The naturalness of Chinese online chatting: Organization and recontextualization

1. Introduction

In the field of pragmatic research, conversational data used for analysis can be authentic or naturally occurring (e.g., those coming from daily face-to-face communication, telephone communication, radio-mediated communication, TV-broadcast face-to-face communication), elicited (such as those collected via the use of questionnaire or discourse complete test), or representational/fictional (such as characters' conversation in novels, plays, or TV series) (Chen, 2019). In the first case, the data reflects what people actually say and how they actually say it in real life; in the second case, the data only stand for what people might or probably say in the projected or imagined situations; in the third case, however, the data only represent what people in the fictional world are designed to say, and how they are designed to say it. Practically, in pragmatics literature, there seems to be an explicit or implicit preference for the use of naturally occurring data over non-natural data (i.e., elicited or representational data) (Chen, 2019), although literary data have their plausible parallel in the real world.

While often explicitly claiming their data to be “naturally occurring,” scholars (e.g., Yu and Li, 2009; Hyeri Kim, 2017; Yu and Drew, 2017) may sometimes take them for granted, without offering or citing any definition of the term. The assumption becomes problematic when we assess the naturalness of the data from non-oral online chatting¹, such as WeChat typed talk (WTT) and QQ typed talk (QTT) among Chinese. Apparently, the form of talk sometimes lacks spontaneity and synchrony, two essential features of face-to-face naturally occurring conversation. For this reason, can we simply disqualify online chatting as “non-naturally occurring,” given that traditional written communication is basically neither spontaneous nor synchronous, either? Alternatively, can we argue that it is still naturally occurring in its own right, given that it represents the natural and genuine way WeChat users communicate with features of both face-to-face conversation and non-face-to-face written communication? This study favors the latter position and seeks to illustrate the “naturalness” of online chatting such as WTT from two perspectives, namely its organization and recontextualization. We incline to argue that while Conversation Analysis (CA) was designed for spoken conversation, it is not inadequate for analyzing WTT interaction as a form of performing social actions, particularly when it takes on some adaptations (as suggested in digital CA). It is hoped that this study may contribute to the understanding of online chatting on the one hand and provide methodological implications for doing online chat research via digital conversation analysis (CA) on the other.

2. Research background

2.1. Defining “naturally occurring”

Loosely speaking, data used in pragmatics research that are not contrived or elicited (as in simulation contexts, such as those collected by means of questionnaire) can be called natural or naturally occurring. However, in CA, what can be called natural/naturally-occurring data generally need to possess at least the following array of defining features, by the standard of the “primordial kind of human verbal communication” (Gruber, 2019, p. 59), i.e., oral, face-to-face conversation:

– Oral (involving paralinguistic features such as pitch, intonation, and voice characteristics).

– Spontaneous (in terms of production).

– Synchronous and immediate (in terms of response).

– Sequentially linear (overlaps kept at a minimum).

– Backchanneled (involving various kinds of immediate mutual feedback).²

– Multimodal (involving the use of non-verbal aspects such as body movements, facial expressions, gestures, etc.)

– Co-presence of communicators in real-life world (Gruber, 2019, p. 59).

With the advent of communication tools such as telephone, radio, Skype, WeChat and QQ, some “affordances” (Hutchby, 2001) of oral face-to-face communication become inactive. For example, in non-visual, non-prepared, non-edited telephone talk, radio-mediated talk, Skype oral talk, WeChat oral talk, and QQ oral talk,³ body movements, facial expressions and gestures are not visible to one's interlocutor(s) and thus cease to function. Also, the communicators are no longer co-present, such that the turn-taking system may be partly disrupted in the absence of mutual monitoring of speaking. Nevertheless, non-face-to-face, non-prepared telephone talk is still generally assumed to be naturally occurring thanks to its remaining features, especially orality, synchrony, and spontaneity. In existing related research adopting CA, the data are often taken from telephone calls and radio-meditated calls, to the neglect of calls supported by Skype, WeChat, and QQ.⁴

By contrast, conversational data from sources such as TV series and novels are generally considered non-natural or not natural enough, although the characters as communicators involved are co-present and the talk is oral, synchronous, immediate, sequential, multi-modal, and backchanneled. The claimed non-naturalness of the kind of conversational data arises from the fact that they are not spontaneous but prepared, edited, and rehearsed. Moreover, conversational data from novels are not oral or multimodal, although the writers may describe some multimodal features such as manners of speaking and emotion as a complement.

Then, how about quasi-conversational forms of Internet-mediated communication such as WTT or QTT?⁵ Although they are not oral, they seem to be a lot more natural than fictional data because they are very much spontaneous and very often synchronous, as we shall demonstrate in section 4.

2.2. Research on features and affordances of online chatting

With the development of Internet pragmatics or cyberpragmatics (Yus, 2011), a growing number of scholars show interest in the study of Internet communication such as online chatting, focusing on the effect of its new affordances.

