“Gua means scrape”: a conversation analysis of identity construction and negotiation in polylogal Wikipedia paratext

ABSTRACT

This article contributes to multidisciplinary research exploring the role played by paratext in the construction of identities. First, it develops a theoretical and methodological framework inspired by recent scholarship on conducting conversation analysis in digital contexts and on the ways in which speakers claim positions of expertise through their turns-at-talk. Next, it presents a case study examining translation-focused discussion forum comments posted within the so-called “Talk” pages that surround the mainspace content of the online user-generated encyclopedia Wikipedia. The analysis highlights the often intensely interactive nature of digital paratext and probes the implications for participants’ identity work: I argue that the polylogal nature of this paratextual space requires the constant negotiation of identities as Wikipedia contributors relentlessly jostle with one another for recognition of their epistemic authority. In the final section, I discuss the implications of this Wikipedia case for broader areas of scholarship on translation and digital paratext.

KEYWORDS:

• Paratext
• identity
• digital media
• Wikipedia
• translation
• conversation analysis

A growing body of research across multiple disciplines has highlighted the significant role played by paratext in the construction of identities. In the field of literary studies, for instance, Lori Ween’s (2003) analysis of ethnic American fiction has made productive use of Gérard Genette’s (1997) theory of paratext to demonstrate how various actors in the contemporary literary scene – including publishers, marketers and reviewers – use both epitextual and peritextual elements such as jacket art, quotations on the back jacket flap, marketing materials and reviews to create “marketable” identities for rising African American authors. In media studies, attention has been drawn to the ways in which paratext is involved in the identity work performed by film producers, directors, screen writers and audiences: Jonathan Gray (2010) has explored the unprecedented wealth of industry-generated paratext that accompanies the extended four-DVD release of The Lord of the Rings: The Two Towers, demonstrating how cast interviews and commentaries in particular serve to lionize Peter Jackson (the film saga’s director) not as yet another Hollywood blockbuster-maker but as a “real Author” of unique artistry, vision and skill. In translation studies, the focus so far has largely been on the construction of gender identities: Alison Martin (2011) has explored the paratext surrounding two botanical texts translated from French in early nineteenth-century Britain, examining how the female translators position themselves in their prefaces and footnotes as amateur botanists, as translators and as women at a time when British society was only just starting to acknowledge the legitimacy of women’s contributions to science. Sang-Bin Lee (2021), by contrast, has investigated the choice of translated title, the design of the target-language book cover and the translator’s preface in a more contemporary selection of overtly feminist literature rendered into Korean. Her analysis shows how the paratext here has served to construct or reveal the translators’ own identities as feminists as well as to express transnational solidarity with feminist authors and readers worldwide.

The case study presented in this article seeks to contribute further to our understanding of the relationship between paratext and identity construction through the investigation of a thread of translation-related discussion forum comments posted within the so-called “Talk” pages that surround the online user-generated encyclopedia Wikipedia. Of particular interest and significance here, I argue, is the “polylogal” nature of Talk pages as spaces set aside by Wikipedia’s community of volunteer contributors for paratext to emerge around the encyclopedia platform’s mainspace content (Giles et al. 2015; Gredel 2017). Like the prefaces, footnotes, book covers, DVD sleeves and film posters on which much paratext-oriented research has previously focused, Talk page comments perform a variety of functions in relation to the Wikipedia articles with which they are associated, including – as we will see in the analysis that follows below – justifying particular contributions to the core text, evaluating the revisions of other participants and thus shaping how the encyclopedia content is received (Ferschke, Gurevych, and Chebotar 2012). Unlike the monological paratext that has so far attracted the bulk of analysis, however, Talk pages typically contain traces of extensive interaction between Wikipedia’s multiple contributors. In this article, I seek to explore how the intensely interactive nature of these paratextual spaces shapes the construction of identities among Talk page contributors. By applying the tools of conversation analysis to examine a Talk page conversation observed between Wikipedians regarding the proper English translation of the Chinese medical term gua sha (刮痧), I show that participants use Talk page comments both to attempt to construct their own identities as credible experts and to contest the authority of others. The affordances of the digital medium, in other words, are shown to have created sites in which the right to be recognized as a legitimate contributor cannot merely be asserted, but must be intersubjectively negotiated through paratextual talk.

