. Although I later wrote, briefly, for the NME too, I never met Baker and so my actual relationship with

48 

D. Giles

him is purely “parasocial”.3 After leaving NME in the early 1980s, Baker became a fairly well-known celebrity in the UK, appearing on BBC radio and television and hosting his own shows, not without a degree of controversy (he was particularly outspoken about football topics, for instance). Like many contemporary celebrities, Baker is an enthusiastic user of Twitter and likes to amuse his large audience (over half a million followers to his principal account) with jocular and irreverent tweets. When the birth of Prince Harry and Meghan Markle’s son Archie was announced in May 2019, he posted a typically facetious remark accompanied by a picture of a young (white) couple hand-in-hand with a chimpanzee. It would almost certainly have gone unnoticed were it not for the ethnicity of the mother and the racist connotation of pairing a monkey, linguistically constituted as a child, with an African-American parent. Following a succession of critical tweets from offended Twitter users, among them several fellow media professionals, Baker issued a series of “apologies” (perhaps more readable as defences) that were not enough to prevent the BBC, whose show he was fronting at the time, from sacking him (BBC, 2019).

 nique Affordances of Twitter and Their U Methodological Implications I have written extensively elsewhere about media affordances (Giles, 2018), and how these must not be reduced solely to technological features. Above all, any given medium develops its own culture, and its interaction patterns should be understood as part of that culture (see also Giles, 2016 and Pederson and Smithson, 2013, for similar ideas in relation to Mumsnet). Twitter affords its users unconstrained access to any other user (unless that user chooses to block a specific user). I can post to, or reply to, the US president’s account at any time I like, and, of course at the risk of being blocked, I can say to them what I like. They  An unbalanced relationship in which partner 1 knows a great deal about partner 2 but partner 2 is oblivious to partner 1’s existence, typically experienced among audiences of broadcast media (Giles, 2002b; see Giles, 2018 for extension to digital culture) 3

3  Context, History, and Twitter Data: Some Methodological… 

49

can even reply to me, and sometimes even the most popular figures reply to other Twitter users of modest status (Giles, 2018; Stever & Lawson, 2013; Usher, 2015). Each tweet arrives in the user’s account with a link to the sender’s account, and we can see who it is that is making contact: what they look like, how they describe themselves, and how many other Twitter users follow them. This is what Twitter affords its users, in a social sense: Twitter also affords the analyst a good deal of information. I can click on a tweet and see when it was posted, how many users have ‘liked’ it or retweeted it, and underneath the tweet I can see a list of all the responses it has generated (along with their timestamps, user details, and all the related information). Helpfully, Twitter chains together series of responses in “conversations”, which act effectively as mini-discussion threads that are analysable in the same way as forum discussions (Giles, 2016, 2017b). Despite what Twitter might appear to afford the analyst, there is surprisingly little literature featuring microanalysis of Twitter interaction. Most published studies of the medium have, at best, relied on analysis of large corpuses of data (e.g. Hardaker & McGlashan, 2016; Scott, 2015; Zappavigna & Martin, 2018). Here, the message content of the tweet itself is extracted from its surrounding contextual detail, which allows researchers to focus solely on linguistic features of the messages in isolation. However, this approach gives us little sense of the interaction that is going on and promotes a kind of content analytic approach to the study of digital culture. One of the challenges of online data analysis is identifying the unit of analysis itself. For corpus researchers like those cited above, this is a simple matter of isolating the message (i.e. the verbal content of the tweet). But tweets are not read in isolation. As discussed earlier, tweets very often generate other tweets: members are responding to one another, albeit not necessarily in neat threads like on a forum. The reading experience at the time of the tweet’s creation will vary according to the device being used, the reader’s personal settings, and the number of tweets appearing at any given moment in the timeline. A reader may well miss altogether critical tweets in an ongoing interaction chain. This latter point applies to forums too. Gibson (2009, para. 18) talks of the need to reconstruct the “reading path” by which forum discussants “demonstrate

50 

D. Giles

through the organisation of their posts how they functionally relate to other contributions”. This is rather harder to do in the case of Twitter, but the timestamp information can allow us at least to gain an idea of the gradual unfolding of the dialogue. In the analysis that follows, I have reconstructed four separate threads from the Danny Baker incident. The first is simply a sequence of tweets that were released from Baker’s account on the evening of 08/05/19, which allow me to analyse his attempts to apologise for the offence caused by the original tweet in relation to existing discursive psychology literature on racism denial and apology. The second is the chronologically arranged sequence of responses to his first ‘apology’; the third and fourth are ‘conversations’ chained together by Twitter itself, where responses to Baker’s tweet have generated a linked thread of tweets, producing the same kind of turn-taking sequence as seen in discussion forums.

Racism Denials and Apologies In Table 3.1, I present a series of tweets that Baker posted on the evening of 08/05/19, beginning with the original tweet at 6.32 pm, and followed by three further tweets on the topic, the second and fourth tweets being explicitly framed as apologies, although readable also as defences. The first thing to note is the speed with which tweet 1 generates controversy: after only 36  minutes Baker has felt obliged to offer his first apology/ defence; after barely an hour he comments that he is now “trending”.4 A further, more elaborate, apology appears later in the evening. Note Baker’s comment in tweet 2 that tweet 1 had already been deleted at that stage. Tweet 2 appears to be a hasty attempt to counter some of the criticism of tweet 1 by issuing a clear apology, opening with what Olshtain and Cohen (1983) call an Illocutionary Force Indicating Device (IFID)—in this case “sorry”—before acknowledging the negative reception of tweet 1 (“has whipped some up”), offering an excuse (“never occurred to me”), and providing evidence of reparation (“down it came”). The concluding  In case anyone is unfamiliar with this term, ‘trending’ refers to the rapid circulation of digital content, such as a specific image or short video, though has become generic for any kind of information (e.g. hashtags on Twitter and Instagram) 4

51

3  Context, History, and Twitter Data: Some Methodological…  Table 3.1   Danny Baker’s tweets following accusations of racism Tweet

Timestamp

Full text of tweet

Other material

08/05/19

Royal baby leaves hospital

ID 1

B&W

photograph

upper-class

6.32PM

of

1920s

style man and woman, the

latter

holding

the hand of a young chimpanzee

attired

in a dress suit 2

08/05/19

Sorry my gag pic of the little

7.08PM

fella

in

the

posh

outfit

has

whipped some up. Never occurred

to

me

because,

diseased.

Soon

well, as

mind

those

not

good

enough to point out it's possible

connotations got in touch, down it came. And that's it. Now stand by for sweary football tweets 3

08/05/19 7.41PM

Ha! I see my silly unthinking gag

pic earlier has me trending. Last

time that happened it had been announced I had head cancer. On balance... 4

08/05/19 10.02PM

Once again. Sincere apologies for

Colour photograph of

earlier. Was supposed to be joke

figure

in

clown

about Royals vs circus animals in

outfit

(Max

Wall?

posh clothes but interpreted as

The wig is similar

about monkeys & race, so rightly

to that worn by the

the

stupid

unthinking

gag

pic

deleted. Royal watching not my

evidently

comic

post-war comedian)

forte. Also, guessing it was my turn in

the barrel.

N.B. each tweet prefaced at top of box by the following material: [profile image] Danny Baker @prodnose [date of tweet]

52 

D. Giles

reference to “sweary football tweets” has the intended effect of closing down any further discussion on the matter, since it signals a return to what Baker wants his audience to understand as ‘business as usual’ (light-­ hearted but profane humour, with topicality limited to sport). As Hagren (2019, p.  6) argues, the main function of denying racism in public discourse is the “maintenance of social positions and relations”. While tweet 2 employs several of the strategies researchers have identified as necessary for racism denials (and apologies) to be effective, it clearly does not work for Baker, because the debate has not been stifled— indeed tweet 1 is “trending” before long, and so much more negative reaction has taken place as the evening progresses that Baker issues a fourth tweet on the topic which contains a further IFID (“sincere apologies”) and another attempt to mitigate the offence (“was supposed to be … but interpreted as”). Van Dijk (1992, p. 91) has argued that racism denial is usually effective as a discursive strategy because it is always difficult to prove “negative intentions”, and throughout, this is what Baker relies upon to maintain his position of innocence. First, in tweet 2 he claims that the monkey/ African association “never occurred” to him because his “mind [is] not diseased”. If racism can be portrayed as irrational, then so can its interpretation: this is Baker’s strategy throughout, starting with the reference in tweet 2 to people who have been “whipped up” (‘into a frenzy’ is hearably the missing part here). In tweet 4 he employs the classic “was supposed to be … but interpreted as” construction, again portraying the critics as either oversensitive or having misinterpreted. Tweets 3 and 4 also include further mitigating material which manages to position Baker as a victim, and the incident to be a routine (or cyclic) one. In tweet 3 he dismisses the criticism as part of the recurrent nature of “trending”, likening it to a false rumour; in tweet 4 as part of a ritual that (presumably) all celebrities need to undergo at some point. Both these suggestions are dismissed by many responders as trivialising the original offence. Baker’s tweets in this sequence are a curious mixture of wounded innocence and dismissive bravado, delivered in short, pithy statements. At no point does he actually deliver an unqualified apology for any offence caused, and one could argue that each of tweets 2, 3 and simply serve to

3  Context, History, and Twitter Data: Some Methodological… 

53

weaken his defence. In this respect he is rather different from the political figures that have typically been studied in examples of racism denial (e.g. Augoustinos & Every, 2010; Hattaka, Niemi, & Välimäki, 2017). A closer comparison would be the Big Brother data studied by Riggs and Due (2010) following the famous confrontation between British reality TV celebrity Jade Goody and Bollywood actress Shilpa Shetty, in which the former was accused of racially abusing the latter. In subsequent discussion of this incident, the show’s presenter, and the interviewees themselves, took great pains to play down the racist aspect of behaviour, reframing it simply as ‘bullying’. The authors argue that, in the context of Big Brother, the requirement to discuss the incident at all was deemed an ‘unfortunate’ diversion from the diet of triviality that the show thrives upon. I argue here that the best way of understanding Baker’s tweets is as a performance that is ultimately geared towards the maintenance of his Twitter persona, itself a digital extension of the persona that he has cultivated over several decades as a broadcaster and celebrity. While aspects of his tweets echo findings from the CA and DP literature on both apologies and racism denials, the successful elements are compromised here by the demands of Baker’s persona and the (actual or imaginary) audience it serves.

Reactions to the First ‘Apology’ Baker’s deletion of his original tweet makes it hard to reconstruct the initial series of responses from other Twitter users, but the first apology (tweet 2 in Table 3.1) was still searchable at the time of writing, along with over 200 of the 370 replies, although several of these have been deleted. In Table 3.2 I have organised these chronologically so that we can see the first 14 replies that were sent. Note that I have stripped all these tweets of their visible details, so that we are simply left with the text itself (and a few references to emojis and hyperlinks). Partly this is to do with space, but largely to do with ethics, since it is only fair to anonymise Twitter data belonging to non-public-facing individuals, that is those with small follower numbers (see Giles, 2017a for more on this), and also

54 

D. Giles

Table 3.2   Initial responses to first apology 7.08

Sorry my gag pic of the little fella in the posh outfit has whipped some up. Never occurred to me because, well, mind not

diseased. Soon as those good enough to point out it's possible connotations got in touch, down it came. And that's it. Now stand by for sweary football tweets 7.09

[GIF/ image macro of actor saying ‘I don’t believe you’ ]

7.10

Is this an apology?

7.10

Perfect response. No hard feelings.

7.10

Nice one

7.10

How old are you Danny?

How could you NOT know the horrible and

horrific historic connotations of that tweet ? 7.12

7.13

Great sense of humour mate. Luckily I nicked it. Don’t worry

about the snowflakes, we know the difference between racism and comedy [three applause emojis]

[emojis communicating applause and ‘face throwing a kiss’]

7.13

Righto, Danny Bakkker.

7.13

I saw it, didn't occur to me to be honest. Just saw a funny pic

7.13 7.13

Too late mate. I've sent it to Limmy and you're getting cancelled.

i feel quite happy with myself as that thought never crossed my mind when i saw it. #ignorethescreamers

7.14 7.15

You bottled it Dan. Was this a gag too?

[image of three-year old tweet from BBC reporter (Danny

Cohen???) about anti-Polish racism in UK and Baker’s sceptical reply] 7.15

So everyone upset by that tweet has a diseased mind. That really is feeble excuse.

because—except where I import this information myself—it is not directly relevant to much of the discussion. It is fair to assume that most, if not all, of these replies were sent by people who are (or were) following Baker’s account at the time, since they would have seen his 7.08 tweet on their timeline and chose to respond within minutes. It is noticeable that, on the whole, the longer the time that elapses between Baker’s tweet and the reply, the more critical or hostile the reply, suggesting that people are receiving retweets of the 7.08

3  Context, History, and Twitter Data: Some Methodological… 

55

tweet (53  in total) or, more likely, retweets of the original ‘joke’ tweet (number unknown). Certainly the proportion of broadly positive replies in this sample (at a rough estimate, at least half ) is not representative of the list as a whole. Nevertheless, they could be regarded as collaborating ‘dialogic partners’ (Condor, Figgou, Abell, Gibson, & Stevenson, 2006) that help to protect the accused from charges of racism, possibly allowing Baker to believe that he has solid support among his followers. What I am labelling ‘Positive’ here does not necessarily mean the respondent endorses Baker’s original tweet, or even his defence of it. For example, the 7.14 tweet “you bottled it Dan” suggests the respondent was not actually offended by the original joke and is disappointed that Baker has seemingly given in to his critics. The 7.13 tweet referring to ‘bakkker’ is more likely a sardonic joke than a serious claim that Baker is a white supremacist. The “Limmy” tweet (also at 7.13) is likewise ambiguous. It is perhaps not surprising that all the positive/supportive tweets (in this sample, not necessarily across the whole list) come from accounts whose owners are visibly white and male. This might be considered merely ‘vernacular’ detail in the offline, Schegloffian, sense, but is surely unavoidable in the present data (and the ethnicity and gender of ‘positive’ fans is directly referenced by a number of later, critical, tweets). Structurally, we might consider these replies as a ‘second turn’, since they are clearly provoked by Baker’s 7.08 tweet, in the same way as responses to an opening post (OP) in a forum thread (Antaki et  al., 2005). Like forum OPs, Baker’s tweet is not hearably addressed to any specific individual, although the convention of ‘followers’ in Twitter might lead us to assume that, in the absence of any personalised information, tweets are addressed largely to those who will receive them directly in their account (500,000 Twitter users in this case). However, those users who objected to the earlier ‘joke’ tweet—and we must assume that many of those will be among those 500,000 followers—are not actually included in Baker’s address, since they are referred to as “some” and “those good enough”, as if the joke had accidentally reached an external audience that has exercised some degree of censorship (the use of Twitter to apparently circumvent gatekeepers is a long-standing part of its appeal for celebrities; Giles, 2018).

56 

D. Giles

If Baker’s 7.08 tweet, then, is addressed to the ‘fan’ element among his audience,5 what about the replies? Again, in the absence of direct addressees, we must assume they are directed at Baker himself, but they are not private messages, and will be received by their own followers in their Twitter accounts. Therefore, they constitute what Bou-Franch et al. (2012) describe as “doubly articulated” contributions—hearable by more than the addressee and ostensibly conceived with the additional audience in mind. Although several tweets mention Baker by name, or use the appellation ‘mate’, some are just statements about the author’s own response to the tweet that could be seen as hearable by the community as a whole. Even the ones referring to Baker by name may be designed for wider consumption (it is hard to imagine ‘you bottled it Dan’ being sent as a private message unless the sender had an established relationship with him). Perhaps the most interesting tweet in this early group is the 7.15 one asking “is this a gag too?” and reproducing an historic Baker tweet in which he appears to be denying racism. This is equivalent to the Mumsnet example discussed in the last section where a member’s previous threads were referred (and linked) to in an attempt to contextualise their current behaviour and to draw dispositional conclusions. Indeed, much of the subsequent debate among fans and commentators more generally concerned the issue of whether Danny Baker is ‘a racist’, irrespective of whether or not the joke was itself racist. Baker’s defenders, particularly those in the media, have pointed to his general anti-fascist stance and early championing of African-American music as a journalist in the 1970s and 1980s (e.g. Stubbs, 2019).

Two ‘Conversations’ Triggered by Replies I next turn to two sets of linked exchanges that Twitter has designated a ‘conversation’ and embedded them in the overall list of replies to the 7.08 tweet. The first, displayed in Table 3.3, begins with a 7.20 reply which is hearably addressed to Baker himself and then proceeds, for a while, like a  The assignation of ‘fan’ to followers in social media is discussed at length in Giles (2017a), where I break down a sample of crime authors’ followers into different categories. 5

3  Context, History, and Twitter Data: Some Methodological… 

57

Table 3.3   Conversation 1 following reply to Baker’s 7.08 tweet OR

7.20

Maybe you could explain why you thought it was either funny or interesting then? Cos I just can't find a good reason for

it tbh Dan. R1

7.22

R2

8.00

OR

7.28

Performing monkeys?

I'd like to hear his explanation. Doesn't seem like he's gonna offer one to me. Pretty sure animals dressed as humans has been a bit of a theme on his shows down the years, tho cant find anything online to back that up. Having listened to him for a long time, i'd be staggered if his intention was what people have concluded.

OR R3

8.09 8.43

Yeh that actually rings a bell. I'd be surprised too tbh. He's run a feature for years on monkeys dressed as famous people set to circus music - it's pretty much exactly as

described and if there was ever any side to it I've been oblivious OR

8.50

R3

11.37

R4

May 9

R1

May 13

R4

May 13

Why didn't he just say that? Not many people recall that. I only vaguely remember it tbf.

[ hyperlink to YouTube video illustrating R3’s previous

tweet]

I’ll go though his tweets for what he sent for the other royal births. The ones without a biracial parent...

Did you see Dan's response? [ hyperlink to video hosted on Twitter, restricted to certain users ] No I didn’t

pseudo-conversational exchange typical of what one might see in a discussion forum: R1 offers a succinct suggestion two minutes later, the original replier (OR) replies six minutes later, and then, at various times, three other Twitter users reply (in each case, to the previous tweet in the chronological sequence). Like a forum thread, the chain spreads across several days: R4 joins in the following day, and R1 then reactivates the ‘thread’ four days later when new evidence comes to light that seems to address the earlier comments. OR’s reply to Baker, while addressing him directly, is clearly doubly articulated, like a question from the member of a TV studio audience asking a performer to account publicly for their behaviour. His replies to

58 

D. Giles

the subsequent tweets continue this sense of being an audience, as Baker is no longer felt to be included in the discussion (being referred to as ‘he’/‘his’ throughout): in any case, unless Baker himself follows any of these users (and he only follows about 1500 accounts), he is unlikely to read any of these tweets. This conversation develops the theme mentioned in the last section about accumulating historical evidence to mitigate the criticism Baker has received over the tweet. While clearly OR does not interpret Baker’s 7.08 tweet as a sufficient apology, he is clearly prepared to forgive him if a sufficient context can be provided: he concurs with both R2 and R3 that racism is inconsistent with Baker’s persona (“having listened to him for a long time” …“I’d be surprised” … the “feature … has run for years”). R4 is even prepared to mine Twitter for historical support (though it’s unclear what form this might actually take). Conversation 2 (Table 3.4) features a short series of dyadic exchanges between a critical follower, who replies sardonically to Baker’s first apology, and a Baker defender. It is not untypical—there were several such sets of exchanges in the list of replies, some much more vitriolic (and, inevitably, shorter). What is interesting about this one is that it spreads across almost an hour and features three tweets from each interactant. R1’s initial reply at 7.44 is quite prickly, with upper case words, and might well have been ignored by OR, except it is clear from her reply six minutes later that she is, like the contributors to conversation 1, prepared to find excuses for Baker (“I don’t think he is racist”). Like the OR in conversation 1, she is evidently a Baker fan who is disappointed at the unconvincing nature of his first apology, and by her final tweet she downgrades her appraisal of Baker’s tweet to ‘bizarre’. Like a lot of the Baker defenders, R1 persists throughout with the “colour blind” defence, even echoing Baker’s “sweary football tweets” sign-off by announcing that he is “off to the watch the [football]”. What prompts these Twitter users to steer away from replying directly to Baker and entering into their own ‘conversations’? While it is evident that few of his followers are convinced by his first apology and, out of dismay have to search for their own excuses for his behaviour, we might also probe a little into the contextual detail for a further explanation. As mentioned before, not all Twitter users are equal. Most of the discussants in these two threads have small numbers of followers, several in double

3  Context, History, and Twitter Data: Some Methodological… 

59

Table 3.4   Conversation 2 following reply to Baker’s 7.08 tweet OR*

7.24

Ah, so everyone else is the racist. Thanks for clearing that

R1

7.44

NOBODY has to be the racist here...unless we insist there

OR

7.50

up.

MUST be one, right? [‘thinking face’ emoji]

He’s the one suggesting that if you thought it could be viewed as racist that you’re the one with the problem. I

don’t think he is racist but his explanation is appalling. I won’t all it an apology because it isn’t. ‘Oops, I didn’t R1

7.57

think. Sorry,’ might have been better.

I actually find the whole thing just quite sad tbh. Surely

the ‘non racists’ (for want of a better phrase) want to see a day where people don’t make such connotations...yet here they are, doin exactly that? OR

8.03

R1

8.21

It’s just a very bizarre tweet to do without realising where

it’s likely to lead.

To a certain extent I agree, but then again, if you

genuinely don’t make that kind of association, how are you to realise where that tweet could lead to? Anyway, we could yap all night. I’m sure Danny can fight his own battles. I’m

off to watch the fitba. Nice talkin, genuinely

figures, but two of them have well over a thousand. The first is conversation 1’s R1, who is an “online marketer” with 4400. The OR in conversation 2 has over 2000 followers. This raises the likelihood that their followers will be drawn into the interaction by seeing their responses (where a user follows both them and Baker). It may also be that the interactants in these conversations have long-standing relationships between one another, though R1’s closing comment in conversation 2 (“Nice talkin, genuinely”) seems to indicate otherwise.

Conclusion It is, perhaps, understandable that conversation analysts and other qualitative researchers have been a little shy about delving into the murky depths of Twitter data. These interactions do not, for the most part, look very much like the sort of conversation that Schegloff and

60 

D. Giles

others have been studying for the last half-century. One might easily be tempted to ignore the messy details and stick to the accumulation of content, using one of many ‘big data’ tools now on the market for combing, mining, and other kinds of broad-brush analysis. I am not arguing against the usefulness of such instruments for identifying network activity, describing trends and charting the spread of easily reducible bits of information. But Twitter is so central now to the way that even politics gets done that it seems unfortunate that we have paid it so little microanalytic attention. In this case study I have, I hope, identified a number of ways in which we can begin to break down the enormity of Twitter data within the context of a specific ‘event’. However, I have also queried some of the assumptions that conversation analysts (and discursive psychologists working in this tradition) might try and bring to bear on such material. Even if we were to strip out the hyperlinks, and all the other visible details, can we really just put on Schegloff’s “mundane coloured glasses” in order to understand the exchanges here? I would argue that only the most mundane features of Twitter interaction would be left, of largely technological interest: useful for CMC perhaps, but nothing that would make a psychologist want to hang around for long. But this is me talking, a psychologist interested in how people interact in social media and why they say the things they do. Ultimately the issue may simply be one of motivated looking (Stokoe & Smithson, 2001): why these data, now, to illustrate what point? What can we assume about the ‘vernacular’ information that it’s Danny Baker talking, or a white man, or a celebrity? Ideology can still be hidden, or ignored, if we are not ‘motivated’ to look at it. The context will always, ultimately, be set by the researcher, even if we can find ample ‘procedural relevance’ in the data. Maybe my interest in framing Danny Baker’s ill-conceived joke as a matter of personal crisis rather than a disgusting example of racist discourse stems from my desire (as a middle-aged white man, etc.) to find an escape route for a former hero. But if this was my chief goal, I could have more profitably used it as an example of how criticism of monarchy and privilege can be sidetracked by other ‘motivated’ concerns. Maybe my choice of this event as a case study of persona defence derives from

3  Context, History, and Twitter Data: Some Methodological… 

61

my fundamental interest in the individual’s history. And so on. Maybe all this is simply another example of Elliott, Fischer, and Rennie’s (1999) maxim about qualitative psychologists owning their perspective.

References Antaki, C., Ardévol, E., Núnez, F., & Vayreda, A. (2005). “For she who knows who she is”: Managing accountability in online forum messages. Journal of Computer-Mediated Communication, 11(1), 114–132. https://doi. org/10.1111/j.1083-6101.2006.tb00306.x Augoustinos, M., & Every, D. (2010). Accusations and denials of racism: Managing accountability in public discourse. Discourse & Society, 21(3), 251–256. BBC. (2019, 10 May). Baker ‘so, so, sorry’ for disastrous tweet. BBC News. Retrieved from https://www.bbc.co.uk/news/entertainment-­arts-­48226247 Billig, M. (1999). Whose terms? Whose ordinariness? Rhetoric and ideology in conversation analysis. Discourse & Society, 10(4), 543–558. Bou-Franch, P., Lorenzo-Dus, N., & Garcès-Conejos Blitvich, P. (2012). Social interaction in YouTube text-based polylogues: A study of coherence. Journal of Computer-Mediated Communication, 17, 501–521. https://doi. org/10.1111/j.1083-6101.2012.01579.x Condor, S., Figgou, L., Abell, J., Gibson, S., & Stevenson, C. (2006). ‘They’re not racist…’: Prejudice denial, mitigation and suppression in dialogue. British Journal of Social Psychology, 45, 441–462. Elliott, R., Fischer, C. T., & Rennie, D. L. (1999). Evolving guidelines for publication of qualitative research studies in psychology and related fields. British Journal of Clinical Psychology, 38, 215–229. Featherstone, M. (2006). Archive. Theory, Culture & Society, 23(2–3), 591–596. Gibson, W. (2009). Intercultural communication online: Conversation analysis and the investigation of asynchronous written discourse [39 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 10(1), Art. 49. Retrieved from http://nbn-­resolving.de/urn:nbn:de:0114-­fqs 0901493 Giles, D. C. (1999). Retrospective accounts of drunken behaviour: Implications for theories of self, memory and the discursive construction of identity. Discourse Studies, 1(4), 387–403.

62 

D. Giles

Giles, D. C. (2002a). Keeping the public in their place: Audience participation in lifestyle television programming. Discourse & Society, 13(5), 603–628. Giles, D.  C. (2002b). Parasocial interaction: A review of the literature and a model for future research. Media Psychology, 4(3), 279–302. Giles, D. C. (2003). Narratives of obesity as presented in the context of a television talk show. Journal of Health Psychology, 8(3), 317–326. Giles, D. C. (2016). Observing real-world groups in the virtual field: The analysis of online discussion. British Journal of Social Psychology, 55(3), 484–498. https://doi.org/10.1111/bjso.12139 Giles, D. C. (2017a). How do fan and celebrity identities become established on twitter? A study of ‘social media natives’ and their followers. Celebrity Studies, 8(3), 445–460. Giles, D. C. (2017b). Online discussion forums: A rich and vibrant source of data. In V. Braun, V. Clarke, & D. Gray (Eds.), Collecting qualitative data: A practical guide to textual, media and virtual techniques (pp.  189–210). Cambridge: Cambridge University Press. Giles, D. C. (2018). Twenty-first century celebrity: Fame in digital culture. Bingley, Yorks: Emerald. Giles, D. C., Stommel, W., & Paulus, T. (2017). The microanalysis of online data: The next stage. Journal of Pragmatics, 115, 37–41. Giles, D. C., Stommel, W., Paulus, T., Lester, J., & Reed, D. (2015). The microanalysis of online data: The methodological development of ‘digital CA’. Discourse, Context and Media, 7(1), 45–51. https://doi.org/10.1016/j. dcm.2014.12.002 Hagren, K.  I. (2019). She has promised never to use the N-word again’: Discourses of racism in a Swedish media debate. Discourse Context & Media, 31, 100–322. Hardaker, C., & McGlashan, M. (2016). Real men don’t hate women’: Twitter rape threats and group identity. Journal of Pragmatics, 91, 80–93. Hattaka, N., Niemi, M. K., & Välimäki, N. (2017). Confrontational yet submissive: Calculated ambivalence and populist parties’ strategies of responding to racism accusations in the media. Discourse & Society, 28(3), 262–280. Jones, E., & Giles, D. C. (under review). Women who remain in relationships with registered sexual offenders: Analysis of forum discussion. Deviant Behavior.

3  Context, History, and Twitter Data: Some Methodological… 

63

Olshtain, E., & Cohen, A. D. (1983). Apology: A Speech Act Set. In N. Wolfson and E. Judd (eds.) Sociolinguistics and language acquisition, pp.18–35. Rowley, MA: Newbury House. Pedersen, S., & Smithson, J. (2013). Mothers with attitude: How the Mumsnet parenting forum offers space for new forms of femininity to emerge online. Women’s Studies International Forum, 38, 97–106. Riggs, D. W., & Due, C. (2010). The management of accusations of racism in Celebrity Big Brother. Discourse & Society, 21(3), 257–271. Schegloff, E. A. (1988/1989). From interview to confrontation: Observations of the bush/rather encounter. Research on Language and Social Interaction, 22, 215–240. Schegloff, E.  A. (1997). Whose text? Whose context? Discourse & Society, 8(2), 165–187. Scott, K. (2015). The pragmatics of hashtags: Inference and conversational style on twitter. Journal of Pragmatics, 81, 8–20. Stever, G. S., & Lawson, K. (2013). Twitter as a way for celebrities to communicate with fans: Implications for the study of parasocial interaction. North American Journal of Psychology, 15(2), 339–354. Stokoe, E. H., & Smithson, J. (2001). Making gender relevant: Conversation analysis and gender categories in interaction. Discourse & Society, 12(2), 217–242. Stommel, W. (2007). Mein Nick bin ich! Nicknames in a German forum on eating disorders. Journal of Computer-Mediated Communication, 13, 141–162. https://doi.org/10.1111/j.1083-­6101.2007.00390.x Stubbs, D. (2019, 9 May). Danny Baker’s sacking by the BBC stinks. Campfire Bugle. Retrieved from https://campfireconvention.uk/bugle/voices/ danny-­bakers-­sacking-­bbc-­stinks Usher, B. (2015). Twitter and the celebrity interview. Celebrity Studies, 6(3), 306–321. Van Dijk, T. A. (1992). Discourse and the denial of racism. Discourse & Society, 3(1), 87–118. Wetherell, M. (1998). Positioning and interpretative repertoires: Conversation analysis and post-structuralism in dialogue. Discourse & Society, 9(3), 487–512. Zappavigna, M., & Martin, J. R. (2018). #communing affiliation: Social tagging as resource for aligning around values in social media. Discourse, Context & Media, 22, 4–12.

4 “It’s time to shift this blog a bit”: Categorial Negotiation as a Local and Cumulative Accomplishment Linda Walz

Introduction Change is a fundamental human experience: every person is likely to go through phases in their life when circumstances change and their established practices, relationships and beliefs are altered. This impacts individuals’ sense of identity (Bridges, 2004) and constitutes an experience of “It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”  Other than having been normalised for spacing, all data are presented verbatim. Alterations due to anonymisation are rendered in square brackets, and hyperlinks are underlined. 1

L. Walz (*) Leeds Trinity University, Leeds, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_4

65

66 

L. Walz

transition that may lead them to discursively negotiate its effects on them over a period of time. Research interest in identity has been growing within the social sciences (Côté, 2006), and it has been recognised that linguistic resources play a key part in identity construction (Bucholtz & Hall, 2010). How individuals make sense of transition through discursive means is a worthwhile area of study, given that today’s dynamic lifestyles subject people to a multitude of changes, and since social media have opened up new discursive spaces for such identity work. This research explores the phenomenon of expatriate blogging. In this discursive practice, individuals who have undergone transnational relocation create their own online space in the form of a blog where they reflect on their experiences of life abroad. The blogs consist of separate posts published over an extended period of time, time-stamped and appearing in reverse chronological order (Heyd, 2017). A medium-specific affordance of blogging is its audience structure. In contrast to a personal diary, a majority of expatriate blogs are publicly accessible, allowing bloggers not only to keep a record of their experiences for themselves, but also to share them with friends, family and indeed anyone else who may be interested in relocating abroad. Another affordance is temporality: because it is a practice that may span months or even years, bloggers can share their unfolding story in instalments. Whilst transition is experienced locally, this is only possible through an awareness of and orientation to what was before. Blogs, therefore, are a medium that is especially suited to engaging with the experience of undergoing personal change, which in turn makes them an attractive phenomenon for analysts aiming to explore the discursive negotiation of transition. This is aided by a further design feature: blogs are less ephemeral than some other forms of computer-mediated communication, as previous posts usually remain accessible in a chronologically ordered archive. Every post can be read separately but is also situated within a broader discursive context. The growing archive in an expatriate blog makes the passing of time tangible by visualising the duration of an individual’s relocation experience and their ongoing narrative as life abroad unfolds. As online social interaction has become firmly established in many people’s lives, so is research increasingly exploring this phenomenon. Whilst this has opened up fruitful new areas of study, it has also engendered discussions on how applicable ‘traditional’ methodologies are for the study of online data (Giles, Stommel, & Paulus, 2017; Giles,

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

67

Stommel, Paulus, Lester, & Reed, 2015). This chapter contributes to such debates through an exploration of what shape an analysis of identity transition in blogging may take and what methodological implications this may have. These revolve around two issues: firstly, the analysis shows that the context of expatriate blogging provides for certain membership categories serving as an inferable discursive resource for identity work such that they do not require an explicit mention. Secondly, because the blogs’ archives grant access to this discursive practice over an extended period of time, researchers are able to explore how categorial work accumulates and fluctuates with individuals’ increasing lived experience. Exemplified on the basis of expatriate blogs, then, the analysis illustrates that blogging is a fruitful medium for studying the dynamic nature of identity construction. In particular, through an adaptation of Membership Categorisation Analysis (henceforth MCA), it demonstrates how this approach can capture both local and cumulative categorial work. Not only does this result in a more thorough understanding of how individuals use online spaces to critically engage with transition in a dynamic world, but also does this serve as an example of how ‘traditional’ methodologies may be applied to account for some of the defining features of online data. The chapter first provides an overview of how research has engaged with the construction of identity in transition. It then outlines the data and analytic framework, distinguishing between category fit and category change. The analysis explores local categorial negotiation, before arguing for transcending the local. The discussion focuses on the challenge of non-explicit mentions of categories compounded with the largely monologic nature of blogging, and on how categorial work is both local and cumulative. The chapter concludes by arguing for more longitudinal work on categorial negotiation in online data.

Identity and Transition Much recent work within the social sciences has come to adopt a constructionist view of identity as the outcome of social positioning and hence emergent from discourse and social interaction (Bucholtz & Hall, 2005). In this view, identity is not a stable essence but a dynamic process, and an understanding of how identity is constructed depends on a close

68 

L. Walz

analysis of the semiotic resources that individuals draw on. Of particular interest are moments of instability, when individuals express a sense that their life is undergoing a change that impacts on them as a person. One way in which individuals engage with undergoing transition is by sharing their experiences online, such as in personal blogs, vlogs or other forms of social media. This not only allows them to receive comments with advice or empathy from readers, but also, in the first place, enables them to make sense of their unfolding experience by reflecting on it in discursive form. Examples include such diverse aspects of life as suffering from cancer (Page, 2012), weight loss (Leggatt-Cook & Chamberlain, 2012) and transgender transformations (Jones, 2019; Raun, 2015). A particular transitional experience negotiated in blogs is engendered by transnational mobility. In such blogs, individuals reflect on how contact with the other in the form of foreign people and places has “transformational effects” on them (Bosangit, Hibbert, & McCabe, 2015, p. 12) and leads to “feelings of change and transition” (Snee, 2014, p. 160). Moving abroad constitutes a “biographical disruption” (Frank-Job & Kluge, 2012, p. 49, author’s translation), and the impact this has on individuals’ sense of identity is evident in the vast number of expatriate blogs that engage with this experience (Walz, 2018). Whilst identity has been analysed through a wide range of approaches, MCA is a framework that particularly lends itself to an exploration of identity from individuals’ own point of view (Fitzgerald & Housley, 2015; Sacks, 1992). Its central tenet is that members in a society can draw on categorisation as a resource to make sense of their social reality. Categories typically perform certain actions and exhibit certain characteristics as well as “rights, entitlements, obligations, knowledge, attributes and competencies” (Hester & Eglin, 1997b, p.  5). These are captured under the umbrella term of category-bound predicates. Yet rather than being fixed cultural properties, predicates can change from one interactional context to the next, as categories are locally occasioned. Categories can be collected in devices. In Sacks’ (1992) famous example, “The baby cried. The mommy picked it up” the categories ‘baby’ and ‘mommy’ are part of the device ‘family’. This is the mechanism by which the mommy is understood to be the mother of this particular baby, and by which the picking up is seen as a reaction to the crying, since it is an activity—or more broadly a predicate—which is here expectably bound to the

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

69

category ‘mommy’ just like crying is bound to ‘baby’. Categorial work is dynamic and constitutes ‘culture in action’ (Hester & Eglin, 1997a). Because of its focus on members’ own sensemaking practices, MCA is an approach that beyond its key analytic concepts remains open to a range of applications: with regard to online data, research has drawn on MCA to explore how new members join a community and in the process negotiate its expectations and the predicates of such a new categorial identity (Giles & Newbold, 2013; Stommel & Koole, 2010). Indeed, its analytic flexibility lies “at the heart of and success of MCA” (Fitzgerald, 2012, p. 308), and it is put to creative use in this research, as discussed below.