Previous scholars are primarily concerned with the differences between online chatting and face-to-face conversation. For example, Shang (2001) finds that unlike real-world face-to-face chatting, online chatting does not afford the possibility of using body language, facial expressions, and the like. Jiang (2006) notes that online chatting has integrated and subverted traditional/ordinary spoken and written language, as it involves limited contextual constraints, hybridity of style, novel and even bizarre expressions, and volatile and unstable language. Adopting CA, Xu and Li (2012) find that the turn-taking features of online chatting in the virtual world are different from those of face-to-face conversation (e.g., one speaker often starts to type and send an utterance before waiting for the other to finish his or her turn, thus resulting in more “chaotic” turn-taking). Dynel (2014) discovers a distinctive participatory framework underlying multi-party interaction on YouTube, a type of asynchronous computer-mediated interaction, involving changes of the users' participatory statuses at both the production and reception ends. Sandel et al. (2018) point out that WeChat interaction platform affords users new methods for creating messages and online interaction. Wang et al. (2016) study the language exchange via WeChat between students in Australia and the U.S. They claim that WeChat affords a “semi-synchronous” interaction, a term they used to describe the speed of the interaction as falling between synchronous and asynchronous. They demonstrate some of the affordances of this technology, including the ability of users to send messages that use text, audio files, photos, and/or emoticons and stickers. They present examples of data, including screenshots of messages, photos, stickers, and audio messages.

By comparison, very few scholars have addressed the commonality between online chatting and oral, face-to-face communication. As an exception, Sun and He (2006) compare face-to-face English conversation with conversation in online chatroom in terms of turn taking. They find that despite some new and distinctive features (e.g., the initiation of talk with “hello” or “Can i talk to u?” and more frequent occurrence of overlaps), online conversation is not fundamentally different from traditional oral communication.

It is worthy of note that online chatting apps are evolving all the time, making available not only some affordances of oral, face-to-face communication but also new possibilities of meaning making, as we shall see in section 4. For example, the use of audio message, which Shang (2001) mentions as lacking, is now partly afforded in WeChat.

2.3. Use of CA to analyze online chatting

Methodologically, the (non-)naturalness of the conversational data may determine whether they are amenable to “pure” or “real” conversation analysis (CA) (Giles, etc., 2015), which supposedly works on naturally occurring conversational data only (Yu and Li, 2009; Wu and Yu, 2017; Yu and Wu, 2020). Given that online chats largely lack some important features of natural conversation such as orality and spontaneity but involve unconventional features such as occasional sequential disruption (Herring, 1999) and asynchrony (Yus, 2011), it has been cautioned that they are not suitable for the use of CA approach. As Giles et al. (2015) note:

We were all, to a greater or lesser extent, successfully using conversation analysis (CA) to conduct analyses of online data and publishing this work; yet, we had all encountered some skepticism from journal reviewers and others, to the effect that what we were doing was not “real” CA. Moreover, none of us was aware of any systematic attempt to develop a methodology for the microanalysis of conversation-like data encountered so frequently in online environments, most commonly in the form of asynchronous discussion forum threads, message feeds in social media, or Gmail chat” (Giles et al., 2015, p. 45).

Despite some doubt about its suitability, recent years have witnessed the growth in the use of conversation analysis (CA) as a method for analyzing interactional practices in online communication (e.g., Gonzalez-Lloret, 2011; Giles et al., 2015; for a comprehensive review of related literature, refer to Paulus et al., 2016). After exploring the participants' methods of interaction, Housley et al. (2017) pay more attention to the primacy of the online interaction phenomenon and demonstrate how useful CA and MCA (multimodal CA) are in exploring these methods in Twitter posts.

To accommodate the special features of online interactions, Meredith (2017) suggests combining CA and the concept of affordances (Hutchby, 2001) to enable analysis of not only the online interaction, but also the effect of its technological context. On the one hand, by analyzing turn adjacency and openings of online interaction, two key interactional practices in light of CA, we can derive an in-depth understanding of it. On the other hand, depending on technological affordances, we get a clear idea of its distinctiveness. Meredith (2017) argues that this kind of so-called “digital CA” (DCA) as opposed to “spoken CA” is a kind of “refracted” CA due to the lens of technological affordances, which can develop a better understanding of interactional patterns across different online platforms. Indeed, some researchers have attempted to adapt CA into DCA in order to analyze online conversation. For example, Meredith (2017) has developed a transcription system for the screen-capture data on the basis of the Jefferson system (Meredith, 2016).