Applying conversation analysis tools to the study of digital paratext

To explore the negotiation of identities in Wikipedia’s paratextual Talk pages, this study makes use of the tools of conversation analysis (CA) (Maynard 2012). CA has enjoyed enormous popularity across the humanities and social sciences, most notably in the fields of linguistics (Kasper and Wagner 2014), sociology (Heritage and Stivers 2013) and interpreting studies (Bolden 2020; Gavioli 2015; Wadensjö 1998). It remains, however, underexploited in most areas of paratext-focused research as well as in translation studies more specifically; Xuemei Chen’s (2022) analysis of discussion forum interactions between Chinese translator Xiao Mao and his readers constitutes a rare exception to this rule. CA was first developed in the 1960s and, as Elizabeth Stokoe (2009, 81) writes, it aims to “examine how, through talking, people live their lives, build and maintain relationships and establish who they are to one another”. CA thus treats conversation not merely as talk about particular topics, nor simply as a vehicle for communicating information, but as a series of actions undertaken by participants in the course of an interaction (Hutchby and Wooffitt 2008). Conversation analysts explore the design and sequencing of utterances, examining how each turn in the conversation shapes the next and in this way identifying how talk is understood and received in its context.

The case for applying the CA approach to the investigation of online discussion forum data might superficially seem straightforward. The very nomenclature that has developed in relation to such web features (“Talk” pages, “discussion” forums) encourages us to draw parallels with the offline conversations that have been the traditional focus of CA. More substantially, messages posted in spaces such as Wikipedia Talk pages are typically “designedly interactional” (Meredith and Potter 2014), meaning they are written to elicit a response from their recipients, and utterances do exhibit many of the typical characteristics of informal oral communication (Gredel 2017). Even so, questions have been raised regarding the applicability of CA methods to the analysis of online discussions, particularly among researchers sceptical of the extent to which internet-mediated interactions can be considered conversations in anything more than a metaphorical sense (Giles et al. 2015).

Certainly, there are important differences between offline and online conversations. First, we must recognize that online exchanges are only “loosely coherent in comparison with the interactional norms of face-to-face conversation” (Herring 1999, 15). Posts may be designedly interactional, but may not necessarily elicit explicit responses: many opening posts go unanswered. Second, while traditional CA assumes that communication is linear “with each turn following another in a strict chronological sequence and within a relatively short time frame” (Giles et al. 2015, 48), online exchanges are considerably less ephemeral than oral communications, the digital medium allowing participants asynchronous access to previous utterances which can then be hyperlinked, quoted or referenced even many years after they were first posted. Third, online discussions are polylogal “with an unlimited number of interactants dropping in and out of the interaction, in some contexts rendering ‘membership’ almost irrelevant” (Giles et al. 2015, 49). As such, Talk page participants are not fully aware of who might receive their utterances; these may be directed at particular individuals but in the public space of Wikipedia, such utterances are always “doubly articulated” (Gredel 2017, 105), simultaneously addressed to any and every Talk page visitor. Finally, Wikipedia discussions largely proceed monomodally: the communicative potential of body language, facial expressions and intonation can only be approximated through punctuation, emoticons and emojis in this virtual environment, generating greater potential for miscommunication and misinterpretation. As we will see in the analysis below, the reduced set of communicative resources available to participants in this virtual environment also holds implications for how participants construct their own identities within the Wikipedia community through talk.

Despite these concerns, this study follows a growing body of research (Giles et al. 2015; Gredel 2017; Meredith and Potter 2014; Meredith, Giles, and Stommel 2021) in suggesting that, with some careful customization attuned to the unique affordances of the online medium, CA does offer a productive toolset with which to conduct analyses of digital interactions, including Wikipedia Talk page posts. As Meredith, Giles, and Stommel (2021) argue, research into many online platforms – including Wikipedia – has tended to pursue “big data” approaches, the ready availability of vast quantities of online content enticing scholars to conduct largely quantitative, wide-angle investigations. For example, David Laniado et al. (2012) have conducted a statistical analysis of the language used in conversations between Wikipedia editors across a dataset of over 870,000 Talk pages and 11 million individual comments, correlating the use of emotionally loaded vocabulary with users’ level of experience and gender to explore how different types of Wikipedian may interact differently within the site. In doing so, however, such macroanalyses have tended to “ignore interactional dynamics at the local level … losing sight of the trees in order to understand the wood” (Giles et al. 2015, 46). CA, by contrast, promotes the value and importance of the microanalysis of digital phenomena, focusing on turn-by-turn examinations of specific conversations (Meredith, Giles, and Stommel 2021). Moreover, unlike discourse analysis and other modes of qualitative (para)text-focused enquiry, CA seems especially suited to the plurivocal and sequential nature of Talk page discussions, given its focus on talk as interaction (Gredel 2017). It encourages us to treat Talk page contributions not as text but as actions, and thus places the emphasis of our analysis not merely on the content of individual turns but also on the “general communicative functions that are achieved through the turn-sequence” (Gavioli 2015, 75). As shown in the analysis below, this approach is especially productive as a means of better understanding the importance of talk for negotiations between Wikipedians seeking to assert and defend their “epistemic authority” over other contributors within the faceless environment of the encyclopedia platform (Heritage 2013). It provides a comprehensive toolset with which to make sense of the ways in which Talk page participants lay claim to relevant knowledge in their paratextual conversations as a means of constructing their own identities as credible experts within the virtual Wikipedia community of pseudonymous volunteers. The components of this toolset are outlined in the following section.