Data and Method: Categorial Negotiation Sampling Expatriate Blogs Individual motivation for relocation encompasses a range of aspects such as work, personal relationships and lifestyle choices. Yet none of the research participants moved to escape hardship, which places them within relatively privileged migration (Amit, 2007). As many of them self-­ identify as expatriates, this term is adopted here despite some negative connotations, such as that of “classed Western whiteness” (Leonard, 2010, p. 2). Indeed, all participants have included their blog in so-called expatriate blog directories, websites that provide links and brief descriptions to blogs on this topic. This positions them within a community of fellow expatriates and other interested readers. Whilst several bloggers emphasise that they began blogging to record their experiences abroad for themselves and to provide their family and friends ‘back home’ with updates, some of them comment on having unexpectedly attracted a wider readership of people they have never met. Although this analysis focuses on the blog posts and does not include any reader comments, awareness that the blogs have a readership is nevertheless important. Whilst bloggers are concerned with providing insight into their lived experience and establishing authenticity (Walz, 2020), this medium also contains a performative element (Davies, 2013). The analysis shows

70 

L. Walz

how expatriate bloggers repeatedly orient to readers, for example by providing hyperlinks to previous posts and including metacomments on the content and frequency of their blogging, such that the audience structure becomes part of how identity work is discursively accomplished. Expatriate blogs are in the public domain, yet reflections on transnational relocation convey a personal experience. Bloggers gave their consent to participate in this research. They were approached following a survey of two expatriate blog directories resulting in 381 accessible blogs. All blogs considered for this research were active at the time of data collection in April 2015, contained content on life abroad, and were written by individuals in English for personal rather than professional purposes. Furthermore, all blogs were begun around the time of the bloggers’ relocation and made reference to this, and were continued for at least one year following the move abroad. The data span 12 blogs, comprising all posts from the beginning of each blog until the one-year anniversary of life in England, which were already part of the blogs’ archives at the time of data collection.

 nalytic Framework: Category Fit A and Category Change This research draws on MCA as an analytic framework (Fitzgerald & Housley, 2015; Hester & Eglin, 1997a; Sacks, 1992), yet with an adaptation. The focus lies on instances where participants negotiate categorial matters. Negotiation is here understood as working through a categorial issue by engaging with the category and its predicates in a way that involves interactional work. It constitutes more than a mere mention of a category: it requires a linguistic effort. Categorial negotiation then comprises instances when members can be seen to make categories and predicates “demonstrably relevant” (Hester & Hester, 2012, p. 567) such that working through what they entail becomes the focus of what members are engaged in at that particular moment. Categorial negotiation is a means of doing identity work. Exploring it therefore grants insight into how members deal with the experience of transition in their blogs. Whilst MCA has always been concerned with how categories are occasioned and used, the contribution of this research lies in the distinction of how such working through happens. Negotiation was found to occur

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

71

in two manners: participants either engage with the fittedness of a category and its predicates, or describe and reflect on a perceived change of category (Walz, 2018). The distinction can be a matter of emphasis, as sometimes the two are closely related, such as when a perceived lack of fit leads to a negotiation of change. In category fit negotiation, the category is examined internally: participants focus on what it involves and consequently whether they are a fitting member. For instance, participants may construct the category or the category-predicate relationship as problematic or in need of an explicit discussion. They may perceive an incongruity between the category and the predicates that are expectably bound to it, or may address a notion that they see other people to hold and that they disagree with. The following example illustrates an engagement with category fit.

Example 1  [Emily—one month before relocation]1 So the talk will soon be ‘we are leaving on such’n’such a date’ and it

will come up very fast indeed. Such is the life of an expat.

So much waiting for confirmation of your move, then wham -bam you’re gone but hopefully not forgotten

Following a long period of waiting for confirmation of their move, Emily suddenly finds herself plunged into preparations and time passes quickly. She orients to the fact that this may appear unusual to somebody who has never moved abroad, yet explains that having to wait for a long time only to then leave on short notice is indeed a typical experience for category members: ‘Such is the life of an expat.’ Whilst for Emily this category-­ predicate relationship is clear, she shows herself aware that it may be problematic for readers, therefore explicitly tying the predicate to the category (Reynolds & Fitzgerald, 2015). The fit between category and predicate is here treated as in need of explanation. In contrast, category change negotiation involves an external ‘working through’ of category membership. Rather than what the category involves,

72 

L. Walz

participants discuss a change of membership from one category to another, which may be past, ongoing or anticipated. For instance, participants may juxtapose past and present regarding their category membership, pointing out that they no longer see themselves as a member of a category they used to identify with, or that they have taken on membership of a new category. An example of a category change negotiation is Claire’s reflection on how living abroad has changed her. Example 2  [Claire—10 months after relocation] In the last year, my travel life, business life, and love life have all gone through dramatic changes. Airports look different to me now.

Instead of moving at breakneck speed towards departure gates or rental car pickups, I travel for love. Flying these days is about reuniting with family and friends or exploring places I’ve never been before.

Before living abroad, Claire’s motivation for travelling was work-related, whereas now, she travels to see family and friends. Consequently, she has become a different kind of traveller. Within the device ‘people who travel’, she is no longer a business traveller, but a traveller for love. Hence, her relocation has caused her to experience and reflect on a change of category. The distinction between category fit and category change negotiation addresses how participants engage with categorial matters. Yet this alone does not take into account what categories are negotiated. This research focuses on categorial negotiation that is treated by participants as linked to transnational relocation. The various categories that an individual may be a member of due to having moved abroad constitute a collection, which this research refers to as the transnational relocation device. Although participants themselves do not name this device, they clearly orient to it as relevant for many instances of categorial negotiation. Transnational relocation is, therefore, an omnirelevant device in expatriate blogs (Sacks, 1992): it is not always evoked, but it may potentially be evoked at any time, and when it is, it makes categorial negotiation readable as pertaining to it. The following analysis aims to illustrate the value of exploring both category fit and category change negotiation within the transnational relocation device, first with a focus on the local and then with the objective of transcending it.

73

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

Analysis Local Categorial Negotiation Participants engage in categorial negotiation throughout their blog. Nevertheless, as their lived experience of relocation grows, this can engender a more extensive focus on the matter. For instance, Sarah devotes an entire post to what being an expatriate means. Example 3  [Sarah—4 months after relocation] As An Expatriate Would they have yachts? Second and third homes all over the world? How would I fit in? Like the first day in a new school, I wondered what my future expat community would be like. […]

I quickly fell into the naive trap of asking others, ‘Where are you

from?’ Sounds innocent, but actually it is quite loaded and can grate on people’s nerves. Let me show you why… One lovely friend is a mum from Lebanon (hi sweetie!). One child was

born in Belgium; one in Ireland. Her husband is Dutch and commutes to

Switzerland from London. I ask you, how are they supposed to answer

the question of where their family is ‘from’ in just a few words? Exactly. Typically a person’s home and car often tell a lot about a person’s economic status. But here, if a middle class family has a robust expat package, they could be living like kings. In contrast, a wealthier family may be living in a small home because they have to pay their

own way. Lives can be dramatically effected by the exchange rate too. If one comes

from

Finland,

everything

would

appear

really

cheap,

and

one

could live a more extravagant life. Yet if one comes from a country

with a weak currency, s/he may have a modest house, travel little, and only buy the necessities. As you can see, natural identifiers of ‘who we are’ can be shed in the

expat world. What crowd you ran with, what groups you joined don’t matter here. It is refreshing to just show up as you are. When else does that happen?

In case you were wondering, I haven’t met anyone with a yacht. It doesn’t seem to be in anyone’s expat package :)

74 

L. Walz

Sarah first describes her feelings before becoming a member of the category ‘expat’. She positions her past self as uncertain about what other expatriates might be like, searching for category predicates and worrying that she might not ‘fit in’. She then elaborates on what she has learnt since becoming a member, now being able to speak ‘As An Expatriate’ based on her lived experience. Negotiating category fit, she discusses a number of unreliable predicates: according to her experiences, expatriates’ origin and economic status cannot easily be pinpointed and impressions may be misleading. She concludes that upon becoming expatriates, individuals ‘shed’ such ‘natural identifiers’ or predicates and are welcomed into the community the way they are: the category comprises individuals from different backgrounds that cease to matter once membership is attained. The function of the post is not only to make an engagement with category fit explicit, as the title suggests, but it also allows Sarah to demonstrate her increased understanding of what category membership entails, and her enhanced position to discuss it based on her lived experience. This change is established retrospectively through a juxtaposition of her current understanding of the category with what she had anticipated. Her former assumptions are repositioned as somewhat absurd and humorous, visually enforced through the concluding smiley. Sarah’s post is thus an example of how categorial negotiation, whilst local, is nevertheless also achieved cumulatively by building on, or in this case distancing from, previous categorial work. Like category fit, negotiation of category change can equally be built on retrospectively modifying previous positions. Example 4   [Jessica—6 months after relocation] 6 Months Check Up!!! I’ve

been

having

so

much

fun

I

actually

missed

our

six

month

anniversary day! March 28th marked 6 months for me in England....and

[husband] is 3 weeks ahead of me. I took a peek back at our 3 month

post and wow, what a complete 180! I think that I was actually writing that I was happier than I really was, maybe trying to cheer myself up a bit? Now that we have settled into our house and our surroundings, and the sun has started to come out more often, things are sooooo much

better. It is totally true, give yourself 6 months in your new country and you will be fine :)

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

75

Jessica’s focus on the passing of time is evident as she counts the months of her life as an expatriate. Juxtaposing her state at three and six months enables her to negotiate how the experience of being an expatriate changes over time. As in Example 3, then, predicates change in the course of category membership. For the categorial negotiation of her change from struggling to being happy to succeed, Jessica needs to backtrack on her previous post, admitting that she had not been completely honest about her emotional state. Yet whilst the orientation to change is made explicit as ‘a complete 180’, the category itself remains unnamed. Nevertheless, it emerges both through Jessica’s orientation to having been in England, her ‘new country’, for six months, and by virtue of the omnirelevance of transnational relocation as the device that organises her categorial negotiation. In her category change negotiation, Jessica draws on the affordances of blogging by hyperlinking to two previous posts. At three months, she posted an ‘expat checkup’ taking stock of her feelings about life abroad, and in a later post she discussed how she was slowly regaining her confidence in her new surroundings. She displays an expectation that readers either will be familiar with these posts or may want to follow the hyperlinks to read up on them. The context on the back of which the present categorial negotiation occurs is thus made demonstrably relevant. This shows that members themselves may orient to the fact that previous categorial negotiation matters for an understanding of their current one. What is achieved locally is therefore also a cumulative accomplishment. An example that brings negotiation of category fit and category change together is a post by Leah, in which she notifies readers of changes she is making to her blog. Example 5 [Leah—10 months after relocation] Also in terms of shift, it also feels like it’s time to shift this blog a bit. I suppose you can be a tourist forever, but if you want to

live somewhere permanently, it becomes detrimental. I’ll continue to write (and hopefully a little more regularly), but rather than being an intrepid reporter for all things *different* between the US and the

UK, I’d like to just write about life as it happens. Things that surprise, comfort, amuse, confuse even. It’s not going to devolve into

a Dear Diary thing (I hope), but it just feels more and more awkward to write in the style I have been for the last few months.

76 

L. Walz

Leah presents the imminent changes as occasioned by a categorial matter. She treats her reporting on all the sociocultural differences that she has been experiencing as a predicate of the category ‘tourist’, of which she has been a member. At the current stage of her relocation experience, however, she has become more accustomed to life in England and she has decided to stay. Therefore, remaining a tourist would be ‘detrimental’. She is now distancing herself from her practice of scrutinising differences, pointing out that it has become increasingly ‘awkward’. This is a means of expressing ongoing category change, on the basis that predicates which used to work well for her no longer seem to fit. Just as she rejects being a tourist, she now rejects behaving in a way that she treats as bound to the category. What used to be appropriate is no longer so.

Transcending the Local So far the analysis has shown that individuals negotiate category fit and category change by establishing comparisons between their previous and current positions. Change is made relevant by invoking it, yet it is established locally. But thanks to previous posts being archived, change can also be analysed with a longitudinal approach by comparing posts by one person over a period of time. Indeed, it is worth considering Leah’s negotiation in Example 5 in relation to her previous engagement with categorial matters. At one week into her life abroad, she posts the following. Example 6 [Leah—one week after relocation] I would like to apologise to the now local ( :D ) people in my life as the anthropologist in me has been having a field day. Coming to a new place

you’ll

always

have

new

experiences

to

process,

(sometimes

happens just crossing America) but instead of just shrugging it off with a “well that’s odd,” I have an intense curiosity to find out why things work that way. It’s my motto to always look at life and new things as interesting, never weird or “not like home.” The world works around things in many different ways, and societies are fascinating because of it.

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

77

Leah orients to her ‘intense curiosity’ about the new experiences she is making abroad as an intrinsic part of herself. This is ‘the anthropologist in [her]’, an aspect that is bound to her in the sense of a character trait. It is so strong as to be her ‘motto’ that ‘always’ shapes her take on things and leads her to engage with her new surroundings in an inquisitive way to understand them in detail. It is only in her later post (Example 5) that such behaviour is recast as bound to a tourist, a category from which Leah aims to distance herself. Unless participants themselves orient to this change, such retrospective modifications (Silverman, 2012, p. 334) only become evident through a longitudinal approach. Considering more than the local instance of categorial negotiation therefore opens up new analytic avenues, and blogging is well suited for such an endeavour. This is further exemplified below. The following three extracts from Emily’s posts across a time span of over ten months show how the predicates of an expatriate are occasioned differently at different stages of her experience. Emily orients to the upcoming move several months prior to her relocation, eventually starting to count down the days. Example 7 [Emily—13 days before relocation] 13 sleeps We’ll be THERE instead of HERE. While things don’t usually change much day to day, these next 13 days will be unusual cos every day something is happening or will happen

that will influence our opinion thereby changing plans. […] I thought it might be fun to blog every day, so you can experience

this final phase of ‘leaving’ and of course, continue it with the

‘arrival’. Life won’t be ‘normal’ again for us for at least 6-8 weeks.

Hope you enjoy the ride.

Emily emphasises that the 13 days before their impending move will be ‘unusual’ as they are making their final preparations for leaving. This involves a level of uncertainty because the details of their relocation may still change. Emily shows herself aware of how moving to England will impact on them, and again no explicit naming of a category is necessary.

78 

L. Walz

Instead, it is pinpointed in physical space, being ‘THERE instead of HERE’, as well as invoked through the change this creates for individuals and their circumstances as they are undergoing the experience of ‘leaving’ and ‘arrival’. Yet beyond this, transnational relocation entails a lack of normality, which Emily expects to endure for considerably longer than the actual physical move. Such lack of normality is here treated as bound to the category ‘expatriate’ at that stage of relocation, and as providing a lot of content for sharing, which accounts for the change in Emily’s blogging practice. As announced, she proceeds to blog daily, counting down the number of remaining ‘sleeps’ in the titles of her posts. However, after two months in England, Emily presents a different picture of her situation. Example 8 [Emily—2 months after relocation] My ‘need’ to blog has lowered, in as much as life has taken on a normalness that you experience every day. The ups and downs of the move have gone til next time. The mini drama’s of finding a house, school ect have gone, til next time.

Emily has overcome the initial challenges of relocating and arranging the necessary logistics. She treats such difficulties as bound to an expatriate at this stage of relocation, as she anticipates them for any subsequent moves. Yet she has moved onto a different stage, in which ‘normalness’ and an everyday routine are back. This does not mean, however, that she no longer identifies as a category member. Rather, she orients to the category as evolving in the sense that different predicates are expectably bound to it depending on whether relocation is imminent or has occurred sometime previously. This engenders continued category fit negotiation, once more accomplished implicitly, where ‘til next time’ refers to transnational relocation as an omnirelevant device. As a consequence of the normality that she once more experiences, Emily posts less often, which is evident to any regular reader even before she accounts for it. This shows that the intricate link to time not only affects what predicates members treat as bound to the category, but also impacts their blogging practice. Accounting for

79

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

this in turn enforces the point that what being an expatriate means to individuals changes over time. Once more, then, the fact that blogs are written and read over a period of time affords discursive identity work. The complexity of the category ‘expat’ is also evident at a time when Emily has spent ten months in England. She writes about the challenges she is faced with, followed by this category fit negotiation. Example 9 [Emily—10 months after relocation] Having

waited

Shouldn’t I?

a

while

before

posting

this

-

thinking

should

I?

Don’t people only want to read the good stuff? Like a friend who travelled to the UK, two European cities, 2 cities in the USA with her

kids (husband stayed behind to work); or several friends who returned ‘home’ for between 4-6 weeks of summer fun; or others who cruised or a

few who managed two different vacations for two weeks each location! Now that’s expat life, isn’t it!

Listing the exciting holidays that other expatriates have enjoyed would appear to treat a life of leisure as category-bound. ‘Now that’s expat life, isn’t it!’ makes this link explicit but only to challenge it: Emily orients to the fact that people may assume such predicates for the category, which she disagrees with based on her own experiences as a category member. Whilst the predicates do not apply to Emily, this does not make her question whether she truly is an expatriate but rather it leads her to negotiate what the category entails. This means that she does not treat common understandings of the category as “protected against induction” (Sacks, 1992, p. 336): if she as a category member experiences life as an expatriate differently, then the predicates that expectably hold from non-­ members’ point of view are questionable. Having been an expatriate for an extended period of time enables her to negotiate category fit in this way, building her understanding as having grown from personal experience. This is comparable to Sarah’s emphasis on her enhanced insight based on category membership (Example 3). Experiencing life as a category member hence trumps seemingly common assumptions held by non-members about what predicates are bound to the category. At the

80 

L. Walz

same time, breaking what is perceived to be readers’ expectations is treated as accountable: Emily acknowledges that she was hesitant to write about matters other than ‘the good stuff’. Viewing this post in the context of previous ones reveals the dynamics of categorial negotiation. Whilst Emily previously explained that the new normality in her life meant that she did not feel a need to post as frequently as before (Example 8), the lack of excitement is now treated as problematic. A comparison of posts written by the same person at different points in time (Examples 5 and 6, and 7 to 9) hence shows that a member’s categorial negotiation does not progress in a linear fashion. As the experience of being an expatriate evolves through different stages (Walz & Fitzgerald, 2020), so do members continuously negotiate what predicates the category entails, and these differ when compared across a series of local instances. Whilst in some cases individuals themselves orient to change through juxtapositions (Examples 3 and 4), change also emerges from an exploration that moves beyond the local to consider the effect of cumulative categorial work.

Discussion The distinction between category fit and category change negotiation emerged from an exploration of how participants engage with categories in the course of their relocation and life abroad, and the above analysis aims to demonstrate the fruitfulness of such an analytic focus. However, drawing on MCA to explore online data encompasses a number of challenges. These revolve predominantly around the identification of categorial negotiation and the analysis of this as both local and cumulative. This section briefly describes these challenges and the approach taken towards them, with the aim to inspire further discussion of these matters as part of the continued exploration of how research may best approach online data and their specific characteristics. Categorial negotiation is not straightforward to identify. This is because a distinction is necessary between a categorial negotiation and an explicit mention of a category. Whilst the naming of a category arguably gives it a certain presence, individuals may not treat it as noteworthy nor engage

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

81

with it any further. At the same time, not every instance of categorial negotiation does in fact name a category term: as shown above, it may involve solely a mention of predicates that the individual in this local instance treats as bound to the category, whereby the category is expectably inferable by virtue of the omnirelevance of the transnational relocation device. One challenge of analysing blogs lies in their nature as largely monologic and often “primarily intra-subjective” data (Brake, 2012, p.  1072). Consequently, they tend to feature relatively few comments that would provide evidence for recipient orientation to a category, as may be possible for online data with a structure more reminiscent of conversational turn-taking such as forum posts (Giles & Newbold, 2013; Stommel & Koole, 2010). Yet whilst readers’ uptake is not evident, this is not meant to suggest that blogs do not involve a degree of recipient design. First of all, individuals use them as spaces to discuss matters pertaining to life abroad, which is in line with raising such expectations through including their blog in an expatriate directory. Further, bloggers orient to what they perceive readers to expect of them, as Examples 8 and 9 illustrate. An analysis of blogs therefore needs to consider their audience, but its validity does not depend on recipients’ confirmation if it explores bloggers’ use of categorial resources. The issue of how legitimately categories can be analysed if participants do not explicitly mention them may be compounded in the case of online data, but it is a wider-reaching matter and has in fact served as a point of criticism of MCA. Whilst participants may treat certain predicates as obviously bound to a category, this ‘obviousness’ constitutes a problem for researchers, who need to be able to show in an ethnomethodological fashion that categories emerge from participants rather than being introduced by the analyst (Schegloff, 2007). However, if researchers restricted themselves to exploring explicit mentions, they could not gain an understanding of the rich identity work participants accomplish through categorial resources. Instead, “the appeal (and danger) of MCA is to try to unpack what is apparently unsaid by members and produce an analysis of their subtle categorization work” (Stokoe, 2012, p. 282, original emphasis). This chapter supports the view that MCA must move beyond the explicit, especially as ambiguity does not appear to constitute a problem for participants—indeed, ethnomethodological research has shown them

82 

L. Walz

to work with the notion of ‘possibles’, different ways in which a particular utterance can be understood (Schegloff, 2006). Expatriate bloggers have established transnational relocation as an omnirelevant device by the very act of devoting their blog to it. Therefore, this research takes the approach that in cases of engagement with an issue that can expectably be read as a categorial negotiation within this device, in the spirit of Sacks’ (1992) viewer’s maxim, such instances will be seen, or rather read, that way. Another issue arises from the finding that categorial negotiation occurs locally, but also with respect to, in a larger context, the transition over a period of time. All negotiation is collected in the blog and is thus visibly contextualised as pertaining to the individual’s experience of transnational relocation. Therefore, a longitudinal understanding of categorial work as building on or moving beyond previous categorisation is not only accessible to researchers as a benefit of this type of data, but can also be presupposed by bloggers for their readership, given that categorial negotiation is situated in time through the affordances of blogging discussed in the introduction. Bloggers themselves monitor the passing of time by counting down the days to relocation (Example 7) and celebrating milestones abroad (Example 4). The situatedness in time is thus a members’ concern as an intrinsic aspect of the building up of lived experience, and bloggers orient to previous posts and sometimes display an expectation that readers are familiar with them. However, this cumulative negotiation of category membership afforded by blogging challenges the locus of analysis traditionally adopted within MCA: because categorial negotiation needs to be understood with regard to previous categorial work, the analysis whilst exploring the local also needs to consider how participants reiterate, modify or distance themselves from previous categorial work. The building of collections has been proposed by Stokoe (2012) as a way of progressing MCA as a methodological approach, and whilst she has demonstrated this with collections of how categories do certain actions, it is an equally promising endeavour in terms of tracing a categorial phenomenon through time, such as in this research the negotiation of identity transition pertaining to transnational relocation. Indeed, recent work has shown the scope of a longitudinal approach to ethnomethodological research using the example of

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

83

conversation analysis (Pekarek Doehler, Wagner, & González-Martínez, 2018). Due to their chronological order and persistence, blogs offer particularly suitable data for such an investigation, and the analysis has demonstrated that categories and predicates are flexible tools and not restricted to exploring single instances. This research therefore argues that analysing the local with a longitudinal view constitutes a promising new path for MCA.

Conclusion This research has shown that participants make sense of their relocation experience by engaging in categorial negotiation. Such identity work can take the form of category fit and category change negotiation, emphasising either what predicates a category expectably entails and how participants themselves fit in, or how their category membership and predicates have changed. Both are accomplished locally, and one means of doing so is by establishing comparisons between the past and the present. However, a longitudinal approach shows that categorial negotiation changes over time, with participants building on or distancing themselves from previous positions. This insight only emerges by adopting a perspective that exceeds the local and considers the wider context in which such negotiation occurs. New media have given rise to a multitude of online data, yet traditional methods may be fruitfully applied to an investigation of such discursive practices. Sacks’ original approach was open to data as different as children’s stories and newspaper articles and hence “broadly multi-modal and ethnographic” in nature (Fitzgerald, Housley, & Rintel, 2017, p. 52), and such an open-minded approach is likely to further benefit MCA in the future. Yet online data present new methodological and analytic challenges. In the case of expatriate blogs, these consist of identifying categorial negotiation in largely monologic data and exploring local instances whilst also taking into account how participants draw on medium-­specific affordances to achieve cumulative categorial work. The analysis has aimed to provide an illustration of how such matters may be addressed. Not

84 

L. Walz

only does this allow analysts to explore the phenomenon of publicly sharing a transitional experience in stages as it unfolds, but it ultimately strives to make a methodological contribution to MCA and to inspire further longitudinal research on how categorial work is accomplished both locally and cumulatively.

References Amit, V. (Ed.). (2007). Going first class? New approaches to privileged travel and movement. Oxford: Berghahn Books. Bosangit, C., Hibbert, S., & McCabe, S. (2015). “If I was going to die I should at least be having fun”: Travel blogs, meaning and tourist experience. Annals of Tourism Research, 55, 1–14. Brake, D.  R. (2012). Who do they think they’re talking to? Framings of the audience by social media users. International Journal of Communication, 6, 1056–1076. Bridges, W. (2004). Transitions: Making sense of life’s changes (2nd ed.). Cambridge, MA: Da Capo Press. Bucholtz, M., & Hall, K. (2005). Identity and interaction: A sociocultural linguistic approach. Discourse Studies, 7(4–5), 585–614. Bucholtz, M., & Hall, K. (2010). Locating identity in language. In C. Llamas & D.  Watt (Eds.), Language and identities (pp.  18–28). Edinburgh: Edinburgh University Press. Côté, J. (2006). Identity studies: How close are we to developing a social science of identity? An appraisal of the field. Identity, 6(1), 3–25. Davies, J. (2013). Discourse and computer-mediated communication. In K. Hyland & B. Paltridge (Eds.), The Bloomsbury companion to discourse analysis (pp. 228–243). London: Bloomsbury. Fitzgerald, R. (2012). Membership categorization analysis: Wild and promiscuous or simply the joy of Sacks? Discourse Studies, 14(3), 305–311. Fitzgerald, R., & Housley, W. (Eds.). (2015). Advances in membership categorisation analysis. London: Sage. Fitzgerald, R., Housley, W., & Rintel, S. (2017). Membership categorisation analysis: Technologies of social action. Journal of Pragmatics, 118, 51–55. Frank-Job, B., & Kluge, B. (2012). Die kooperative Konstruktion von Identitäten im virtuellen Kommunikationsraum des Web 2.0: Blogs zum Thema Migration nach Québec. In A. Gerstenberg, C. Polzin-Haumann, &

4  “It’s time to shift this blog a bit”: Categorial Negotiation… 

85

D. Osthus (Eds.), Sprache und Öffentlichkeit in realen und virtuellen Räumen (pp. 47–78). Bonn: Romanistischer Verlag. Giles, D., Stommel, W., & Paulus, T. (2017). The microanalysis of online data: The next stage. Journal of Pragmatics, 115, 37–41. Giles, D., Stommel, W., Paulus, T., Lester, J., & Reed, D. (2015). Microanalysis of online data: The methodological development of “digital CA”. Discourse, Context & Media, 7, 45–51. Giles, D. C., & Newbold, J. (2013). ‘Is this normal?’ The role of category predicates in constructing mental illness online. Journal of Computer-Mediated Communication, 18(4), 476–490. Hester, S., & Eglin, P. (Eds.). (1997a). Culture in action: Studies in membership categorization analysis. Washington, D.C.: International Institute for Ethnomethodology and Conversation Analysis and University Press of America. Hester, S., & Eglin, P. (1997b). Membership categorization analysis: An introduction. In S. Hester & P. Eglin (Eds.), Culture in action: Studies in membership categorization analysis (pp.  1–23). Washington, D.C.: International Institute for Ethnomethodology and Conversation Analysis and University Press of America. Hester, S., & Hester, S. (2012). Categorial occasionality and transformation: Analyzing culture in action. Human Studies, 35(4), 563–581. Heyd, T. (2017). Blogs. In C. R. Hoffmann & W. Bublitz (Eds.), Pragmatics of social media (pp. 151–171). Berlin: Mouton de Gruyter. Jones, L. (2019). Discourses of transnormativity in vloggers’ identity construction. International Journal of the Sociology of Language, 256, 85–101. Leggatt-Cook, C., & Chamberlain, K. (2012). Blogging for weight loss: Personal accountability, writing selves, and the weight-loss blogosphere. Sociology of Health & Illness, 34(7), 963–977. Leonard, P. (2010). Expatriate identities in postcolonial organizations: Working whiteness. Farnham: Ashgate. Page, R. E. (2012). Stories and social media: Identities and interaction. Abingdon: Routledge. Pekarek Doehler, S., Wagner, J., & González-Martínez, E. (Eds.). (2018). Longitudinal studies on the organization of social interaction. London: Palgrave Macmillan. Raun, T. (2015). Video blogging as a vehicle of transformation: Exploring the intersection between trans identity and information technology. International Journal of Cultural Studies, 18(3), 365–378.

86 

L. Walz

Reynolds, E., & Fitzgerald, R. (2015). Challenging normativity: Re-appraising category bound, tied and predicated features. In R. Fitzgerald & W. Housley (Eds.), Advances in membership categorisation analysis (pp.  99–122). London: Sage. Sacks, H. (1992). Lectures on conversation (G.  Jefferson, ed., vol. 2). Oxford: Blackwell. Schegloff, E. A. (2006). On possibles. Discourse Studies, 8(1), 141–157. Schegloff, E.  A. (2007). A tutorial on membership categorization. Journal of Pragmatics, 39(3), 462–482. Silverman, D. (2012). Beyond armed camps: A response to Stokoe. Discourse Studies, 14(3), 329–336. Snee, H. (2014). A cosmopolitan journey? Difference, distinction and identity work in gap year travel. Farnham: Ashgate. Stokoe, E. (2012). Moving forward with membership categorization analysis: Methods for systematic analysis. Discourse Studies, 14(3), 277–303. Stommel, W., & Koole, T. (2010). The online support group as a community: A micro-analysis of the interaction with a new member. Discourse Studies, 12(3), 357–378. Walz, L. (2018). The linguistic construction of identity in transnational relocation narratives: examining discursive practices in expatriate blogs. PhD thesis, University of Leeds. Retrieved from http://etheses.whiterose.ac.uk/20853/. Walz, L. (2020). ‘You wanted to know about expat life’: Authenticating identity in blogs about transnational relocation. Text & Talk, 40(3), 377–397. Walz, L., & Fitzgerald, R. (2020). A stranger in a foreign land: Identity transition in blogs about transnational relocation. Discourse, Context & Media, 36, 1–9.

5 The Radio Host Cried, the Facebook Users Identified: Crying as an Action Linked to ‘good people’ Elisabeth Muth Andersen

Introduction When witnessing someone crying, people tend to respond by comforting the person crying and displaying support and empathy using a range of linguistic actions and embodied resources, including touch and gesture (Hepburn & Potter, 2007, 2012; Kupetz, 2014). These verbal and embodied actions are produced on a turn-by-turn level with systematic sequential positionings in the course of an activity (Kupetz, 2014). Through their design and positioning following a turn-at-talk and/or an “It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

E. M. Andersen (*) University of Southern Denmark, Odense, Denmark e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_5

87

88 

E. M. Andersen

embodied action, responses such as facial expressions, reaction tokens, evaluations and understanding displays accomplish communicative functions such as displays of an affective stance, understanding, alignment and empathy in mundane talk-in-interaction (Kupetz, 2014). Interactional studies of crying in institutional settings show that professionals tend to handle crying somewhat differently, orienting to institutional goals of the conversation and asymmetries between participants, for example, by treating crying as a disruption of the conversation by using so-called take-your-times (Hepburn & Potter, 2007, p.  89) that license the disruption and thereby work as an affiliative action, or by offering a formulation of the mental state of the person crying. Some types of online social interaction afford the possibility of displaying intense emotions such as crying using embodied resources through using a video format that may be witnessed by other people that are not co-present. These witnesses may respond by commenting on the video at a later point in time to the actual occurrence of the crying or the witnessing of it. In such contexts, the participant configurations are often very different than in ordinary face-to-face-interaction in the sense that the interaction may be public and include an indefinite number of potential participants that may participate through various means such as watching the video, liking or sharing the video or commenting it (Boyd, 2014; Dynel, 2014). Since the crying is not ongoing when users comment on a video in which crying occurs, the question is how crying may be responded to in such a context. How may social norms related to crying be managed by users who respond to a public video that includes crying, and what is accomplished thereby? In this chapter I analyze comments made in response to a public video on a Facebook site. The video shows a radio host who starts crying while broadcasting. The comments are overwhelmingly supportive and include positive assessments and personal stories. Using Membership Categorization Analysis (MCA), I show how users account for and identify with the crying by making reference to common-sense-knowledge about people and what kinds of conduct particular types of people engage in. More specifically, I show how person identifications (Hauser, 2011) and action formulations (Schegloff, 2007) construct the situated crying in the video as an example of a generalized pattern of behavior linked to ‘good people’. I show instances of how this praise occurs in comments in which Facebook users acknowledge a disposition to react to the crying in a potentially quite unexpected

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

89

and problematic way, namely by laughing. I also show how indications of personal experiences are used to claim membership of the category of people who cry in particular situations (i.e., category-bound activities; Stokoe, 2012, p. 281). Users thereby praise the radio host and identify with her using personal stories, while avoiding overt self-praise. In fact, the analysis shows that the stories are built using a range of devices, including humor, to downplay the stories as instances of self-praise.