By and large, there is still much to do with the development of DCA with regard to various methodological questions such as to what extent spoken CA findings apply to the analysis of online interaction (Meredith, 2017). On the conceptual side, however, an even more significant question is whether we should confine our attention to them at all, especially as online interaction platforms are evolving so fast. We should not only look for different or new manifestations of spoken CA features, but also embrace new, natural mechanisms of digital conversational organization and recontextualization. The latter, while discussed in discourse analysis (DA) (e.g., van Leeuwen, 2008; Culpeper and Haugh, 2014; Gruber, 2019), has been underexplored in CA. In this direction, the present study on WTT, which has been underexplored, serves as an illustration. Specifically, it will address the following research questions:

a. How is WTT organized for performing social actions in terms of turn taking, adjacency, repairing, etc.?

b. How is recontextualization realized in WTT?

3. Data sources and analysis

For convenience sake and from an emic perspective, all the data of this study were excerpted from the recent WeChat interaction that all involved the first author (10 one-on-one private chats and 15 group chats) but only occasionally the second author (5 group chats). The chats varied in length of turns (from 4 to 16). Of the 25 chats, 18 were more or less synchronous and continuous, whereas the remaining 7 chats involved long breaks that significantly affected the continuity of the chatting. Also, 23 of the chats were interpersonally oriented, whereas 2 chats related to work.

Apart from the first author (and the second author), other people involved in the WeChat groups included family members, colleagues, students, or friends.⁶ The version of WeChat app used was the latest one at the time of writing this paper (i.e., 8.0.22) for mobile phone and for computer (3.6.0.1000), both launched in April, 2022. WTT⁷ varies in its composition of modalities. It can be purely typed language; it also can be a mixture of typed language and emojis, or a mixture of typed words and voice message(s), or a mixture of typed language, emojis and voice message(s). This study only concerned typed talk with or without emojis.

To answer the two research questions, the first author looked into how WTT interactants organized and recontextualized information, bearing in mind the notions of CA (such as turn taking, adjacency pairing, and repairing) and recontextualization (notably features of intertextuality such as quoting). In recognizing spontaneity and continuity of turn taking, we depended on the technological 5-min threshold of WeChat beyond which breaks will be automatically marked by a new time message. Meanwhile, he paid close attention to finding whether any new manners of organization and recontextualization have emerged or occurred. As no quantitative analyses would be attempted, he just detected varied instances of WTT organization and recontextualization. He would take screenshots whenever he ran across relevant cases such as spontaneity, (a) synchrony and mechanism of turn-taking, as well as modes of recontextualization.

In the presentation of the data for analysis and discussion below, the authors removed personal data to comply with ethical requirements. The data were not transcribed with a strict CA format (because features such as pauses, pitch, and overlaps were not recognizable) but presented in the same manner and sequence as they actually showed on the screen, with emojis retained where they were, in a way very much similar to the use of screenshots neatly practiced by Sandel et al. (2018). Yet the dialogues on the screen were converted into the format of exchange we see in regularly transcribed conversation.

4. Findings

This section presents the findings in relation to the two research questions.

4.1. The organizational features of WTT

Very often, WTT as found in our data is a synchronous, jointly and continuously completed chat. This occurs especially when the communicators are performing a particular task. Asynchronous WTT occurs often when one or two sides are occupied or distracted by some other task(s). Occasionally, an asynchronous response is ascribable to a negative attitude or a reluctance to reply immediately. Yet, sometimes, it is hard to tell which. For this reason, WTT, like Facebook chat, is quasi-synchronous (Garcia and Jacobs, 1999). In other words, for WTT, sometimes, “synchronicity is not a fixed feature but rather is something that participants negotiate through the course of the interaction.” (Meredith, 2017, p. 53). Most explicitly, the synchrony is technically marked by the signaling of “对方正在输入…” (XX is inputting …) on the top line of the screen, as demonstrated in Ex. 1:

Ex. 1

1 陈:  忙啥呢?

2 吴:  你猜呢?

3 陈:  我猜你在写文章吧?

4 吴:

5 陈:  你说今天能完成的哦

6 吴:  呃

7 陈: ? ?

        [对方正在输入…]

8 吴:  好吧

1 Chen: What are you busy with?

2 Wu: Guess?

3 Chen: You're writing the paper, I guess?

4 Wu:

5 Chen: You said you would finish it today

6 Wu: Uh

7 Chen: ? ?

    [Wu is inputting …]

8 Wu: Oaky.

In Ex. 1, after Chen shows puzzlement (line 7) over Wu's use of a hesitation marker “uh” (line 6), Wu is shown to be working out a response, technologically indicated by “Wu is inputting ….” The signal does not occur anytime when somebody is offering a response. Rather, it occurs only when the recipient responds by beginning to type in the input space within 10 s after receiving the speaker's message. It immediately disappears after the message being typed is sent out. The affordance of the app can be shut down manually, though.

In WeChat platform, a response is technologically recognized as synchronous as long as it follows a previous message within 5 min. Thus, the notion of synchrony is very much stretched, although responses often are almost equally immediate. Most often, WTT is sequentially well-organized in the turn-taking format, very much similar to face-to-face conversation is, as shown in Ex. 2:

Ex. 2

1 陈:  你好！周五下午有课吗?