Territories of knowledge, epistemic status, epistemic stance

It is a truism that knowledge is unevenly distributed across society: some individuals and groups are more knowledgeable about certain topics, while other individuals and groups are less knowledgeable. In CA, this fact is accounted for through the concepts of “territories of knowledge” and “epistemic status” (Heritage 2012). When two speakers interact in conversation, there is typically an imbalance in what is known about the topic at hand with some element of knowledge falling more within one participant’s “territory” than it does in the other’s (Goodwin 1979). The speaker to whom the knowledge primarily belongs can then be described as having “higher epistemic status” within the conversation, while their interactant can be described as having “lower epistemic status” (Heritage 2013). The higher status speaker may initially enjoy an “absolute” epistemic advantage over the other if the latter starts out from a position of total ignorance, but it is important to recognize that the relative positioning of interactants along the epistemic gradient of possessing-lacking knowledge will normally change through the course of the interaction as participants exchange information. As previous research by Heritage (2012, 31) has shown, participants are “exquisitely sensitive to their epistemic positions relative to their addressees”, designing each of their turns-at-talk based on what they think the other person knows or does not know at that moment in time.

One’s epistemic status is not, however, simply a matter of knowing or not knowing. Rather, in conversation, this status is intersubjectively negotiated, ultimately involving the different parties’ collective awareness and acknowledgement of their comparative access to a particular domain of knowledge, as well as their differing rights and responsibilities to possess and share that knowledge (Heritage 2013). To give an example, a teacher may not always know the answer to a question asked by a student, even if the topic falls within their area of specialism; the teacher nevertheless occupies a position of superior epistemic status within the interaction because both speakers acknowledge the question sits more within the teacher’s epistemic territory. The teacher is expected by the student to know the answer and may feel an obligation to provide a constructive response despite lacking the requested knowledge. Clearly, then, epistemic status has as much to do with the construction and maintenance of personal and social identities as it has to do with the possession of knowledge (Jacknick and Avni 2017); it is inextricably intertwined with “being recognised as a certain kind of person, in a given context” as well as with participants’ own views about who they are relative to others (Gee 2000, 99). It is for this reason – illustrated in the analysis below – that, when challenged, interactants tend to police the boundaries of their epistemic territories and defend their epistemic status so vehemently. Epistemic territories are, in effect, “territories of the self” (Goffman 1971, 28).

Interactants in face-to-face communication make assumptions regarding the identities, epistemic statuses and knowledge territories of their addressees based on a wide range of visual and audible markers of social identity (Enfield 2006). We may recognize, for instance, a healthcare professional in a hospital setting at least partly by their uniform, this being the most efficient means of communicating their professional identity, institutional affiliation and possession of medical expertise. If we enter into conversation with a uniformed member of staff, assumptions made on the basis of this identity marker will inevitably influence how we are likely to receive any knowledge that our interactant imparts.¹ In online forums, however, the pseudonymous, virtual nature of interactions means that language – or more specifically conversation – becomes the primary mechanism by which participants establish their own epistemic status and assess that of others. Social traits such as gender, age, ethnicity, class and profession are much less visible in computer-mediated communication (Abd Rahman 2018). Consequently, it is mainly through their talk that users make claims to expert identities and garner recognition of their privileged access to particular areas of knowledge (Jacknick and Avni 2017). An individual Wikipedian’s epistemic status with respect to topics such as translation can thus be seen to emerge through their Talk page interactions rather than as a result of their peers’ prior awareness and recognition of their socio-cultural background, relevant experiences, professional qualifications and/or institutional accreditations.

The linguistic resources through which language users attempt to claim a particular epistemic status are theorized in CA with reference to the concept of “epistemic stance” (Heritage 2012, 2013). While epistemic status has to do with who knows what, as well as who has the right and responsibility to possess and articulate that knowledge, epistemic stance relates to the ways in which these relationships are actually expressed and encoded in the design of each conversational turn. For example, from the perspective of epistemics, there is a world of difference between saying “you’re wrong” and “I’m not sure that’s correct”: while the propositional content of the two statements is essentially the same, the first encodes a high degree of certainty and authority whereas the second acknowledges the possibility that the speaker could be mistaken. Exploring the interplay between interactants’ epistemic statuses and stances has been shown to provide a fruitful line of inquiry in previous studies (Gavioli 2015). As Heritage (2013, 377) notes, speakers do generally “act so as to preserve … consistency between the epistemic stance they encode in a turn-at-talk and the epistemic status they occupy relative to a topic” but it may sometimes occur that speakers might seek to appear more knowledgeable than they really are to convince others of their point of view.