Methodology: Membership Categorization Analysis There are different approaches to investigating how people deal with witnessing emotional expressions such as crying. Models of emotion expressions that are grounded in attribution theory provide explanations for observer responses with reference to cognitive processes such as “attributions inferred from characteristics of the crying episode and the characteristics of the observer” (Becker, Conroy, Djurdjevic, & Gross, 2018, p.  133). Such theoretical models are often not grounded in naturally occurring interaction, (e.g., Becker et al., 2018; Culpeper, 2001; Yzerbyt, Dumont, Wigboldus, & Gordijn, 2003), and do not offer insights into how actions such as crying responses are accomplished interactionally, and how attribution may be oriented to and used by participants as a part of “social life, or ‘norms-in-action’” (Housley & Fitzgerald, 2009, p. 346) In this chapter I employ MCA for this purpose. The outcome of such analyses is accounts of categorization (Hauser, 2011) or categorial practices (Stokoe, 2012), that is, the categorization work that participants carry out “in relation to the local accomplishment of social and moral organizational order” (Garfinkel, 2002, cited in Housley & Fitzgerald, 2009, p. 346). In order to do so, analyses, drawing on Sacks’s (1992) ground-breaking work, attempt at showing how participants build an ‘apparatus’ of categories of person and related actions and other predicates in order to explain, account for and make sense of social conduct in social interaction. As outlined by Schegloff (2007) who sketches Sacks’s work focusing on the famous ‘the baby cried’ paper (Sacks, 1972), person (or membership) categories may be organized into collections, referred to as membership categorization devices (MCDs), and MCDs may have alternative

90 

E. M. Andersen

collections of categories depending on what is indicated or inferred from the context of the usage. When person categories are used, common-­ sense knowledge about members of society is inferred, that is, categories are ‘inference-rich’ and an attributed member of a category is a presumptive member of the category (Schegloff, 2007, p. 469). This common-­ sense knowledge for example includes characteristics and actions that are linked to categories. MCA focuses on the linking practices and what they accomplish in interaction, for example how activities may be constructed as ‘category-bound’ (Schegloff, 2007, p. 470; Stokoe, 2012, p. 281) and how action formulations or the doing of a category-bound action can allude to specific category membership of persons (Schegloff, 2007). As an example, Sacks (1974) analyzed a story recounted by a child as “The baby cried. The mommy picked it up”. In this example each person in the story is identified in terms of a membership category (“baby” and “mommy”) that carries with it certain inferences. Sacks (1974, p. 219) argues that both “baby” and “mommy” can be seen to be categories from the same MCD, the device ‘family’ to which other categories such as “father” also belong. The fact that they can be heard as belonging to the same device, according to Sacks (1974), explains why most people would presumably take “the mommy” to be the mother to the baby and not somebody else’s mother. Sacks goes on to suggest that “baby” is also heard as a member of the MCD ‘stage of life’ through the activity the baby is described to perform, “cry”, arguing that crying is taken to be a characteristic of babies, that is, a category-bound activity. As also pointed to by Schegloff (2005, pp. 469–470), someone’s conduct can activate a category, and when they cannot be seen as members of that category, they may be seen to imitate the actions bound to a certain category, c.f. expressions such as ‘crying like a baby’ or ‘acting like a child’.

Data This analysis focuses on online comments to a video clip posted on Facebook. The video clip is shared by a radio show, is less than two minutes long and is a video recording of the radio studio while two hosts, Maria and Christian, are broadcasting on the radio. The video starts with

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

91

an encouragement by Maria. She invites viewers to navigate to YouTube and find a ‘tearjerker’ clip, that is, a clip that is emotional and might make you cry and explains that it has a relaxing effect and that she does it once a week. Maria then provides the name of the clip that has the best effect on her in terms of triggering tears; a scene from Disney’s animated movie “Dumbo” in which Dumbo stands outside of the cage in which his mother is locked. While Maria explains details of the scene, she starts crying. During the telling, the other host, Christian, responds to her crying by laughing, by comforting her, by making a noticing about her crying, “DU BEGYNDER A GRÆDE NU A HAHAHAHA” (‘you begin crying now a hahaha’), and by directing Maria to move closer to the microphone while telling her story, clearly orienting to the audience as the primary recipient of the telling (Jautz, 2014). Even though it is a professional setting and Maria is performing as a radio host, Christian does not question the authenticity or the sincerity of the crying, and neither do the users in the comments analyzed. In the course of about one year, the video got more than 27,000 likes (in the form of thumbs up symbols and other preselected emoticons), more than 4000 users shared the video, and it got more than 23, 000 comments. Christian’s responses to Maria’s story and the intense and emotional delivering of it differ from how interactants ordinarily respond when they are recipients to a telling that involves crying (Hepburn & Potter, 2007, 2012; Kupetz, 2014). However, the whole activity of encouraging the audience to seek out stimuli for the purpose of triggering tears also differs from settings in which crying has been found and analyzed, such as in calls to a child protection helpline and mundane interaction. The analysis concentrates on users’ comments to the video. Initial observations showed that comments may be constructed as comments to the video prompt (see Adami, 2014; Benson, 2015; Herring, 2013 for accounts of the term), or they may be constructed as responses to previous comments (see Bou-Franch, Lorenzo-Dus, & Blitvich, 2012 on the interactional organization of YouTube comments). In order to focus the analysis on how the crying in the video is treated, the comments analyzed in this chapter are all comments to the video prompt.

92 

E. M. Andersen

 erson Identification and Action Formulations P to Establish the Situated Action as an Example of an Underlying Pattern As mentioned, whereas it is not relevant or possible for users to comfort Maria, very many comments align with Maria and indicate support for her. A recurrent comment type is evaluation and in particular, various actions that communicate positive assessments, praise and thanks. In this section I focus on how constructions of evaluations of the host, Maria, as well as characterizations and formulations of the crying contribute to building a shared understanding of Maria’s response to her own story (crying) as tied to specific characteristics and specific occasioned membership categories (Hauser, 2011) such as ‘a good person’, ‘a real human being’ and ‘someone with a big heart’.

Person Identification and Characterization Some comments respond to the video prompt by characterizing Maria. Thereby these types of comments specify which aspects of Maria as a person are relevant in order to make sense of her reaction, and they may be seen as a practice of identification (Hauser, 2011). Noticeably, although the comments respond to a situated action of crying while telling a specific story in a specific setting, the identifications are often formulated in generalized terms. One example of this can be seen in Example 1:

Example 1 HG du har jo bare et stort hjerte Maria Respekt herfra MEEN lidt sjovt blir det alligevel you just simply have a big heart Maria Respect from here BUUT it becomes a bit funny anyway

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

93

In this example, Maria is addressed (“you”) and implicitly categorized as someone with “a big heart”. This is done by making a claim that Maria possesses “a big heart” which refers to a body part often used figuratively to refer to the center of emotion, especially alluding to compassion, affection and love. The emoji at the end, “ ”, may be seen as a way of indexing this symbolism of the heart (Gibson, Huang, & Yu, 2018). Alternatively, or complementarily, the heart emoji might be interpreted as an expression of affection for Maria. The use of this formulation constructs this characteristic Maria is claimed to have as an internal feature of her personality and emotional life. There are indications that the statement that assesses and categorizes Maria is to be understood as an explanation for her action of crying in the video. First of all, its sequential placement as a response to a video in which Maria starts crying and is responded to by the co-host with outbursts of laughter makes the categorization work recognizable as doing the work of an explanation. Second, the statement entails the Danish modal particles (Christensen, 2009) “jo” and “bare”, here translated to “just” and “simply”. Both particles have various possible potential meanings depending on the context of their use. Dictionary lookups (ordnet. dk/) of the words show that “jo” may be used to imply explanation and that “bare” may be used as part of a response to trivialize a prior event, and, hence, they mark the categorization work done as explaining the conduct in the video. Further, they mark the categorization of Maria as an obvious explanation for her behavior. The comment continues with an acknowledgment (“Respect from here”) and a modification “BUUT it becomes a bit funny anyway ”. Through these actions, the events in the video are referred to (“it”), and the person categorization is linked to that event. Further, Maria’s role as a performer and entertainer is also hinted at as the user characterizes the event as “a bit funny anyway ”, and this modification may function as an alignment with the co-host, Christian, who responds by laughing. Another example of generalized identification can be seen in Example 2 in which formulations such as “Maria is X” occur:

94 

E. M. Andersen

Example 2 JGB Seriøst, jeg tror Maria Fantino er en af de bedste radioværter jeg længe har oplevet. Hun er bare naturligt sjov, og hun er

fuldstændig frygtløs når hun snakker om sig selv, ligesom hun gør her. Rigtig god underholdning!

Seriously, I think Maria Fantino is one of the best radio hosts I experienced in a long time. She is just naturally funny, and she is completely fearless when she talks about herself as she does here. Really good entertainment!

As mentioned, in this example Maria is characterized in generalized terms through attributing qualities to her using the format “Maria is X”. However, these characterizations are embedded in other categorization work. The user first categorizes Maria in terms of her professional role in the clip (i.e., radio host) and continues to characterize her performance as a radio host by indicating two statements entailing predicates with positive connotations (“naturally funny”; “completely fearless”). The user then connects this general characterization of her to her ‘performance’ in the video by adding “as she does here”. Further, the membership category radio host is linked to a category-bound activity, talking about herself, which is also constructed as recurrent with the construction when-then. Hence, it is in her capacity as a radio host talking about herself that Maria is characterized, and it is these characteristics that become evident through her performance that is praised as “Really good entertainment!”. This categorization work accomplishes at least two things of relevance to the argument in this chapter: (1) Maria is categorized as a person with certain characteristics and (2) the context of her action of crying is accounted for with reference to a membership category and a category-bound activity. Thus, implicitly, the crying is separated from the category-bound crying of members of other categories, such as babies and whining children as well as imitations of those categories (Schegloff, 2005). Moreover, in order to make this generalization about Maria, the user, JGB, implies that he bases the assessment on a comparison between different hosts and recurrent experiences with Maria as a host (c.f. the when-then construction).

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

95

L inking Action Characterization and Perception Another categorization practice that occurs in the data and deals with how to make sense of the crying that occurs in the video is action characterization using predicates that denote aspects of emotional capacity and reactions related to emotional perception. This occurs, for example, in elliptic constructions such as “(it is) so sweet, so sensitive and at the same time extremely funny” in Example 3 below. In the Facebook comments, these constructions tend to link action (i.e., crying) in the video to psychological characteristics of Maria and to the observer’s perception of the action and its relation to person characterization. Thereby, these constructions display an understanding of the actions in the clip as an instance of both authentic emotional presence and expression and as an instance of humor that one may legitimately respond to with laughter. Example 3 shows one instance of this: Example 3 ME så sødt, så følsomt og samtidig voldsomt sjovt. ser frem til hver eneste dag når i går på.... I er de rette arvtagere for Anders og Anders samt Espen og Peter konge program

so sweet, so sensitive and at the same time extremely funny. looking forward to every single day when you go on .... You are the right heirs for Anders and Anders as well as Espen and Peter Wonderful show

The comment starts with a characterization: “so sweet, so sensitive and at the same time extremely funny”. As already mentioned, these action characterizations refer both to emotional capacity of the person doing the action characterized (i.e. “so sweet, so sensitive”) and to observers’ emotional reactions to the action (“at the same time extremely funny”) and thereby link the observer’s analysis of the action-doers emotional capacity to the observer’s emotional reaction as a witness to the event. Hence, as

96 

E. M. Andersen

in Example 1, the comments align with both Maria’s crying and Christian’s laughter in response to it. Example 3 continues with an elaboration of the positive characterization in which the user implies that his praise of the clip may apply more broadly to the show as such, that is, it is another instance of generalization. In this case the Facebook user addresses both of the hosts which underlines the partnership and interaction between the hosts as the source of the humor. Example 4 is another example of how actions, emotional capacity of the ‘action performer’ and the observer’s emotional reactions to having witnessed the event, are linked:

Example 4 AC Springfyldt med empati - jeg er vild med det - men nøøøj manner, det er også virkelig sjovt Filled with empathy – I love it - but boooy, it's also really funny

As in Example 3, the Facebook user in Example 4 starts with a characterization using a predicate in an elliptic construction: “Filled with empathy”. The characterization, again, focuses on emotional capacity, this time using the term “empathy” which denotes “the action of understanding, being aware of, being sensitive to, and vicariously experiencing the feelings, thoughts, and experience of another of either the past or present without having the feelings, thoughts, and experience fully communicated in an objectively explicit manner” (https://www.merriam-­webster.com/dictionary/empathy). This characterization is linked, using dashes, to a description of the user’s perception of the clip (“—I love it—but boooy, it’s also really funny ”). Since one type of humor involves disparagement of a target at whose expense others are amused (Dynel, 2013), this double linking practice of ascribing a high degree of emotional capacity to Maria who is crying and acknowledging the crying as a source of humor for observers works to counter a potential accusation that people who respond to the crying by laughing, including Christian, are unsympathetic to Maria.

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

97

 ction Formulations and Establishment A of Membership Categories A type of categorization work related to action characterization that also accounts for the situated crying in the video is statements that link a membership category to an action type. One instance of this can be seen in Example 5. Example 5 PR Jeg synes kun at det er sympatisk ☺, hvis man kan vise sine følelser, og græde en gang i mellem, det viser jo bare at man er et godt menneske ☺

I think it's simply sympathetic ☺ if you can show your feelings and cry once in a while, it just simply shows that you are a good person ☺

This comment is packaged as a personal and positive stance toward the actions in the clip which are associated with the action doer’s emotional capacity by characterizing the actions as “sympathetic”, that is, it is an instance of linking action characterization and perception as described in the previous section. The comment (as the one in Example 6), however, entails some features that the instances analyzed in the previous section did not, as this comment accounts for the positive characterization (“sympathetic”) by formulating the actions in the video as an instance of specific action types associated with a membership category introduced in the comment. In Example 5 it is suggested that the ability “to show emotion and cry once in a while” is indicative of being “a good person”, that is, it follows that you are a good person if you do the actions formulated repeatedly over the course of time. The membership category is logically bound to the actions formulated through an if-then construction (Sneijder & te Molder, 2005, p. 684). As in Example 1, the Danish modal particles “jo” (just) and “bare” (simply) are used to present the linking between actions and a membership category in order to explain the crying and present the explanation with reference to membership categorization as self-evident. One more aspect of this type of categorization work is worth mentioning, namely the fact that the membership category “a good person” is

98 

E. M. Andersen

constructed as a subset of a broader, and not specified group introduced using the pronoun “man”. According to a Danish dictionary, “man” is used to refer to an indefinite person as a representative of a group or to people in general (ordnet.dk), and it may be translated to “you”, “one”, “they” or “people”. Thus, Danish “man” may be seen as an inclusive device, similarly to how “you” may be used in English in some contexts (Housley & Fitzgerald, 2009), and as a consequence, it possibly includes many people besides Maria, the person who actually cried in the video. The use of the device in this context works to make “a good person” relevant as a membership category and thereby imply that not every person is necessarily “good”. As the construction is formulated in generalized terms, using “man”, action formulations and a membership category, the comment provides an instruction for how to understand the situated practice in the video, including how Maria should be identified and categorized (Hauser, 2011), at the same time as it suggests how the events may be applied to other contexts and other people. Example 6 is another instance of accounting for the conduct in the video by suggesting how a membership category is logically bound to a particular type of action, again by constructing an occasioned membership category as a subset of a larger group with “man” and this time using a when-then formulation as a linking device. Example 6 NLLK I min optisk er man først et rigtigt menneske når man tør og kan vise tårer, uanset hvor latterligt det måtte synes for andre. In my view, you are only a real human being when you dare and can show tears, no matter how ridiculous it may seem to others.

In this comment, the membership category “a real human being” is suggested as bound to particular actions that are indicated as necessary requirements for membership of that category. Such an understanding is achieved using the formulation “you are only a real human being when …”. Hence, it is implicitly suggested that one may be a human being without being “real”. The tie between particular actions and being “real” suggests that ‘realness’ is associated with emotional capacity: the

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

99

ability and courage to “show tears”, expressed through the Danish modal verbs “kunne” (can) and “turde” (dare). The action formulation “show tears” rather than for example “cry” is interesting as well. First, it points to the fact that emotion may be made visible in the form of tears, and second, it may imply that the action, “dare” to “show emotion”, may be a choice associated with willpower or, as mentioned, a matter of ability and/or capacity (“can show tears”). The comment also specifies that this link between action type and membership category applies even in conditions in which others might ridicule you for crying: “no matter how ridiculous it may seem to others”. This specification may be seen as a formulation of the co-host’s reaction in the video in which he starts laughing and mocking her and at the same time work to further specify aspects of what “being a real human being” is meant to imply. The specification of difficult conditions under which crying should still occur for you to belong to the category of ‘real persons’ suggests certain inferences such as the fact that real human beings have integrity and stay true to themselves, even when they are not met with social and emotional support. As with Example 5, the comment in Example 6 is packaged as a personal stance, in this case by starting the comment with “in my view…”. Constructing a pattern of behavior by linking a membership category to action types as a personal viewpoint suggests that this generalization has been arrived at on the basis of more than the situated activity from the video clip, and that it is a part of the Facebook user’s view of humans as such that he may apply to different situations in his everyday life.

 ersonal Experiences to Establish P Identification and an Underlying Reaction Pattern Associated with ‘good people’ The previous analytic section presented three categorization devices used to account for the crying in the video in ways that formulated the situated action of crying as a certain action type that was characterized in terms of the observer/user’s perceptions of it, including the observer’s

100 

E. M. Andersen

perception of the action doer’s emotional capacity, and constructed the crying as linked to particular occasioned membership categories, accomplishing to suggest that Maria’s behavior should be made sense of with reference to person characteristics. Another strategy for accounting for the action of crying in the clip involves identification through indications of personal experiences. In this section I present three recurrent ways these experiences are formatted, and I outline how Facebook users thereby imply belonging to the same membership category as Maria, construct actions indicative of a pattern and protect themselves from accusations of ridicule on the one hand and self-praise on the other using a range of self-deprecating strategies.

 stablishing Crying as an Unavoidable Reaction E to Specific Stimulus One way of constructing identification with Maria in terms of her crying in the clip is through statements that claim a pattern of behavior as a response to a specific ‘stimulus’; the stimulus being either the same as the one Maria reacts to telling about (a scene from the animated Disney movie “Dumbo”) or one that shares features with it (e.g. other dramatic and emotional scenes from animated Disney moves). In Example 7 below, this is indicated through the formulation “I can’t help but”. In this comment the Facebook user constructs crying as an unavoidable consequence of being exposed to a specific ‘stimulus’, namely the same film scene as Maria had selected (referred to with “THAT”). The formulation “can’t help but” accomplishes this, and choosing the exact same clip as Maria makes a claim of identification with her and of possibly belonging to the same group of people as her, based on the assumption that this type of action is bound to a membership category. As was shown in the previous section, on the one hand, Facebook users tended to praise Maria for her reaction and for the way she performs her role as a radio host, including categorizing her as belonging to the group of people with praiseworthy emotional and human capacities. By implication, then, identifying with Maria may be heard as self-praise. On the

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

101

Example 7 THT [name tag KN] SE HER, jeg er ikke den eneste der har lige netop DET, som den ting i verden jeg ikke kan lade være med at græde til fandmer sindssyg børnefilm [name tag KN]

SEE HERE, I'm not the only one who has exactly THAT

as the thing in the world I can't help but crying too bloody crazy children’s movie

other hand, many Facebook users account for her crying, among other things distinguishing it from crying as a baby or a child. There are indications that the Facebook user in Example 7 protects herself from both these implications (i.e., self-praise and acting like a child). This is done by (1) identifying a specific stimulus as the cause of crying; (2) presenting the stimulus as an extreme case (Edwards, 2000) using a predicate often used as an invective and adding a swearword to the statement that categorizes the film: “bloody crazy children’s movie”; and (3) indicating several emojis to the statements in positions that make them understandable as a display of how the statements should be understood (Gibson et  al., 2018, p. 95), that is, as statements that are meant to invite emotional reactions, including laughter. Hence, through these means, the personal experiences indicated by the Facebook user are presented as an instance of humor, and the action of crying as a reaction to specific stimulus, and, thus, implicitly not as something one might use to manage emotion in all sorts of everyday contexts or as a means to manipulate others and so on.

Extreme Cases as Identification Devices Small stories about specific instances in which the Facebook user started crying are another way of displaying identification with Maria. These stories tend to dramatize the event by using formulations of crying constructed as nonliteral exaggerations and by describing the ‘stimulus’ that is indicated to trigger the crying using dramatizing devices. Example 8 illustrates such an instance of constructing identification by presenting a case in which a particular stimulus led to an emotional reaction of crying:

102 

E. M. Andersen

Example 8 PKP [nametag] det her, det kunne ikke være mere mig. Jeg så en video i morges, af en mand, der reddede en lille bitte hvalp fra at drukne. Der kom måske et par tåre, eller 395035

[nametag] this, it couldn't be more me. I watched a video this morning,

showing a man who saved a tiny puppy from drowning. Maybe a few tears came, or 395035

The Facebook user in this instance explicitly formulates the comment as a claim of identification (“this, it couldn’t be more me”) and continues the comment with a small story that describes events occurring the same morning as the comment is posted (“this morning”). The fact that the Facebook user is able to identify a recent experience constructed as recognizably similar to Maria’s hints at the possibility that she has similar experiences regularly, and, thus, that she belongs to the same membership category as Maria in terms of personality and emotional capacity. In any event, the story works to substantiate the claim of identification made initially. As mentioned, these stories often include devices that dramatize the singled out ‘crying’ instance. First, the action of crying may be formulated using extreme case formulations (Edwards, 2000). In Example 8, the crying is formulated as “Maybe a few tears came, or 395035”, that is, two alternative descriptions of the number of tears that came during the crying incidence are provided. As both of the alternatives may not accurately quantify the crying as they are not compatible, and as the amount “395035” is clearly recognizable as an overstatement, the description signals a “nonliteral intent” (Edwards, 2000, p. 370); that is, the description should be taken as ironic. The comment also entails a description of the ‘stimulus’. The description identifies a main event and main ‘characters’ in it: a man saving a tiny puppy from drowning. Using a specific action formulation, “save from drowning” (rather than, e.g., “take” or “lift” out of the water), the membership category of the man in the story as ‘savior’ may be generated, and, hence, a role as hero of the story may be inferred. The dog that undergoes the saving is referred to as “a tiny puppy”, thereby categorizing

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

103

the dog in terms of its stage of life (Andersen & Rathje, 2019; Stokoe, 2012), allowing readers to make inferences about its abilities to handle the situation and how it appears. By selecting an instance in which watching video clips led to crying, the user accomplishes the telling of a so-called second story. Second stories are described as “stories told in a series in which later stories are designed to achieve a recognizable similarity with the first (or previous) story” (Arminen, 2004, p. 319). Based on a study of second stories used in AA meetings, Arminen (2004) suggests that second stories have several functions such as providing support for the prior speaker, amplifying the core ideas and offering interpretations of the first story. Comments with personal stories, including formulations of crying and descriptions of the stimulus that triggered the crying, establish, through their recognizable similarities to the incidence in the video, one example of how particular types of people respond in an extreme way to particular types of very emotionally charged stimuli. The extreme case formulations, as mentioned, also mark the story as intentionally nonliteral (Edwards, 2000), humorous, self-directed and self-deprecating (Hay, 2001) and thereby invite others to treat is as such. This contributes to the identification and accounting work that is accomplished through these comments. By this I mean that the users, through telling personal stories that include extreme case formulations, imply an understanding of themselves as someone who is not afraid to expose themselves, which, according to Zajdman (1995), may be perceived as indicative of having strength (e.g., in the sense of having a surplus of mental resources). This is also reflective of the video itself in which crying was suggested to be healthy, and in which crying was responded to by the other host with laughter, among other things.

 eaction Indications as Identification R and Empathy Displays A third type of identification is users’ descriptions of their reaction to watching the video clip or listening to the segment in the radio. In these comments, crying is described as a direct response to Maria’s emotion

104 

E. M. Andersen

displays in formulations such as “I cried because she cried”. In Example 9, such an understanding is indicated with the construction “get totally carried away by”:

Example 9 HEJ Bliver jo helt revet med af hendes stemning, lidt tårer i øjnene får jeg og imens kan jeg alligevel ikke lade være med at grine I surely get totally carried away by her sentiment, a few tears in my eyes I get and at the same time I cannot help laughing

In these comments, crying (and other reactions such as laughing) tends to be constructed as triggered by the emotion displays of people in the video. In Example 9 this is constructed with the formulation “get totally carried away by”. Through such means, Facebook users commenting with reaction indications like this one imply being empathetic (see definition of empathy above) and sensitive toward other’s feelings to an extent that they visibly display the same feelings. Thereby reaction descriptions work as an empathy display device. It may be argued that what makes the video both extreme and funny for many people, and what makes it relevant that Maria is treated as a particular type of person in terms of her emotional capacities, is the fact that Maria starts crying, not while watching a movie, but while telling about it. Reaction indications may be seen as a way of expressing a reaction which signals similar emotional capacities and thereby imply identification with Maria and a shared membership in terms of emotional capacity while positioning themselves as a member of the category observer/viewer/listener to the segment in which Maria performs as a radio host. As we have seen, in other comments, Maria’s behavior in the clip is expressed as indicative of extremely sensitive and empathetic people. Unlike comments that present similar experiences to Maria’s, reaction comments are not presented to be taken as nonliteral. In Example 9, Maria’s emotional reaction is indicated in very general terms (“her sentiment”), whereas the Facebook user’s visible emotion display is described

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

105

in more detail (“a few tears in my eyes”) and described as something occurring simultaneously or as highly coordinated with Maria’s. This is done with a present tense construction “I get” and through the description of an additional emotional reaction occurring “at the same time”. In fact, the reaction comments tend to indicate both crying and laughing as reactions they had during the viewing or listening to the event. The reaction comments also tend to include emojis depicting faces laughing and/or crying, either working to index the descriptions or working to invite other users to treat the reaction indications in particular ways or both (Gibson et al., 2018). These emojis may be central means of indicating to other users that the emotional reactions are not treated by the user as problematic and should not be treated as such by others, that is, as a way of implying self-reflection and emotional capacity as person qualities.

Discussion Using MCA as a methodology, this chapter has attempted to show how attribution and categorization are related to the accountability of the person doing the evaluation, for example, how laughing at someone crying may be treated as being possibly problematic and dealt with by praising the person crying, and how identifying with someone who is being praised may be dealt with by using humoristic, but self-deprecating, devices. MCA has traditionally been used in the analysis of video-recorded interaction, but also of news media discourse. MCA on online social interaction is sparse (but see Andersen & Rathje, 2019; Gibson & Roca-­ Cuberes, 2019; Stommel & Koole, 2010). As this chapter suggests, however, categorization practices are indispensable for making sense of social conduct, and systematic analysis of the practices supply knowledge about the culture of a community that is built using online platforms. Data analyzed in this chapter, online videos with comments, are naturally occurring and not elicited by researchers and may be a new, but potentially, rewarding site for such analysis, since in many cases it will be possible to build a collection of comments that all deal with making sense

106 

E. M. Andersen

of, understanding and accounting for the same social conduct, that is, that presented in the video clip. Another aspect of this type of data concerns the fact that the data has come into being through complex processes many of which the researcher does not have access to when only accessing the comments that have been posted at a specific point in time. Further, the comments are organized according to the possibilities and limitations in the interface set up by a social media platform. As regards comments to video prompts on Facebook and YouTube, this, for example, means that the interactional organization of comments is quite complex, a topic in its own right (BouFranch et al., 2012), and that an important part of the data collection and the initial analytical work consists of getting a sense of how the social media users’ comments are related to other comments, the stimulus or other external references. This paper, for example, only focuses on comments that respond to the video and not to other comments as several ‘side sequences’ developed that did not relate to the crying in the video.

Conclusion The comments analyzed in this chapter, from a video distributed on Facebook by a radio show, testify to the fact that the show and its hosts mainly address and possibly expand an existing audience, some of which may even behave as fans. The findings suggest that Facebook users align with both Maria’s crying by treating it as an indication of a high degree of empathy and Christian’s laughter as a reaction to it by acknowledging the clip as funny. In fact, I have found no comments that question the authenticity of Maria’s crying or accuse Christian of being unsympathetic. As mentioned, the handling of the crying in the video by the participants themselves is very different from how crying is handled in other mundane and professional settings. The radio audience is encouraged to ‘seek out a weep’, and when the host, Maria, starts to cry, it is treated as extreme, but as a source for laughter by the co-host and as part of the performance of the radio show, also evident from the fact that the clip is chosen for distribution on Facebook.

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

107

The Facebook users’ handling and understanding of the crying in the video are also noticeable and are inextricably linked to the affordances available to them for responding. First, users’ responses to the crying are composed (asynchronously in relation to the actual crying) through written means and are deliberate in a sense that responses to crying displayed through talk are not. Consequently, actions such as reaction indications in the written comments are clearly performances with interactional and social purposes rather than “inadvertent manifestations of individual bodily processes” (Wilkinson & Kitzinger, 2006, p. 150). Second, users do not treat crying as a disruption of the activity in the video, but instead treat it as a topic, as a source of humor and treat crying as a disposition that can be linked to personal qualities of people, that is, the topic of this chapter. Thus, perhaps counter-intuitively when considering how crying is managed in other contexts, crying is used as an identification device for Facebook users in this particular case.

References Adami, E. (2014). “Why did dinosaurs evolve from water?”: (in)coherent relatedness in YouTube video interaction. Text & Talk, 34(3), 239–259. Andersen, E.  M., & Rathje, M. (2019). Age and stage of life categorizations used to moralize in online social conflict. Discourse, Context & Media, 28, 19–26. Arminen, I. (2004). Second stories: The salience of interpersonal communication for mutual help in Alcoholics Anonymous. Journal of Pragmatics, 36(2), 319–347. Becker, W., Conroy, S., Djurdjevic, E., & Gross, M. (2018). Crying is in the eyes of the beholder: An attribution theory framework of crying at work. Emotion review, 10(2), 125–137. Benson, P. (2015). YouTube as text: Spoken interaction analysis and digital discourse. In R. H. Jones, A. Chik, & C. A. Hafner (Eds.), Discourse and digital practices (pp. 81–96): Routledge. Bou-Franch, P., Lorenzo-Dus, N., & Blitvich, P. G.-C. (2012). Social interaction in YouTube text-based polylogues: A study of coherence. Journal of Computer-Mediated Communication, 17(4), 501–521.

108 

E. M. Andersen

Boyd, M. S. (2014). (New) participatory framework on YouTube? Commenter interaction in US political speeches. Journal of Pragmatics, 72, 46–58. Christensen, T. (2009). Tag nu bare økomælken. Om imperativ og modalpartikler i dansk. Dramatikken i grammatikken: Festskrift til Lars Heltoft, 51–67. Culpeper, J. (2001). Language and characterisation: People in plays and other texts. Singapore: Pearson Education Limited. Dynel, M. (2013). Humorous phenomena in dramatic discourse. The European Journal of Humour Research, 1(1), 22–60. Dynel, M. (2014). Participation framework underlying YouTube interaction. Journal of Pragmatics, 73, 37–52. Edwards, D. (2000). Extreme case formulations: Softeners, investment, and doing nonliteral. Research on Language and Social Interaction, 33(4), 347–373. Gibson, W., Huang, P., & Yu, Q. (2018). Emoji and communicative action: The semiotics, sequence and gestural actions of ‘face covering hand’. Discourse, Context & Media, 26, 91–99. Gibson, W., & Roca-Cuberes, C. (2019). Constructing blame for school exclusion in an online comments forum: Membership categorisation analysis and endogenous category work. Discourse, Context & Media, 32, 100331. Hauser, E. (2011). Generalization: A practice of situated categorization in talk. Human Studies, 34(2), 183–198. Hay, J. (2001). The pragmatics of humor support. HUMOR: International Journal of Humor Research, 14(1), 55–82. Hepburn, A., & Potter, J. (2007). Crying receipts: Time, empathy, and institutional practice. Research on Language and Social Interaction, 40(1), 89–116. Hepburn, A., & Potter, J. (2012). Crying and crying responses. In A. Peräkylä & M.-L.  Sorjonen (Eds.), Emotion in interaction (pp.  195–211). Oxford: Oxford University Press. Herring, S. C. (2013). Discourse in Web 2.0: Familiar, reconfigured, and emergent. In D. Tannen & A. M. Trester (Eds.), Discourse 2.0: Language and new media (pp. 1–25). Washington, DC: Georgetown University Press. Housley, W., & Fitzgerald, R. (2009). Membership categorization, culture and norms in action. Discourse & Society, 20(3), 345–362. Jautz, S. (2014). Who speaks and who is addressed in radio phone-ins? Journal of Pragmatics, 72, 18–30. Kupetz, M. (2014). Empathy displays as interactional achievements— Multimodal and sequential aspects. Journal of Pragmatics, 61, 4–34. Sacks, H. (1972). On the analyzability of stories by children. In D. H. J. J. Gumperz (Ed.), Directions in sociolinguistics: The ethnography of communication (pp. 325–345). New York: Holt, Rinehart and Winston.

5  The Radio Host Cried, the Facebook Users Identified: Crying… 

109

Sacks, H. (1974). On the analysability of stories by children. In R. Turner (Ed.), Ethnomethodology; selected readings (pp. 216–232). Penguin Books. Sacks, H. (1992). Lectures on conversation (G. Jefferson, ed. with Introduction by EA Schegloff, volumes I and II). Blackwell. Schegloff, E. A. (2005). On integrity in inquiry… of the investigated, not the investigator. Discourse studies, 7(4–5), 455–480. Schegloff, E.  A. (2007). A tutorial on membership categorization. Journal of Pragmatics, 39(3), 462–482. Sneijder, P., & te Molder, H.  F. (2005). Moral logic and logical morality: Attributions of responsibility and blame in online discourse on veganism. Discourse & Society, 16(5), 675–696. Stokoe, E. (2012). Moving forward with membership categorization analysis: Methods for systematic analysis. Discourse studies, 14(3), 277–303. Stommel, W., & Koole, T. (2010). The online support group as a community: A micro-analysis of the interaction with a new member. Discourse studies, 12(3), 357–378. Wilkinson, S., & Kitzinger, C. (2006). Surprise as an interactional achievement: Reaction tokens in conversation. Social Psychology Quarterly, 69(2), 150–182. Yzerbyt, V., Dumont, M., Wigboldus, D., & Gordijn, E. (2003). I feel for us: The impact of categorization and identification on emotions and action tendencies. British Journal of Social Psychology, 42(4), 533–549. Zajdman, A. (1995). Humorous face-threatening acts: Humor as strategy. Journal of Pragmatics, 23(3), 325–339.

6 “On that note I’m signing out”: Endings of Threads in Online Newspaper Comments Joanne Meredith

Introduction This chapter focuses on how threads of online newspaper comments end. Online newspaper comments, also known as ‘below-the-line’ comments, are posted by readers on a newspaper’s website in response to a specific news or opinion item. These comments are written in response to an item by a professional writer, who will rarely respond to the comments, but commenters in the thread can, and do, interact with one another on the “It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

J. Meredith (*) Department of Psychology, University of Wolverhampton, Wolverhampton, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_6

111

112 

J. Meredith

topic of the news item (Graham & Wright, 2015). The interactivity which newspaper comment threads afford has allowed users to become more involved in consuming, and sometimes producing, news (Ksiazek, Peer, & Lessard, 2016). Some comments left on newspaper websites are uncivil, with the potential for comments which are abusive, libellous or racist to be posted (Hughey & Daniels, 2013; Singer, 2009). On the counter side, discussions in online newspaper comment threads can be relatively deliberative and can sometimes enhance journalistic knowledge and practices (Graham & Wright, 2015). Previous research on newspaper comment threads has tended to focus on the content of the discussions (e.g., Cooper, Green, Burningham, Evans, & Jackson, 2012), with less research focused on the interaction between posters. Coles and West (2016) took a discursive approach to analysing issues around trolling in online comments, showing how participants defined and invoked notions of trolling in this setting. Meredith and Richardson (2019) used discursive psychology to explore how the categories of Brexiter and Remainer are constructed and resisted in online newspaper comments. While we explored the language used and some of the discursive practices, we did not explore the interactional implications of communicating in this medium. These implications have been explored for other asynchronous media, including online forums, with analysis showing how individuals maintain coherent interaction in a multi-party forum (e.g., Gibson, 2009), and how posters orient to sequential issues in online forums (e.g., Stommel & Koole, 2010). This chapter explores endings in online newspaper comment threads, which have not previously received much attention in the literature. In these massively multi-party interactions, the interactional constraints are different to those found in spoken interaction. Participants in online interaction are not necessarily expected to be engaged throughout an entire thread, but rather may simply post one message in the entire interaction (Giles, 2016). There is also the possibility that each individual thread of interaction could simply continue, even many months after the initial post, until such a point as moderators choose to close the thread (Giles, 2016). In the specific context of newspaper comments many of the threads are, what we may call, conflictual (Hutchby, 1996), and as such, the endings of these threads may also be the ending of the conflict (or at least, that portion of the conflict). As such it is of interest to explore how threads end when there is some significant conflict (in this case around

6  “On that note I’m signing out”: Endings of Threads in Online… 

113

Britain leaving the European Union) and how this might have implications for public understanding and debate of highly contested topics.