2 李:  老师您好！没有课的,  一定来听讲座。

3 陈:  想请你去南京农业大学翰苑宾馆接下黄老师,  下午2点到那儿接。

4 李:  好的老师!

5 陈: 136xxxx0530

6 陈:   他的电话

7 李:   好的,  谢谢老师！我和黄老师联系。

1 Chen: Hi, do you have class this Friday afternoon?

2 Li: Hi, professor. No, I'll have no class. I'll come to attend the lecture!

3 Chen: I'd like you to pick up Professor Huang from Hanyuan Hotel of Nanjing University of Agriculture. Two o'clock.

4 Li: Alright, professor.

5 Chen: 136xxxx0530

6 Chen: His phone number

7 Li: Ok. Thank you, professor. I'll contact him.

In Ex. 2, Chen and Li speak alternately in the first four lines, resulting in a joint construction of the conversation. For both the fifth and sixth lines, Chen apparently occupies two turns. He could have typed the two messages in one single turn. In reality, he sends the two messages one after another. This has happened perhaps as an afterthought or an increment because he wants to clarify that the phone number is that of Professor Huang, which partly evidences the spontaneity of WTT. Yet, more often than not, sending a series of broken or fragmental messages in continuous turns rather than a complete message in one single turn occurs in very relaxing chatting between familiar or intimate interlocutors (it can also be an idiosyncratic practice, though), as shown in Ex. 3:

Ex.3

1 妈妈:@Peter 你的胃好些了吗

2 儿子: 吃了药以后连续两天不疼了

3 儿子: 很有效果

4 妈妈: 养胃是一生的事

5 妈妈: 咖啡、可乐等不能多喝

6 妈妈: 牛奶、稀饭养胃

7 妈妈:  白开水是最好的饮料

8 妈妈:  烟酒都不能沾

9 妈妈:  作息有规律

10 儿子:  是的是的

11 儿子:  马上就睡了

12 儿子:  嗯嗯

1 Mom: @Peter Does your stomach feel any better?

2 Son: It does not pain after I've taken medicine for 2 days on end.

3 Son: Very good effects.

4 Mom: Nourishing the stomach is a lifelong thing.

5 Mom: Can't drink much coffee, kola and the like.

6 Mom: Milk and porridge nourish the stomach.

7 Mom: Boiled water is the best drink.

8 Mom: Can't smoke or drink alcohol.

9 Mom: Work and rest need to be regular.

10 Son: Yes, yes.

11 Son: Going to bed soon.

12 Son: Okay, okay.

In Ex. 3, the son could have finished his first utterances (lines 2–3) in one single shot (i.e., sent out with one single press or click of the button “send”). The same applies to the mom's utterances in lines 4–9. There is another possible reason for the mom not to finish her utterances at one shot: she is adding up new points as she is thinking. Since she is chatting with her son, she may find it more comfortable and natural to talk like this than produce a long message after it is completely thought over and ready for articulation. This is just another manifestation of the widely and often distributed spontaneity of WTT (the production of these utterances is completed within a short span of time and, apparently in fast succession; after all, there is the 5-min threshold for technological recognition of it as continuous chatting on the WeChat platform).⁸

Owing to non-co-presence, participants to WTT cannot monitor each other's speaking time in the operation of turn-taking. This inability may lead to apparently messy turn taking in WTT.⁹ For example, in Ex. 3, the son's response “Yes, yes” (line 10) is a response to his mother's views (lines 4–9) on how to keep a healthy stomach; his response of “Going to bed soon” is targeted specifically at her point “Work and rest need to be regular” (line 9); and his last utterance “Okay, okay” shows the son's compliance with his mother's advice, including “Can't smoke or drink alcohol” (line 8).

Yet, despite the occurrence of such disrupted synchrony, we should not refute the basic fact that people are still doing turn taking in their interaction and that the resultant WTT is still sequentially organized. Example 3 involved what Herring (2013) termed “loosened relevance” of adjacency pairs in CMC. The notion may be evoked to explain how interactants make sense of apparent non-adjacency. After all, participants can read back and forth and make sense of the apparently disorderly messages. In other words, we can still identify adjacency pairs from sometimes disorderly exchange of utterances.

As in face-to-face conversation, participants to WTT can repair their messages immediately, which is another important signal of spontaneity. Consider Ex. 4:

Ex. 4

1 曹:  老师好,  恭祝教师节快乐！我们几个已经年老的学生想请您和师母与同门聚聚,  不知您何时有空啊?

2 陈:  不用 keq

3 陈:  客气

4 曹:  应该的啊,  老师千万莫推辞哦。

1 Cao: Dear professor, happy teachers' day! As your already old students we want to gather with you and your wife. When will it be convenient for you?

2 Chen: No need to keq

3 Chen: stand on courtesy (keqi)

4 Cao: It's what we should do. Please do not decline.

In Ex. 4, Chen makes a mistake (line 2) when using Chinese pinyin to type the expression “客气.” Actually, he is using the English input mode to type the characters.