“Gua means scrape”

The next two sections present the findings of a case study in which I make use of CA tools to investigate the construction and contestation of identities in Wikipedia’s paratextual discussion forums. The vast scale of the Wikipedia platform poses significant challenges for researchers working with more qualitative modes of enquiry and the process of identifying suitable passages of talk in amongst the many millions of online conversations taking place within the encyclopedia space is not straightforward (Jones 2022); it was necessary therefore to develop effective strategies for narrowing the scope of the potential dataset. Inspired by Sonya Pritzker’s (2014a) work examining offline translation-related discussion (dubbed “translation talk”) between American students and teachers of Traditional Chinese Medicine (TCM), it was judged reasonable to assume similar conversations might be found in the English Wikipedia’s Talk pages connected with this same topic. This assumption was quickly proved correct through perusal of Talk pages categorized within the encyclopedia as relevant to TCM (Wikipedia Category: Traditional Chinese Medicine): as in Pritzker’s (2014a) study, Wikipedians too can be seen to struggle with the challenges of communicating ideas and realities originating in a distant time and culture to English-speaking readers. Indeed, the data for this case study comes from a particularly interesting passage of translation talk revolving around the proper interpretation of the TCM term gua sha (刮痧).

Gua sha refers to a common practice in TCM involving the patient’s skin being firmly rubbed or scraped with a blunt tool, creating dark, bruise-like marks. According to its practitioners, this process can help to extract disease from the patient, particularly in the case of respiratory illnesses. The English Wikipedia article on gua sha (Wikipedia: Gua Sha) was created on 6 January 2004 but discussion of the English translation of the term gua sha does not appear on the associated Talk page (Wikipedia Talk: Gua Sha) until 2 November 2006. At 1:06,² User 1³ opens the paratextual discussion with the comment documented in Figure 1 and thus sets the agenda for the ensuing conversation.

Figure 1. User 1’s opening post.

As suggested by the use of the adjective “alternate”, this opening post is made in direct response to the main article text, which – since 8 June 2006 – had included “to scrape for cholera” as a gloss translation for gua sha (Figure 2). Given that User 1 had never previously contributed to this Wikipedia article or its Talk page, we can assume the motivation for posting their opening comment must have arisen as they engaged with the encyclopedia as a reader: on seeing a translation of gua sha that jarred with their understanding of the term (“extract sand”), they decided to raise the issue on the Talk page. This case thus illustrates: 1) the capacity of any Wikipedia reader to engage directly in discussion with others regarding the contents of the encyclopedia; and 2) the paratextual function of Talk pages already alluded to in the introduction. These discussion forums offer a space in which readers can comment on the encyclopedia text and thus influence how the text is viewed by other Talk page participants. More specifically, User 1’s aim with this intervention is to highlight a lack or even error in the existing content and to suggest an alternative gloss for gua sha might be included in the Wikipedia entry.

Figure 2. Opening paragraph of Wikipedia mainspace entry on “Gua Sha” as it was on 2 November 2006.

Examining this opening post with the tools of CA brings to our attention the low epistemic stance encoded in User 1’s comment. The phrase “I’m not Chinese but” immediately signals an attempt to position this topic towards the peripheries of User 1’s territory of knowledge. User 1 recognizes the importance of source-language proficiency for effective translation and seeks to establish explicitly that they are not claiming a significant epistemic advantage over other members of the Wikipedia editing community on the basis of some native speaker-level expertise with Chinese. As they explain, their only source of expertise regarding the translation problem is drawn from their broader subject knowledge as a “student of Oriental Medicine”. Further hedging strategies include the use of the verb phrase “I’m guessing” which signals that this knowledge claim is not built on firm epistemic foundations but rather on conjecture. The two ellipses (“ … ”) similarly indicate an absence of certainty: this punctuation is commonly used in written digital communication to signal unassertiveness and can therefore be taken as a further marker of low epistemic status (Vandergriff 2013). Finally, we might consider relevant the fact that User 1 has chosen to post a comment in the paratextual space of the Talk page rather than inserting their proposed translation directly in the mainspace of the encyclopedia itself as Wikipedians are otherwise encouraged to do (Wikipedia: Be Bold): although this opening post is not formulated as a question,⁴ its very placement in the discussion forum signals lower confidence and a desire on the part of this user to have this knowledge claim confirmed by other Wikipedians of higher epistemic status.