Closings in Spoken and Online Interaction Schegloff and Sacks (1973) note that for speakers there is a ‘closing problem’; that is, how to get to a point where the operation of turn-taking can be suspended with it being understood as the closing of an interaction, rather than as a speaker’s silence. They note that closings are coordinated sequences, which require both parties to ‘agree’ on the closing of the interaction. However, studies of closings in online interaction have found some subtle differences to those in spoken interaction. Pojanapunya and Jaroenkitboworn (2011) focused on closings in Second Life, a virtual world where individuals can create a virtual body, or avatar, and interact with other users. They found that the closings tended to comprise of four elements: a pre-closing (such as “well I’ve got to go now”), insertions, a terminal closing turn and an afterclose. However, only the terminal turn (i.e., the goodbye) was obligatory, with other components not appearing in all closings. Two patterns for closing sequences have been found in instant messaging interaction (Raclaw, 2008). The first pattern was similar to a spoken closing sequence, with the pre-closing most often providing an account for leaving the conversation. At the end of the terminal exchange, users most commonly set their status to ‘away’, or on rare occasions logged out, to signal to their interlocutor that they would no longer be responding to messages. The second pattern was described as a ‘partially automated closing’, because the setting of the status and the automated message that this triggered often replaced the terminal exchange. Markman (2009) noted that getting to the terminal exchange could be more difficult in online interaction, because of the inability to coordinate turn-taking in this kind of interaction. In other words, even if one participant posted a pre-closing, another participant could, at the same time, post a continuation of the topic, or a new topic, which would mean that the interaction would not move into closings. In e-mail interaction there is not always a clear closing, with users sometimes finishing e-mails without a specific ‘sign-off’ (Waldvogel, 2007). In online counselling interactions, it has been found that a sequence where the counsellor provides advice, and the client provides an acknowledgement of that advice can work

114 

J. Meredith

as a pre-closing (Stommel & Te Molder, 2015). However, in text chat the advice acknowledgement may not always be forthcoming and so this can have implications for the closing. Previous research, therefore, indicates that although there may be some cases where online closings look similar to those in spoken interaction, there are notable differences in terms of how interactants move into and complete closings online. In contrast to the contexts above, online newspaper comments often involve more conflictual interaction. There are challenges in defining precisely what a conflict or dispute is (see Joyce, 2019), but we can generally understand a conflict as comprising oppositional utterances which disagree with some aspect of the speaker’s prior turn. Hutchby (1996) notes that any utterance can potentially be an arguable, and so it is an oppositional turn to the ‘arguable’ that results in a conflictual interaction. Vuchinich (1990) finds a number of ways in which conflict talk can be resolved in spoken interaction: (1) submission or compromise, where one party effectively backs down; (2) stand-off, where neither party backs down but the conflict ends, usually through a topic shift; (3) withdrawal, where one or both parties withdraw from the conflict, by walking away, for example, without there being any resolution and with a subsequent end to the interaction. Vuchinich found that stand-offs are the most common form of endings of conflict talk and suggests this may be because submission and withdrawal may lead to the speaker losing face. In an online context, Bou-Franch and Blitvich (2014) found that in YouTube comments the strategy of ‘withdrawal’ was the most commonly used; that is, the poster simply did not return to the comment thread. They very rarely found cases of stand-offs and found no cases of compromise, which indicates that in these kinds of online conflicts, coming to an agreement or resolving a conflict appears to be very rare. When examining the endings of threads in more conflictual online interaction, it is important to consider the broader context of the interaction. Newspaper comment threads are asynchronous, and so users cannot see whether someone else is posting a message until it is posted. As such, when conducting analysis of such threads, it is not possible to know the exact contextual information a user had when they posted their message (Giles, 2016). For example, an analyst will not know whether the person read all of the previous messages or whether previous responses were even visible when they started constructing their message. In this multi-party environment, it is also important to consider the various recipients of online messages. There

6  “On that note I’m signing out”: Endings of Threads in Online… 

115

may be a direct recipient of a message, that is, the person to whom a message replies or is directed at through the use of an address term. There may also be an indirect recipient, who may be those who have previously been involved in that thread of interaction but not the direct recipient. BouFranch and Blitvich (2014) also note that if a post is directed at a specific group (in their example at people who are gay) then anyone in the thread who identifies as gay may orient to being the indirect recipient of those messages. Finally, there are overhearing or lurking participants, who may be more passive observers of the interaction. Dynel (2014) notes that it seems inconceivable that there could be such a thing as an ‘overhearing’ participant in YouTube interaction, as overhearing implies that there is some sort of illegitimate hearing of the interaction which is not the case online. However, I would argue that lurking participants are distinct from those who are direct or indirect recipients in that the commenter is not aware of these participants in the same way that they would be of active participants. The following sections in this chapter will describe the methods used for collecting and analysing the data. The analysis section will lay out the various sequences which lead to threads ending. As will be shown below, there are very rarely coordinated closings in these threads, and as such, they will be referred to as endings rather than closings throughout.

Methods Data Collection The data for this chapter come from a study about the UK referendum to leave the European Union, known as Brexit. This referendum was held on 23 June 2016, and 52% of voters voted to leave the EU. The study originally focused on issues of political identity in online newspaper comment threads (see Meredith & Richardson, 2019). We collected data from four national newspaper websites, both tabloids and broadsheets, which allowed comments and were not behind a paywall. We wanted to ensure that we had a sample of comments from both pro and anti-Brexit newspapers, and as such, the newspaper sites chosen were Daily Mail/Mail on Sunday, The Independent, The Guardian/The Observer and The Express/Sunday Express. Of these, The Independent, Mail on Sunday and

116 

J. Meredith

The Guardian/Observer backed remaining in the EU. The Express, Sunday Express and Daily Mail supported leaving the EU. As there was a large amount of discussion about Brexit both during and after the referendum campaign, we aimed to collect a sample of data that was relatively focused in order to keep the dataset manageable. We chose to examine any stories which focused on a particularly prominent Vote Leave campaign slogan ‘We send £350m a week to the EU. Let’s fund our NHS [National Health Service] instead’. Focusing on this claim allowed us to collect articles from both before and after the referendum date as the slogan was discussed for a prolonged period of time. Data was collected from the date the referendum was announced in February 2016 until October 2016. Articles were excluded if they did not relate directly to Brexit, if they were live blogs, and if they did not allow comments. We sampled threads (a collection of posts, generally related to the same topic) rather than individual posts, as threads are the key analytic unit when analysing such data (Giles, 2016). In order to ensure relatively even representation from all newspapers, whilst maintaining a manageable data set, we chose to limit the number of threads sampled for each item to 100, which gave us a dataset of 2586 threads. We used NVivo capture to collect the data and then transferred it to Microsoft Word so it would be searchable. The study was approved by the University of Salford ethics committee. We did not get consent from participants due to the fact that data were taken from public comment sites and so could be considered in the public domain (BPS, 2017). In order to protect participants all names or pseudonyms have been changed to provide an extra layer of anonymity.

Data Analysis All threads were read, and the ending sequence of each thread, as determined by the sequential and topical organisation of the thread, was manually collected. The endings of threads are retrospective in the sense that the posters themselves will not be aware when making their posts that they will not get a response. By the time we had collected the data, all comment sections on these news items had been closed so no further posts have been made since data collection. It is worth noting that because we were not able to collect time stamps for all the data it’s unclear whether threads may

6  “On that note I’m signing out”: Endings of Threads in Online… 

117

simply ‘fizzle out’ because there are long delays between posts, or whether these articles have simply been ‘buried’ by new articles and so do not have the prominence on the homepage of the newspaper (Giles, 2016). However, the question is why these threads ‘fizzle out’ at the points that they do, and whether there are any common features of such posts. Conversation analysis (CA) was used to analyse the data (see Introduction to this volume for more detail). Broadly, the goal of CA is to describe and explicate “the competences that ordinary speakers use and rely on in participating in intelligible, socially organized interaction” (Heritage & Atkinson, 1984, p. 1). By examining conversation, it is possible to understand the “actions and activities through which social life is conducted” (Drew, 2005, p. 75). In other words, through examining the norms and practices that occur in everyday interaction, we can begin to see how it is that participants maintain ‘intersubjectivity’, or a shared understanding of the world (Schegloff, 2006). Digital CA (Giles, Stommel, Paulus, Lester, & Reed, 2015) uses the analytic approach of CA, but applies it to online interaction. The aim is to focus on the way in which the interaction unfolds and how participants manage the potential affordances of the interaction, without presuming any a priori influence of the medium. To conduct the analysis, all collected threads were then read multiple times, and the sequence of the thread ending was analysed. The turn-by-turn sequence was investigated in detail, and similar practices were identified across these. Collections of interactional practices were created, and extracts for each of these practices were analysed in more detail.

Analysis There were no examples identified of mutually coordinated closings as found in spoken interaction. Equally, there were no examples of the kinds of sign-offs that might be found in e-mail or counselling talk (such as “with best wishes, Jo”; see Waldvogel, 2007; Stommel, 2012). As such, it appears that online comment threads are treated by users as not routinely needing a sign-off or closing sequence. It is, though, important to note that when individuals post a message, because of the multi-party nature of the chat, they cannot know for sure whether there will be a response.

118 

J. Meredith

There were a few deviant cases where there was a unilateral sign-off, an example of which is shown below. This post comes at the end of a long discussion between Mario87 and another poster about tax issues. The final portion of Mario87’s final post is shown below. Extract 6.1  (From The Guardian: Brexit campaign ‘writing cheques they know will bounce’, says Cameron, 4th June 2016) Mario871 Mario871 [lines omitted] 1

To be fair, this is a complex argument involving a massive ecosystem

3

haven't even touched on the effect of a drop in interest rates on

5

camp's economists have belittled themselves by over-simplifying

2 4 6 7

of overlapping fiscal and monetary multipliers and policy changes (we

spending and tax receipts). My point still stands - the Remain

forecasts of the possible consequences of BREXIT.. and on that note I'm signing out. Enjoy the referendum!

At line 1, Mario brings their2 post to a close with a turn which in some way offers a compromise, which may be indicative of a pre-closing of this stretch of interaction (Vuchinich, 1990). In brackets at lines 2–4, Mario indicates some potential other topics which have not been discussed so far, but it appears they are declining to fully raise them as new topics as they are put between brackets, which may also indicate that this is moving towards a closing. They then reiterate their stance or position (lines 4–6), which may also function as a pre-closing (Hutchby, 1996). Finally, at line 7 they offer an explicit closing statement “on that note, I’m signing out”, which provides some kind of account for why they will no longer be responding, which is also evident in other online pre-closings (Raclaw, 2008). The final component of their post “enjoy the referendum!” implies that this is a terminal component and they will not be returning to the thread, particularly as the referendum which occurred on 23 June was  All names used are changed from the original name used online. However, we attempted to keep the same feel and style as the original name. 2  It is not possible to know the gender of online commenters, even if the name may seem to be gendered. Therefore, the pronouns they and their are used throughout when referring to commenters. 1

6  “On that note I’m signing out”: Endings of Threads in Online… 

119

still over a week away from the date of posting. Such sign-offs were incredibly rare in the dataset with only three cases of some kind of ‘ending’ identified. In this example, it may be that because this interaction took place between two individuals over a prolonged period of time, the interaction was oriented to as a one-to-one interaction, which may be understood as requiring a closing. Mario also had a slightly different type of ‘closing problem’; that is, they needed to leave the interaction, but the conflict had not been resolved. Therefore, providing an account for leaving the interaction indicates that they are not ‘backing down’, but rather there is a stand-off, with the conflict not being resolved (Vuchinich, 1990). The interactional features identified as potentially leading to a non-­ response and so an ending are (1) directly named recipient not responding, (2) idioms used in the final post, and (3) a shift towards personal comments and meta-talk. Examples of each type of ending sequence will be shown below. Extract 6.2  (From Daily Mail: Boris Johnson accused of hypocrisy as Brexit tour bus is German, 11th May 2016) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Geoff the Grizzly What a stupid article we don't hate Europe the continent or their goods you idiot we just hate the Eu I love European holidays but hate being ruled by a mercel Mary Actually not stupid. Faux pas by Boris. If you love your European holidays and we leave the EU expect less Euros for your £ for a considerable time. Make sure you have private insurance and pay your tax on goods being brought back into the UK. We take so much for granted and everyone is going to have hell of a shock and say " oh I didn't realise that would happen." Bob Malaga Mary . Ever since we have been in the eu the pound has gone down.In 1973 it was about 8 dm to the pound it would now be under 2. I think you reasoning is wrong errant_punctuation Mary ## Or holiday in the UK or outside the EU. It's their loss, especially if more EU citizens holiday in the cheaper UK. If the EU wants to throw their toys out of the pram in a fit of pique by not agreeing fair exit terms, then they'll be the ones saying "oh I didn't realise that would happen." endoftheworld Mary, their desperation to keep us in is telling us everything we need to know

120 

J. Meredith

Directly Named Recipient Not Responding There are different types of non-response identified within the ending sequences. The most common lack of response is when a question is asked of a specific recipient, who does not return to the thread to respond to that question. The interest in this post is the three responses to Mary (lines 15–30), who herself had posted a counter-argument to the initial post by Geoff the Grizzly. The three posts by Bob Malaga, errant_punctuation and endoftheworld are all directly addressed to Mary, indicated by the use of her name at the start of the post (Stommel & Koole, 2010). On the Daily Mail website threading only occurs in response to an initial post in the thread, with all subsequent responses being shown as responsive to the first post. Therefore, in Extract 6.2 the use of the name “Mary” at the start of each post is a way of ensuring that messages are understood as being addressed to that specific person (Werry, 1996), rather than to the initial post. The address term not only maintains coherence but also indicates that a response may be expected from Mary. In this way, the commenters are effectively allocating the next speaker, which in spoken interaction would mean there is an obligation for them to take the next turn (Sacks, Schegloff, & Jefferson, 1974). However, in online multirecipient contexts, there is the possibility for someone else to take a turn instead of the named recipient, and as such, there may not be such strict allocation of speakership norms. In this thread Mary does not return to respond to the comments; there are no new oppositional moves and so the thread ends. It is, of course, possible that someone else could respond to one of the posts which is directed at Mary, but the fact that they do not may indicate a preference for the selected speaker to take the next turn (Stivers & Robinson, 2006). In the following extract, which responds to a newspaper article about Diane Abbott, the next speaker is tacitly selected through a question directed at a specific speaker:

6  “On that note I’m signing out”: Endings of Threads in Online… 

121

Extract 6.3  (From Daily Express: Question Time audience member puts down Diane Abbott as she attacks Farage for NHS ‘lies’, 22nd June) 1 2 3 4 5 6 7 8 9 10 12 13 14 15 16 17 18 19 20 21

JamieMac She made a comment that the labour parties priority was black and Asian communities, I'm from the north east Middlesbrough which has the highest number anywhere in the uk for asylum seekers per head of population, when it see her on the TV and listen to Corbyn I feel physically sick at what they have done to my town. For her to turn round everything on immigration by the leave campaign shows the woman is a racist and has no concept of what White working class people go through on a daily basis, I dont think this woman has every been to the North east we don't matter to the labour luvvies in London , my town would vote a monkey in a red rosette before this referendum , I'm hoping this has changed their minds. I hate the Labour Party for what it's done to my country and most of all to my town. Gordon the gopher Well said JamieMac. Also very accurate. MattMattJones And exactly what is it that you have to go through?

JamieMac (JM) posts their views on asylum seekers and immigration, ending with a summary of their view of the Labour party. There are many potential issues which could be responded to in the post, such as their town, the Labour party, Diane Abbott, immigration and so on. The initial response by Gordon the gopher affiliates (Lindström & Sorjonen, 2012) with the stance of the post, and adds their own assessment that the content is “very accurate”. MattMattJones’s (MMJ) post responds to JM’s post, asking a direct question which requests clarification, specifically about lines 8 and 9. MMJ links their post to the original post through the use of the conjunction “and”, as well as using the pronoun “you” which indicates that it is directed at that person (we could imagine that he could instead use the category of white working class). In this way, the post is designed to select JM as the next poster, but JM does not return to the thread. As with the previous extract, there seems to be a preference for the selected poster taking the next turn, and as a result, the thread ends. In the extracts above, there is no response when one might be expected due to the next poster being selected by the previous one. In the following extract, in which two users are involved in an exchange about

122 

J. Meredith

Extract 6.4  (From Daily Mail: Brexit campaigners should not be trusted about NHS claim top doctors, 14th June 2016) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

englishalways Just because we will be out of the EU does not mean that no one from the EU will be able to work here. They are making it up as they go along. The NHS was staffed by people from all over the world before we joined the EU. Including people from all over Europe. Vote out. Barbara Windsor Yeah, but what happens to Brits who want to, or are currently working in the EU? No one ever mentions that. [2 posts omitted] Englishalways Barbara Nothing will happen to them, they will be able to stay put. They will be protected. British people work all over the world without any problems so it will be no different. I am married to a German national who won¿t be going anywhere. Vote out. Barbara Windsor Really? So you're up to date with carnets, and work permits are you? Things we don't need at the moment. You're also aware of the concerns many (educated) ex pats have concerning Brexit? English always Yes I am up to date believe it or not! I know someone that deals with that sort of thing. I understand the concerns of ex pats! I have a few members of my family living and working abroad. Do not believe all of the scare stories the in campaign say. They have their own agenda Living in the EU will get you token health cover but not much, you still need health insurance. Unlike hear where most get it for free wherever you come from.

working in the EU after Brexit, we see an example of how a response may not be relevant and so the thread ends. This thread effectively functions as a series of adjacency pairs. As Giles (2016) notes, the original post on a discussion forum will set the topic for the discussion and deviations from this may well be removed. In the case of online newspaper comments, the news item itself sets the topic and is effectively a ‘prompt’ for the first message (Herring, 2013). The first post by EnglishAlways (EA) is a first pair part (FPP) which responds to the ‘prompt’ of the news item, but specifies a specific topic around working in the NHS. Barbara Windsor’s (BW) next turn is therefore the second pair part (SPP), which presents a disagreement with the FPP posted by EA. After two unrelated posts, EA posts a message which could be deemed a sequenceclosing third. In other words, it responds to the SPP but seemingly ends the

6  “On that note I’m signing out”: Endings of Threads in Online… 

123

sequence with a mobilisation of support for leaving the EU (“vote out”). However, BW continues the sequence (lines 18–21) with another oppositional turn (Vuchinich, 1990). Within BW’s turn there are two potential turn-constructional units which could be read as first pair parts (“so you’re up to date on carnets, and work permits are you?”—lines 19–20; and “you’re also aware of the concerns many (educated) expats have concerning Brexit?“—lines 20–21). In EA’s response they address each of these FPPs in the order in which they were posted. This message, therefore, provides an SPP to both of the FPPs in BW’s post, and so even without a sequence closing third, we could argue that a response is not necessarily expected. There is, of course, no coordinated closing sequence and in fact EA attempted to end the exchange. However, as long as the thread remains open, there could be a continuation of the thread by other posters, but the non-response is not necessarily a noticeable absence. In terms of the dispute, we could argue that this is a stand-­off, in that neither party has offered concessions, but the interaction has come to an end (Vuchinich, 1990).

Idiomatic Endings Another practice identified in the endings of threads is that the final post may often include an idiom. The following extract is the final post in a thread in which there has been a dispute about the amount of money being paid to the EU. Extract 6.5  (From The Independent: Vote leave wipes NHS £350m claim and rest of its website after EU referendum, 27th June 2016) 1 2 3 4 5 6 7 8 9

Thinker4 We pay more in overseas aid than we pay to the EU - for good reason it helps those countries strengthen and fosters partnerships with them. Its a huminatarian thing, but the real hard business reason is that it benefits our trade. The figures quoted by both sides now are different to yours. You have also forgotten money that is directed to different areas of the UK economy in grants... that is to science, medical research, agriculture etc etc. Sadly we are a nation of Turkeys who just voted for Christmas.

Thinker4 starts their post with a formulation of their argument (line 2) and then provides an account for this. At the end of their turn, they include a

124 

J. Meredith

three-part list (Jefferson, 1990) which is commonly treated as if it is conveying a more general point rather than it being the complete list. The poster emphasises that the list is not complete by also including “etc etc”. The final formulation of the turn includes an idiom (lines 8–9). Literature on topic shifts (Drew & Holt, 1998) has found that the use of an idiomatic summary or formulation can sometimes be a boundary of a topic change. This kind of idiom does not add anything new to the topic, but rather is backwardslooking in the sense that it summarises the previous topical sequence before moving onto the next one. In the extract above, the idiom summarises the commenter’s stance towards the final part of the post, that is, that the UK has voted for something that will make it worse off. While it is not always the case that an idiom leads to the ending of a thread, idiomatic phrases may be more difficult to offer a counter-argument to, precisely because they are idiomatic and as such this may lead to a non-response. In the next extract we see a similar practice, which is also at the end of a discussion about the amount of money sent to the EU. Extract 6.6  (From Daily Express: VOTE LEAVE VICTORY! Official figures reveal UK billed MORE than £350m each week by EU, 2nd August)

1 2 3 4 5 6 7 8 9 10 11 12 13

Kojak87 Using your example, our "contribution" is £20, our rebate is £6. We, as a country, physically transfer only the £14. That £6 is never transferred. By not being part of the club we won't have to send that £14 anymore but the Leave campaign were quoting the £20 which we never paid in the first place. We won't get £20 to spend on the NHS because £6 we kept and have already invested into the NHS... We have £20 to spend, not £26 (£20 plus the 6 we already held onto) which is the message the leave campaign were sending. Typical constituent Whether, which way, it is the same thing. None so blind as those who refuse to see, or are too dim to see.

In the first post Kojak87 presents a counter-argument to a previous poster’s analogy about the amount of money sent to the EU (data not shown). In their response, Typical constituent seemingly dismisses Kojak’s explanation of the differences as “it is the same thing”. They then offer an idiomatic formulation based on “none so blind as those who will not see” meaning that they cannot be forced to understand the point if they refuse to. They also add a sub-clause to the formulation, inferring that it is also

6  “On that note I’m signing out”: Endings of Threads in Online… 

125

that people are too unintelligent to understand the point. As with the previous extract, this idiomatic summary comes at the end of the post and has a finality to it, not least because the meaning of the idiom suggests that there is no point in continuing with the discussion. Therefore, both the idiomatic nature of the final message and its meaning indicate that further posts may not be relevant.

Personal Comments and Meta Talk In this final section of the analysis, I turn to exploring how posts end due to the topic of the interaction shifting from Brexit to more personal (or ad hominem) comments on the commenter. Extract 6.7 is responsive to a news article about a reporter presenting Boris Johnson with a cheque for £350 million.

Extract 6.7  (From The Independent: Boris Johnson ambushed with ‘£350m for NHS cheque, 4th October 2016) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Iago We aren't out of the EU yet, so this is a silly stunt. The £350 million figure was an exaggeration which many Brixiters didn't go along with (including Farage), but once the UK's contributions cease that will free up almost £200 million to be used as the elected government sees fit, including on the NHS. (1 post omitted) Conservative Idiots I don't know who is the biggest crackpot you or Boris. Donald Always cheap abuse from you lefties Never proper argument No facts figures and data Try going to debate without heckling Whack That's daft. You've constantly ignored the experts to side with self serving political wonks who feed you the lies you like and you gobble them up and regurgitate them with regularity. Donald Proving my point beautifully Whack You invite condescension and then you whine about it, Donald.

126 

J. Meredith

The initial post responds to the news article and comprises an assessment of the news story and an account for that assessment. Conservative Idiots (CI) responds to the initial post, not with a second assessment related to the topic, but instead with an implicit assessment of the poster themselves. Therefore, the topical content of the thread has shifted from the £350 million figure to the commenter. Donald’s post (lines 11–14) does not take up the topic, but rather highlights the action that the previous poster has engaged in; that is, they note the shifting of the type of talk away from the topic of Brexit to attacking the poster. In this way, Donald engages in ‘meta-talk’ where they are discussing the nature of the talk, rather than discussing any specific topic (Cannon, Meredith, Speer, & Mansell, 2019; Jager & Stommel, 2017). Donald’s post does not, though, simply address CI but also ascribes this type of behaviour to anyone in the category of “leftie” (see Meredith & Richardson, 2019). In the subsequent post the meta-talk itself is topicalised by Whack, who seemingly shifts the talk away from groups of people (Lefties) and back towards individual commenters, although “you” in lines 17 and 19 could be more general than just Donald. Donald’s response (line 22–23) continues with this meta-talk, with a slightly idiomatic formulation noting that Whack’s post does not include facts and figures. Donald does not, then, add any further content to the interaction, and instead their turn is somewhat backwards looking (Holt, 2010). In spoken interaction, topics and conversations may come to an end when no new topical information being added (Holt, 2010). Therefore, this potentially indicates that Donald is orienting to topic closure and thread ending. It is notable, though, that this does not close the topic as Whack posts a final message. In this post they again shift towards attacking Donald personally, and by characterising Donald as “whinger”, which implies that their complaint is potentially not valid (Edwards, 2005). At the end of this thread, then, there has been little discussion of Brexit as a topic, but rather the meta-talk has been topicalised and there have been ad hominem comments about posters. The ad hominem comment in the final post about the individual poster indicates that no new information is being added to the topics of ‘meta-talk’ or Brexit, and the thread ends potentially because there is no ‘on-topic’ information.

6  “On that note I’m signing out”: Endings of Threads in Online… 

127

Discussion This chapter has shown that, in contrast to spoken interaction and some forms of online interaction, there are no mutually coordinated and achieved closings in online comment threads. Where there are very rare closings, these tend to be unilateral (see Hutchby, 1996). Because there are no mutual closings it can sometimes be difficult to know why threads end, particularly as the threads remain open for comments unless closed by moderators. Therefore, it is not possible in all cases to draw conclusions based on the sequential and contextual nature of the thread. However, the analysis has shown that threads tend to end when there is the close of a sequence and so no response would be expected. Threads may also end when turn allocation fails, and a chosen next commenter does not return to the thread. This may indicate that there is a preference for the allocated next speaker to take the next turn, or at least to respond to messages which are specifically directed at them. Idioms can also sometimes lead to a thread ending, and in this way, they may function in a similar way to idiomatic phrases as topic boundaries (Drew & Holt, 1998). Finally, when the interaction shifts towards meta-talk, that is, commenting on the posters or the interaction itself, this may indicate that the substantive topic is exhausted which leads to the thread ending. The analysis demonstrates that online comment threads have different expectations in relation to closing than spoken interaction and also some other forms of online interaction (e.g., Pojanapunya & Jaroenkitboworn, 2011; Raclaw, 2008; Stommel & Te Molder, 2015). There are potential differences in affordances which may offer an explanation. Firstly, online newspaper comment threads are one-to-many interactions, with many potential participants. Therefore, there may not be a norm to ‘sign-off’ when it is not a one-to-one interaction. It is, in fact, notable that when sign-offs do occur it is when the interaction becomes more like a one-to-­ one interaction (i.e., it becomes a debate between just two posters). Secondly, the interaction itself tends to be conflictual, and conflicts do not necessarily end in a closing sequence (Vuchinich, 1990), but rather there may be a withdrawal from the interaction. Previous research on online conflicts has supported the notion that withdrawal rather than

128 

J. Meredith

compromise is a much more prevalent ending for an online conflict (Bou-Franch & Blitvich, 2014). Therefore, when commenters do not return to threads, we can argue that this constitutes a withdrawal from the dispute. One potential implication of this is that participants do not necessarily have to resolve disputes, and so facts, claims or assertions that may be untrue are left on the site either unchallenged or unresolved. In online forums it has been found that there is the potential for a community to develop, where posters know each other and can also look through the digital archives to access old posts (Giles, 2016). However, in online newspaper comment threads, this type of community seems to be less prevalent, at least in the data on Brexit. This may, of course, be an artefact of the fact that this is a particularly salient news story, and so there was likely to be a much wider group of individuals posting on these types of stories. It may be that if endings had been studied on, say, the online comment threads on The Guardian’s crosswords (see Cresci, 2015), the findings may have been different. However, on a much-discussed topic such as Brexit, there are many potential participants. This also means there are many potential overhearing recipients to these posts, and so unchallenged or unresolved disputes may have implications for public understanding of controversial issues. There are some challenges with analysing this kind of data using CA. CA has most commonly investigated spoken interaction, with studies of online interaction a lot rarer. Therefore, the knowledge we have about how online comment threads function is a lot less comprehensive than the knowledge we have about spoken interaction (although see Gibson, 2009; Stommel & Koole, 2010 for examples of the analysis of asynchronous online forums). Equally, much work focuses on one-to-one interactions, which show different norms and expectations to one-to-­ many interactions, such as online forums or comment threads. With this in mind, it can be a challenge to examine something like endings, as the knowledge we have of endings comes predominantly from the basis of the canonical closing (Schegloff & Sacks, 1973). There is a temptation to simply focus on how online closings or endings deviate from the canonical closing rather than looking at how they function in their own right

6  “On that note I’m signing out”: Endings of Threads in Online… 

129

(see Giles et al., 2015). When the norms and affordances of online interaction allow for significant differences in ending sequences compared to spoken interaction, it can be tempting to see them as in some way deficient or even rude. However, in this context withdrawal from the conflict seems to be a relatively ‘safe’ way of managing the ending of a thread. In online comments it seems that conditional relevance is much weaker than in talk, and so the escalation of conflicts may be avoided due to participants simply withdrawing from the interaction. Therefore, although it may seem that online news sites might be more conflictual or argumentative (Hughey & Daniels, 2013), it also seems that the escalation of conflicts may be avoided through withdrawing from the dispute.

Conclusion This chapter has examined the ways in which online newspaper comment threads end. Based on the analysis of the endings of over 2500 threads, it has shown that mutually coordinated closings do not occur in this context, with unilateral closings similarly rare. It is much more common for threads to simply end. When we examine the sequential context, we can see that some endings are non-responses to direct questions or failure of turn allocation. However, there are also cases where a sequence is closed and as such the non-response is not a noticeable absence. Finally, there are particular discursive practices which might be more common at the end of threads, such as idiomatic formulations. Similarly, a move away from a substantive topic towards commenting on posters themselves can also create a sequential environment in which a thread may end. The analysis shows how in this conflictual context disputes are rarely resolved, with a withdrawal from the dispute much more likely. Analysing such data using conversation analysis can present challenges due to the different affordances and interactional norms. However, analysing data in this way allows for an exploration of these practices and also to consider the implications that this might have for public debates on controversial topics.

130 

J. Meredith

References Bou-Franch, P., & Blitvich, P. G.-C. (2014). Conflict management in massive polylogues: A case study from YouTube. Journal of Pragmatics, 73, 19–36. British Psychological Society. (2017). Ethics guidelines for internet-mediated research. INF206/04.2017. Leicester: Author. Retrieved from https:// www.bps.org.uk/news-­a nd-­p olicy/ethics-­g uidelines-­i nternet-­m ediated­research-­2017 Cannon, C., Meredith, J., Speer, S., & Mansell, W. (2019). A conversation analysis of asking about disruptions in method of levels psychotherapy. Counselling and Psychotherapy Research. https://doi.org/10.1002/capr.12243 Coles, B. A., & West, M. (2016). Trolling the trolls: Online forum users constructions of the nature and properties of trolling. Computers in Human Behavior, 60, 233–244. Cooper, G., Green, N., Burningham, K., Evans, D., & Jackson, T. (2012). Unravelling the threads: Discourses of sustainability and consumption in an online forum. Environmental Communication: A Journal of Nature and Culture, 6(1), 101–118. Cresci, E. (2015, August 10). Get down with the Guardian’s crossword community, wonderful across the board. The Guardian. Retrieved from https:// www.theguardian.com/crosswords/crossword-­b log/2015/aug/10/ get-­d own-­w ith-­t he-­g uardians-­c rossword-­c ommunity-­w onderful-­ across-­the-­board Drew, P. (2005). Conversation analysis. In K. L. Fitch & R. E. Sanders (Eds.), Handbook of language and social interaction (pp.  71–102). Mahwah, N.J: Lawrence Erlbaum. Drew, P., & Holt, E. (1998). Figures of speech: Figurative expressions and the management of topic transition in conversation. Language in Society, 27(4), 495–522. Dynel, M. (2014). Participation framework underlying YouTube interaction. Journal of Pragmatics, 73, 37–52. Edwards, D. (2005). Moaning, whinging and laughing: The subjective side of complaints. Discourse Studies, 7(1), 5–29. Gibson, W. (2009). Intercultural communication online: Conversation analysis and the investigation of asynchronous written discourse. Forum Qualitative Sozialforschung/Forum: Qualitative Social Research, 10(1). Retrieved from http://www.qualitative-­research.net/fqs/fqs-­e/rubriken-­e.htm

6  “On that note I’m signing out”: Endings of Threads in Online… 

131

Giles, D., Stommel, W., Paulus, T., Lester, J., & Reed, D. (2015). Microanalysis of online data: The methodological development of “digital CA”. Discourse, Context & Media, 7, 45–51. Giles, D. C. (2016). Observing real-world groups in the virtual field: The analysis of online discussion. British Journal of Social Psychology, 55(3), 484–498. Graham, T., & Wright, S. (2015). A tale of two stories from “Below the Line” comment fields at the Guardian. The International Journal of Press/Politics, 20(3), 317–338. Heritage, J., & Atkinson, J.  M. (1984). Introduction. In J.  M. Atkinson & J. Heritage (Eds.), Structures of social action: Studies in conversation analysis (pp. 1–15). Cambridge: Cambridge University Press. Herring, S. C. (2013). Discourse in web 2.0: Familiar, reconfigured, and emergent. In D. Tannen & A. M. Trestor (Eds.), Discourse 2.0: Language and new media. (pp. 1–26). Georgetown University Press. Holt, E. (2010). The last laugh: Shared laughter and topic termination. Journal of Pragmatics, 42(6), 1513–1525. Hughey, M. W., & Daniels, J. (2013). Racist comments at online news sites: a methodological dilemma for discourse analysis. Media, Culture & Society, 35(3), 332–347. Hutchby, I. (1996). Confrontation talk: Arguments, asymmetries, and power on talk radio. Routledge. Jager, M., & Stommel, W. (2017). The risk of metacommunication to manage interactional trouble in online chat counselling. Linguistik Online, 87(8). https://doi.org/10.13092/lo.87.4179 Jefferson, G. (1990). List construction as a task and resource. In Interaction competence (pp. 63–92). Irvington Publishers. Joyce, J.  B. (2019) The sequential and moral (dis)order of public disputes: How speakers resist, partition and do being reasonable in talk-in-interaction. Unpublished thesis, Loughborough University Ksiazek, T.  B., Peer, L., & Lessard, K. (2016). User engagement with online news: Conceptualizing interactivity and exploring the relationship between online news videos and user comments. New Media & Society, 18(3), 502–520. Lindström, A., & Sorjonen, M.  L. (2012). Affiliation in conversation. In T.  Stivers & J.  Sidnell (Eds.), The handbook of conversation analysis (pp. 250–369). Wiley Blackwell. Markman, K. M. (2009). “So what shall we talk about”: Openings and closings in chat-based virtual meetings. Journal of Business Communication, 46(1), 150–170.

132 

J. Meredith

Meredith, J., & Richardson, E. (2019). The use of the political categories of Brexiter and Remainer in online comments about the EU referendum. Journal of Community & Applied Social Psychology, 29(1), 43–55. Pojanapunya, P., & Jaroenkitboworn, K. (2011). How to say “good-bye” in second life. Journal of Pragmatics, 43, 3591–3602. Raclaw, J. (2008). Two patterns for conversational closings in instant message discourse. Colorado Research in Linguistics, 21, 1–21. Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the organization of turn-taking in conversation. Language, 50(4), 696–735. Schegloff, E. A. (2006). Repair after next turn: The last structurally provided defense of intersubjectivity in conversation. In P. Drew & J. Heritage (Eds.), Conversation analysis, volume 1: Turn-taking and repair (pp.  215–266). London: Sage Publications. Schegloff, E.  A., & Sacks, H. (1973). Opening up closings. Semiotica, 8(4), 289–327. Singer, J. B. (2009). Separate spaces: Discourse about the 2007 Scottish elections on a national newspaper web site. The International Journal of Press/ Politics, 14(4), 477–496. Stivers, T., & Robinson, J. D. (2006). A preference for progressivity in interaction. Language in Society, 35(3), 367–392. Stommel, W. (2012). Salutations, closings and pronouns: Some aspects of recipient design in online counselling. Communication & Medicine, 9(2), 145. Stommel, W., & Koole, T. (2010). The online support group as a community: A micro-analysis of the interaction with a new member. Discourse Studies, 12(3), 357–378. Stommel, W., & Te Molder, H. (2015). Counseling online and over the phone: When preclosing questions fail as a closing device. Research on Language and Social Interaction, 48(3), 281–300. Vuchinich, S. (1990). The sequential organization of closing in verbal family conflict. In A. D. Grimshaw (Ed.), Conflict talk: Sociolinguistic investigations of arguments in conversations (pp. 118–138). CUP Archive. Waldvogel, J. (2007). Greetings and closings in workplace email. Journal of Computer-Mediated Communication, 12(2), 456–477. Werry, C.  C. (1996). Linguistic and interactional features of Internet Relay Chat. In S. C. Herring (Ed.), Computer-mediated communication: Linguistic, social and cross-cultural perspectives (pp. 47–64). Amsterdam: John Benjamins.