Another sign of spontaneity in terms of WTT organization is the recalling of messages. Technologically, WeChat app affords the opportunity to recall one's message within 2 min, as indicated in Ex. 5.

Ex. 5

1 蔡:  各位好,  晚上有人打牌吗?

2 徐:  我

3 蔡:  呵呵 还要两人呀

4 游: @sky 如果缺人,  我可以参加

5 刘:  我可以补缺

6 蔡:  好啊 五点结合！

7 刘:  啥地方

8 游:  都是响应会长的号召啊

       $\text{sky 撤回了一条消息}$

9 蔡:  会长定的地点

10 刘:  可以啊

11 徐:

1 Cai: Hi, all. Wanna play cards this evening?

2 Xu: Me

3 Cai: Hehe, we need two more

4 You: @sky I can if needed

5 Liu: I can fill a vacancy

6 Cai: Okay. Let's meet at five.

7 Liu: Where

8 You: We're all responding to the president's call

      $\text{sky has recalled a message}$

9 Cai: The place fixed by the president

10 Liu: Alright

11 Xu:

When recalling occurs, the platform will leave a trace of recalling, as shown by the shadowed part in Ex. 5. Sometimes, we do not just recall a message, but go on to repair it and resend a revised message. In this case, the platform will leave the trace “你撤回了一条消息 重新编辑” (You have recalled a message re-edit) on the sender's screen above the resent message. Note that repairs, as well as restarts and reruns, although frequently done, are not visible in WTT unless the repaired message has been sent and then recalled within 2 min. This might undermine the degree of our claim about the spontaneity of WTT and thus its suitability for CA analysis, as some important information inferable from the repairs, reruns and restarts is lost.

One remarkable aspect of turn-internal organization in WTT signaling spontaneous “monitoring” is bracket annotating (Song and Chen, in preparation). Consider Ex. 6:

Ex. 6

1 A: 你们有无看好的馆子

2 A: 一直想去的

3 B: 我好像不知道有啥

4 C: 唯一一家看好的是澡堂

4 C: (bushi)

5 A: 哈哈哈哈哈哈

6 A: 快分享一下

1 A: Do you have any good restaurant to recommend?

2 A: One you always want to choose

3 B: I'm afraid I don't know any

4 C: The only one I favor is a bathhouse

5 C: (no)

6 A: Hahhhh

7: A: Share it right now

In Ex. 6, in response to A's request to recommend a good restaurant, C jokingly recommends a bathhouse instead (line 4). Then, she immediately follows up with a clarification “No” within the brackets. Such form of interactional practice in turn-internal organization, which is frequent among young Chinese netizens, serves to clarify the speaker's jocular intent (as in Ex. 6) or perform some other pragmatic functions (Song and Chen, in preparation).¹⁰

Another function worthy of special mention is tickling. Specifically, a member of a chat group can tickle someone else during a chat, without saying anything. For example,

Ex.7

1 A: 今晚约球

2 B:

3       $\text{C拍了拍B}$

4 B: 今晚有事,  改天约

1 A: Let's meet this evening and play table tennis

2 B:

3       $\text{C tickles B}$

4 B: I've got an appointment today. Let's meet some other day.

In Ex. 7, by means of tickling, C addresses B directly to convey some unsaid message. In this context, C probably was hinting at B against playing table tennis with A, as indicated by B's change of attitude ( is an emoji that often signals a positive response). People may tickle each other for other purposes such as greeting.

Other organizational features of face-to-face conversation such as the occurrence of insertion sequence and back sequence (Levinson, 1983) are also found in WTT. As they are easily recognizable and identical with those in face-to-face interaction, we choose not to spare any space to illustrate them here.

4.2. Recontextualization features of WTT

As in the case of discourse organization, WTT also demonstrates both common and distinctive features in terms of discourse representation or, more generally, recontextualization in discourse.

By recontextualization, we mean “the insertion of a discourse element from a source context into a target context” (Gruber, 2019, p. 61). As a ubiquitous discursive practice (van Leeuwen, 2008), it may take on various forms. In traditional spoken and written communication, it may be realized as either direct or indirect reported speech. In written communication, direct reported speech, generally a verbatim representation of a chunk of the source text, is marked by the use of quotation marks, whereas indirect reported speech is a paraphrase of the represented content (usually coupled with the source information in both cases). In oral communication, we may mark our direct¹¹ or indirect quoting of others' utterance by metapragmatic expressions (Culpeper and Haugh, 2014; Chen, 2020, 2022) such as “as you said,” “somebody once said,” and “in somebody's words.” In both written and oral communication, we can quote our speech or cite our own text by using metapragmatic expressions such as “as I said earlier” and “as I mentioned in Section 2”.