By contrast, User 2’s response (Figure 3), posted one hour and 51 min later at 02:57, adopts a considerably higher epistemic stance. User 2 begins by reaffirming User 1’s lack of linguistic expertise (“you say you are not a student of Chinese”) and continues by making explicit the significance of this deficiency: “[you] have no idea of just how many characters can share the same, identical pronunciiation [sic]”. These comments are then followed by a point-by-point rebuttal of User 1’s suggestion that gua sha might be translated as “extract sand”. Here, there is no hedging or uncertainty in the explanation, only outright refutations and self-confident assertions: “Sha in Gua Sha does not mean sand”, “Gua means ‘scrape’” and “the technical term can mean nothing more or less than (literally) ‘to scrape for cholera’”. Indeed, such is this user’s desire to highlight the extent of their own linguistic and domain-specific knowledge that they appear to violate Grice’s (1989) conversational maxims of quantity and relevance, including an excess of information in their response, some of which is not directly relevant to the question at hand (e.g. “the Qing is the same character as in the acupuncture point BL.1.”):

Figure 3. User 2’s response to the opening post.

Clear explanations of the rationale lying behind certain decisions are of course essential to the effective coordination of collaborative encyclopedia building (Ferschke, Gurevych, and Chebotar 2012). However, the rather bullish tone and excessive length of User 2’s comment might be explained by three factors: first, consulting the Wikipedia entry’s Revision History shows that it was in fact User 2 who originally added the translation of gua sha as “to scrape for cholera” to the encyclopedia text in June 2006. On this basis, we can understand that User 2 might view User 1’s opening post as a challenge to their own self-image as a knowledgeable expert. By suggesting that gua sha might instead be translated as “extract sand”, User 1 effectively threatens User 2’s claim of ownership over this area of knowledge and their self-assigned right to contribute to the Wikipedia article text with valid information about the TCM term gua sha. Second, it is important to remember that online conversational turns are typically “doubly articulated” (Gredel 2017): while the prevalence of second person pronouns would suggest this comment is directed at User 1, we should not forget that this conversation is taking place in the relatively public space of Wikipedia and that it can be read by any and all visitors to the Talk page. User 2’s comment thus not only functions to establish User 2’s epistemic advantage over User 1, but also serves to position User 2 as the primary authority on the translation of gua sha within the wider Wikipedia editing community. By pointing out User 1’s errors of misunderstanding in such detail, User 2 “simultaneously, albeit subliminally, projects knowledge, identity and power, conferring distinction” on themself (Heritage 2013, 372). It is through this translation talk that User 2 is able to construct their own identity as a valuable member of this virtual community of otherwise unaffiliated strangers; although they do not explicitly present themselves as “an expert translator”, they do clearly want to be seen as someone who enjoys privileged access to relevant sources in the original Chinese language, and who can identify suitable translations in the target language, rather than as someone who has to rely on the “distorted” translations of others. Third, it is helpful to consider User 2’s comment from the perspective of what Dan Sperber et al. (2010) term “epistemic vigilance”. In CA, epistemic vigilance can be defined as the need for everyone to “track what others know and the reliability of their claims … as a means of determining whether and how to act on what is being asserted” (Heritage 2013, 392). In this light, the purpose of User 2’s comment is to reduce the risk that User 1 might go ahead and act on what they think they know by including mention of an “alternate translation” in the main article text; perhaps more importantly, User 2 also seeks to make sure that other contributors reading the Talk page will similarly not support the view that the main article text is in some way deficient and erroneously attempt to change its presentation of the translation of gua sha.

User 1 does not respond immediately so, 21 h later, User 2 takes the opportunity to further develop recognition of their own epistemic status in a second post published at 23:38 (Figure 4):

Figure 4. User 2’s follow-up post.

The manifest aim of this comment is “[t]o make the answer above a little clearer” and to make User 2’s argument “a little more easily understood”. In doing so, User 2 not only helps coordinate the community’s translation efforts, but again seeks to emphasize their epistemic advantage over User 1, inferring that – given the latter’s unfamiliarity with the topic – they might need additional help from User 2 to make sense of the Chinese term. Moreover, by recasting the discussion using more technical, academic terminology (“citation meaning”, “context meaning”), User 2 seeks to further demonstrate the extent to which this topic lies substantially within the boundaries of their epistemic territory and thus to consolidate their identity as a credible translator and their ability to speak from a position of authority. By deploying this language, User 2 foregrounds the depth of their understanding of Chinese and Chinese-English translation, and seeks to bolster their fellow contributors’ recognition of their right to contribute to the encyclopedia’s English-language presentation of gua sha.

