Extending Prior Posts in Dyadic Online Text Chat

Abstract

This study explores whether chat users are able to extend prior, apparently completed posts in the dyadic online text chat context. Dyadic text chat has a unique turn-taking system, and most chat softwares do not permit users to monitor one another's written messages-in-progress. This is likely to impact on their use of online extensions as an interactional resource. For example, intonation is one interactional device used in spoken conversation to indicate imminent turn closure and project a response from interlocutors, thus creating a possibly complete turn-constructional unit (TCU) and transition relevance places (TRP). The speaker, however, may add further talk to that TCU, thus producing an extension and redoing the TRP. If prosody and monitorability of recipients' online talk production are unavailable as interactional resources, can participants still extend prior posts during online textual interaction? The analysis of chat logs of geographically dispersed speakers of Italian provides new evidence on the fundamental role of semantics, pragmatics, syntax, and punctuation in users' deployment of extending devices with various forms and interactional functions in dyadic textual online chat, as resources for social action.

Acknowledgments

This study was inspired by data sessions of the Talk-in-Interaction research group, under the direction of Kwang Kwong Luke, at the Nanyang Technological University's Division of Linguistics and Multilingual Studies during my term there as Senior Research Fellow in 2011. Thank you to students of Italian 3A at the University of South Australia for making their online interactions available for research. I also express my appreciation to Paul Drew and especially the anonymous reviewers who provided such detailed and insightful comments on an earlier version of this study.

Notes

¹ The practice of splitting possible TCUs is one exception to this tendency, because multiple posts are required to complete a TCU that participants orient to as a “turn” because it is possibly complete and therefore provides a potential TRP. Garcia and Jacobs' (1998, 1999) studies of group chat also indicate that it is difficult to know for certain which posts are part of the same turn and which are not, although dyadic chat may be slightly clearer than a multiple user environment.

² Chat interactions are transcribed verbatim, preserving the conventions and variations of chat usage, with translation not necessarily reflecting nontarget language forms. The data consist of the printouts of the chat logs as recorded by the computer, although some printouts (Dyad 1) did not replicate emoticons. Participants are referred to as either NS (native speaker) or NNS (non-native speaker) to assist readers with no knowledge of Italian in the differentiation of participants' linguistic backgrounds, which would otherwise be unidentifiable in the translation to English. The researcher was the coordinator of the course where students were required to “chat” as part of their language learning and assessment. Students completed their studies either online or on campus, with the NNS in Dyad 1 an online student and the NNS in Dyad 2 an on-campus student. In accordance with University of South Australia ethics procedures, only participants who formally agreed to participate in this study were included in the data set, which consists of chat logs submitted to the researcher. Any names used by participants in the interactions are fictional.

³ Overlap occurs both in dyadic and group chat, as participants may type and post simultaneously, although this study was not designed to access the message input process (see The Data, above).

⁴ Recent versions of software used by participants in this study, such as Microsoft Instant Messenger (MSN Messenger), however, do provide some timing information and a window that indicates when an interlocutor is typing a message, thus permitting the enactment of “overlap avoidance strategy” (Pasfield-Neofitou, 2013, p. 136) and monitoring of interlocutors' post production in real time.

⁵ I wish to acknowledge the anonymous reviewer who pointed out these two possibilities.

⁶ Herring (1999) reports on an adaptive practice in group chat: “One IRC user reported to the author that he and a regular IRC conversational partner had devised the practice of typing a ‘%’ symbol at the end of a message to indicate that they were not yet ready to give up the floor, and thus that the other should wait before taking the next turn” (p. 8).

⁷ In this case, there are three list items plus the summary statement, unlike prototypical three-part lists observed by Jefferson (1990) in spoken interaction.

⁸ Couper-Kuhlen and Ono (2007) define “add-ons” as TCU continuations that “replace” a prior TCU and involve “prosodically disjunct added-on material which replaces or repairs one or more elements in the host” (p. 519). Extract 1 provides an example of an add-on (in bold) from spoken conversation:

⁹ Extract 2 provides an example of a free constituent NP (in bold):Although there is no syntactical link to the host TCU (that guy was (dreaming)), the extension here, indicated in bold, is linked semantically and pragmatically, as may also be possible in written talk.

¹⁰ Couper-Kuhlen and Ono (2007) define glue-ons as nonrepairing extensions that are distinguished from insertables in terms of whether they are grammatically fitted to the end of the host (glue-ons) or not properly fitting the end of the prior unit but belonging somewhere within it due to syntactic closure of the prior turn (insertables), as in Extracts 3 and 4, which provide examples from spoken language of a glue-on and insertable extension, respectively (in bold).In Extract 3, lack of response from Dan allows Gor to in a sense “recomplete” prior TCUs (turns 1–3) by using syntax to link or glue one TCU to the other. In Extract 4, the “insertable” is produced after turn 2, which is itself a continuation of turn 1. Or not is produced two turns later but is interpretable as part of turn 1 as tch Are you gonna drive inor not.

¹¹ Extract 5 provides an example from spoken talk of a syntactical adverbial extension of a prior turn from Ford et al.'s (2002) data (extension in bold).The speaker here comes to a place of possible completion at the end of hour. At this point the utterance is audibly complete syntactically (a complete clause), prosodically (low falling intonation), and pragmatically. However, the speaker subsequently produces with the two of ‘em on it, an adverbial continuation of the prior TCU.

¹² The fact that posts 4 and 5 are in fact a list of three items (1, due volte al giorno; 2, a colazione; 3, alla sera) suggests the NS may be projecting subsequent items at post 4, although this is unclear because of the possible completeness of this post.

¹³ At the time of writing this article, I am unaware of studies of increments in spoken Italian that would permit comparison with the online context.