In online communication, Internet communicator tools (ICTs) provide built-in program features that enable users to insert an entire previous message (or parts of it) automatically into a current message (Gruber, 2019, p. 57). The built-in quoting and sharing features make exact quoting possible (for a relatively complete review of research on different forms and practices of discourse representation in various media and genres, refer to Gruber, 2019).

Like other apps, WeChat app provides its users with effective affordances for doing recontextualization or discourse representation. One affordance is the citing function. It was introduced into WeChat (Edition 7. 0.9) in 2019. By long pressing the target message (or photo) or right-click the message (or photo) on the mobile phone,¹² the citing function will appear along with other functions on the screen. Communicatively speaking, the function is particularly necessary and useful, because it is sometimes difficult to respond to an earlier utterance without repeating it verbally. Since it has been intervened by pieces of more recent utterances, understanding based on adjacency or coherent pairing sometimes becomes difficult or even impossible. With the affordance of the citing function, the earlier utterance or coherence is selected and then “inserted” in the current context in a marked way as shown in Ex. 8. In so doing, adjacency is recovered so that the new utterance could be comprehended in the context of this inserted former utterance. Consider the shadowed part of Ex. 8, which indicates the citing function:

Ex. 8

1 陈:  请写一个本次会议的报道

2 孙:  好的,  我知道了,  写完后发在公众号吗?

3 孙:  后面可能还是要弄个网站。网站和公众号还是不一样的

4 陈:  应该有网站吧

5 陈:  报道写好后让我看一下

6 孙:  好的

7 孙:  学院主页上中心有个链接,  现在打不开

        $\text{陈: 应该有网站吧}$

8 陈:  问一下

9 孙:  好的

1 Chen: Please write a report on the meeting

2 Sun: Okay. Got it. Publish it on our official accounts after finishing it?

3 Sun: Perhaps we need to build a website afterwards. Websites and official accounts are different

4 Chen: I guess we've already got one

5 Chen: Let me have a look after you finish the report

6 Sun: Alright

7 Sun: There's a link to the Center on the School's website, but it can't be accessed

        $\text{Chen: I guess we've already got one}$

8 Chen: Ask someone

9 Sun: Okay.

In Ex. 8, Chen not only mentions that there should be a website for the Center (line 4), but also asks Sun to let him have a look at the report when it is ready (line 5). Sun first responds to Chen's request. Then, with the citing function, she quotes his guessing utterance in line 4 and responds to it with what she has found in relation to the Center's website.

Unlike the case of the citing function, which enables verbatim representation of the cited content and form, WTT enables partial representation or paraphrase of the source message. Consider Ex. 9:

Ex. 9

(Context: Chen shared a lecture poster)

1 曹:  @陈 老师早！谢谢分享

2 袁:  老师好,  谢谢分享!

3 杨:  老师好,  谢谢分享!

4 程:  谢谢老师分享!

5 沈:  谢谢老师分享!期待了

1 Cao: @Chen Morning, professor. Thanks for sharing

2 Yuan: Hi, professor. Thanks for sharing

3 Yang: Hi, professor. Thanks for sharing

4 Cheng: Thank you, professor, for sharing

5 Shen: Thank you, professor, for sharing Looking forward to it.

In Ex. 9, after Cao greets Professor Chen and thanks him for sharing the lecture poster (line 1), several former doctoral students of Professor Chen follow suit (lines 2–5). It is quite evident that Yuan, Yang and Cheng have more or less duplicated or reworded Cao's utterance (including the emojis) in one way or another. Shen not only borrows exactly what Cheng says but also goes a step further by adding that she is looking forward to the lecture (line 5).

A less explicit way of recontextualization in WTT is to resend or forward a post or utterance(s) from one chat group to another or other groups. The outcome is that the forwarded post or utterance(s) extracted from the source context has acquired a new context and may be processed similarly or differently in terms of its meaning or function in the new context. Note that Gruber (2019) calls such online practice “decontextualizing” (p. 75) rather than recontextualizing, in the sense that the post or utterance(s) is removed out of its original context but resent to someone else or members of other group(s).

It is worth mentioning that recontextualization by way of traditional means such as the use of metapragmatic expressions of citations (e.g., “as you said,” “sb. once said”) is still applicable in WTT.

5. Discussion

From the illustrations of the organization and recontextualization of WTT, we can find that it is a form of digital conversation that is naturally occurring in its own way. Notably, we mean that it is not only similar to face-to-face interaction in many ways, but also different in several ways.

On the one hand, WTT is an immediate, spontaneous, synchronous, sequentially completed chat¹³ to a significant extent, with more or less frequent exceptions. In WTT, we find a variety of face-to-face conversational features and mechanisms of organization, such as immediate response, turn taking, repairs, fragmental speech, overlapped speech, and reported speech. Yet, we need to interpret the notions of immediacy and synchrony not as purely temporal notions but rather sequential ones. Although sequence in WTT equals immediacy in time on many occasions, it does not always denote real immediacy. As mentioned earlier, on WeChat platform, a follow-up message is displayed on the screen as an immediate one as long as it is sent within 5 min since the preceding message by oneself or the other interactant was sent.