7 Similarities and Differences Across Settings: The Case of Turn Continuations in Instant Messaging Anna Spagnolli, Sonia Genovese, and Mattia Mori

Introduction When studying technology-mediated conversations, one frequently notices some resemblance to practices observed in spoken interaction. The question arises whether this noticing represents a beneficial insight, or instead, projects similarity across settings that are in truth very “It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

A. Spagnolli (*) • S. Genovese • M. Mori Dipartimento di Psicologia Generale, Università degli Studi di Padova, Padova, Italy Human Inspired Technology Research Centre, Università degli Studi di Padova, Padova, Italy e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_7

133

134 

A. Spagnolli et al.

different. On the one hand, applying familiar concepts to newly analysed phenomena is a welcome habit in science because it facilitates intersubjectivity and allows scholars to position a new study properly in the body of already accumulated knowledge (Given, 2008). On the other hand, conversational practices should be understood in terms of the “constraints and affordances” of the local context in which they occur (Schegloff, 2009, p.  375). Are the constraints and affordances of a medium too quintessentially unique, and are apparent similarities across media too superficial, to be pursued in the analysis? In their analysis of the historical innovations from perspective painting to photography, film, and television, Bolter and Grusin (2000) observe that the innovation process is rather one of re-mediation than of revolution: “New media take a property from one medium and reuse it in another. With reuse comes a necessary redefinition” (p. 45). Indeed, new technology around us does not overwrite past technology but retains a relation with it, such as electric lamps in chandeliers, personal computers incorporating the typewriter keyboard, and virtual reality being displayed on eyeglasses. Users first rely on the possibility of recognizing familiar affordances1 (Gaver, 1991) in new devices to understand what to make of them, and then adjust them to the affordances of the medium, thereby operating some degree of transformation of the original practice that might not alter it completely. In web chats, for example, one can find old-fashioned repair initiations stopping the progression of a sequence and highlighting a problem in conversation; at the same time, due to the affordances of the medium, one cannot find selfcorrections in first position, because they occur in the private screens of the individual users performing them (Schönfeldt & Golato, 2003). Re-mediation represents a balanced framework to appreciate the relationship between different generations of technical devices. In this chapter, we adopt this framework to understand the novelty of a widespread practice in instant messaging; that is delivering a contribution to the conversation  Affordances are properties of a certain medium that invite a specific action on the users’ part. Scholars of human–computer interaction no longer talk about how the technical properties of a medium affect the users’ behavior; they prefer to talk about how the affordances of a medium take shape in the encounter with the users’ needs. An affordance is a property that makes sense when a person finds a way to exploit it within a specific action (Gaver, 1991; Hutchby, 2001). Whittaker (2003) offers a classic discussion of the affordances of communication media. 1

7  Similarities and Differences Across Settings: The Case of Turn… 

135

across several subsequent messages so that one message becomes the extension of the previous message (Baron, 2010). This practice bears some resemblance with the conversational practices of turn continuation, and we would like to compare the two practices to consider how they unfold, what their interactional implications are, and how they relate to the affordances of the setting in which they occur. A turn is the individual contribution-at-talk of a participant in a conversation and ends when speakers alternate. A complete turn is semantically, intonationally, and pragmatically complete, and its end represents a transition relevance place (TRP; Clayman, 2013; Schegloff, Jefferson, & Sacks, 1974). Turn continuations are talk that recognizably extends a previous turn, or “host”; it is often (but not exclusively2) produced by the same speaker, and occurs after the TRP (Ford, Fox, & Thompson, 2002; Schegloff, 1996; Schegloff, 2016) and after a prosodic break from the host.3 Turn continuations can add new material to the host (increments) or take material from the host and replace it with other material (replacements; Vorreiter, 2003). Although all turn continuations are semantically integrated with the host, they can be syntactically integrated with it (“glue-ons” and “insertables”) or separate from it (“free constituents”; Couper-Kuhlen, 2012; Vorreiter, 2003). From an interactional perspective, turn continuations have been reported as ways to deal with issues of comprehension as well as alignment (Walker, 2004), including lack of response (Schegloff et al., 1974). Thus, in this paper we address whether extending a message in a chat is tantamount to continuing a turn. In addition, we examine whether a series of extended messages form a turn, so that transition to a different speaker only becomes relevant at the end of the series. We start by describing the practices of extending a message found in our collection and see what they imply about turn completion and turn transition.  Increments initiated or completed by a person other than the speaker have been reported (e.g. Lerner, 2004; Sidnell, 2012). 3  Ford (1993) also noticed the presence of a final intonation in the talk preceding postcompletion extensions. In addition, Couper-Kuhlen (1996) noticed the absence of a declination reset, namely of the increase in the tone frequency characterizing the start of a new utterance, between the host and some cases of turn continuation. 2

136 

A. Spagnolli et al.

The Design of Message Extensions The examples included in this chapter are extracted from a collection of 12 mediated conversations in Italian. In each conversation, four students sitting in different rooms talked via textual chat (Skype) about aspects of taking a short journey, such as choosing the destination, schedule, lodging, mode of transportation, and attractions. The type of textual chat used here allows messages to become visible to all participants once sent to the chat window by pressing the enter key; however, they are not visible while the user is typing. In addition, Skype makes it possible for multiple users to type or enter their messages simultaneously, in which case each message appears on a different line in the chat window with the same timestamp. We collected the log files of the conversations (10,093 lines), as well as the video recordings of the activity on the participants’ screens (96 hours), as suggested by Garcia and Jacobs (1999) and Markman (2006). For this chapter, we singled out the exchanges in the group conversation in which a message was extended by a subsequent message, finding 516 messages that were extended by one or many messages after them. All extracts included in this chapter are in Italian and are accompanied by an idiomatic interlinear translation to English.4 In our collection, messages are extended in two ways. The first way is undoubtedly similar to the methods with which we are familiar in face-­ to-­face interaction (as demonstrated by Tudini, 2015). An example is provided in Extract 7.1, beginning at a point of the conversation in which the four participants discuss their journey destination.  We would like to address the possible concerns (Kendrick, 2017) about our data being collected in a laboratory setting, where the researcher occasioned both the event and the task. Our analysis does not focus on the social event or on the decisions made in the task. The focus is on turn construction and turn-taking without interest in the related social psychological phenomena, such as power, status, identity, group performance, social influence, or linguistic phenomena (e.g. speech quality, lexicon, formality, and politeness). With respect to the practices considered in the present study and to the object of the study, we can then say “the natural living order of social activities as they are endogenously organized in ordinary life” (Mondada, 2013, p. 34) is not compromised by having collected the conversations in a laboratory. Moreover, the practices we analysed have already been reported in studies of naturally occurring conversations (e.g. Baron, 2010; Tudini, 2015), in addition to being very familiar to any user of instant messaging services. Baron observed what she called “utterance breaks” during chats between students. Tudini (2015) observed them during out-­ of-­class dyadic chats between acquainted or unacquainted classmates. Differently from the current study, none of Baron and Tudini’s chats focused on a specific activity other than chatting, which suggests that turn instalments are not specific of long, focused activities such as the ones in our chats. 4

7  Similarities and Differences Across Settings: The Case of Turn… 

137

Extract 7.1  (Chat e7) 1 → [10.40.27]

Massimo: india?

2

Giulio: va bene anche l’india... basta decidere...

India [10.40.54]

India is fine as well... just let’s make a decision 3 → [10.41.00]

Massimo: o anche egitto... or Egypt as well

Massimo’s message in Line 1 contains a proposal for a possible journey destination (“india?”), which is complete and has reached a TRP. In Line 3, he produces a second message, which starts with the conjunction “or”, adds a second possible journey destination (“or Egypt”), and acknowledges the prior proposal (“as well”). The extended message (or host) looked complete to all respects and another participant, Giulio, replied to it. The extension is semantically and syntactically connected with the host and performs the same action as the host, adding new material to it. Therefore, this episode looks like a classic turn continuation, more precisely a post-other-speakertalk increment (Schegloff, 2000); the two messages together represent a turn. The second type of extended message in our data, as we mentioned above, looks quite different from the retrospective type just considered, and is very familiar to every user of instant messaging services. Extract 7.2 provides an example. The participants are deciding upon their journey destination.

Extract 7.2  (Chat 15) 1

[12:08:18]

cippa: bella raga, qualcuno ha qualche idea di dove andare?

2

[12:08:21]

twin:

hey guys, any ideas of where to go? propongo la spagna I’ll propose Spain 3 → [12:08:23]

pippo: se ce la facciamo con i soldi if we can make it with the money

4 → [12:08:30] pippo: io propongo il Sudamerica I’ll propose South America 5

[12:08:39]

gigi: con 1000 euro fuori dall’Europa? outside Europe with 1000 euros?

138 

A. Spagnolli et al.

The extract starts with the request for journey proposals (Line 1) and the response from a participant called Twin, who proposes Spain as a possible destination (Line 2). In the third line, another participant, Pippo, writes, “if we can make it with the money”. Pippo’s message represents an incomplete action and as such, it can be taken either as an extension of a previous message or as the initial part of an action that will be extended in a subsequent message. Seven seconds later, Pippo sends a second message (“I’ll propose South America”; Line 4), extending his previous message both syntactically and semantically. It is now clear that the two messages represent one action (i.e. a new proposal). The connection between the two messages is also apparent to a user involved in the conversation, Gigi, whose message addresses both parts of Pippo’s turn (i.e. the budget issue mentioned in Line 3 and the proposed destination mentioned in Line 4: “outside Europe with 1000 euros?”). This second extract represents the type of extended message that is more commonly associated with mediated communication. While in the turn extension exemplified by Extract 7.1 the extending message is incomplete, in this second type of extension it is the extended message that is incomplete. With this practice, the production of subsequent messages that will complete the turn is projected. We will refer to this second type of extensions as prospective extensions and the various subsequent messages that contribute to composing the same turn are called instalments. In other words, instalments are messages designed to look like incomplete turns, to be completed in a forthcoming message from the same speaker. There are several ways to make a turn look incomplete. In Extract 7.2, the turn was designed to look incomplete by syntactical incompletion (“if we can make it with the money”); however, users have additional ways of designing a message to look incomplete. These ways rest on the need to complete the action carried out in the turn. Extract 7.3 provides an example.

7  Similarities and Differences Across Settings: The Case of Turn… 

139

Extract 7.3  (Chat 9) 1 → [13.09.12] rondine: 4°g: Che dedicheremo interamente al quartiere di # Notting Hill passando per

4th day. which will be entirely devoted to Notting Hill neighbourhood via PortobelloRoad (dove ci sono centinaia di bancarelle)

2

Portobello Road (where there are hundreds of stands) e Hide Park.Fine Giornata.Cena a PiccadillyCircus.

3

and Hyde Park. End of the day. Dinner at Piccadilly Circus. 4 → [13.09.38] rondine: 5°g si parte 5th day we leave 5

[13.10.00] mary:

che ricordo a tutte è il 24 non il 25... which I will remind every one of you is the 24th not the 25th

6

[13.10.00] rondine: e si torna a casa con l’aereo delle 17.40 and we return home with the 17.40 flight

The participant nicknamed “Rondine” produces one long message spanning across Lines 1 to 3 and listing the activities on the fourth day of the journey. Dinner is the last activity so the description of the fourth day is complete. The larger activity of summarizing the journey, however, cannot be considered complete because the fourth day is not the last day of the journey. Indeed, a new message is produced soon afterward, regarding the fifth day of the journey (“5th day we leave”) and representing a further instalment of the overall recapitulation activity. A very long turn at talk is in production here, and the turn is made visible piece by piece. In the previous part of the conversation, not reported here, Rondine summarized the first, second, and third days of the trip, each day in one message and each message starting with the number of the day. The design of her ongoing activity is then clear: each message—although complete in itself—will not complete the multiunit turn she has embarked on, until all five days are recapitulated. In Schegloff’s words, “what is at issue here is a course of conduct being developed over a span of time (…) to which co-participants may become sensitive, which may begin to inform their inspection of any next sequence [to] start to see whether or

140 

A. Spagnolli et al.

how it relates to the suspected project” (Schegloff, 2007, p.  244). The speaker has designed her activity to project that it will span across several messages. In sum, instalments are messages that do not complete a turn, because they are designed to look syntactically or pragmatically incomplete and to project the occurrence of subsequent messages completing the turn.5 In the next section, we will reflect on the interactional consequence of this practice.

Interactional Implications of Messages that Look Incomplete In Extracts 7.2 and 7.3, Pippo and Rondine took several messages (instalments) to complete their turn at talk; these messages appeared in the shared chat window one after the other, a few seconds apart. No technical issue truncated the first instalment, and users were not limited in the number of characters that a message could contain. Thus, we shall examine why Pippo and Rondine broke their turns across subsequent messages, running the risk of interruption before completing their contributions.6 Risky as it sounds, using instalments has several interactional advantages in this mediated setting, because instalments make parts of the turn available to the other interlocutors without waiting for the whole turn to be typed. From a conversation analysis perspective, the invisibility of turn production characterizing instant messaging is a limit because the interlocutors cannot promptly display issues of comprehensibility or alignment and because other speakers can take a turn before the typing is complete (Sacks, 1987; Schegloff et al., 1974). Delivering a turn in instalments can partially circumvent these limits, getting the speakers as close as possible to making their turn production accessible to their interlocutors. We shall analyse the advantages of this practice one by one.

 Installments are not necessarily TCUs; in some cases they represent possibly complete turns (when they are part of multiunit turns), while in other cases they are not possibly complete turns. 6  The feature that nowadays shows someone is typing was not yet available in the version used by our participants. 5

7  Similarities and Differences Across Settings: The Case of Turn… 

141

First, if interlocutors can access turn production during a conversation, then they can display their reception of the spoken turn at a proper time (back-channelling; Duncan & Fiske, 2015) and signal issues of comprehension or alignment. Something similar happened in Extract 7.3 above: at Line 5, Mary immediately addresses one of Rondine’s turn instalments (“5th day we leave”) by repairing it (“which I will remind everyone of you is the 24th not the 25th”). That is Mary replaces the ordinal reference Rondine (5th) used with a specific date (24th) that was the object of debate in a previous discussion (not reported in the extract). A second advantage of accessing the turn production in a conversation is claiming the turn for oneself and the action performed in it. This is useful in case other interlocutors are expected to try to claim the same action. Extract 7.4 provides an example of this risk and how using instalments can help in the competition for the turn. The extract starts with Fabrizio’s proposal of recapitulating the decisions made. In the first few lines of this extract (Lines 1–5), the interactants agree on recapitulating the decisions made thus far about the journey. Who will be in charge of the recapitulation is not clear, but a message in which someone in the group will offer a recapitulation is relevant, possibly starting from the departure decision as suggested by Fabrizio. Monica takes the turn and recapitulates departure information in three subsequent messages at Lines 7, 8, and 9. Subsequently, Nicola takes the turn to rephrase Monica’s information with more accuracy and continues by adding the recapitulation of other decisions (Lines 10, 11, and 13). Then, at Line 20, Monica adds a further instalment, which seems to conclude the recapitulation project (“then visiting what we said above”). The decisions made in the group are recapitulated in the same temporal order in which the journey would unfold (i.e. leaving the city to the airport, taking off, arriving at the destination, having dinner, and visiting attractions). Overall, the recapitulation is prepared in Lines 1–5, but takes place in Lines 6–11 and then Lines 13, 14, and 20. In this case, each instalment of the recapitulation is a single phrase or single clause TCU (Schegloff et al., 1974) of a multiunit turn (Schegloff, 1982, 2007) so that the occurrence of a new message continuing the recapitulation is projected as relevant until the recapitulation is finished.

142 

A. Spagnolli et al.

Extract 7.4  (Chat 6FB) 1

[14.51.36] fabrizio: allora ricapitoliamo la situazione?

2

[14.51.43] nicola: se bisogna ok

3

[14.51.52] carlotta: chi racapitola?

4

[14.51.54] fabrizio: partenza?

5

[14.51.56] monica: vai!

6

→ [14.51.58] nicola: dunque

7

→ [14.52.11] monica: prtenza ore 8.30 pd

8

→ [14.52.14] monica: taxi

9

→ [14.52.23] monica: volo parte ore 13.30

now shall we recap the situation?

ok if needed

who does the recap? departure?

go ahead! so

departure at 8.30 Padova taxi flight leaves at 13:30

10 → [14.52.29] nicola: aereo alle 13.20 da venezia marco polo

flight at 13:20 from Venice Marco Polo

11 → [14.52.39] nicola: arrivo alle 18 circa a mosca

arrival at about 18:00 in Moscow

12

[14.52.40] monica: ok

ok

13 → [14.52.47] nicola: mercoledi 30 aprile wednesday april 30

14 → [14.53.01] monica: cena da zar al gran Ïhotel

dining like czars at grand hotel

15

[14.53.06] nicola: perfetto

16

[14.53.07] fabrizio: perfetto

17

[14.53.08] monica: + vodka

18

[14.53.09] fabrizio: esatto

19

[14.53.23] nicola: partita a scacchi con kasparov

perfect

perfect

+ vodka

correct

chess match with kasparov

20 → [14.53.25] monica: poi visitare quel che si è detto prima then visiting what we said above

21 → [14.53.27] carlotta: mi piace mi piace

I like it I like it

7  Similarities and Differences Across Settings: The Case of Turn… 

143

Extract 7.4 shows how concrete the risk is that another speaker could claim the same action that one wants for herself. Nicola starts at Line 6 with “so”, which is a turn initial particle used to take the turn; the message ends there, in the middle of the TCU, projecting the rest of the recapitulation in a subsequent message. Monica pre-empts Nicola and continues the recapitulation herself. Later on, as soon as she finishes recapitulating the departure, Nicola takes the turn and continues the recapitulation. Being quick at sending messages by breaking a turn into instalments, instead of typing the whole recapitulation in private before sending it, is therefore a decisive way to take charge of an activity. A third advantage of using messages that are designed to look incomplete is that they give the speaker a “TRP waiver” (i.e. they deprive the end of the message of its value as a TRP). Indeed, what makes instalments so expedient is that they mimic the accessibility of turn production to the interlocutors while deactivating the relevance of a turn alternation after each instalment. Its strategic use is shown in Extract 7.5, in which the speaker claims the turn before being even ready to perform the whole action. The extract starts with a discussion about airport transfer. Alis proposes renting a car and a fast-paced exchange follows. The sequence of interest starts at Line 15. Alis’s message in Line 15 is an announcement, which informs the interactants about the model of a car, the duration of the rental, and the nature of the information to be provided next (i.e. the cost). The message is syntactically incomplete, recognizably projecting the upcoming subsequent message with the missing piece of information about the cost. The video recording of Alis’s activity on the screen, which is also available in our collection along with the textual logs of the chats, reveals that she does not have that information when she sends the first instalment: Alis inspects the rental-agency price table in her Internet browser before sending the second message containing the cost information. Therefore, by sending an instalment of a longer action, Alis takes advantage of the TRP waiver, and fetches the missing information while the interlocutors withdraw from intervening. To summarize, instalments are message designed to look like incomplete turns and to project completion in subsequent messages from the same speaker. As such, they allow the speaker to mimic the sharing of

144 

A. Spagnolli et al.

Extract 7.5  (Chat 14) 1 [15:09:15] bru:http://www.prague-airport-transfers.co.uk/transfers-it.htm 2

15:09:22] [ bru: spostamenti transfer

3 [15:10:04] alis: e se ci noleggiamo una macchina la? what about renting a car there? 4

[15:10:18] adel:

si la macchina!!!

5

[15:10:39] alis: http://www.andel3w.dk/praga-trasferimenti-it.shtml

yes, a car!!! 6 [15:10:43] bru:

io direi una Maserati cabrio I’d say a Maserati cabrio

7 [15:10:50] bru: 8

(clap)

[15:10:54] adel: il migliore you’re the best

9 [15:10:57] laulu: ci sto anche IO!!!! I am in!!! 10 [15:11:21] alis:

ma c’è nell’elenco? but is it in the list?

11 [15:11:32] bru:

no skerzavo no I was kidding

12 [15:12:24] laulu:

ma per spostarci dall’aereoporto? to move from the airport?

13 [15:12:33] alis:

che macchina scegliamo?

14 [15:12:40] alis:

in teoria si..

what car shall we chose? in principle yes 15 → [15:13:17] Alis:

per otto giorno la ford transit costa a ford transit for eight day (sic) costs

((the rental fairs table in Alis's browser opens and covers the instant messaging window)) 16 → [15:13:27]

Alis:

197 euro..

197euros

turn production to a certain extent, without waiting for the whole turn to be typed. This has the advantage of collecting timely signs of alignment or understanding from the interlocutors, as well as claiming the action before other interlocutors do. At the same time, by projecting an upcoming message that contributes to completing the turn, instalments

7  Similarities and Differences Across Settings: The Case of Turn… 

145

avoid having the end of a message being taken as the end of the turn and making speaker alternation relevant after it, except for displaying alignment or repair initiations. In the next section, we reflect on the end of the instalment series and then on the TRP waiver they involve.

Negotiating the End of the Instalment Series When does a turn look incomplete? In conversation analysis, a turn is signalled as complete when the speaker (a) syntactically, pragmatically, and semantically reaches a possible completion and (b) stops talking (Schegloff et al., 1974). Furthermore, turn completion needs to be recognized by the interlocutors to become consequential in a conversation. In this section, we will see that completion after instalments can be ambiguous, and we reflect on the interlocutors’ role in defining it. Once a user has claimed the right to keep the turn past a single message by using a prospective extension, he or she has deactivated the transition relevance places for as many instalments as needed to complete the turn. Because this number is not determined, constrained, or specified in advance, a user can defer reaching the transition relevance place and add more messages. Pippo in Extract 7.6 provides a case-in-point when providing information about a flight. Pippo, in Line 1, proposes a course of action (i.e. looking for the cheapest flight). One minute later, he enacts that course of action by sending a message that contains information about a flight; that is its schedule and code (Line 3) but not yet the price. Four seconds later, he adds another message that is syntactically and semantically continuous to the previous message and provides the information about prices. Afterward, Pippo adds more messages about prices (Line 6: “outbound”, Line 8 “on the way back they rip us off though”). From Line 9 (“maybe it is more affordable to buy one way only”) onwards, Pippo sends messages that elaborate upon the price information already provided without adding any new information (Line 10: “and then we take advantage of the offer”; Line 11 “from stansted”). The first message sent from a different user is Cippa’s in Line 12 (“and for the return flight, we’ll figure out once we’re there”). Cippa agrees with his interpretation of Pippo’s recommendation at Line 9 (i.e. not buying the return flight). Cippa’s message suggests that he treated Line 9 as the end of Pippo’s

146 

A. Spagnolli et al.

Extract 7.6  (Chat 15) 1

→ [12:28:19]

2

[12:28:47]

pippo: vediamo i voli più economici quando ci sono let’s see when are the cheapest flights cippa: ma se ci dividessimo i compiti e poi ci consultiamo? what if we split tasks and then discuss?

3 → [12:29:22]

pippo: Sat, 15 Jul 06

Flight

FR 799

Sat, 15 Jul 06

Flight

FR 799

4 → [12:29:26]

pippo: 20 euro

5

twin: e la città di destinazione?

20 euros [12:29:30]

and the city of destination? 6 → [12:29:30]

pippo: andata

7

cippa: ci sono le tasse da aggiungere

outbound [12:29:41]

there are taxes to add 8 → [12:29:43]

pippo: al ritorno ci spennano, però

9 → [12:29:52]

pippo: forse conviene per andare prendere solo andata

10→ [12:30:02]

pippo: e poi si approfitta dell’offerta

on the way back they rip us off though maybe it is more affordable to buy one way only and then we take advantage of the offer 11→[12:30:05]

pippo: da stansted

12 [12:30:10]

cippa:

from stansted (sic) e al ritorno ci penseremo la and for the return flight, we’ll figure out once we’re there 13 [12:30:18] gigilatrottola: non mi sembra saggio it does not sound wise 14

[12:30:19] pippo: no, dobbiamo organizzare ora no, we must organize it now

proposal and then as a TRP.  The screen activity confirms this (Fig.  7.1): Cippa starts typing the reply when Pippo’s message at Line 9 is available (“maybe it is more affordable to buy one way only”), before the subsequent instalments. The subsequent instalments are actually left unaddressed. This extract shows the grey area of the TRP. In the context of instant messaging, as we showed in the previous section, the end of a message cannot be a TRP. The habits and the familiarity users have with this medium allow

7  Similarities and Differences Across Settings: The Case of Turn… 

147

9 pippo: forse conviene per andare prendere solo andata maybe it is more affordable to buy one way only 10 e poi si approfitta dell’offerta and then we take advantage of the offer 11 da stansted from stansted 12 cippa: e al ritorno ci penseremo la and for the return flight, we’ll figure out once we’re there

Fig. 7.1  The Skype window of one participant in the group discussion, Cippa, magnified from the recording of his screen activity

them to know that the speaker might not have finished the turn when one message is sent and might need more messages to finish it. On the other hand, the absence of an immediate answer can encourage the current speaker to keep sending messages, according to the “current speaker continues” practice (Schegloff et al., 1974). When the recipients call an end and take the next turn is a grey area, with ample space for negotiation. A similar case of a recipient stepping in to take the turn is seen in Extract 7.7. In Lines 1 to 5, the user Vecchiaspugna sends an initial message with a general request to the other users (“girls let’s make a decision”). She also sends three messages enacting such request (“Barcellona”, “Vienna”, “London”) and two further messages clarifying a general criterion for carrying out the request (“someplace not too far away”, “or we’ll waste two days just to travel”). Vecchiaspugna sends these six messages a few seconds apart from each other. Simultaneously with the fifth message, the user Vecchiaruga intervenes to propose some destinations, which means she treated the third example of a possible destination provided by Vecchiaspugna as a TRP.  The request was clear enough, and then Vecchiaruga could take the turn and respond to it. The last two extracts illustrate that, in this mediated context, what counts as TRP can be open to negotiation. The user doing the extensions can take advantage of the deactivated TRP and continue extending the turn until one of the recipients treats one of the instalments as completing the turn and starts talking.

148 

A. Spagnolli et al.

Extract 7.7  (Chat e5) 1 → [10:49:44] vecchiaspugna: ragazze decidiamoci girls let’s make a decision 2 → [10:50:03] vecchiaspugna: barcellona barcellona 3→ [10:50:09] vecchiaspugna: vienna vienna 4→ [10:50:16] vecchiaspugna: londra london 5→ [10:50:32] vecchiaspugna: qualcosa non troppo lontano some place not too 6

[10:50:33] vecchiaruga:

far away

helsinky se cosi si scrive oppure barcellona helsinky (sic) if this is how they spell it or barcellona

7

[10:50:34] maregiulia:

rivotiamo ...ma ad Amsterdam non ci vuole andare nessuno?! let’s vote again...but what about Amsterdam doesn’t anybody want to go there?

8 → [10:50:45] vecchiaspugna: altrimenti ci mangiamo 2 giorni solo di viaggio or we’ll waste two days just to travel 9 [10:50:49] vecchiaruga:

per me va bene anche amsterdam I’m ok with Amsterdam as well

10 [10:51:07] franvacanza:

ok per amsterdam Amsterdam then

11 [10:51:19] maregiulia:

..è vicina e ci si sta bene! it’s close and pleasant

12 [10:51:24] vecchiaspugna:

ok, ma a visitare cosa??? ok but to visit what???

Conclusions The practice described in this chapter consists of prospectively extending a message with an unspecified series of subsequent messages. More specifically, it consists of the following: • sending messages designed to look like incomplete turns and to project completion in subsequent messages from the same speaker;

7  Similarities and Differences Across Settings: The Case of Turn… 

149

• delivering subsequent instalments of the same turn until the latter is completed or the recipients treat it as such. The interactional implication of this practice is to take the turn and claim the action in one message, while deferring the completion of both turn and action to a later message. The incompleteness of each instalment minimizes the risk of interruption, because it shows that turn alternation is not relevant after each instalment except for displaying alignment or repair initiations.

Novelty of the Practice Although prospective turn continuations have not been the object of much research, practices such as the one we described in the previous paragraph can easily be found, perhaps under different names. For instance, McHoul (1978) studied intra-turn pauses in teacher’s talk, as shown in the following extract.

McHoul (1978: Extract 30) T: So we have a concentration (0.2)

T: of commercial activities (0.2)

T: in the heart of the city (0.2) T: then of course we must have smaller regional (0.4)

T: er shopping centres or shops (0.4)

T: e::r satisfying customers T:

(0.2)

on the outskirts of town

The teacher in the extract delivers her turn in instalments, pausing after each one. Carr and Smith (2014) reported a similar practice in motivational interviews and thought that it gave the interviewer’s talk a sort of

150 

A. Spagnolli et al.

Carr & Smith, 2014, Extract D T: And sometimes following a reflection (1.2) we need to allow a little bit of time (.9) right, for it to sit.

client-centred directiveness. The extract below is from a training activity in which interviewers practised a form of “managed disfluency” when talking. Here, we considered the practice of prospectively extending a turn as a type of turn continuation. The practices of turn continuation reported in the literature occur after a turn has already recognizably reached its completion (i.e. retrospectively; Schegloff, 1996, Ford et al., 2002, Vorreiter, 2003), much like the case we examined in Extract 7.1 in our collection and those Tudini (2015) examined in her study of instant messaging. Despite a difference in timing, however, both prospective and retrospective continuations consist of (a) producing talk that is designed to be semantically, pragmatically, and, mostly, syntactically integrated with previous talk and (b) deactivating the TRP of the continued talk.7 In prospective continuations, the TRP is deactivated in advance, whereas in retrospective continuations, the TRP is deactivated afterward. However, both practices allow speakers to revise the ends of their turns and keep the floor. Moreover, prospective continuations can smoothly leave room for retrospective ones in the same turn, as we observed in Extracts 7.6 and 7.7. Figure 7.2 offers a schematic representation that elaborates on Vorreiter’s (2003) taxonomy of turn continuations to include prospective ones.  In some cases, the speaker addresses these issues immediately by continuing to talk (“next-beat increment”, Schegloff, 2000, or “extensions of first pair part”, Ford, 1993). In other cases, the speaker continues after a gap, treated as anticipating issues of comprehension or preference (“post-­ gap increments”, Schegloff, 2000, or “prompted by pause”, Ford, 1993). Finally, turn continuations can occur after a recipient’s response, which again is treated as signaling issues of preference or comprehension (“post-other-speaker-talk increments”, Schegloff, 2000, or “accounts of dispreferreds”, Ford, 1993). In all cases, the turn continuation retrospectively transforms an inter-turn gap into an intra-turn gap, moving the TRP to the end of the extension. 7

7  Similarities and Differences Across Settings: The Case of Turn… 

151

turn continuation prospective

retrospective

(installments)

(host + increments)

Fig. 7.2  An extension of Vorreiter’s (2003) turn continuation taxonomy

In sum, prospectively extended messages are not a novel practice; they can be found in face-to-face conversation (e.g. Carr & Smith, 2014; McHoul, 1978) and they share the core features of a larger family of practices known as turn continuations.

 eculiarities of Prospective Continuations P in Mediated Settings In instant messaging, prospective turn continuations have an interactional result peculiar to this context: they mimic accessibility of turn production by posting a turn piecemeal. This result is possible because an affordance of the medium, the “send” button, makes a message visible to recipients. Visibility relates to two other affordances of this medium: the invisibility of message composition and the possibility of more than one participant composing messages simultaneously. Therefore, although prospective extensions are also present in face-to-face interaction, their usefulness in this mediated setting increases because they allow users to overcome a technical constraint. Also, compared with prospective extensions produced in face-to-face settings, turn continuations performed in mediated settings are more discernible. Whereas instalments are “only” separated by silence (intra-turn pauses) in face-to-face settings, in instant messaging they are more conspicuous as a delimited unit: each instalment is a message with a sender, time stamp, and content spatially separated from all other messages. This makes them more discernible, and permanently so. Messages are an interesting unit: they can accommodate a complete turn, and be followed by a TRP; or, through prospective or retrospective

152 

A. Spagnolli et al.

continuations, they are not complete turns and no TRP follows them. The status of the message in this setting testifies that the length of conversational units is not predefined but answers to local contingencies (Ford, 2004). In conclusion, extended messages in the context of instant messaging are the re-mediation of a familiar practice (i.e. turn continuations), with interactional implications that are accounted for by the specific affordances available in the mediated setting. The comparative study of mediated and familiar settings shows the constant work of adjustments and creativity that characterizes conversational practices.

References Baron, N.  S. (2010). Discourse structures in instant messaging: The case of utterance breaks. Language@ Internet, 7(4) Retrieved from https://www.languageatinternet.org/articles/2010/2651 Bolter, J. D., & Grusin, R. (2000). Remediation: Understanding new media. The MIT Press. Carr, E. S., & Smith, Y. (2014). The poetics of therapeutic practice: Motivational interviewing and the powers of pause. Culture, Medicine, and Psychiatry, 38(1), 83–114. Clayman, S. E. (2013). Turn-constructional units and the transition-relevance place. In J. Sidnell & T. Stivers (Eds.), The handbook of conversation analysis (pp. 150–166). Wiley-Blackwell. Couper-Kuhlen, E. (1996). Intonation and clause combining in discourse: The case of because. Pragmatics, 6(3), 389–426. Couper-Kuhlen, E. (2012). Turn continuation and clause combinations. Discourse Processes, 49(3–4), 273–299. Duncan, S., & Fiske, D. W. (2015). Face-to-face interaction: Research, methods, and theory. Routledge. Ford, C. E. (1993). Grammar in interaction: Adverbial clause in American English conversation. Cambridge: Cambridge University Press. Ford, C.  E. (2004). Contingency and units in interaction. Discourse studies, 6(1), 27–52. Ford, C. E., Fox, B. A., & Thompson, S. A. (2002). Constituency and the grammar of turn increments. In C. E. Ford, B. A. Fox, & S. A. Thompson (Eds.), The language of turn and sequence (pp. 14–38). Oxford: Oxford University Press.