18 hours later, at 18:06 on 3 November, User 1 finally responds with the comment reproduced in Figure 5. When viewed from the perspective of epistemics, this post is interesting for a number of reasons. To begin with, there is the hint of sarcasm implied in the opening sentence (“Thank you for the illuminating discussion”). Sarcastic intent is notoriously difficult to recognize in written communication given the lack of orally conventional cues such as tone of voice (Kreuz 1996; Thompson and Filik 2016). Nevertheless, considering this opening statement in the context of the remainder of User 1’s post suggests that User 1 is not seeking here to express genuine thanks towards User 2, nor do they sincerely view the preceding discussion as “illuminating”. Instead, it seems more likely that User 1’s post should be read as an attempt to undermine the epistemic hierarchy previously established by User 2, claim back some degree of authority and defend their right to participate in improving Wikipedia’s coverage of gua sha and TCM more broadly. This reading of User 1’s intentions is supported by the fact that they go on to directly contest User 2’s characterization of User 1 as someone without any relevant linguistic knowledge (“And yes I am aware … ”), highlighting instead their prior familiarity with the “difficulty of standardizing translations of Chinese Medical terms” and “the prevalence of homophones in the language”. In doing so, they attempt to push back the boundaries of their individual area of expertise, relocating them from where they had been set by User 2 and reincorporating this knowledge as part of their claimed territory. Through multiple references to their ongoing education, to their “professors who are native speakers” and to key contributions to Anglophone scholarship on TCM (“Flaws, Wiseman, etc.”), User 1 further seeks to accrue social capital and distinction through association with these high-status sources of authority. Finally, User 1 adopts a higher epistemic stance in this post by blaming the inadequate translation “extract sand” not on their own knowledge deficiencies but on “a source offered to me in the context of my education”: User 1 thus effectively distances themselves from their original knowledge claim, attributing fault with the textbook and perhaps even with the teacher who recommended this source.

Figure 5. User 1’s response.

What comes through most strongly in the analysis of this comment, however, is the extent to which it demonstrates how acutely User 1 felt the attempt to wholly dismantle their epistemic authority in the public space of the Talk page discussion forum. User 1 acknowledges that they might have drawn less heated criticism had they adopted a lower epistemic stance in their opening post and framed their comment instead “in question form”, but they nevertheless seem to feel they have been unfairly treated. User 1 accuses User 2 of “unwarranted vindictiveness” in their attack on the initial knowledge claim and, in the final sentence, explicitly frames User 2’s response as an “insult”. This reaction is not surprising when we consider the extent to which the epistemic status that a person enjoys within a community is inextricably intertwined with identity and power relations. User 2’s outright rejection of User 1’s claims undermines the latter’s credibility as a valued member of the Wikipedia-editing community; it has the effect of placing them in a subordinate position of power and influence over the ongoing collaborative construction of the encyclopedia text.

“The verb ‘Gua’ does not mean ‘to scrape for’”

With the phrase “In the future … ”, User 1 effectively closes down the conversation. By indicating how they intend to frame their comments in Talk page activity elsewhere, User 1 makes clear they wish to pursue the present matter no further, and they are not expecting any further action or interaction from User 2 or any other co-participant. Nevertheless, this discussion thread receives four additional contributions over the following two years to November 2008 when the thread does finally go cold. The first (Figure 6) is posted by User 3 on 20 December 2006, seven weeks after User 1’s opening comment.⁶

Figure 6. User 3’s post.

At first glance, it is difficult to identify what might have motivated User 3 to post this comment. The information that User 3 provides can be seen to complement and clarify User 2’s earlier explanation to some degree, and therefore help coordinate the community’s translation work by ratifying the existing solution. But the discussion thread has been inactive for well over a month, and neither of the original two participants appear to think that further consideration of this topic is necessary. The comment does not shed light on any new aspects of the translation problem, nor does it raise any further questions or uncertainties. Moreover, to date, no contributor has ever commented on or otherwise responded to User 3’s intervention, either in the Talk page or in the main body of the Wikipedia article itself; even User 3 has not made any edits to the encyclopedia entry that can be connected directly to this paratextual comment. Attention to the epistemic stance encoded in this talk-as-action, however, suggests a possible rationale for this post: User 3 might see this discussion thread at least in part as an opportunity to garner greater epistemic authority within the community as a credible contributor-translator. Through reference to a dictionary entry whose interpretation requires knowledge of Chinese, along with discussion of the specific graphical components (“the sickness radical”) of the Chinese characters, User 3 demonstrates their familiarity with this language and consequently their privileged access to reliable sources relevant to the topic at hand. Like User 2 and – to a lesser extent – User 1, User 3 seeks to claim this topic within their own territory of knowledge, to make known their relevant expertise in this area, to distinguish themself from their peers and to assert their entitlement to evaluate possible translations of TCM terms.

This attempt to use translation talk in order to achieve wider recognition of higher epistemic status seems to be a common feature of all the posts which appear in this paratextual discussion thread after November 2006. Figures 7–9 present the three comments published by one further contributor, User 4, two years later in November 2008.⁷

Figure 7. User 4’s first post.

Figure 8. User 4’s second post.

Figure 9. User 4’s third post.