On the other hand, WTT is natural in many unique ways. To start with, the spontaneity of WTT is technologically afforded in various ways. For example, it is sometimes indicated by “对方正在输入…” (XX is inputting …) on the top line of the screen. It is sometimes shown by the practice of recalling a message marked by “XX has recalled a message” at the turn where it takes place. Also, it is sometimes manifested by the use of bracket annotation for various purposes, which has little or even no equivalent in face-to-face interaction or written communication (but sometimes in drama scripts).¹⁴ Secondly, WeChat platform affords the chance to “listen” back in WTT, thanks to the persistence and storability of interactions (Gruber, 2019, p. 55). As the interlocutors are not co-present, overlapped speech seems to occur more frequently than in face-to-face interaction. This may more or less disrupt the turn-taking system characterized by sequential organization. Yet the change to “listen back” reduces the risk of missing messages and enhances the chances of “deep” reading. Finally, in terms of recontextualization, WeChat platform affords quick and accurate reported speech owing to the embedded copying function, and convenient text quoting thanks to the embedded citing function.

With the invention and application of various social media, our way of communication has undergone significant changes. We communicate not only in face-to-face physical world but also in non-face-to-face, non-co-present virtual world. Communication in the latter world has made us gradually but fundamentally accustomed to interaction involving varying degrees of non-immediacy, non-synchrony and spontaneity. Partly because we often perform multi-tasks today and partly we try not to engage others who we reckon may be occupied at the moment, we take occasional or even frequent short-term delayed responses to be normal and natural and respond to them as if they were synchronous. WTT reflects the new natural form of interaction. Neither the reception of messages nor subsequent feedback is or needs to be immediate or synchronous. It can be synchronous and non-synchronous alike. In cases of little or no urgency, we choose to exchange short-text messages, synchronously or non-synchronously, instead of making audio/visual phone calls, or going offline to interact. In the new form of interpersonal interaction, when we cannot or even need not immediately capture the thinking process of others manifested in restarts, reruns or repairs, we do not always care about it (although it significantly helps us capture our interlocutor's intention and attitude). Rather, we care more about what he or she has actually and eventually “said” and how it is “said.” Although delayed response sometimes suggests a passive attitude (e.g., a refusal, a disagreement) in face-to-face interaction, we generally do not take the delay to be negative but rather to be a consequence of non-availability of the interlocutor at the moment. To avoid potential misunderstanding, we do sometimes provide an explanation to excuse ourselves from making a long delayed response. When it comes to recontextualization, the affordances of WTT enable novel recontextualization formats for representing extracts from previous utterances in their actual utterances (Gruber, 2019, p. 55–56). They satisfy the need of economy and sometimes that of alignment (Lee and Tanaka, 2016). For one thing, when following other group members in responding to a post sharing good or important news, directly copying a prior response more or less verbatim and then posting it saves time and improves efficiency. At the same time, the practice contributes to alignment among group members by way of complete or partial identical response. In addition, the citing function facilitates intertextuality and can maximally rescue the disruption of turn-taking system caused by non-co-presence-induced non-synchrony, while guaranteeing the degree of accuracy of discourse representation (Gruber, 2019, p. 55).

All the unique features of WTT, functionally motivated and technologically afforded, are “genre constitutive practices” (Gruber, 2019, p. 55). Based on these features, we may conclude that it constitutes a distinctive sub-genre of cyber discourse. Given the primacy of phenomenon, we need to “recognize some of the wider potential for reading and applying Sacksian principles in ways that move beyond a sole focus on sequential matters within talk and interaction” (Housley et al., 2017, p. 15). Instead of strictly following the hard-line of pure CA analysis, “a wider interpretation of Sacks' work can be mobilized and in ways that transcend any “narcissism of small differences” amongst the ethnomethodological, conversation analytic and wider interactionist Communities” (Housley et al., 2017). After all, Sacks used to take an interest in the phenomena per se, observing, describing and documenting how members do social life in “any possible places and ways” (Fitzgerald, 2018, p. 7) it occurs. Actually, as he focused on social actions, he could engage in whatever data he had available, including social media data, in ways that enable systematic, repeatable and inspectable analyses (Fitzgerald, 2018). This means that we can, sometimes at least, use CA to directly handle online conversation, which, like traditional face-to-face conversation, falls within the “scope of humans activities as methodical” (Fitzgerald, 2018, p. 8). We are not supposed to doubt about the suitability of subjecting WTT to CA analysis, only because some important information inferable from the reruns and restarts is lost. Instead of rejecting CA for online conversation, we can extend, enrich and modify the current version of CA to capture the new affordances made possible by the new platform of interaction, as suggested by Giles et al. (2017):