7  Similarities and Differences Across Settings: The Case of Turn… 

153

Garcia, A., & Jacobs, J. (1999). The eyes of the beholder: Understanding the turn-taking system in quasi-synchronous computer-mediated communication. Research on Language and Social Interaction, 32(4), 337–367. Gaver, W.  W. (1991). Technology affordances. In Proceedings of the SIGCHI conference on Human factors in computing systems (pp. 79–84). ACM Press. Given, L. M. (Ed.). (2008). The SAGE encyclopedia of qualitative research methods (Vols. 1–2). Thousand Oaks, CA: SAGE Publications. Hutchby, I. (2001). Technologies, texts and affordances. Sociology, 35(2), 441–456. Kendrick, K.  H. (2017). Using conversation analysis in the lab. Research on Language and Social Interaction, 50(1), 1–11. Lerner, G. H. (2004). On the place of linguistic resources in the organization of talk-in-interaction: Grammar as action in prompting a speaker to elaborate. Research on Language and Social Interaction, 37(2), 151–184. Markman, K. M. (2006). Computer-mediated conversation: The organization of talk in chat-based virtual team meetings (Doctoral dissertation). The University of Texas at Austin. McHoul, A. (1978). The organization of turns at formal talk in the classroom. Language in society, 7(2), 183–213. Mondada, L. (2013). The conversation analytic approach to data collection. In J. Sidnell & T. Stivers (Eds.), The handbook of conversation analysis (pp. 32–56). Malden, MA: Blackwell. Sacks, H. (1987). On the preferences for agreement and contiguity in sequences in conversation. In G. Button & G. R. E. Lee (Eds.), Talk and social organisation (pp. 54–69). Clevedon: Multilingual Matters. Schegloff, E. (2009). One perspective on conversation analysis: Comparative perspectives. In J. Sidnell (Ed.), Conversation analysis. Comparative perspectives. Cambridge: Cambridge University Press. Schegloff, E., Jefferson, G., & Sacks, H. (1974). A simplest systematics for the organization of turn-taking for conversation. Language, 50(4), 696–735. Schegloff, E. A. (1982). Discourse as an interactional achievement: Some uses of ‘uh huh’and other things that come between sentences. In D. Tannen (Ed.), Analyzing discourse: Text and talk (pp. 71–93). Georgetown University Press. Schegloff, E. A. (1996). Turn organization: One intersection of grammar and interaction. In E. Ochs, E. A. Schegloff, & S. A. Thompson, (Eds.), Interaction and grammar (pp. 52–133). Cambridge University Press. Schegloff, E. A. (2000). On turns’ possible completion, more or less: Increments and trail-offs. Paper delivered at the 1st Euroconference on Interactional Linguistics, Spa, Belgium.

154 

A. Spagnolli et al.

Schegloff, E. A. (2007). Sequence organization in interaction: A primer in conversation analysis I (Vol. 1). Cambridge University Press. Schegloff, E. A. (2016). Increments. In J.D. Robinson (Ed). Accountability in social interaction (pp. 239–263). Oxford University Press. Schönfeldt, J., & Golato, A. (2003). Repair in chats: A conversation analytic approach. Research on Language and Social Interaction, 36(3), 241–284. Sidnell, J. (2012). Turn-continuation by self and by other. Discourse Processes, 49(3–4), 314–337. Tudini, V. (2015). Extending prior posts in dyadic online text chat. Discourse Processes, 52(8), 642–669. Vorreiter, S. (2003). Turn continuations: Towards a cross-linguistic classification. InList (Interaction and linguistic structures) No. 39. Walker, G. (2004). On some interactional and phonetic properties of increments to turn in talk-in-interaction. In E.  Couper-Kuhlen & C.  E. Ford (Eds.), Sound patterns in interaction (pp. 147–169). John Benjamins. Whittaker, S. (2003). Theories and methods in mediated communication. In: A. Graesser, M. Gernsbacher & S. Goldman (Eds.), The handbook of discourse processes (pp. 253–293). Erlbaum.

8 The Spectre of ‘Ghosting’ and the Sequential Organization of Post-match Tinder Chat Conversations Christian Licoppe

Introduction Mobile dating applications such as Tinder—which is mostly, though not exclusively, used by heterosexuals—are changing the way we meet potential partners for ‘romantic’ relationships. The use of dating apps involves selecting people with profiles of interest and getting in touch with them (in Tinder, this is mediated by the swipe right/match device). Then, users may chat online on the platform, which they do on their smartphones, and from there they may move on to a first face-to-face encounter if, and

“It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

C. Licoppe (*) Department of Social Science, Telecom Paristech, Palaiseau, France e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_8

155

156 

C. Licoppe

when, they feel like it. The chat conversation is therefore a crucial feature in the development of an encounter through dating apps. While a significant amount of research on Tinder (and dating apps in general) has focused on profile selection and swiping (see for instance David & Cambre, 2016; Ranzini & Lutz, 2017), post-match Tinder conversations have been little studied, in spite of the importance Tinder users grant them. From outside and from a technical outlook, this form of “textual interaction” (Reed, 2001), a particular type of “mediated interaction” (Arminen, Licoppe, & Spagnolli, 2016), looks a bit like various other forms of interpersonal texting, messaging and chatting which “creolize” writing and talking (Baron, 1998). It can also be construed as “online talk”, which as such may be analysed through the use of conversation-­ analytic methods (Meredith & Potter, 2013). However, my aim here is more oriented towards getting an anthropological understanding of this phenomenon and using a particular type of CA-inspired sequential analysis to show how distinctive these Tinder conversations are, and in particular how certain recognizable and specific concerns are embedded within their sequential organization. It is often the case that large data sets of textual interaction are retrievable from technology-mediated archives or resources. The data used here are more limited in size and have been made available by Tinder users from their conversation archives (as text) or as screenshots. This has important implications for analysis, for such data are mostly text, and sever most of the links to the context of their production in time. What is lost is the way in which such textual conversation threads are embedded in off-screen and on-screen activities and ecologies, that is texting and chatting as ‘lived work’. Various studies have tried to use audio and video recordings of typing and texting practices to show the sequential implications of the situational embeddedness of text-based interactions as they unfold in time (see for instance Meredith & Stokoe, 2014; Reeves & Brown, 2016). While such studies keep more in line with ethnomethodogical and conversationanalytic concerns regarding the situatedness of textual interaction, they cannot be deployed on the time scale of Tinder conversations, for this would raise unmanageable practical and privacy issues. Some form of CA-inspired analysis may still be done on such decontextualized textual threads, however, using CA concepts “heuristically to assist us in the initial and tentative explication of our data, as ‘aids to a

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

157

sluggish imagination’” (Greiffenhagen & Watson, 2009, p. 71). The trick is to use a reductive version of context, in which the relevant context is taken to be the previous message, as if it were just occurring. Such a move is made possible by the textual design of the chat interface in which messages are presented as dialogic threads of messages to be read one after the other. Messages therefore have to achieve some form of sequential relatedness (Reed, 2001) which is why they are routinely understood and described by participants as ‘conversations’. They appear experienced as organized around mutual accountability and sequential implicativeness. It is this orientation which makes some concepts of CA relevant to the analysis of textual interaction threads (for a review of the ways in which CA has been used with respect to online interaction, see Paulus, Warren, & Lester, 2016), especially when it comes to understanding addressivity and coherence as an accomplishment (Stommel & Koole, 2010). Such an orientation provides an entry point for the application of what we could call CA-inspired micro-analysis of online discourse (or MOOD, see Giles et al., 2015). Such a CA-inspired approach has been used for online textual interaction (see for instance Rintel, Mulholland, & Pittam, 2001; Antaki, Ardévol, Nunez, & Vayreda, 2005; Stommel & Te Molder, 2015). I will use here two specific CA analytic resources to carry out my analysis. The first is ‘next position proof ’: to achieve relatedness and the ‘conversational’ character of their threads, Tinder chatters orient to next messages as responses, which can be scrutinized for the kind of understanding of previous messages that they display (see also Vayreda & Antaki, 2009). The second is adjacency pair organization: I will show how participants orient to adjacency as a powerful resource to achieve relatedness and other things. When speaking of adjacent pairs here, I will only use their formal properties (as stated in Sacks, Schegloff, & Jefferson, 1974), namely that the production of a recognizable first pair part makes conditionally relevant certain types of responses.1 However, ‘responses’ are produced very differently in talk-in-interaction. First, messages are  This differs from the way adjacent pairs may work in talk-in-interaction in two respects. First, the links between adjacency pairs, turn-taking and the emergent character of talk as unfolding, are severed in this type of electronic, message-based communication. Second, there is a loosening of the way participants orient towards ‘adjacency’ and ‘nextness’. With chats, a recognizable second pair part is expected in the next contribution, but not necessarily in the next message, or in first position in the next message. 1

158 

C. Licoppe

often accessed sometime after their sending, as a self-contained bit of text to read, with almost no knowledge of the circumstances of their production. Second, while actual responses may be produced unreflexively and in the heat of reading the last incoming message, responses may also be produced later, after a lot of thinking and editing. Goffman described conversational interaction turns as “interaction moves” and adjacent pairs as “interlocking interaction moves” (Goffman, 1981). He was criticized by CA researchers for arbitrarily separating talking and listening, which does not work for ordinary conversation, where talking and listening are coordinated moment-by-moment (Goodwin & Goodwin, 2004). With textual interaction however, production (writing) and reception (reading) are separate in time, and Goffman’s perspective becomes more relevant, as well as other aspects of his work. Bringing Goffman into the fore allows us to use the distinction he proposes between ‘systemic’ and ‘ritual’ constraints in the production of interaction. Systemic constraints refer to the participants’ expectations that conversational actions are mutually intelligible, and that one person (and only one) is speaking. It can also refer to the need of “sustaining whatever is felt to be appropriate by way of continuity of topic and tone from previous speaker’s statement to current speaker’s” (Goffman, 1981, p. 18). Sequential matters, which are of special interest to CA, are an integer part of systemic constraints. Participants are not only held accountable with respect to such ‘systemic constraints’, but also with respect to ‘ritual constraints’. Ritual constraints refer to the notion that conversational actions can be, and are, scritinized with respect to how they achieve a proper kind of deference to co-participants or not, how they impinge or preserve their face and their ‘personal territories’, etc. What is actually being done at the systemic level (the formal description of which is the self-professed goal of conversation analysis) is unavoidably accountable at the ritual level, and this joint accountability is sensitive to the particulars of situations, settings and ongoing activities. A potential break in the fluency of the conversational flow, and in particular an extended silence which might be taken as a non-response, is potentially accountable both as a sequential problem and as a ritual offence. Similarly, an abrupt leave-taking where one participant ‘unilaterally’ deserts the interactional scene raises sequential issues, both at the

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

159

systemic and the ritual level. At the systemic level, it eschews the proper collaborative accomplishment of closings and pre-closings. At the ritual level, since the way in which one leaves an interaction can be treated as a kind of comment regarding the said encounter, interactional desertion can appear as an offensive kind of withdrawal in which the leave-taker “betrays his encounter … by taking leave in a precipitous fashion”, and even as a rejection of the other party (Goffman, 1971, p.188). So, what can we say then of the sequential organization of Tinder chat ‘conversations’? How distinctive are these with respect to other kinds of online textual interactions?

Data and Method In the Tinder study we first recruited a sample of twenty users in Paris (Ten female and ten male), through a mix of personal relationships, word of mouth and ads. This provided us with a typical sample of young urban professionals (between the ages of twenty and forty), with all of the twenty participants being educated at college level or more. The study involved two hour-long qualitative interviews during which the participants navigated the app and commented on it, and when they agreed to it, some of the Tinder conversations they had on their mobiles were captured for further analysis (as anonymized electronic conversations).2 We thus gathered about twenty-five post-match conversations, of varying length (with the shortest being a few messages long, and the longest about one hundred messages long) principally from six Tinder users, three male and three female. I will focus my analysis on this corpus of chat conversations, and more specifically on the aspects of their organization which only the use of naturalistic data (i.e. actual archives of conversation giving some access to the way in which these exchanges actually unfolded) can reveal.

 The reasons for which these conversations were archived were highly variable, so that a classification would be difficult). Some had just been left there (also, not all users were aware they could archive some chat conversation), some were kept for their interest. 2

160 

C. Licoppe

 nalysis: The Organization of Post-match A Tinder Online Conversations Participants mostly describe their post-match chats as ‘conversations’. However, such exchanges seem to stray far from talk-in-interaction. Communication is done by the means of written messages, and at such a pace that responses often take tens of minutes to be composed. In what way can it still be thought of as a ‘conversation’ then? First, there is a recognizable alternation of communicative actions (in a sense that we will describe later). Second, participants orient towards an interactional ideology of conversation, in which conversation is taken as a form of dialogic, potentially rapport-building, interaction. Its prototype is the interpersonal face-to-face or phone conversation between friends, and they seem to consider that what they do has a family resemblance with the latter in that respect (Licoppe, 2019). Finally, the design of chat interfaces presents the ongoing interaction as a spatialized, vertical transcript of successive contributions, which projects reading it as written dialogue. So when a participant reads an incoming message, it is presented to him/ her and enacted in his/her reading as the next turn in an ongoing, dialogic interaction. When the participant reads the latest message received he/she reads it as if the other participant had just produced it, although the message may in fact have been sent an hour or more earlier. Reading messages in such a spatially organized presentation makes the message ‘alive’ as the latest in a stream of dialogue. It projects responding to as if it had just happened, bringing the chat experience closer to what happens in talk-in-interaction.

Interweaving Adjacent Pairs in Multi-component ‘Turns’ Let us consider the beginning of a rather typical conversational opening (J is male and F is female). Each numbered line corresponds to a message. There is no timing for later messages, but they are written within the same evening.

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

161

The conversation is initiated with a greeting message. This has been made possible by a match on Tinder. When both participants ‘swipe’ one another right, the Tinder algorithm treats that as a match and notifies both participants. The greeting can be understood as a response to such a match (thus treated as a mutual appearance, opening the possibility of an encounter). The next message (message 2) is made of three interactional moves: a return of the greeting and a question followed by another question which immediately reformulates the preceding one. The greeting is oriented retrospectively and recognizably as a second pair part (it is a return greeting), while the questions can be read as first pair parts, when considered within the lens of adjacent pair organization. The next contribution (messages 3 and 4) keeps to the same kind of organization which combines backwardlooking and forward-looking interactional moves, designed as adjacent pair parts: a recognizable answer and another question designed as ‘reciprocal’ (‘and you?’). Here the two types of moves are done in two distinct, successive messages, but they are linked through the connective ‘&’. The CA-inspired analysis of textual interactions tends to equate “turns” with messages (Herring, 1999; Antaki et al., 2005). In the conversation above, we have a rather strict alternation of contributions,3 which can be made up of one or several messages containing distinctive messages or groups of messages involving multiple interactional moves (a single message can contain one or several of them), the majority of which are designed as clearly recognizable adjacent pair parts. For the purposes of our analysis, then, in order to retain the idea that participants respond after the ‘turn’ of another, we need to treat as ‘turns’ these alternating compound contributions, whether they are made up of one message or more. Apart from sequential considerations, the coherence of such turns usually resides in design features such as grammatical resources, adjacent pair organization and topical relevance. What makes turn transition points recognizable as such is a more complex issue. We will not be able to treat this point here. Timing seems important, but also the way an ongoing turn can be understood as having treated the previous turn (completing all its first pair parts) and initiating a new move, usually through a question.  There is very little interweaving of messages, such that such an alternation is preserved even with multi- message turns. This may be due to the pace of the messages, multi-message turns are usually done within a minute, while answers usually come minutes or tens of minutes later. 3

162 

C. Licoppe

Extract 8.1  Tinder Users Jacques and Federica 1. 12h08.

F. : Hello

2. 21h15.

J. : Hello. Comment vas tu? Sympa ton we?

3. 22h08.

F. : Yep. Il a fait beau c'était cool

Hi Hi.

How are you?

Was your weekend nice?

Yep. The weather was nice it was cool 4.

& toi ? & you ?

5. 22h23.

J. : Yes. Tu viens d ou Federica? Joli comme prénom

6.

F. : Merci, mes parents sont italiens.

Yes. Where do you come from Federica? Nice first name Thank you, my parents are Italian 7.

Et toi Jacques d'où viens-tu? And you Jacques, where do you come from?

8.

J. : Bah je suis francais de souche comme on dit. Mais je suis née à Rome. Alors je me sens un peu italien Well my origins are French as I said. But I was born in Rome. So I feel a bit Italian

9.

Tu es né la bas? You were born there?

10.

F. : Non, seulement mes parents. No, only my parents.

11.

Comment ça se fait que tu es né à Rome? Tu y as vécu un peu? How come you were born in Rome? Did you live there for a while?

The next turn in Extract 8.1 (message 5) is made of a single message, with an initial response to the question in the earlier turn, followed by a question initiating a new topic (Federica’s first name). It is a bit more complex, pragmatically, in the sense that the question about the origin of the name is followed by a positive assessment of it. In her next, two-­ message turn, Federica addresses all those interactional moves in a way that is consistent with adjacent pair organization. She first thanks Jacques, showing that she has understood the assessment as a compliment, gives an account of her first name related to her family origins and in the next

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

163

message, returns the question, asking about his own origins. A rapid inspection of later messages shows that this pattern of combining backward-­looking and forward-looking interactional moves designed as adjacent pair parts seems to hold for the rest of the conversation. Turns usually involve quite recognizable adjacent pair parts, that is at least a second pair part, replying to the previous turn, and a first pair part, mostly a question. Beyond this example, this organization of the chat conversation in such an alternate sequence of compound contributions is a recurrent feature in the data gathered among our sample of educated Tinder users. What might such an organization actually do? It ensures that almost all contributions will contain at least a response and a question, thus using adjacent pair part organization to tie one contribution to the previous one, while projecting strongly a next turn, as the crucial property of adjacent pair organization is to make conditionally relevant some types of response. Users then seem to harness the systemic properties of adjacent pair organization as a powerful resource to sustain the flow of conversation. While the first move is oriented backwards and responds to the previous turn (it works as a second pair part), the second move looks forward. As a first pair part, it makes conditionally relevant a response. This sequential work makes the absence of a response more salient and accountable at the systemic level. It also hints at a potential and implicit concern, that is that no further turn may be coming, which would be like a form of desertion of the interactional scene in which a participant ‘just’ ceases to respond, and which would not only transgress sequential expectations, but also have negative implications at the ritual level. While non-­ response is a common feature of chat-based interaction, and is usually noticeable (Rintel, Pittam, & Mulholland, 2003) and occasionally consequential (Stommel & Te Molder, 2015), the salience of such a concern is a distinctive feature of Tinder chat conversations.

An Orientation Towards Providing Elaborate Answers Elaborate answers are also a common feature of such conversations. Even when single word answers would fit the type of question (as with Y/N questions), they are rare. There is one occurrence in Extract 8.1, but the

164 

C. Licoppe

token of agreement is immediately followed by a question initiating a new topic (message 5). Elsewhere we get a lot of elaboration, as with message 3, in which the question about her weekend gets a ‘yes’ answer, and a further elaboration about her enjoyment. Similarly, J elaborates while answering the question about his birthplace, hinting that he lived in Rome for a while in message 8. This is oriented to as a kind of topical

Extract 8.2  Pauline and Valentin 1. PAU 2. VAL 3. 4. 5. PAU

Valentin, enchantée. Qu’est-ce que tu explores en ce moment ?

Valentin, nice to meet you. What are your exploring now?

Salut Pauline ! En ce moment c’est des stylos connectés lol Hi Paula ! Right now, it’s connected stylos lol Enchanté :)

Nice to meet you :) Et toi ?

And you ? Intéressant ! Dans un labo ou pour une boîte ? Moi je suis payée pour lire, un autre type d’exploration..

Interesting ! In a lab or for a firm? Me I’m paid to read, another type of exploration… 6. VAL

C’est cool ça aussi :) comment as-tu géré pour être payer de

7.

Journaliste ? Critique…. ? :)

8. VAL

Dans une boite semi gouvernementale, dont les actionnaires sont des

la

sorte ?

That’s cool as well :) how did you manage to get paid that way ? Journalist? Critic…. ? :)

universités

In a semi-governmental firm, the stakeholders being universities 9. PAU

Je suis journaliste, pour une émission de débat. Je lis, et j’en tire des idées de sujets et d’invités :)

I am a journalist for a talk show. I read, and from that I get ideas for topics and guests 10. PAU

Et tu conçois les stylos du coup ?

11. VAL

Ah c’est frais ça :)

12.

Non pas du tout, c’est un designer qui est venu nous voir

And are you designing the pens then ? Ah it’s cool that

No not at all, it’s a designer who came to see us

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

165

‘offering’, which Federica picks up on in her next turn, in which she initiates a line of questioning about his stay in Rome. In conversational pragmatics, Grice’s maxim of quantity states that as a conversationalist, one should try to be as informative as one can, and give as much information as is needed, but no more (Grice, 1989). Elaborations are therefore noticeable, and should be considered for what they do. Extract 8.2 provides us with other instances of such an orientation towards elaborate answers. In this new conversation, we observe again the same pattern of ‘turns’, made with a short burst of two messages or more, and combining retrospective and prospective interactional moves. It is almost systematic here. Message 1 is not just a single greeting as was the first message in Extract 8.1, it already combines a greeting (retrospectively oriented to the match and prospectively to a greeting sequence), and a question about self which obliquely refers to the profile (the main resource for initiating a first topic4). This question will occasion a paired sequence of reciprocal question and answers, with the reciprocal question being asked in message 4. Let us now consider more closely the Q/A sequence involved in messages 7 and 10. In message 7 Valentin asks Pauline about what she does, adding candidate answers formulated as professional categories, which strongly projects a category as an answer. An answer of this type is produced at first position in her next turn (“I am a journalist”, message 7). It provides exactly the kind of answer which was projected by the previous question. However, it is followed by further elaborations. According to Sacks’ economy rule in membership categorization analysis (Sacks, 1992), which relates to some extent to Grice’s maxim of quantity, a single category would be referentially adequate (Sacks, 1992), which makes elaborations potentially noticeable and available for scrutiny with respect to what they do. In her elaboration Pauline adds that she works for a talk show (so on TV), and then she provides a description of what she actually does (“I read, and from that I get ideas for topics and guests”). Altogether, this is a rich description, as well as an elaboration in which she divulges  Studies of face-to-face interactions have shown that in encounters between strangers, first topic initiation was to be found in the particularities of the mutually available context (Maynard & Zimmerman, 1984). In online dating, online profiles constitute the main source of mutually available context. 4

166 

C. Licoppe

information about herself that lies in her personal territory. It therefore provides many elements for a recipient to comment and on which to ‘rebound’ with new questions, and it is the way in which such elaborations are powerful resources in these Tinder conversations. They provide opportunities for further talk by allowing participants to move away from reciprocal, short info-oriented question-answer sequences and into a kind of topical talk which may lead to relational build-up. Elaborate answers may be used as a kind of sequential and topical fuel, and such utterances display the commitment of their author to sustain conversation, while it simultaneously hints at the fact that the latter might stop, that interactional desertion is always a possibility. Such a disappearance actually occurs here. Though her description of her work was quite elaborate, he responds with a short third position assessment (‘it’s cool’) which orients to her elaborated description as a whole, without picking out anything in particular. He follows that up with a question, keeping in line with the retrospective-prospective turn organization they have kept to so far. However, without any obvious hint that it might happen (and particularly no pre-closing or closing sequence as would be expected in face-to-face or phone conversations), PAU vanishes from the Extract 8.3  Pauline and Jonathan 1. PAU

C’était une très bonne année oui ! Ah oui “IEP 5 de Bordeaux” j’avais tout sous les yeux. Tu fais quoi maintenant ?

It was a very good year, yes ! Ah yes “IEP de Bordeaux” I had everything under my eyes. What are you doing now? 2. JON

je bosse dans la communication (j’avais pris ça comme master spé) I’m working in communication (I had done this as a master pro)

3. PAU

Mais encore ? And what else?

conversation, forever. Such unheralded disappearances from the interactional scene do happen indeed, and it seems to be the main way in which such conversations stop, at least when no move proposing a face-to-face encounter or to shift the conversation on to another chat platform has been made.

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

167

Extract 8.3, taken from another conversation, near its beginning, shows how such elaborations can actually be expected. At the end of her message, Pauline asks a question about Jonathan’s current job. He provides an answer, stating a professional domain, but in the next message, though relevant, Pauline treats that answer as inadequate. She explicitly displays her expectation that further talk about what he does would have been relevant, that is, an orientation towards elaborate answers. Pauline’s alertness to what she treats as a lack of elaboration may also be reinforced by the lack of a question addressed to her in the previous message. Tinder involves encounters with strangers. Participants usually have no pre-existing relationship, and no shared history beyond the Tinder match situation. The main topical resources at the beginning of chat conversations are either the online profile, with possibly some texts and pictures, and which may provide some clues as to some personal experiences and tastes, and a few (a very limited number of ) shared circumstances regarding the encounter (the date, day or hour of the exchange for instance, or the fact that they are both on Tinder). For a conversation and/or a relationship to develop, these scant topical resources have to be made relevant as a topic and turned into common ground. That some degree of intimacy may emerge from what starts as a textual interaction between strangers is the root of what we could call ‘topical pressure’ which bears on participants. Tinder conversationalists need to find and try a stream of relevant topics to talk about that may contribute to the creation of relationship-relevant common ground, at a sustained enough tempo to avoid the risk of the chat stopping. Otherwise, the conversation runs the risk of exhausting itself and stopping ‘abruptly’ (one of the participants becoming silent without any explicit conversational closing sequence): “Often there is no more conversation topic. So we drop it or we stay in touch without really staying in touch” (Male, 18yo). Topical pressure thus also contributes to the risk of a co-participant abruptly becoming silent and deserting the encounter to follow other matches and interests. The orientation towards elaboration can be taken as a resource to try to relieve that topical pressure by making interactional moves inferentially rich, and to show one’s commitment to moving the conversation forwards. In its apparent breaching of the maxim of quantity, which, according to Grice,

168 

C. Licoppe

governs the organization of ordinary conversations, it shows that such ‘conversations’ may not be completely ordinary. It is as if Tinder chats were haunted by concerns regarding the risk of interactional desertion, with their peculiar sequential organization in turns and their orientation towards elaboration, trying to postpone or counter that spectre, from inside the unfolding conversational process itself.

The Open Question/Closed Question Device In the light of this discussion, if we reconsider message 6 in Extract 8.2, Valentin’s question about Pauline’s job has an interesting design. The prospective part of the message involves two successive questions. The first is an open question (‘how did you manage to get paid that way?’), which projects an elaborate answer. However, it is followed by two questions that are framed as candidate answers. So the initial open question is sort of reformulated as a Y/N question, projecting a brief answer. Such cases of “one question after another” display a sensitivity of

Taken from Extract 8.1 11. F.

Comment ça se fait que tu es né à Rome? Tu y as vécu un

peu?

How come you were born in Rome? You lived did you live there for a while?

Taken from Extract 8.2 6. VAL

C’est cool ça aussi :) comment as-tu géré pour être payer

de la

sorte ?

That’s cool as well :) how did you manage to get paid that way? 7.

Journaliste ? Critique…. ? :) Journalist ? Critic….? :)

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

169

the speakers with respect to the action of questioning, and some potential implications of the first question (Raymond & Heritage, 2013). A similar design can be observed in the last message of Extract 8.1. The interactional move in initial position in the message is an open question first projecting explanations about his birthplace. It is reformulated in the same message as a Y/N question about his having sojourned for some time in Rome. As in our prior example, an open question is reformulated as a Y/N question, and we have repeated instances of this device throughout the corpus. In legal settings, differences in such types of question have been interpreted in terms of coerciveness (Danet & Bogoch, 1980), with open questions putting less constraint on the answerer than Y/N questions. In our setting, in view of the orientation towards elaboration we have described, open questions, which precisely project elaborate answers in a rather unconstrained way, should be fine. Therefore asking one (closed) question after another (open) one suggests another concern being in play in Tinder conversations. If we consider the situation from the viewpoint of the recipient, providing elaborate answers to open questions takes time and effort to compose in writing. Therefore turning the open question into a yes/no question, possibly with a candidate answer, may be a way to show that the questioner is aware of that, aware of it enough to transform the question so as to project an answer that might be less burdensome to compose. Such a questioning device may then be understood as a form of recipient design which displays an orientation of the questioner towards facilitating a response and therefore towards maintaining the conversation alive. It is likely to be particularly relevant in an interactional context in which the abrupt ending of a conversation and the disappearance of a co-participant is very much a possibility.

Accountability and Desertion Let us consider an instance in which the male participant actually deserts the interaction, or, to use a contemporary, but polysemic, expression, ‘ghosts’ his female co-participant. In Extract 8.4 below, Vincent, the male participant, provides no answer to Maureen’s last turn (messages 13 to 15),

170 

C. Licoppe

Extract 8.4  Maureen and Vincent 1.

17:28

MAU

Hey ! Est ce qu’on apprend des choses intéressantes à l’université de technologie de Troyes ?

Hey ! Does one learn interesting things at the university of Troyes? 2.

17:52

3.

17:53

VIN

Lol oui il me semble mais j’ai tout oublié ! Lol yes I think so but I have forgotten everything ! Tu as fait une ut ? Did you attend an ut6 ?

Next day 4.

20:02

MAU

Non pas du tout et je n’ai qu’une vague idée de ce qu’on y fait :)

No not at all and I have only a vague idea of what ones does there :) 5.

20:03

Tu habites vers où ? Je vois qu’on est à une énorme dis de 19km ! Whereabouts do you live? I see we are separated by a huge distance of 19 km

6.

20:23

VIN

Sans voiture ça peut faire loin oui Without a car that can be far yes

7.

Dans le 14e

20:23

In the 14th7 8.

21:07

MAU

9.

21:08

MAU

Effectivement tu t’esrapproché..

Indeed you have gotten closer..

Je connais mal… juste vaguement vers Denfert I don’t know that district well… just vaguely around Denfert

10. 21:08

MAU

11. 22:19

VIN

Tu aimes bien ton coin ?

Do you like your neighbourhood ? Ça ça tu habites où toi ? It’s ok and where do you live ?

12. 22:19

Ça va It’s ok

13. 22:43

MAU

Ds le 11e In the 11th

14. 22:43

T’es un grand bavard j’ai l’impression You’re quite the talker I see

15. 22:44

T’as été au Japon ? Have you been to Japan ?

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

171

though these messages included two forward-looking interactional moves, first an assessment, then a question, which are both first pair parts that would make a response conditionally relevant. Vincent simply vanishes from the interactional scene. What is interesting here is that a potential problem seems to be noticed by Maureen, Vincent’s female co-participant before he disappears. This rests on her assessment of him in message 14: “you’re quite the talker I see”. In such a categorical and ironical formulation of the kind of conversationalist he is, she implicitly also expresses her assessment of the conversation so far. Through membership categorization analysis, it can be understood as the kind of conversation that is relevant to a participant who can be categorized as ‘not a talker’, within this particular speech-­exchange context. She makes him explicitly accountable for the way he has been involved in their conversation. We may also surmise that either his conversational behaviour and/or her making him accountable for it might be related in some way to his disappearance, which occurs just after this. Can we analytically find some grounds for her ironic formulation of his conversational behaviour so far, and for her characterization to be understood as ironic as well? The pace of their conversation is slow, but that is true on both sides. As a hypothesis, I will treat what I have observed in the conversations I have analysed so far as the expected, proper way to display: (a) concern with keeping the conversation going (at the systemic level); and (b) a commitment with respect to the ongoing encounter (at the ritual level), in the sense that interactional moves display the proper kind of conversational deference to the kind of person the other person is taken to be. When we consider Maureen’s turns, she appears to play that kind of interactional game. All her turns are made of responses and questions (message 1, then messages 4 and 5, 8–10 and 13–15). She often volunteers additional elaborations of her interactional moves, in a way that offers opportunities for further topical development (see her avowal of ignorance regarding things with respect to which he is supposedly knowledgeable in messages 4 and 9, or her assessment of their system-based mutual distance, in message 5). Finally, she initiates or shifts previous topical lines in many of her questions. The same cannot be said of him. He strays a lot from the response/question model, often producing

172 

C. Licoppe

responses only (for instance in the turn made of messages 6 and 7). His few questions are reciprocal “return questions” (message 3, message 11), which makes it her lot to initiate new topics. Some of his responses are noticeably unelaborate and inferentially poor (“it’s ok”, message 11, as a response to the question “do you like your neighbourhood?”). These systemic cues also automatically trigger ritual concerns, such as a potential lack of commitment to the encounter, which his eventual ghosting then seems to retrospectively confirm. In sum, Vincent does not play the same interactional game, so that from the normative standpoint we have suggested, which it is plausible to surmise is Maureen’s as well, he does not display the same level of interest in sustaining the conversation (sequentially and topically). His conduct thus opens him up to a categorization such as ‘not the talker’. More generally, participants are held accountable for not being ‘into the conversation’ (and therefore not involved enough in the emergence of a relationship) in two main respects. First, this is the case when they appear noticeably slow to respond (a perception which opens them to inferences about the fact that they might be engaged in multiple conversations5). Second, it is also the case when they provide what may look like ‘simple’ or non-elaborate answers (such behaviour being then accountable as that of someone who is not very ‘talkative’ in this context). Because of the relational stakes (getting an intimate enough relationship to emerge) and the risks (abrupt desertion and the implicit sense of rejection which might go with it) which are involved here, participants are particularly alert to such cues (as in Extract 8.4 above). Such a heightened sensitiveness is a characteristic feature of such conversations and part of the lived experience of Tinder chatting.

Conclusion In this chapter, I used a CA-inspired approach to analyse the Tinder chat conversations of a sample of educated users in Paris. Though such chats seem to stray far away from ordinary conversation and talk-in-interaction  This is occasionally topicalized in these conversations and confirmed in interviews.

5

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

173

in terms of their pace and written character, they are still experienced as a sort of conversation, perhaps because they occur on interfaces designed with that purpose (which present such messages such that they may be read as a dialogue). All of this makes a CA-inspired approach of such a form of textual interaction relevant. I have shown a distinctive sequential organization in which participants: (a) design turns as a combination of backward-looking interactional moves (recognizably responsive to parts of previous messages) and forward-looking ones (often questions, and almost always moves designed as first pair parts of an adjacent sequence); (b) orient towards providing significant elaborations of their moves, made noticeable by the way they seem to run against the ‘maxim of quantity’ and (c) use the ‘one question after another’ device, in which an open question is immediately reformulated (in the same message or in the next one) as a Y/N question. I showed that participants were accountable for not using such message design resources systematically. When messages were not designed in this way, something was seen as ‘missing’, commitment-wise, and the relevant participant ironically categorizable as ‘not a talker’, that is not adequately involved in the chat conversation. Conversely, it suggests that the presence of these design features is treated as displaying a proper involvement in the Tinder conversation. More specifically, since a recognizable function of such resources is to reinforce the way in which successive turns project an answer, their use displays a concern that the conversation might stop (as we saw by one party vanishing from the electronic interactional scene), and correlatedly, the commitment of their users in keeping it going. What emerges in this way as a distinctive feature of such conversations, which make them distinct from other types of online interactions discussed so far in the literature, is that they seem especially haunted from the inside by the possibility of interactional desertion. In the way the design of contributions seems to harness sequential resources to project more strongly a response, they display an underlying concern with the possibility of ‘ghosting’ and the heightened significance of non-response and desertion in this particular online courtship context. The resources deployed by our sample of educated users operate at the systemic level, to make ghosting less easy, but also more salient and

174 

C. Licoppe

noticeable when it happens. Ghosting simultaneously operates at the ritual level. It negates the underlying project of getting some personal relationship to emerge from such a conversation between strangers. As a form of conversational breaching, it opens the ongoing conversation to retrospective, more encompassing inferences, such as the ghosted party feeling inadequate or even rejected. In showing this alertness of the participants to such a possibility, as displayed by the sequential resources they use to make ghosting slightly less easy, our CA-inspired approach gives us a sense of the distinctive character of such conversations, that is the way they are done and experienced as ‘lived work’. They appear as a kind of conversational undertaking in which the interaction is reflexively haunted from the inside by the spectre of interactional desertion or ‘ghosting’. Analysing the way participants craft their interactional moves in a given online setting, and the implications of such sequential design with respect to the sustaining of the conversation, opens up the way to an ‘Ethnography of Online Interaction’. It is important to note that the kind of organization we have identified here requires some high level of literacy and writing skills, and that our sample is limited in that respect. While it is recurrent in the conversations provided by our sample, there is no reason to believe it would be deployed or used in the same way in a less educated sample of Tinder users, though we may still expect the concern with ghosting to be operative there as well. While there is no room here to elaborate these points, an ongoing study with a less educated sample seems to support this hypothesis.

References Antaki, C., Ardévol, E., Nunez, F., & Vayreda, A. (2005). ‘For she who knows who she is:’ Managing accountability in online forum messages. Journal of Computer-Mediated Communication, 11(1), 114–132. Arminen, I., Licoppe, C., & Spagnolli, A. (2016). Respecifying Mediated Interaction. Research on Language and Social Interaction, 49(4), 290–309. Baron, N. (1998). Letter by phone or speech by other means: The linguistics of email. Language and Communication, 18, 133–170.