As with User 3’s comment above, it is initially not easy to understand why these posts have been made. In offline, face-to-face conversation, a delay of two years between turns is unthinkable; it is only the affordances of the digital medium that make possible the extension of the interaction over such a long period of time (Giles et al. 2015). Even so, nothing of relevance to the topic of this exchange appears to have changed in the main Wikipedia article to provoke renewed interest in November 2008 in the question of how gua sha might be rendered in English. User 4’s second comment (Figure 8) seems particularly unwarranted, apparently having little bearing on the specific translation problem at hand and being only loosely connected to a point made earlier by User 2 (Figure 3). Recourse to the theoretical framework of conversation epistemics again proves helpful, therefore, in suggesting that each of these posts can be viewed as part of an attempt to usurp User 2’s position of epistemic superiority within the community of Wikipedians focused on improving the encyclopedia’s coverage of TCM. This challenge is perhaps most explicit in Figure 9 in which User 4 self-confidently dismisses User 2’s original claim that gua can be understood as “to scrape for”; but the challenge to User 2’s epistemic status as primary authority on this topic within the community is also implied in the proposal of another solution for the translation of sha in Figure 7 (“‘Sha’ in this case simply means a ‘disease’ that is unseen but that could be felt through its effect”). It is in large part the use of the adverb “simply” which encapsulates the strong epistemic stance adopted here: User 4 frames this translation activity as untaxing, highlighting their own mastery of the skills and knowledge required to interpret the source term for anglophone readers. Moreover, by insinuating that User 2 may have misunderstood such a “simple” source-language term, User 4 hints at a perceived discrepancy between User 2’s performed epistemic stance and their actual epistemic status: the suggestion is that User 2’s possession of this area of knowledge may not in fact be as total as their two comments might have led others previously to believe.

With their second comment posted 22 min later (Figure 8), User 4 then seeks to further establish the extent of their own territory of knowledge and to promote their identity as expert in this field. Focused on two further TCM terms, qing-re and re-qi, this post constructs an image of this Wikipedian as someone concerned with broader questions targeted far beyond the original scope of the discussion and associated with the relationship between language, translation, public health and education. This identity is subtly reinforced by the first-person plural pronoun “us” in the first line of the turn, which includes User 4 as a member of a community of experts positioned as distinct from and epistemically superior to wider publics of lay TCM users, and charged with modernizing the language of Chinese medical practice.

In the absence of any opposition from their peers, User 4 implements a change to the first sentence of the mainspace gua sha entry on 24 December 2008, as documented in Figure 10.

Figure 10. User 4’s revised version of the opening paragraph of the Wikipedia article “Gua Sha”, published 24 December 2008.

The paratextual translation talk can thus be seen to directly shape the evolving Wikipedia text: this revision not only incorporates User 4’s suggested translation, discarding User 2’s rendering of gua sha as “to scrape for cholera” in favour of “scrape away disease”; but it also paraphrases User 1’s original suggestion that this TCM term can be understood as “extract sand” (Figure 1).

Discussion and conclusions: implications for studies of translation and digital paratext

That the debate over the meaning of these Chinese characters should have come full circle is perhaps unsurprising given the plurality of translations of TCM terms that circulate more widely within anglophone texts on this topic: as Pritzker (2014b) has explored, in the absence of any single organization with the authority to oversee the globalization of Chinese medical practice and the standardization of its English-language vocabulary, discussions as to which translations are most appropriate extend far beyond the Wikipedia space. Nevertheless, as this analysis has demonstrated, the study of Wikipedia paratext offers a valuable site in which to observe such metalinguistic discussions and to better understand how individuals interactively establish their own identities as TCM experts with the authority to promote certain interpretations over others.

Looking more broadly, the article has made a number of contributions to wider fields of research on translation and digital paratext. First, the analysis promotes translation talk as an important yet so far under-explored site of paratext worth examining across a broader range of online and offline contexts. While research in this area has tended to focus on monological forms of paratext, a focus on translation talk brings to our attention the ways in which texts may often be framed and identities constructed through the paratextual interactions of multiple participants. Conversation analysis – and Heritage’s (2013) work on epistemics in particular – has proven a valuable framework through which to investigate multi-actor translation-related discussions in Wikipedia, and future work could usefully build on this approach to research similarly polylogal phenomena occurring, for example, on social media (e.g. Twitter threads, Facebook discussions, YouTube comments), or in more traditional media settings (e.g. interviews published in newspapers and literary magazines).