“As more researchers in the social sciences and humanities are turning to digital phenomena as their substantive objects of interest, it is becoming increasingly clear that traditional methods of inquiry need considerable adjustment to fully understand the kinds of interaction that are taking place in online environments.” (Giles et al., 2017, p. 37)

As a genre variant of digital communication, WTT exhibits the combined effect of “pull factors” and “push factors” (Gruber, 2019, p. 57). The former have to do all kinds of Internet Communication Tools (ICTs) as new medium affordances; the latter have to do with the communicative needs of a social group faced with a communicative problem (Gruber, 2017). Thus, digital CA needs to embrace the idea that online chatting like WTT is “digitally” naturally occurring, and take into account all the technological affordances of organization and recontextualization when analyzing collected typed talk from social media platforms.

Moreover, digital CA can still be improved, at least by benefitting from the use of input log software. Clearly, one disadvantage of digital CA is its inability to capture the speaker's languaging (Li, 2011) process. Although we find conversational repairs in WTT, they probably occur with a lesser frequency in face-to-face interaction, as observed from the perspective of the interlocutor. Actually, our own experience of WTT is that as “speakers,” we repair very frequently. However, in WTT, pauses, reruns and restarts are not visible.¹⁵ Although this is still not technologically shown, the record of our “languaging” process via an input log software would tell the whole story. If that is done, we will be able to record and analyze the spontaneous and significant features of WTT even though the process and result of reruns and repairs are not shown on the screen of the interlocutor's computer or smart phone.

6. Conclusion

The goal of this paper has been to demonstrate the naturalness of WTT as a form of digital conversation in terms of interactional organization and recontextualization in particular. Based on a close examination of some details of WTT, we have concluded that WTT is natural in both common and distinctive ways. On the one hand, its naturalness consists in partial or even predominant spontaneity and synchrony characteristic of face-to-face conversation. On the other hand, it is technologically marked in some noticeable ways, such as the recalling of a message, the act of “X is inputting …,” the act of bracket annotating, the citing of a prior message, and the act of copying. These features shown on the screen may impress us as if we could feel the co-presence of our interlocutors. Although interactants occasionally overlap in some disorder, owing to the non-co-presence of the interactants, WeChat platform allows them to read back to sort out the adjacency issue. In view of the above-mentioned features of WTT, which make it a unique sub-genre of digital conversation, it is clear that while we can apply canonical CA in some cases, we can enact a modified version of CA in other cases. For this reason, the assumption that CA is not applicable to social media data, while cautious and helpful, may risk oversimplifying the picture of online chatting and overlook its ubiquitous and sometimes distinctively “natural” features. Instead, it is sensible that we adopt an alternative view of the notion of “(non-) naturalness.” Specifically, we need to bear in mind the “natural” features of online chatting made possible by various technological affordances. In so doing, we are truly embracing Sacks' insistence on the primary focus of the phenomena, and have great chances of understanding online interaction “members' social action and their methods in undertaking that social action” (Fitzgerald, 2018, p. 9).

The present study, first of its kind to relatively systematically examines patterns of WeChat organization and recontextualization, adds to the existing literature on messaging on other platforms, mostly in English. Some issues to be addressed in the future regarding WeChat talk may include the possible reasons and interpersonal consequences of delayed response as opposed to immediate response, message recalling and emoji/emoticon use, the manifestations of interactional style choice between involvement vs. detachment (such as occurrence of overlaps, use of backchanneling, use of emoji, and short, fragmental turns) and the like. The issue of emoji literacy is worthy of special mention, as discrepancy in understanding the emojis may mean possible digital “generation gap” and “gender gap” and lead to severe interpersonal consequences. Finally, the issue of inadequate contextual support in WTT due to its being non-situated calls for the reconceptualization of context in a different way so that the analysis of WTT could benefit from it.

This study suffers various limitations. First, in terms of data collection, the data size and extent were limited. Second, the selection of instances of “spontaneity,” “mechanism of turning taking” and “modes of recontextualization” might not be systematic in the absence of a principled scheme. Third, the issue of recontextualization only receives limited discussion in this study although it could be much broader/more complex. Finally, we only addressed the organization and recontextualization in WTT, to the neglect of other social media. Different platforms vary in their technological affordances and thus may enable different forms of organization and recontextualization. Different ICTs provide quite different possibilities for representing and integrating reported utterances into a report utterance (Gruber, 2019). The differences and variations may constitute a key challenge for analyzing digital communication (Meredith, 2017). How to capture the naturalness of other digital communication remains to be explored.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

XC provides the general structure and basic ideas of the paper. TC helps to develop the draft of the paper and supply needed references. All authors contributed to the article and approved the submitted version.

Funding

This work was funded by National Social Science Fund of China (20&ZD299).

Acknowledgments

The early version of the study was presented at the First International Conference on Internet Pragmatics held in Fujian Normal University (FNU) on September 21-23, 2018.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

References