8  The Spectre of ‘Ghosting’ and the Sequential Organization… 

175

Danet, B., & Bogoch, B. (1980). Fixed fight or free-for-all? An empirical study of combativeness in the adversary system of justice. British Journal of Law and Society, 7(1), 36–60. David, G., & Cambre, C. (2016). Screened intimacies: Tinder and the swipe logic. Social Media & Society, 2(2), 1–11. https://doi. org/10.1177/2056305116641976 Giles, D., Stommel, W., Paulus, T., Lester, J., & Reed, D. (2015). Microanalysis of online data: The methodological development of ‘Digital CA’. Discourse, context and Media, 7, 45–51. Goffman, E. (1971). Relations in public. Microstructure of the public order. New York: Harper & Row. Goffman, E. (1981). Forms of talk. Philadelphia: University of Pennsylvania Press. Goodwin, C., & Goodwin, M. (2004). Participation. In A.  Duranti (Ed.), Handbook of linguistic anthropology (pp. 222–244). London: Blackwell. Greiffenhagen, C., & Watson, R. (2009). Visual repairables: Analysing the work of repair in human-computer interaction. Visual Communication, 8(1), 65–90. Grice, P. (1989). Studies in the ways of words. Harvard: Harvard University Press. Herring, S. (1999). Interactional coherence in CMC. Journal of Computer-­ Mediated Communication, 4(4). https://onlinelibrary.wiley.com/doi/ full/10.1111/j.1083-6101.1999.tb00106.x Licoppe, C. (2019). Liquidity and attachment in the mobile hookup culture. A comparative study of contrasted interactional patterns in the main uses of Grindr and Tinder. Journal of Cultural Economy, 10.1080/17530 350.2019.1607530. ⟨hal-02166542⟩. Maynard, D., & Zimmerman, D. (1984). Topical Talk, Ritual and the Social Organization of Relationships. Social Psychology Quarterly, 47(4), 301–316. Meredith, J., & Potter, J. (2013). Conversation analysis and electronic interactions: Methodological, analytic and technical considerations. In H. Lim & F. Sudweeks (Eds.), Innovative methods and technologies for electronic discourse analysis (pp. 370–393). Hershey (PA): IGI Global. Meredith, J., & Stokoe, E. (2014). Repair: Comparing Facebook ‘chat’ with spoken interaction. Discourse & Communication, 8(2), 181–207. Paulus, T., Warren, A., & Lester, J. (2016). Applying conversation analysis methods to online talk. A literature review. Discourse, Context & Media, 12, 1–10. Ranzini, G., & Lutz, C. (2017). Love at first swipe? Explaining Tinder self-­ presentation and motives. Mobile Media and Communication, 5(1), 80–101.

176 

C. Licoppe

Raymond, G., & Heritage, J. (2013). One question after another: Same-turn repair in the formation of yes/no type initiating actions. In G. Hayashi & J.  Sidnell (Eds.), Conversational repair and human understanding (pp. 135–171). Cambridge: Cambridge University Press. Reed, D. (2001). ‘Making Conversation:’ Sequential integrity and the local management of interaction on Internet newsgroups. Proceedings of the 34th Hawaii International Conference on System Sciences—2001. Retrieved February 18, 2004, from http://csdl.computer.org/comp/proceedings/ hicss/2001/0981/04/09814035abs.htm. Reeves, S., & Brown, B. (2016). Embeddedness and Sequentiality in Social Media. In Proceedings of the 19th ACM Conference on Computer-Supported Cooperative Work & Social Computing (CSCW ’16) (pp.  1052–1064). New York: ACM Press. Rintel, S., Mulholland, J., & Pittam, J. (2001). First Things First: Internet Relay Chat Openings. Journal of Computer-Mediated Communication, 6(3). https:// onlinelibrary.wiley.com/doi/10.1111/j.1083-6101.2001.tb00125.x Rintel, S., Pittam, J., & Mulholland, J. (2003). Time will tell. Ambiguous non responses on internet relay chat. The Electronic Journal of Communication, 13(1). http://www.cios.org/EJCPUBLIC/013/1/01312.HTML Sacks, H. (1992). Lectures on Conversation. Cambridge: Cambridge University Press. Sacks, H., Schegloff, E., & Jefferson, G. (1974). A simplest systematics for the organization of turn-taking for conversation. Language, 50(4), 696–735. Stommel, W., & Koole, T. (2010). The online support group as a community: A micro-analysis of the interaction with a new member. Discourse Studies, 12(3), 357–378. Stommel, W., & Te Molder, H. (2015). Counselling online and over the phone. When preclosing questions fail as a closing device. Research on Language and Social Interaction, 48(3), 281–300. Vayreda, A., & Antaki, C. (2009). Social support and unsolicited advice in a bipolar disorder online forum. Qualitative Health Research, 19(7), 931–942.

9 Participation of Companions in Video-­ Mediated Medical Consultations: A Microanalysis Wyke J. P. Stommel and Martijn W. J. Stommel

Introduction Microanalytic studies of online data have primarily dealt with written interaction in contexts such as forums (e.g., Giles & Newbold, 2013), Facebook (e.g., Meredith & Stokoe, 2014), chats (e.g., Stommel & Molder, 2015), YouTube comments (e.g., Bou-Franch & Garcés-Conejos “It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

W. J. P. Stommel (*) Center for Language Studies, Radboud University, Nijmegen, The Netherlands e-mail: [email protected] M. W. J. Stommel Department of Surgery, Radboud University Medical Center, Nijmegen, The Netherlands © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_9

177

178 

W. J. P. Stommel and M. W. J. Stommel

Blitvich, 2014), etc. (see also Paulus, Warren, & Lester, 2016 for a review). Such is also the case for the other empirical chapters in this book. In this chapter, we focus on what may seem a fundamentally different type of mediated interaction in which talk-in-interaction is the dominant communication mode: video-mediated interaction. Naturally occurring video-mediated interaction can be seen as analogous to written online interaction with regard to participation. In both environments, participation is negotiated and organized by the participants as the interaction unfolds. Obviously, the resources to do this vary: for example, participation to online forum discussions is constructed in the exchange of posts with potentially specific forms of address (Bou-Franch, Lorenzo-­Dus, & Garcés-Conejos Blitvich, 2012; Stommel & Koole, 2010), while in video-mediated interactions participation is strongly linked to visibility (Licoppe, 2017; Licoppe & Morel, 2012). Our microanalytic examination of participation of companions in video-consultations underscores the relevance of perceptibility, either visually or audibly, for participation. The data of our study consist of medical video-consultations with patients who had undergone surgery around two weeks prior to the consultation, with the purpose of discussing the histopathological test results of the removed tissue and discussing their recovery. Traditionally, when these patients visit the hospital for the consultation, they are accompanied by a family member, frequently a partner or child. Similarly, when the consultation is conducted through video—surgeon at the hospital, the patient at home—a companion may be present on-site with the patient and be involved in the interaction. This chapter examines this participation by companions microanalytically, both in terms of verbal and ostensible (visual) practices and with special attention to participation shifts.

Companions in Medical Consultations It is generally recognized that companions may play an important part in medical consultations (Laidsaar-Powell et al., 2013). They can offer support in making a (medical) decision, offer new perspectives, help recall relevant information, and elaborate or clarify issues (Fioramonte & Vásquez, 2019). However, a companion alongside the patient may also

9  Participation of Companions in Video-Mediated Medical… 

179

substantially alter the organization of the interaction. With a companion, the communication consists of a triadic, rather than a dyadic setting. This may increase the complexity of the communication compared to dyadic doctor-patient interaction (Beisecker, 1989; Huber et  al., 2016). Companions may be rather passive, but may also speak on behalf of the patient or be so dominant that the physician accepts them as the primary conversational partner (Beisecker, 1989; Hasselkus, 1992). To the best of our knowledge, it has not been examined yet how companions participate in medical consultations when the consultation takes place through a video connection.

Medical Video-Consultations Tele-consultations or video-consultations have become increasingly introduced in health settings, and they are generally considered beneficial (Zilliacus et  al., 2011). When the consultation takes place through a video connection in order to involve a distant expert, a local health professional (GP, nurse) may be present on-site with the patient. Some aspects of the role of such on-site professionals have been addressed in previous studies (Miller, 2003; Pappas & Seale, 2009, 2010). The professional was found to act as a facilitator, moderating the interaction between the specialist on the other end and the patient locally. It is unknown, but rather unlikely, whether family members would similarly take up a facilitating role. In our data, the companions do not act on behalf of the patient (because the patient cannot speak or is incoherent), but rather act alongside the patient (cf. Fioramonte & Vásquez, 2019). Given the specific characteristics of the video technology and cameras, acting alongside the patient may look quite different from the situation in the consultation room. Questions that arise are whether companions are on-camera when they speak and whether companions who are off-camera do participate. In this chapter, we examine the participation of companions in video-­ consultations specifically in relation to their visibility and/or audibility. We explore both cases of minimal and more elaborate participation and analyze explicitly how the affordances (Hutchby, 2001a, 2001b) of the video channel are exploited for and/or restrict participation on a moment-­ by-­moment basis.

180 

W. J. P. Stommel and M. W. J. Stommel

Video-Interaction and Participation Conducting a medical consultation through a in video-consultation, with the patient being in his/her private domain and the physician at the hospital, has several implications. Here, we focus on the issue of visibility and participation. Studies of video-interaction have revealed a major orientation of speakers to the visibility of their interlocutor(s) (Mondada, 2010). In dyadic video-interaction, the “talking heads” configuration is oriented to as the default organization for the interaction (Licoppe & Morel, 2012). Overall, mutual visibility is regarded a precondition for interaction. However, visibility gets more complex when participants are accompanied by others. Once mediated interaction involves more than one party in one space the interaction is affected by interaction between participants among each other in the local space. That is, participants appear to orient to multiple interactional spaces (cf. Wasson, 2006). However, rather than as separate entities with stable structures, these spaces are interactionally coconstructed by the participants themselves (Oittinen, 2018). In videoconsultations, the position of the companion in the scene in relation to the camera seems critical for his/her interactional playing field.

Participation Framework A useful point of departure for the analysis of participation is the notion of participation framework (Goffman, 1981). Goffman (1981) distinguished between ratified and unratified participants, including—in the category of hearers—bystanders or overhearers who are not ratified but perceivable to the official participants and even eavesdroppers who secretly overhear the interaction. In the category of ratified participants, the concept of hearer is broken down to unaddressed and addressed recipients. Furthermore, it is recognized that ratified participants and unratified participants can engage in subordinate communication, including: • “byplay” by ratified participants, for example, two participants chatting privately during a conversation with multiple parties at a dinner table

9  Participation of Companions in Video-Mediated Medical… 

181

• “crossplay” by ratified and unratified participants, for example, a director addressing an actor during a play • “sideplay” by unratified participants, for example, talk among bystanders of an interaction A typical feature of subordinate communication, which does not challenge the dominant communication, is managed using resources like gestures and voice (whispering). Goodwin (2007) criticized the relatively static conceptualization of speaker and hearer central to Goffman’s (1981) model. Instead, he showed that participation can be analyzed as a temporally unfolding process through which participants demonstrate to each other their ongoing understanding of the interaction by building actions that contribute to the further progression of these very same events. This includes many other signs than speech, such as smiling, laughter, and gaze. Thus, participation is not just laminated but organized moment-by-moment, on the part of both the hearer and the speaker. A conceptualization of ratification as an action-in-interaction rather than as a static participant status allows for a more dynamic understanding of participation by third parties, both as hearers and as speakers. In the medical setting, the prototypical situation is that the patient and physician interact while the companion is positioned both physically/visibly and metaphorically at the patient’s side. The companion may speak or remain silent. Whether companions are ratified is not pre-­determined, but a local achievement of the interactants. Additionally, their participation status may evolve and shift during the consultation; for instance, companions who have been silent as, and treated as, bystanders may at some point be addressed by the physician. In this chapter, we examine the relation between visibility/audibility and ratification, that is, whether and how visibility/audibility is oriented to in talk, who addresses (and thus ratifies) whom and how orientations may shift or conflict in the interaction.

Data and Method For this chapter we draw on a set of 22 post-operative video-­consultations with patients who had undergone abdominal or esophagus tumor resection approximately two weeks before the consultation. Their condition

182 

W. J. P. Stommel and M. W. J. Stommel

and the stage of illness varied, but most operations were discussed as (relatively) successful. At the stage of the video-consultation—two weeks after surgery—all patients tended to have minor functional problems or pain, as full recovery usually takes six weeks or more. Thirteen patients were female, nine male, mostly over fifty-five years of age. The three surgeons (all male) conducted consultations with the patients they operated. The calls started with the software signaling the patient present in the virtual “waiting room”; then, the physician pressed a button to initiate the call. The overall organizational structure of the consultations was first delivery and explanation of the test results, which was announced as the reason for the consultation and involved mainly talk by the physician and then a patient self-report on the recovery, frequently elicited step by step by physician questions like “how is eating and drinking going?” and “is the wound healing well?” (but sometimes in the reverse order, see Stommel, Van Goor, & Stommel, 2019). The video-consultations lasted on average 12 min 20 sec and the language used was Dutch. The patients were instructed to use the video software through a leaflet. The text recommended patients to make sure their face was visible onscreen (for instance by overturning the table or phone), but no instruction or advice was given on how partners or companions could take part in the interaction. The recordings were made in the consultation room of the physician, with one camera directed at the physician and one at the screen; hence, we had no camera at the scene of the patient and no (other) information about the presence/absence of companions than what we heard or saw on the physician’s screen. As the recordings were made by several medical students, there was substantial variation in tripod position and camera angle, which explains the different kinds of images we show in the findings section. In 13 out of 22 conversations, a companion appeared to be present at the patient’s scene, which was always a domestic setting. We analyzed the data using microanalytic, conversation analytic techniques, mainly focusing on the organization of sequences (Sidnell & Stivers, 2013). First, the consultations were transcribed in detail (Jefferson, 2004). Then, sequences in which the companion was addressed or spoke were identified, upon which we examined the video-recordings in detail with regard to the appearance of the companion in relation to the talk. Then, we examined how the companion’s appearance elicited an action at talk, where in the sequence companions spoke and doing what with their

9  Participation of Companions in Video-Mediated Medical… 

183

turn and when, by whom (patient or surgeon), and how the companion was addressed. We used the participation framework including various participant statuses as proposed in Goffman (1981) amending these by focusing on ratification by (indirectly) addressing a party to characterize some of the interactional dynamics of companion participation in video-consultations.

Findings A first remarkable finding was that companions rarely sat next to the patient on-screen in talking head configuration (3 out of 13 cases). Instead, they were frequently minimally visible, that is, an arm, a shoulder, a part of head, and talked minimally (11 out of 13 cases). This is notable given the “default” situation in face-to-face consultations. In the consultation room, physicians face two or more conversational partners (see Fig.  9.1, taken from an additional data set of out-patient, post-­ operative hospital consultations), while in the video-setting they

Fig. 9.1  Physician facing patient and companion (#2)

184 

W. J. P. Stommel and M. W. J. Stommel

frequently face just the patient (see Figs. 9.4a, 9.4b, 9.5, 9.6, 9.7, and 9.8). Relatedly, in the video-setting the patient’s talking head frequently “fills” the screen, hardly leaving room for the companion (see Fig. 9.4a and 9.4b), which is also an orientation to the patient as the primary speaker. As a consequence, it may not always be self-evident from the onset of the consultation whether a companion is present at all, and how the companion may take part in the interaction. We found that companions were involved in the interaction in various ways, either as fully ratified conversational partners (2 out of 13 consultations) or as bystanders of the patient, meaning the patient and physician acted as the primary participants (11 out of 13). We will not further examine the two cases with the companion’s full participation while appearing on-screen as this triadic participation framework appeared to be relatively stable throughout these two consultations. We will also not further explore the consultations in which a companion appeared on-­ screen only very briefly, but never talked and was never addressed by either patient or physician (4 out of 13). More interesting are the cases (7 out of 13) in which companions acted as bystanders most of the time, but were sometimes briefly involved in the interaction. Generally speaking, companions’ verbal contributions predominantly served as prompts and thus displayed knowledge of the patient’s health situation and recovery specifically. These prompts served to speak alongside (rather than on behalf of ) the patient (cf. Fioramonte & Vásquez, 2019). Companions were rarely involved in the openings of the consultations. However, the bystander role was susceptible to the moment-by-moment unfolding of the “video-in-interaction” (Licoppe & Morel, 2012) and thus interchangeably and temporarily transformed into a more active role. Companions may self-select to say something, their appearance on-screen or crosstalk by the patient may elicit ratification by the physician as a co-­ participant in the consultation. We first discuss an example of a companion’s self-selection as a bystander, prompting the patient. Then, we show how the physician can ratify a hitherto off-screen companion, either overtly with a noticing or more covertly by responding to the companion. Next, we analyze a case of re-ratification of the companion by the physician toward the consultation closing, and at last we discuss two instances of moderate

9  Participation of Companions in Video-Mediated Medical… 

185

interactional trouble surrounding a companion’s turn, namely one case of conflicting orientations and one case of negotiated participation of a second companion in the room. Central to our analyses of these sequences is the interrelation between on-screen appearance and verbal participation.

Companion as Bystander Overwhelmingly, the physician-patient dyad is oriented to as the primary participation framework in video-consultations. We found that if companions do self-initiate a turn, this contribution consists of crossplay with the patient, which serves to prompt the patient to report or ask something. Consider Extract 9.1, taken from a consultation with an 81 years old patient. The companion, who is the patient’s adult son, is clearly visible on-screen in talking head configuration, sitting next to the patient. At the beginning of the consultation, the physician reported the visibility of both patient and his son (“I can see you both well,” data not shown, but see also Figs. 9.2 and 9.3), but until the beginning of the extract, the physician has not directly addressed the son yet. In line 1, the companion produces a turn addressing his father with a second-person pronoun (“you”) and meanwhile pointing and gazing at him (lines 1–2, Fig. 9.2), thus prompting his father to ask the doctor about whether he can attend a wedding and have dinner. By addressing his father rather than directing the question to the

Extract 9.1  (#31, 03:17:11) 1

Comp: morgen heb jij vijftig jarige BRUILOFT;

2

(0.2) eventueel, (0.3) met DINER; (0.2) possibly , (0.3) with DINNER;

3 4

(0.3) Pat:

OJA morgen heb ik een BRUILOFT, (0.4) een vijftigjarige OH YES tomorrow I have a WEDDING,(0.4) a golden

5

bruiloft met een DINER (.) kan

ik daaraan deelnemen?

wedding with a DINNER (.) can I take part in that?

186 

W. J. P. Stommel and M. W. J. Stommel

Fig. 9.2  Pointing and gazing, line 1, Extract 9.1

Fig. 9.3  Line 4, Extract 9.1

physician, the companion orients to his father as the primary conversational partner and acts as a facilitating bystander. The patient reciprocates these participation roles by first treating the prompt as a relevant reminder (“OH YES,” line 4) and then phrasing the prompt as a question for the physician (lines 4–5) (cf. Norrick, 1987). Meanwhile, the companion has shifted gaze away from his father toward the screen (Fig. 9.3), watching the physician while waiting for his father to finish the question. Hence, even when a companion in a video-consultation appears on-screen, he may act, and be treated by the patient, as a bystander. In this setting, the bystander role involves “doing facilitation.”

From Bystander to Ratified Participant Using the video-connection, physicians lack full perceptual access to the patient’s setting (Heath & Luff, 1993), including information on the absence or presence of a companion. When a companion is not

9  Participation of Companions in Video-Mediated Medical… 

187

Extract 9.2  (#4; 00:55:29) 1 Phys:→

IK ZAG EEN RANDJE VAN UW MAN in beeld (.)zag ik. I SAW A FRINGE OF YOUR HUSBAND on-screen (.) I saw.

2

(2.9)

3 Pat:

wat zegt u?

4 Phys:

•h IK ZAG NET

what are you saying? EEN RANDJE VAN HET HOOFD VAN UW

•h I JUST SAW AN 5

in beeld.(0.5) die is t’r ook bij denk on-screen .(0.5) he is also there I think

6 Pat: 7 8 Phys:

MAN (.)

EDGE OF THE HEAD OF YOUR HUSBAND (.) [ik of] niet; [or

] not;

[oh;

]

[oh;

]

(1.3) [ja ] [yes]

9 Pat:

[ja ] die zi- hij zit hier naast mij; [yes] he si- he is sitting here next to me;

perceivable to the physician, we have the somewhat paradoxical situation that the companion is acknowledged by the patient, but unratified by the physician. Appearance on-screen, even very short or just partly, may change this. It may spark a recognition of the companion’s participation by the physician. That is, the companion’s appearance on-screen may elicit a noticing of the companion by the physician. By noticing, the physician brings into attention the companion as an aspect of the interactional environment by way of which he marks it as relevant for the ongoing interaction (Schegloff, 2007, p. 219) and thus ratifying his/her participation. Extract 9.2 is an example of noticing a companion by the physician. Figure 9.4a shows the physician’s screen which precedes the beginning of the extract. This image is the perceptual basis for the noticing. Figure 9.4b occurred at line 1. The physician notices the patient’s husband, referring to what he saw on-­screen, thus marking the noticing as specific to the video-mediated aspect of the interaction. In fact, this noticing is based on the partner’s

188 

W. J. P. Stommel and M. W. J. Stommel

Fig. 9.4a  Physician’s screen preceding Extract 9.2

Fig. 9.4b  Line 1, Extract 9.2

“appearance-­for-the-first-time” (Licoppe, 2017), which is an important cue for the initiation of interaction through a video connection. The patient’s other-initiated repair elicits a reformulation of the noticing (line 4), which is expanded by the question whether the husband is there too (line 5). This noticing reflects that not only non-headshots are scrutinized for their relevance to the ongoing interaction (Licoppe & Morel, 2012), but also so far unseen co-participants appearing on-screen, even very brief and partial appearances. At the same time, the noticing holds the patient accountable for whatever appears on-screen and could have made relevant a “correction” toward full talking head appearance. The patient’s response in line 6 is a change-of-state token (Heritage, 1984), which indicates that she was unaware of the on-screen appearance of her partner. She confirms and provides an account (“he is sitting here next to me,” line 9). The question verifying the partner’s participation (“he is also there or not,” line 5) could also have been taken up as an invitation for the companion to enter the frame in talking head position. The fact that this is not what happens and that this is not treated by the physician as lacking suggests that the physician aligns with the companion as merely a bystander. Hence, the on-screen appearance of the

9  Participation of Companions in Video-Mediated Medical… 

189

companion served as the ground for the physician’s ratification of the partner as an overhearing co-participant and created the environment for the companion to take up a more active role, but the companion declines this. Ratification may also be elicited by a turn by the patient rather than a visual cue. This is the case when talk between the patient and the companion signals to the physician that a companion is overhearing. Hence, crossplay between a ratified and an unratified participant may turn the companion into a perceivable participant for the physician. When the companion then directly addresses the physician, this constitutes a shift from overhearer to co-participant. Extract 9.3 is an example of such a shift. Figure 9.5 corresponds with line 1; the companion has not been visible or hearable for the physician yet until the beginning of the extract. In this extract, the patient initiates crosstalk with the companion (line 6) to enquire about the injections mentioned in line 1 (probably whether he has sufficient injection in stock). The companion, in turn, does not direct her response to the patient, but she addresses the physician directly, referring to the patient with the third-person pronoun “he” (line 6; Fig. 9.6). Interestingly, the companion does not move toward the camera to appear on-screen for the physician, despite the relevance of visibility for interaction (Goffman, 1981) and that of the talking head configuration for video-interaction specifically (Licoppe & Morel, 2012). Despite the fact that the spouse speaks from the invisible background, the physician responds to the new information (lines 8–9) and thus acknowledges the partner as a co-participant. Later on, even after sequence closure by the patient (“okay” line 16) the companion self-selects again to expand the sequence about the injections (line 17) and again, the physician responds directly (line 18). Hence, companions may shift from bystanders to more active co-participants, even without appearance on-screen. These examples show that the shift in participation status may be triggered by either visual perceptibility or by talk. In the Extract 9.2, it was the companion’s on-screen appearance which elicited a noticing from the physician; in Extract 9.3, it was crosstalk initiated by the patient, which invited the companion to talk, which in turn elicited a response from the physician to the companion directly. In both cases, the physician’s turn ratified the participation of the companion in the consultation.

190 

W. J. P. Stommel and M. W. J. Stommel

Extract 9.3  (#3; 08:52:00) 1

Pat:

EH HOE LANG MOET IK MET DE TROMBOSEPRIKKEN DOORGAAN EH HOW LONG MUST I CONTINUE THE THROMBOSIS INJECTIONS (0.4) die el[lende]

2

(0.4) that mi[sery ] 3

[vier ] (.) vier weken tot

Phys:

[four ] (.) four weeks until until four (.) four weeks na de operatie.

4

after the operation. (1.5)

5 6

Pat:

7

Comp: →

8

Phys:

(kan ik nog-- ik heb)((gazes at companion, image 6)) (can I

still -- I

have)

hij heeft drie weken d’r meegekrege’;= he has received three weeks;= =JA : oke da - da’s- drie weken is ook voldoende. =YES: okay tha - that’s- three weeks is also sufficient. (1.0) dat eh=

9

(1.0) that eh= 10

Pat:

=EHM

11

Phys:

dan hebben we die

12 13

=UHM

Pat: Phys:

[dagen]

then have we those [ days ] [mhm

]

[mhm

]

hè de- die dagen in het ziekenhuis al meegeteld (.) huh the- those days in the hospital already included (.) dan kom je onge veer op vier weken uit dus da’s goed.

14

then you have ap proximately four weeks so that’s good. (1.1)

15 16 17 18

Pat: Comp: → Phys:

okee. okay. hij heeft er wel eentje per ongeluk vergeten. he has forgotten one by accident. nah is is geen eh (.) g(h)een d(h)oodzonde eerlijk gezegd; nah is is not eh (.) n(h)o m(h)ortal sin to be honest ;

9  Participation of Companions in Video-Mediated Medical… 

191

Fig. 9.5  Line 1, Extract 9.3

Fig. 9.6  Gazes at companion, line 6, Extract 9.3

Re-ratifying the Companion Generally, companions who have been ratified are not addressed by the physician with initiating actions, such as questions. However, toward the closing of the consultation the physician may direct a turn at the companion. Extract 9.4 is an illustration of this. Earlier during the consultation, there has been some direct interaction between physician and companion (see Extract 9.3 which was taken from the same consultation). Although the companion has remained invisible on-screen for the physician during the whole consultation (e.g., see Figs. 9.5 and 9.6), the physician produces a question which indirectly addresses her toward the end of the consultation. After a “thank you” from the patient which moves the interaction toward closure (Schegloff & Sacks, 1973), the physician asks whether the companion—the patient’s spouse—has any other questions (lines 2 and 4). As the companion has spoken before during the consultation, this is an indirect invitation for her to speak again and present additional questions. This kind of invitation only occurred in consultations in which the

192 

W. J. P. Stommel and M. W. J. Stommel

Extract 9.4  (#3; 13:50:00) 1

Pat:

bedankt voor de goeie zorgen;

2

Phys: had uw

thanks for the good care; [ vrouw ] nog

did your [ wife ] still have 3

Pat:

4

Phys: had had uw (.) vrouw nog vragen,

[okee ] [okay ] did did your (.) wife still have any questions, >of niet,
or not,< (1.1)

6 7

Comp: nee hoor ik heb me - rustig mee zitten [luist’re’.] no [hoor] I have been

8

Phys:

9

Comp:

[listening] along quietly [okee [okay

(h)ik ben zee(h)r tevr(h)eden.

10 Phys: (h)I am ver(h)y cont(h)ent.

[hehehe

]

[hehehe

]

] ]

[hartstikke] goed; [excel

] lent;

companion had already spoken before; silent bystanders do not seem to be invited to present additional questions. As the patient and physician already produced various pre-closings (data not shown, but note the thanking in line 1 and the “okay” in line 3), this seems a “better late than never”—inserted move that works to re-ratify the companion. With this question, the physician acknowledges the companion’s contribution despite the fact that she never appeared on-screen. However, it should be noted that the physician addresses the companion indirectly using a third-person reference (“your wife”) in line 2 and again in line 4. The indirect address seems related to the companion’s non-appearance and thus potential absence. Addressing someone who might have left the scene is socially risky. Nevertheless, if she is still present, re-ratification is a pro-social move which may be particularly relevant

9  Participation of Companions in Video-Mediated Medical… 

193

in video-consultations (cf. Stommel et al., 2019). Indirect address of the companion via the patient (“your wife”) serves as a solution to that dilemma. A further noticeable aspect of this extract is that the physician’s turn in lines 2 and 4 receives a response from the companion directly. The companion claims she has no questions (line 7), and interestingly, she makes her status as a bystander explicit, whose role is to “listen […] along quietly” (line 7). She expands her answer with a positive evaluation produced with laughter particles (“I am ver(h)y cont(h)ent”), marking the ironical character of this evaluation (cf. Drew, 2009). This evaluation simultaneously serves as an account for her limited participation: being content there is no reason to speak. The physician’s re-ratification of the companion, however, shows that the companion’s participation is oriented to as relevant, even when they participate “in the background” during video-consultations.

Navigating Companion Participation Statuses Physicians and patients may differ in their orientation to the companion’s participation status, treating the companion either as a bystander or as a full conversational partner. This is what happens when a companion directs a question to the physician, which is then repeated (cf. second speaker repetition, Norrick, 1987) by the patient. Thereby, the patient treats the companion as an unratified participant and the companion’s turn as crosstalk. Nevertheless, the physician may respond to the companion directly and thus treat the companion as a ratified participant. Such conflicting orientations can be observed in Extract 9.5, which is taken from the same consultation as Extract 9.1. Figure 9.7 corresponds with line 1. Although the companion spoke as a bystander earlier in the conversation (see Extract 9.1), he now directs a question to the physician directly, acting as a full co-participant: “he can just take a shower right,” line 1. The physician immediately confirms (line 2) and thus reciprocates the companion’s participation. Nevertheless, the patient rephrases the question (line 4), treating his son as a bystander. The silence in line 3 and the

194 

W. J. P. Stommel and M. W. J. Stommel

Fig. 9.7  Line 1, Extract 9.5

Extract 9.5  (#31; 07:52:21) 1 Comp:

hij kan gewoon douchen he, he can just take a shower right,

2 Phys:

JA YES

3 4 Pat:

(0.4) KAN IK OOK GEWOON DOUCHEN, CAN I ALSO JUST TAKE A SHOWER,

5 6 Phys:

(0.4) ja u mag zeker gewoon douchen, yes you may certainly just take a shower,

increased volume in line 4 suggest the patient has not heard the physician’s answer in line 2 and hypothesizes that the companion’s question was produced too softly to be hearable for the physician. Hence, the communication medium and potential restricted audio transmission interferes with the negotiation of participation statuses. The physician next produces a full response rather than a minimal confirmation, addressing the patient directly (“you”) and thus avoiding an orientation to the repeated nature of this response or to the now disregarded companion’s role in the sequence. Hence, participants’ orientations to the status of the companion in the conversation may be conflicting and shift during video-consultations. Although physicians may sometimes work explicitly to ratify the companion as a conversational partner (see Extract 9.4), at other times they settle with an orientation to the patient as the primary co-participant.

9  Participation of Companions in Video-Mediated Medical… 

195

In one of our recordings, the patient was accompanied by two companions, the son and partner. The default talking head configuration for video-interaction (Licoppe & Morel, 2012) may be even more complex for three participants at one end as they have to move away from the screen/device to capture a sufficiently wide camera angle (cf. Mondada, 2007, 2010). The last extract shows how two companions take up different positions, one on-screen and the other off-screen, while both engaging in the interaction with the physician. This extract comes from the same consultation as Extract 9.1 and 9.5, but the transcript shows two companions participating in the interaction: the patient’s son (on-screen, see Fig. 9.8) and his spouse (off-screen). Like in Extract 9.5, the son speaks as a ratified participant. He reports a potential problem (smelling urine) and directs a turn to the surgeon, referring to his father with third-person reference (“his urine” lines 2–3). As the physician acknowledges the problem (“okay” line 2), the patient’s wife self-selects in overlap to expand the son’s problem account with “and he drinks little” (line 6). The design of this turn is ambiguous with regard to who is addressed, the son or the physician. Hence, the off-­ screen participant does not disambiguate whether her turn is crosstalk with the son, prompting him to add this relevant information, or whether she addresses the physician directly: she leaves it to the son and physician to decide. As she does not appear on-screen to talk while the son is visible, she does not seem to solicit ratification and thus merely acts as a facilitator. There is a short silence in which the physician could have responded to the spouse’s turn (line 7). As he does not, the son first repeats and then upgrades the problem of drinking little (“too little” lines 8-9), thus treating his mother’s turn as a prompt, that is, as crosstalk. In response, the Fig. 9.8  Physician’s screen, Extract 9.6

196 

W. J. P. Stommel and M. W. J. Stommel

Extract 9.6  (#31; 04:03:17) 1

Son:

[van de] week ging het goed he met (0.5) uhm (0.3) mijn [this

2

] week it went well right with (0.5) uhm (0.3) my

vader is ook naar de wc geweest heb jij eig - h >uh< zijn father also went to the toilet do you act - h >uh< his

3

urINE (0.3) die uh die RUIKT urine (0.3) it

4 5 6

uh it

nogal (.) dat wel;

SMELLS rather (.) really;

(0.7) Phys: Wife:

[okee

]

[okay

]

[en hij] drinkt weinig; [and he] drinks little;

7 8

(0.2) Son:

9

Phys:

10

Son:

en hij drinkt weinig (.)

[te ]

and he drinks little (.)

[too] [ja ] [yes]

weinig eigenlijk inderdaad ja . little actually

11

Phys:

indeed

[( )]

yeah. [( )] [ja ] dan zou mijn advies [yes] then my advice

12

zijn om te beginnen met wat meer te drinke', would be to start with drinking somewhat more,

physician does not explicitly orient to the wife’s contribution, merely producing a response token (“yes,” line 9) and an explicit advice to drink more (line 11–12). Thus, a companion whose turn opens up the possibility for ratified participation may turn out to remain a bystander. This seems related to the fact that her contribution was candidate crosstalk and to the fact that she did not appear on-screen. These two examples show that participation of companions in video-­ consultations is navigated or negotiated by the participants, involving both the physician and the patient, and potential of other companions. This may sometimes involve conflicting orientations or create “micro-­ friction” on the level of the interaction.