Second, the study lends further weight to the view that in order to serve as a productive tool in investigations of digital culture, the concept of paratext must be expanded beyond Genette’s overly narrow theorization, especially with regard to Genette’s (1997, 407) insistence that “the main issue for the paratext is … to ensure for the text a destiny consistent with the author’s purpose”. However, the Wikipedia case does not merely suggest that it is important in the context of digital culture to consider the influence of other forms of paratext produced, for example, by fans and consumers, as other translation and media scholars have argued (Batchelor 2018; Desrochers and Apollon 2014; McCracken 2013). Instead, it also shows that Genette’s assumptions around authorial intention are largely meaningless in some new media contexts given the fluid characteristics of digital authorship itself. As we have seen, the question of who might constitute “the author” of a Wikipedia text is problematic due to the collective nature of Wikipedia editing (Kennedy 2016): Wikipedia articles do not have a single author but an endless stream of readers-turned-contributors, authorized by the encyclopedia platform to create and edit its pages. Moreover, the above analysis shows that even those readers-turned-contributors who do not impose specific changes on the Wikipedia article text itself may still substantially influence the ongoing evolution of that text through their paratextual conversations. No one individual takes responsibility for this collaborative activity and each Wikipedian may have a different set of motivations and agendas driving their production of text and paratext; there is certainly no single and uncontested authorial intention which might be promoted and preserved through paratext.

Finally, the case study presented here challenges existing theorizations of the addressee of paratext, again suggesting a need to rethink Genette’s framework in the context of digital culture. While paratext theory has traditionally assumed paratext is created for either public or private consumption (Genette 1997, 9), the intended addressee of Wikipedia paratext is evidently not the general public, nor is it even general Wikipedia readers; at the same time, the analysis presented here suggests that creators of Talk page comments are aware of the fact that their posts are potentially accessible to any and every Wikipedia user, meaning these cannot be categorized as “intimate” paratext (Genette 1997, 9) intended for only a closed group of recipients, even when a comment appears superficially to be directed at a specific individual (as in Figure 3 above). Instead, in order to better understand the functions and forms of digital paratext, our theories need to recognize that digital communication may often be “publically private” (Lange 2007), freely available but not widely accessed, doubly articulated towards known and unknown addressees.

Acknowledgements

The author would like to thank Kathryn Batchelor, Chiara Bucaria, Jan Buts, Maeve Olohan and the two anonymous peer reviewers for their detailed and constructive feedback on earlier drafts of this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

All data analyzed in this article can be accessed directly on the Wikipedia platform at the following links: “Gua Sha” Wikipedia article: https://en.wikipedia.org/wiki/Gua_sha; “Gua Sha” Talk page: https://en.wikipedia.org/wiki/Talk:Gua_sha.

Additional information

Funding

This work was supported by the UK’s Arts and Humanities Research Council under Grant AH/V013203/1.

Notes on contributors

Henry Jones

Henry Jones is Lecturer in Translation Studies at the University of Manchester, UK. He is a co-coordinator of the Genealogies of Knowledge Research Network and co-editor of the Routledge Encyclopedia of Citizen Media (2021). He is also Principal Investigator on the WikiAltMed project, funded by the UK’s Arts and Humanities Research Council (2021–2023). His work has been published in international journals such as Translation Studies, Target, Globalizations and Humanities and Social Sciences Communications, and in edited volumes including the Routledge Encyclopedia of Translation Studies (2020) and the Routledge Handbook of Translation and Media (2022). His current research interests lie in the medical and health humanities, media theory, digital culture and corpus-based methodologies.

Notes

1 There are interesting parallels here with Genette’s (1997) suggestion that factual information can have paratextual value: Genette argued that knowing the gender of an author, for example, may often influence how we receive their text. Similarly, a professional’s uniform can be seen to function paratextually as a “consciously crafted threshold” (Batchelor 2018, 142) shaping how their interactants in conversation (patients or colleagues) interpret and design turns-at-talk.

2 Wikipedia uses Coordinated Universal Time (UTC).

3 Wikipedians contribute pseudonymously but, to add an additional layer of anonymity, I have replaced all usernames with numbers.

4 As Gavioli (2015) notes, asking a question typically signals lower epistemic status and an attempt to attribute the interlocutor with higher epistemic status.

5 When User 1 published this post, they were not logged in to their Wikipedia account. This means this post is attributed to an anonymous IP address instead of User 1’s username. From the content of the post it is clear, however, that the anonymous IP editor and User 1 are the same person.

6 Rather than placing this comment sequentially below the latest comment in the thread as is conventional in Wikipedia, User 3 inserts this information in the middle of User 2’s first post, between entries 1 and 2 in the numbered list. This unusual placement would suggest that User 3 is seeking to respond not to User 1’s opening post directly, but to the specific passage of User 2’s reply in which the meaning of the Chinese character 痧 (sha) is discussed.

7 Unlike User 1, User 2 and User 3, User 4 is a so-called “IP editor”, meaning they have made these contributions anonymously, without registering a user account and logging into the Wikipedia platform; as such, the creator of all three of the posts reproduced in Figures 7–9 is recorded only by their IP address. The IP address for Figure 9 is different to that in Figures 7–8 but it can reasonably be assumed these are the same individual based on the patterning of their contributions.