9  Participation of Companions in Video-Mediated Medical… 

197

Discussion First of all, it is remarkable that participation by companions in video-­ consultations is not self-evident. Many companions stay in the background, avoiding to be seen by the physician or refraining from speaking. Although we did not make a direct comparison of companion participation in the video- and the face-to-face setting, this tendency to stay “in the background” in video-consultations does seem different from the face-to-face setting in which the companion sits next to the patient and is continuously visible for the physician. We would even like to suggest that the companion in the video-setting normatively takes part as someone listening in and facilitating when relevant, but not (fully) appearing on-screen and not talking. This explains that when they do talk, their contribution often works as, or is invited by, crossplay with the patient, prompting the patient to address some issue, or inviting the companion to add information. Hence, companions in our data played a relatively modest part in the interaction. Now one may wonder why this is the case. We argue that this is related to the dominant set-up of the talking head position in video-interaction (Licoppe, 2017; Licoppe & Morel, 2012). Although we have seen instances of the companion speaking while being off-screen, this is rare and marked. Physicians may respond to the invisible third party, but they may also avoid to do so and leave it to the patient (or other companion) to rephrase the turn. This raises the question why companions do not “just” enter the frame next to the patient. Possibly, the norm for video-­ interaction does not only involve that available speakers appear in talking head configuration, but also that persons in talking head position should speak. Therefore, companions may avoid to appear in talking head position. Hence, it could be a medium factor that makes extensive participation of the companion less likely and more problematic in video-consultations. Merely facilitating companion participation may be beneficial in some cases, but detrimental in others. We found that physicians do not ratify the companions’ participation in the opening of the consultation, while in face-to-face consultations this is conventionally carried out as early as

198 

W. J. P. Stommel and M. W. J. Stommel

the opening with a handshake and mutual gaze. Instead, we only found explicit ratification by noticing the companion (one case) and in the consultation closing, but only in those cases in which the companion had already spoken before. Hence, in case companion participation is regarded beneficial, some enquiry and subsequent ratification of companions in the video-consultation opening may be a key to achieving this. This may be taken as a practical implication of our analysis. Relating our findings to previous studies of video-mediated interaction, a number of issues arise. First, like in informal video-interactions, we found that on-screen appearances may elicit talk (Licoppe & Morel, 2012) also in medical video-consultations. Thus, typical and normative video-in-interaction behavior occurred in the institutional setting. We would like to suggest that this should be seen in the context of affordance-­ related interaction that trickles into, and can be a resource for, institutional interaction, similar to how mundane interactional practices feature in institutional interaction (cf. Benwell & McCreaddie, 2016; Drew & Heritage, 1992). Furthermore, it is notable that we found no instances of byplay between companions, for instance between the patient’s wife and son. In multi-party videoconferences, participants in the local scene recurrently talked between one another, not involving the other end of the video connection (Oittinen, 2018). In our data, the local space interfered less with the video-mediated space, possibly because the local space included a limited number of participants (mostly two, in one case three) and because the overall dominant framework consisted of the physician-­ patient dyad, the meeting’s agenda and related duration. Goffman’s (1981) participation framework proved useful for the analysis of companion participation in video-consultations, although Goodwin’s (2007) critique was valid in the sense that the different speaker and hearer roles appeared dynamic rather than static during the interactions. The concept of crossplay recurrently applied to the interactions in accordance with Goodwin’s point that participation is organized moment-­ by-­moment in unfolding interaction. Although we adopted Goffman’s label of “bystander” for the analysis, the term “facilitator” may do more justice to the interactional contribution of companions in video-­ consultations. The companions do not “happen to be around” (as

9  Participation of Companions in Video-Mediated Medical… 

199

bystanders), but they reduce their own relevance to the interaction and orient to a status as potential facilitators. Overall, our analysis has shown that video-consultations may not just be considered and analyzed as a different institutional setting. When studying video-mediated interaction several issues arise in terms of data collection and analysis that are somehow distinct from research on co-­present interactions. First, critical choices have to be made with regard to recording: when recordings are made on both ends the analysis can take into account orientations to the local space. In case of one-end recordings, the analysis should be sensitive to the fact that relevant information from the other local space might be missed, such as whether or not a companion is present in the room. This implies that the analysis is inevitably focused on one ends’ perspective (cf. Olbertz-Siitonen, 2015). Second, some resources available in co-present face-to-face interaction, such as gaze, are lacking or impaired in video-mediated interaction (Heath & Luff, 1993). This means that analytically, gaze is not always usefully considered. Hence, the analysis of institutional video-interaction involves medium-specific questions and sensitivities, for participants as well as analysts. A potential limitation of our analysis is that our data included consultations from one specific medical setting: post-operative consultations. This means it is unclear whether our findings are representative for medical video-consultations in general. Typical for the consultations in our data set is that the physician explained test results, which tended to involve a lot of talk by the physician. A relatively minor part of the consultations consisted of a patient report of the recovery; most companion contributions were related to this activity. Possibly, companions have a more active role and default on-screen appearance, for example, in consultations focused on treatment recommendations (shared-decision making).

Conclusion Our microanalysis revealed how video-in-interaction plays out in the specific institutional setting of post-operative video-consultations. The analysis focused on the patients’ intimates, who were present at, and sometimes contributed to, the consultation. We examined the

200 

W. J. P. Stommel and M. W. J. Stommel

participation of companions in 13 video-consultations, focusing on those consultations and specifically sequences in which a ratification of the companion took place. In most cases, the companion served as a facilitating bystander, invisible for the physician and thus maintaining the physician-patient dyad as the dominant framework for the interaction. This raised the question when and how companions “join in” and how this is related to the video-mode of communication. We found that companions as bystanders are rarely invited to talk unless they appear on-screen and are noticed by the physician. Self-initiated turns by companion consist of crosstalk with the patient, prompting them to report or ask something, while companions may also directly address the physician with relevant information. Physicians, in response, ratify companions, but they rarely actively invite them to participate. In fact, they only sometimes (indirectly) initiate interaction with companions toward the end of the consultation inviting them to raise additional questions. Overall, companions’ on-screen appearance seemed crucial for their participation, which suggests that the role of companions in video-consultations is fundamentally different from how they participate in face-to-face consultations.

References Beisecker, A. (1989). The influence of a companion on the doctor-elderly patient interaction. Health Communication, 1(1), 55–70. Benwell, B., & McCreaddie, M. (2016). Keeping “small talk” small in health-­ care encounters: Negotiating the boundaries between on- and off-task talk. Research on Language and Social Interaction, 49(3), 258–271. https://doi. org/10.1080/08351813.2016.1196548 Bou-Franch, P., & Garcés-Conejos Blitvich, P. (2014). Conflict management in massive polylogues: A case study from YouTube. Journal of Pragmatics, 73, 19–36. https://doi.org/10.1060/j.pragma.2014.05.001 Bou-Franch, P., Lorenzo-Dus, N., & Garcés-Conejos Blitvich, P. (2012). Social interaction in YouTube text-based polylogues: A study of coherence. Journal of Computer-Mediated Communication, 17, 501–521. Drew, P. (2009). Po-faced receipts of teases. Linguistics, 25(1), 219–253. https:// doi.org/10.1515/ling.1987.25.1.219

9  Participation of Companions in Video-Mediated Medical… 

201

Drew, P., & Heritage, J. (1992). Talk at work: Interaction in institutional settings. Cambridge: Cambridge University Press. Fioramonte, A., & Vásquez, C. (2019). Multi-party talk in the medical encounter: Socio-pragmatic functions of family members’ contributions in the treatment advice phase. Journal of Pragmatics, 139, 132–145. https://doi. org/10.1016/j.pragma.2018.11.001 Giles, D., & Newbold, J. (2013). ‘Is this normal?’ The role of category predicates in constructing mental illness online. Journal of Computer-Mediated Communication, 18, 476–490. https://doi.org/10.1111/jcc4.12022 Goffman, E. (1981). Forms of talk. Philadelphia: University of Pennsylvania Press. Goodwin, C. (2007). Interactive footing. In E. Holt & R. Clift (Eds.), Reporting talk: Reported speech in interaction (pp.  16–46). Cambridge: Cambridge University Press. Hasselkus, B. (1992). The family caregiver as interpreter in the geriatric medical interview. Medical Anthropology Quarterly, 6(3), 288–304. Heath, C., & Luff, P. (1993). Disembodied conduct: Interactional asymmetries in video-mediated communication. In G. Button (Ed.), Technology in working order: Studies of work, interaction, and technology (pp. 35–54). London: Routledge. Heritage, J. (1984). A change-of-state token and aspects of its sequential placement. In M.  Atkinson & J.  Heritage (Eds.), Structures of Social Action (pp. 299–345). Cambridge: Cambridge University Press. Huber, J., Streuli, J., Lozankovski, N., Stredele, R., Moll, P., Hohenfellner, M., … Peters, T. (2016). The complex interplay of physician, patient, and spouse in preopreative counseling for radical prostatectomy: A comparative mixed-method analysis of 30 videotaped consultations. Psycho-Oncology, 25, 949–956. https://doi.org/10.1002/pon.4041 Hutchby, I. (2001a). Conversation and technology: From the telephone to the internet. Cambridge: Polity Press. Hutchby, I. (2001b). Technologies, texts and affordances. Sociology, 35(2), 441–456. https://doi.org/10.1177/S0038038501000219 Jefferson, G. (2004). Glossary of transcript symbols with an introduction. In G.  Lerner (Ed.), Conversation analysis: Studies from the first generation (pp. 14–31). Amsterdam and Philadelphia: John Benjamins. Laidsaar-Powell, R. C., Butow, P. N., Charles, C., Lam, W. W. T., Jansen, J., McCaffery, K.  J., … Juraskova, I. (2013). Physician-patient-companion communication and decision-making: A systematic review of triadic medical consultations. Patient Education and Counseling, 91(1), 3–13. https://doi. org/10.1016/j.pec.2012.11.007

202 

W. J. P. Stommel and M. W. J. Stommel

Licoppe, C. (2017). Skype appearances, multiple greetings and ‘coucou’. The sequential organization of video-mediated conversation openings. Journal of Pragmatics, 27(3), 351–386. Licoppe, C., & Morel, J. (2012). Video-in-interaction: ‘Talking heads’ and the multimodal organization of mobile and skype video calls. Research on Language and Social Interaction, 45(4), 399–429. https://doi.org/10.108 0/08351813.2012.724996 Meredith, J., & Stokoe, L. (2014). Repair: Comparing Facebook ‘chat’ with spoken interaction. Discourse and Communication, 8(2), 181–207. https:// doi.org/10.1177/1750481313510815 Miller, E. A. (2003). The technical and interpersonal aspects of telemedicine: Effects on doctor-patient communication. Journal of Telemedicine and Telecare, 9, 1–7. Mondada, L. (2007). Imbrications de la technologie et de l’ordre interactionnel. L’organisation de vérifications et d’identifications de problèmes pendant la visioconférence. Réseaux, 5, 141–182. Mondada, L. (2010). Eröffnungen und Prä-Eröffnungen in medienvermittelter Interaktion: Das Beispiel Videokonferenzen. In L. Mondada & R. Schmitt (Eds.), Situationseröffnungen: Zur multimodalen Herstellung fokussierter Interaktion (pp. 277–334). Tübingen: Narr. Norrick, N. (1987). Functions of repetition in conversation. Text, 7, 245–264. Oittinen, T. (2018). Multimodal accomplishment of alignment and affiliation in the local space of distant meetings. Culture and Organization, 24(1), 31–53. https://doi.org/10.1080/14759551.2017.1386189 Olbertz-Siitonen, M. (2015). Transmission delay in technology-mediated interaction at work. PsychNology Journal, 13(2-3), 203–234. Pappas, Y., & Seale, C. (2009). The opening phase of telemedicine consultations: An analysis of interaction. Social Science and Medicine, 68, 1229–1237. Pappas, Y., & Seale, C. (2010). The physical examination in telecardiology and televascular consultations: A study using conversation analysis. Patient Education and Counselling, 81, 113–118. Paulus, T., Warren, A., & Lester, J. N. (2016). Applying conversation analysis methods to online talk: A literature review. Discourse, Context & Media, 12, 1–10. Schegloff, E. (2007). Sequence organization in interaction: A primer in conversation analysis. Cambridge: Cambridge University Press. Schegloff, E., & Sacks, H. (1973). Opening up closings. Semiotica, 8(4), 289–327. Sidnell, J., & Stivers, T. (Eds.). (2013). The handbook of conversation analysis. London: Blackwell Publishing.

9  Participation of Companions in Video-Mediated Medical… 

203

Stommel, W., Van Goor, H., & Stommel, M. (2019). Other-attentiveness in video-consultation openings: A conversation analysis of video-mediated versus face-to-face consultations. Journal of Computer-Mediated Communication, 24(6), 275–292. https://doi.org/10.1093/jcmc/zmz015 Stommel, W., & Koole, T. (2010). The Online Support Group as a community: A micro-analysis of the interaction with a new member. Discourse Studies, 12(3), 357–378. Stommel, W., & Te Molder, H. (2015). Counselling online and over the phone: When pre-closing questions fail as a closing device. Research on Language and Social Interaction, 48(3), 281–300. https://doi.org/10.1080/0835181 3.2015.1058605 Wasson, C. (2006). Being in two spaces at once: Virtual meetings and their representation. Journal of Linguistic Anthropology, 16(1), 103–130. Zilliacus, E. M., Meiser, B., Lobb, E. A., Kelly, P. J., Barlow-Stewart, K., Kirk, J. A., … Tucker, K. M. (2011). Are videoconferenced consultations as effective as face-to-face consultations for hereditary breast and ovarian cancer genetic counseling? Genetics in Medicine, 13(11), 933–941. https://doi. org/10.1097/GIM.0b013e3182217a19

10 Conclusion: Future Directions in Analysing Digital Interaction Janet Smithson

This book showcases a variety of new developments in analysing digital interaction, both in terms of methodological innovation and in terms of application of microanalysis of talk to new technological and social contexts. The chapters comprise a detailed set of examples of how to use Conversation Analysis (CA), Membership Categorisation Analysis (MCA) and related microanalysis techniques to understand developing technological social interaction. They exemplify some of the key challenges in this area and highlight a few key directions for future work. In summarising the nine chapters, I draw attention to these aspects of each

“It’s noted that the style guide is that extracts are presented with a grey background. However, there is inconsistency across chapters, with chapters 3, 4 and 5 having extracts with no grey background, but all the other chapters do have this. I suspect this is because in chapters 3, 4 and 5 the extracts are called examples or presented in tables. I would much prefer it if all extracts could be presented as those in chapters 4 and 5, with a white background with a box around them.”

J. Smithson (*) Washington Singer Laboratories, University of Exeter, Exeter, UK e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7_10

205

206 

J. Smithson

separate chapter, before drawing out some common themes and issues, and ideas for future research directions. In Chap. 1, Meredith, Giles and Stommel set out the rationale behind this book, and raise several key issues around research in this area. These issues include the question of what counts as interaction, in a digital context. Using the construct of recipient design, drawing on Goffman’s theory of participation frameworks, they explain that online “interaction” is more complicated to define and establish than in tradition contexts. In face-to-face (physical) interaction, it is much clearer who an utterance is aimed at—who are the participants—and also whether they have received the information, evidenced by the next turn. Establishing the existence and nature of digital interaction is also less straightforward than more traditional technology (television, film, radio), where there is typically a speaker and an audience. In digital interaction, the participation framework can be hard to specify; often there is a mixture and crossover of ratified participants (those who are aimed at and respond) and non-ratified readers (lurkers). This is an interesting area for further study, as well as an issue to note in the study of digital “interaction”. We can see attention to it in the empirical chapters in this book. A second area covered in this introductory chapter is the development and adaptation of CA and related methods to online talk and text. The notion of affordances, or how the design of everyday objects, including technological objects, affords particular uses, is introduced as having historical importance in CA and ethnomethodology, and as being central to an understanding of online interaction. Here the CA focus of examining the interaction in detail, first, and using this to understand the details of the medium, is discussed. A related issue, and a matter of ongoing debate in microanalytic research, is noted in the introduction chapter: the question of what counts as context, and a mediated context, when analysing digital interaction. One of the interesting facets of studying humans and technology is the mediation between “normal” human behaviour, the design of suitable tools for humans to use and resulting amendments to human behaviour from tool use. As noted by Thorne and Black (2011), technological tools do more than merely provide a setting for interaction: they re-­ mediate human activity. The interdependence of human and technology therefore produces a revised form of interaction, and new types of human actions, expectations and identities.

10  Conclusion: Future Directions in Analysing Digital Interaction 

207

In Chap. 2 Ditchfield outlines some of the key areas of discussion for ethics for digital research. Notable aspects include the public/private distinction, a historic distinction which matters to ethics theorising and systems, but which makes less sense in fluid online boundaries. Similarly, the concept of informed consent is complicated in digital research, as user expectations do not always align with legal definitions of public activity. Other recent challenges to research ethics due to technological change include the concept of searchability. In early digital research it was possible to mask data or use data which would be unlikely to be linked to the original source. Now a quote can be traced very quickly via search engines. This is particularly an issue for microanalytic approaches; as Ditchfield points out, amending data to make it less traceable does not fit with a microanalytic approach, where the exact context and wording is crucial. One possible solution is limiting visibility of the data, limiting access to researchers but not all readers. Ditchfield reminds us that ethics is a matter of developing guidelines, rather than rigid rules, but it is still essential to have (preferably published) guidelines and frameworks to work within, especially for new researchers needing to gain ethical approval for their studies. The rest of the book comprises empirical examples of microanalytic studies of a variety of interaction types occurring in digital media. In Chap. 3, Giles focuses on the often-thorny issue of context in microanalytic research, using Twitter threads as a case study. As he points out, there is a surprising lack of microanalytic research into Twitter interaction, given its salience in a wider culture, “central now to the way that even politics gets done”. A challenge for a conversation analyst in studying Twitter is the messiness of the interactions: determining which tweet is a response to which other tweet, when tweets are not set out neatly in threads or conversations, so conversations need “reconstructing”. Giles draws on the notion of affordances, introduced previously in Chap. 1, and suggests that this is central to an understanding of online interaction. Taking a Discursive Psychology approach, Giles argues that affordances should not be reduced solely to technological features, but also to wider contextual information that the interaction is embedded in. Specifically, he argues that the analysis of Twitter communication needs external cultural or contextual knowledge to understand the data. In Chap. 4, Linda Walz focuses on new types of social communication, using Membership Categorisation Analysis (MCA) to study

208 

J. Smithson

identity negotiation through blogging and blogs, “a medium that is especially suited to engaging with the experience of undergoing personal change”. While, like Twitter, blogs are a new form of writing enabled by recent technology, unlike Twitter the format of the blog entries is easily adaptable to existing analytic methods. Walz demonstrates some similarities to traditional diaries or letters in the longitudinal nature of blog posts, or, given the public audience of blogs, the similarities to a weekly column in a traditional print newspaper. In Chap. 5 Anderson also uses MCA to study how Facebook users respond emotionally to a video clip of chat show hosts. Analysis focuses on a traditional CA interest, the morality work done by hosts and followers/users, highlighting how both co-host and users do not appear to question the “authenticity” of the host’s displayed emotions—specifically, when hosts produce tears, this is considered as authentic emotion, not as part of a performance. This activity of “encouraging the audience to seek out stimuli for the purpose of triggering tears” is shown to be qualitatively different from what occurs in other studies of crying—for instance, in calls to a child protection helpline. This chapter considers users’ evaluation of the host as a good person, genuine and caring—a “big heart” in Danish. The novel aspect which this analysis demonstrates is how people typically post as though they are talking directly to the radio show hosts, even when posting below a video clip. Andersen relates this to the concept of “context collapse”—although users at one level know that they are posting after the event, not directly to the presenters, that is, at a different communicative level, they post as though they were on the same communicative level, in the same time frame, as the show hosts. In this study, traditional MCA methods, and a traditional MCA focus on morality talk in chat shows, illustrate new knowledge about digital communication formats and how people adapt to them. In Chap. 6, Meredith draws on the literature in CA in ending conversations, including previous work on online closings. Looking at comments “below the line” in online newspapers, she demonstrates how participants “end” an online dispute by just not returning, or by posting an idiomatic ending (“Like turkeys voting for Christmas”) which indicates that no further comments are needed, rather than by providing an explicit closing. As does the previous chapter, this chapter highlights differences between face-to-face and online communication, and the

10  Conclusion: Future Directions in Analysing Digital Interaction 

209

concept of affordances is used to explain this difference. An online argument can look like a one-to-one discussion but actually has many possible members, either participating or in the audience. This is an example of the tension between online interaction as public or private talk in terms of ethics, as this sort of interaction has features of both. Even where only two posters debate, there are also elements of performativity, of awareness of providing information or entertainment for a large, unseen audience. In Chap. 7, Spagnioli uses the core CA interest in turn continuation, focusing on the similarity between practices observed in new mediated conversational settings and familiar spoken conversation. This similarity represents both a methodological issue and a conceptual issue related with the nature of mediated interaction: Remediation. Spagnioli gives an example of chat users modifying their online turns, explicitly using a particular format—“turn continuations”, while in face-to-face conversation, they would be holding the floor more visibly/audibly to the audience. In a text conversation they need to make it clear that they are working on a longer, extended turn, without losing attention. This “constant work of adjustments and creativity” by which people amend their interactional activity to new contexts and formats demonstrates the flexibility which participants bring to new technologies and also reminds us of the distinctive human pleasure in playing around with linguistic and interactional changes. As Meredith (2019) concluded, the study of online interaction offers fascinating insights into human interaction more generally. In Chap. 8, the focus is slightly different from previous chapters, in that Licoppe’s interest is in a new type of behaviour, ghosting, which is enabled or emerging from online interaction possibilities. Licoppe uses Tinder as a case study, but the phenomenon of ghosting has been noted in other social media. This supports the suggestion in the previous chapter that digital culture is not just communication about what happens offline but something new. Licoppe aims for an “anthropological understanding” of the Tinder chats, using “CA-inspired sequence analysis”. Similarly to Giles’ analysis of Twitter threads, Licoppe shows the non-­ sequential nature (or at least complex in determining the sequentiality) of Tinder “conversations” and draws on ethnomethodological concepts of situated talk to argue that “textual conversation threads are embedded in off-screen and on-screen activities and ecologies, that is texting and chatting as ‘lived work’”. Licoppe’s question about the distinctiveness of the

210 

J. Smithson

sequential organisation of Tinder chat conversations brings us back to one of the key issues arising in this book and points to possible future directions for microanalytic research into digital communication. In the final empirical Chap. 9, Stommel and Stommel analyse video-­ mediated interaction: another increasingly common form of communication, but with little research conducted on it, both in CA/DP circles and more widely. The chapter describes a study of medical consultations between a health professional and patients, who often have a companion present. The analysis focuses on the notion of visibility, drawing on Goffman’s (1981) concept of participation frameworks, and they illustrate how new normative expectations are emerging with this new type of online communication. For instance, they found that the appearance of companions on screen seemed to be central for their participation in the consultation, so effective consultation including companions needs to ensure that companions are actively included in the frame, by either patient or companion or medical consultant. Stommel and Stommel argue that, therefore, the role of companions in video-consultations is fundamentally different from their role in face-to-face consultations. Reading these chapters overall, several aspects are noticeable. Some authors focus on using “traditional” social science research methods and techniques to study new forms of technology-mediated communication. Others focus on new methodological approaches suitable for emerging modes of communication. One key question for this volume and future directions is this: can we apply existing CA, MCA and related microanalytic practices to new terrain, and do we need to adapt the methods and concepts to deal with the new terrain? Giles, Stommel and Paulus (2017) argued that “traditional methods of inquiry need considerable adjustment to fully understand the kinds of interaction that are taking place in online environments”, but several authors here have demonstrated how the traditional tools of CA and MCA can be utilised both to illustrate similarities with traditional communication formats and to clarify the details of what is going on in these modes of communication. For example, Walz’s blog analysis has implications for gender dynamics, relationship expectations, social influence and hierarchies. Authors have also highlighted how the use of traditional microanalytic tools can illustrate new interactional patterns emerging in digital contexts—“remediation”

10  Conclusion: Future Directions in Analysing Digital Interaction 

211

as Spagnioli terms this. There does therefore appear to be plenty of potential for the traditional methods of CA and MCA in the study of online data. A related question is, in what ways do the new forms of communication illuminate pre-existing social norms, behaviours and interaction patterns, and in which ways do new forms lead to interactional/behavioural change, or remediation of human activity? For example, does the practice of “ghosting” reflect a tendency already existing in face-to-face relationships, but make it easier? Or is it a primarily online behaviour, emerging from the technological possibilities? If this is so, might this emerging new online behaviour then mutate into offline relationship behaviour and expectations? Several of the chapters have implications for our understanding of how morality and emotions are enacted via digital communication, and how social norms, display and expectations may be changing as people increasingly utilise digital media for an ever-wider number of social and business interactions. A persistent question in microanalytic research, historically and also evident in this volume, is to what extent does context matter including the nature of the technology being used for interaction? We can see the DP/CA distinction here in the various author perspectives, with Giles, in particular, focusing on some of the limitations of traditional CA for understanding Twitter conversations, arguing that without understanding the context and the links to other Twitter discussions the interaction is incomprehensible. Licoppe’s analysis similarly focuses on the importance of knowledge about the non-digital context in understanding Tinder communication. Another use of microanalysis, illustrated in this book, is the focus not on what new technology features are initially designed, or marketed to do, but on how users do in practice utilise features or not. Close study of how users actually use these technologies, together with study of the impact of human interactional expectations on utilisation, can inform subsequent design, or redesign of technologies. Overall, the volume suggests a variety of directions for further research and new areas to develop. This book focuses particularly on social interaction, understood broadly, for example, including health professional-­ patient interactions. Ditchfield suggests that a reason for ethical

212 

J. Smithson

uncertainty in online research is that “the internet is still in a “formative phase” with new phenomena continually emerging”. However, the internet is unlikely to “settle” and new technologies will keep emerging, making earlier knowledge or practices incomplete. So, one enduring feature needs to be the ability to, or expectation of, adapting approaches and methods as technology does change. A key issue here is not just how microanalytic researchers think about ethical issues but how to manage this within the wider constraints and guidelines of General Data Protection Regulation (GDPR) and ethics committees. In 2020, due to the Covid-19 global crisis, major developments in online interaction were forced into being, with the sudden and unplanned shift of much of the world’s work and social interaction onto online technologies. “Since lockdown began, the online world has ignited with activity. From Tik Tok videos and the Houseparty app, to Instagram parties and Zoom quizzes and yoga sessions, it appears many of us are more sociable than ever now that government policy has restricted our physical contact” (Liddell, 2020). This has created many opportunities and also challenges for microanalytic research in this field. One challenge is research keeping up with the speed of technological change. Most of the chapters in this book, with the exception of Stommel and Stommel (Chap. 9), are based on written text online (blogs, comments pages, texting) but, as Meredith notes in the introduction (Chap. 1), the Covid-19 crisis has led most obviously to a rise in videoconferencing videochat use. Clinicians, school teachers, university lecturers, charities, local councils and many organisational teams have shifted to working via virtual platforms (Zoom, MS Teams, Skype), while online socialising has utilised these same virtual conference platforms, and also less formal videochat apps (WhatsApp, Houseparty) and gaming platforms. A fruitful area for research is the ways in which previously distinct age-­ related preferences, patterns and usage have collapsed somewhat, so technologies previously favoured by teenagers and young adults—the digital natives (Suh, Bentley, & Lottridge, 2018), and those used by some types of workforce, such as universities, have been more widely taken up, with people of all ages swiftly learning to shop online, to talk to friends and family by videochat, and attending churches, mosques, medical appointments and social activities by Zoom.

10  Conclusion: Future Directions in Analysing Digital Interaction 

213

As noted by Ditchfield (Chap. 3), changes in technology as well as changing understandings of privacy and security (most notably, GDPR) mean that the field of online research has had to attend to fast-changing ethical expectations, and Covid-19 crisis has added a new layer to this. Emerging ethical concerns for users of online communication systems include worries about Big Data, including minimising the risk of personal data breaches and reduction in participant rights and liberties in new hurriedly rolled out systems, and attending to further use of research data, either due to open access policies or large organisations breaching regulations (Hesse, Glenna, Hinrichs, Chiles and Sachs, 2019; Sula, 2016). There are also recent organisational and personal concerns around videoconferencing security (Marhefka, Lockhart, & Turner, 2020). These domains need attending to in microanalytic research, but also might be relevant domains for study. Another possible focus of microanalytic research is around the increasing use of medical apps such as wellbeing apps, or the current focus on coronavirus testing and tracing apps. Research and development of these apps has been increasing in medical and clinical centres, but until now there has been little CA or Discourse Analysis (DA) study of user interactions in apps (one exception being Thomas and Lupton’s Discursive Psychology study of pregnancy apps, 2016). Future research could include active consideration of who is given prominence or enabled in new media, and who is minimised or silenced. One area to note is people who are unable to afford or use new technology and risk being left behind. As the Covid-19 crisis demonstrated, it is far easier for people with space at home, fast broadband and good technology to work, study and socialise online, and this digital divide may be inadvertently replicated in research without a specific consideration of who benefits, who loses out and in whose interests it is to enable certain forms of communication and prioritise some types of technology over others. Microanalytic approaches have tended to consider that all relevant issues for CA study about power and hierarchies are “demonstrably relevant” in the detail of the talk, and certainly CA can be utilised to highlight inequalities (Stokoe & Speer, 2017). However, within the studies of users of various technological media in different contexts, we might consider, who is prioritised, who is ignored and how does this relate to existing societal hierarchies?

214 

J. Smithson

What is the developing shared cultural context within which participants are working? Who is not speaking in these formats, and who is not even on these technologies? This book was focused on the increasing role of digitally mediated interaction in our lives. Suddenly, this expertise is widely needed and relevant to a much bigger audience, as individuals, organisations and government policymakers look to reshape many forms of interaction into online systems. Some of the benefits and risks of technology-mediated communication may be made more evident through this Covid-19 crisis, and the benefits (but hopefully not the risks) may be retained and expanded. The expertise of how to study online interaction, and the implications and possibilities which this field offers, is absolutely crucial to the current and emerging priorities across the globe, as people move on from this crisis but also develop robust technological systems to cope with future changes, both planned and unplanned.

References Giles, D. C., Stommel, W., & Paulus, T. (2017). The microanalysis of online data: The next stage. Journal of Pragmatics, 115, 37–41. Goffman, E. (1981). Forms of talk. Oxford: Blackwell. Hesse, A., Glenna, L., Hinrichs, C., Chiles, R., & Sachs, C. (2019). Qualitative research ethics in the big data era. American Behavioral Scientist, 63(5), 560–583. Liddell, F. (2020, May). Finding value in the challenges of COVID-19. Cultural Practice. University of Manchester: Institute for Cultural Practices. Retrieved from https://culturalpractice.org/valueandcovid-­19. Marhefka, S., Lockhart, E., & Turner, D. (2020). Achieve research continuity during social distancing by rapidly implementing individual and group videoconferencing with participants: Key considerations, best practices, and protocols. AIDS and Behavior, 1, 1983–1989. Meredith, J. (2019). Conversation analysis and online interaction. Research on Language and Social Interaction, 52(3), 241–256. Stokoe, E., & Speer, S. (2017). Conversation analysis, language and sexuality. In K. Hall & R. Barrett (Eds.), The Oxford handbook of language and sexuality. Oxford: Oxford University Press.

10  Conclusion: Future Directions in Analysing Digital Interaction 

215

Suh, M., Bentley, F., & Lottridge, D. (2018). ‘It’s Kind of Boring Looking at Just the Face’ How teens multitask during mobile videochat. Proceedings of the ACM on Human-Computer Interaction, 2(CSCW), 1–23. Sula, C. A. (2016). Research ethics in an age of big data. Bulletin of the Association for Information Science and Technology, 42(2), 17–21. Thomas, G. M., & Lupton, D. (2016). Threats and thrills: Pregnancy apps, risk and consumption. Health, Risk & Society, 17(7–8), 495–509. Thorne, S.  L., & Black, R.  W. (2011). Identity and interaction in Internet-­ mediated contexts. In C. Higgins (Ed.) Identity formation in globalizing contexts: Language learning in the new millennium (Vol. 1, pp. 257–278). Berlin: de Gruyter.

Name Index1

B

G

Baron, Naomi, 135, 136n4, 156 Bou-Franch, Patricia, 56, 91, 106, 114, 115, 128, 178

Giles, David, 5, 8, 12, 13, 23, 25, 27–29, 32, 34, 35, 37, 41–44, 46, 48, 48n3, 49, 53, 55, 56n5, 67, 69, 81, 112, 114, 116, 117, 122, 128, 177, 206, 207, 209–211 Goffman, Erving, 8, 158, 159, 180, 181, 183, 189, 198, 206, 210 Goodwin, Charles, 158, 181, 198 Grice, Paul, 165, 167

D

Drew, Paul, 117, 124, 127, 193, 198 E

Edwards, Derek, 15, 34, 101–103, 126

H F

Fitzgerald, Richard, 68–71, 80, 83, 89, 98

Heritage, John, 117, 169, 188, 198 Herring, Susan, 12, 91, 122, 161 Housley, William, 15, 25, 29, 35, 68, 70, 83, 89, 98

 Note: Page numbers followed by ‘n’ refer to notes.

1

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7

217

218 

Name Index

Hutchby, Ian, 13, 23, 112, 114, 118, 127, 134n1, 179

R

Reed, Darren, 5, 12, 15, 23, 41, 67, 117, 156, 157

J

Jefferson, Gail, 7, 9, 10, 120, 124, 135, 157, 182 M

Markham, Annette, 24–26, 33 Meredith, Joanne, 6–8, 12–14, 24, 25, 29–32, 36, 37, 112, 115, 126, 156, 177, 206, 208, 209, 212 P

Paulus, Trena, 5, 14, 15, 23, 41, 67, 117, 157, 178, 210 Potter, Jonathan, 7, 8, 15, 29, 31, 37, 87, 88, 91, 156

S

Sacks, Harvey, 7, 9, 10, 68, 70, 72, 79, 82, 83, 89, 90, 113, 120, 128, 135, 140, 157, 165, 191 Schegloff, Emanuel, 7, 9, 10, 42–44, 59, 60, 81, 82, 88–90, 94, 113, 117, 120, 128, 134, 135, 137, 139–141, 145, 147, 150, 150n7, 157, 187, 191 Stokoe, Elizabeth, 9–11, 14, 24, 26, 29, 32, 36, 43, 60, 81, 82, 89, 90, 103, 213 Stommel, Wyke, 5, 14, 15, 23, 26, 32, 33, 37, 41, 46, 67, 69, 81, 105, 112, 114, 117, 120, 126–128, 157, 163, 177, 178, 182, 193, 206, 210, 212

Subject Index1

A

Accountability, 34, 105, 157, 158, 169–172 Affordances, 13, 31, 42, 48–50, 66, 75, 82, 83, 107, 117, 127, 129, 134, 134n1, 135, 151, 152, 179, 206, 207, 209 Anonymity/anonymous, 8, 13, 24, 31–35, 37, 46, 47, 116 B

Blogs/blogging, 6, 7, 45, 46, 65–84, 116, 208, 210, 212 C

Chat, 8, 14, 29, 30, 32, 33, 36, 114, 117, 134–140, 136n4, 142,

143, 145, 147, 155–174, 177, 208–210 Context, 5, 8–13, 23, 26, 27, 30–33, 35, 37, 41–61, 66–68, 75, 80, 82, 83, 88, 90, 93, 94, 98, 101, 107, 112, 114, 120, 129, 134, 146, 147, 151, 152, 156, 157, 165n4, 169, 171–173, 177, 198, 205–211, 213, 214 Conversation analysis (CA), 5, 6, 9–16, 23–28, 31–35, 37, 38, 42, 43, 53, 83, 117, 128, 129, 140, 145, 156–158, 205, 206, 208–211, 213

 Note: Page numbers followed by ‘n’ refer to notes.

1

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 J. Meredith et al. (eds.), Analysing Digital Interaction, Palgrave Studies in Discursive Psychology, https://doi.org/10.1007/978-3-030-64922-7

219

220 

Subject Index

D

P

Discursive psychology (DP), 9, 11, 15, 23–28, 31–35, 37, 38, 42, 50, 53, 112, 207, 210, 211, 213

Participant framework, 8, 180–181, 183–185, 198, 206, 210 Public/private, 24–32, 34–37, 42, 45, 46, 52, 56, 70, 88, 113, 116, 128, 129, 134, 143, 180, 207–209

E

E-mail, 11, 113, 117 Ethics, 16, 24–26, 38, 53, 116, 207, 209, 212

R

Recipient design, 7, 8, 81, 169, 206 Relevance, 42, 43, 94, 129, 143, 161, 178, 188, 189, 199

F

Facebook, 7, 13, 25, 27–31, 36, 87–107, 177, 208 Forums, 14, 15, 32, 35, 36, 46, 49, 50, 55, 57, 81, 112, 122, 128, 177, 178 I

Informed consent, 24, 25, 28–31, 35, 37, 207 Instant messaging, 7, 8, 14–16, 113, 133–152

S

Sequence, 3, 9, 11, 12, 14, 15, 50, 52, 57, 113, 115–117, 119, 120, 123, 124, 127, 129, 134, 139, 143, 163, 165–167, 173, 182, 185, 189, 194, 200, 209 Social media, 2, 5, 7, 13, 15, 33, 36, 42–45, 56n5, 60, 66, 68, 106, 209 T

M

Membership categorization analysis (MCA), 9–12, 16, 67–70, 80–84, 88–90, 105, 165, 171, 205, 207, 208, 210, 211 Microanalysis, 1–16, 49, 177–200, 205, 211

Threads, 7, 46, 47, 49, 50, 55–58, 111–129, 156, 157, 207, 209 Twitter, 1–3, 7, 8, 13, 15, 25, 27, 28, 34, 35, 41–61, 207–209, 211 V

Video-mediated, 6–8, 15, 177–200, 210

N

Naturally occurring, 23, 89, 105, 136n4, 178

Y

YouTube, 7, 45, 91, 106, 114, 115, 177