Relationship building in L2 telecollaboration: examining language learner closings in online text-based chats

ABSTRACT

Online exchanges among learners are an important form of communication in language learning contexts. This is due to the affordances this medium of communication creates for learning as well as the growing interest in internationalization of curricula. In fact, online communication can foster connections between people, including learners, across the globe. This paper is based on a qualitative analysis of online chat communication between Canadian learners of German and German learners of English. We discuss how these learners negotiate closing sequences in their chats, considering that the patterns for closing sequences may be different in different languages or be culture-specific and learners need to make choices on a spectrum between business focus (i.e. staying on task) vs. personal focus (i.e. addressing personal information). Hence, learners must negotiate interactive patterns that they may not be familiar with in creative ways that establish and maintain interpersonal relationships. Research about online communication helps to better understand learner communication, especially in an international context. In addition, it helps raise language awareness for teacher training and provides impetus for sensitizing teachers and learners to the ways in which interaction works and how relationships are maintained in communication.

KEYWORDS:

• Chat communication
• transcultural pragmatics
• closing sequences
• German
• North American English

Introduction

Online exchanges among learners have served as learning activities in language classes around the world for several decades now (Kern, Ware, and Warschauer 2004; Warschauer 1997; for more recent research, see Çiftçi and Savas 2018). O’Dowd (2011) and Belz (2007) argue that online interaction should play an even bigger role in foreign language classes, and it can be assumed it will, considering continuing classroom linkages across the globe and increasing interest in the internationalization of curricula (Liebscher, Mueller, and Griffiths 2017). The growing research on this form of communication, especially in second language learning contexts, attests to the interest in the ways in which this form of communication has become an important medium of classroom interaction as well as the affordances it creates for learning.

Telecollaboration, or ‘global communication networks in foreign language education’ (Belz 2002, 61), has been a common kind of online learning engagement between language learners, often as part of a tandem arrangement for learning each other’s first languages, which was also the case in our project. In this article we focus on one aspect of online text-based chat communication: closing sequences or leave taking in communication between Canadian learners of German and German learners of English. Considering that leave taking is an important part of negotiating community building (Darhower 2007), we are interested in how closing sequences are sites for the creation of (different kinds of) relationships through language (Enfield 2013) or, in other words, how participants position themselves in interaction, as they make different aspects of their identities relevant at different times (Bucholtz and Hall 2010). Closing sequences, especially for language learners, are not a trivial matter for several reasons: a) the patterns for closing sequences may be (or participants assume them to be) different in different languages or be culture-specific (Bardovi-Harlig et al. 1991; Harren and Raitaniemi 2008), b) learners need to make choices on a spectrum between business focus (i.e. staying on task) vs. personal focus (i.e. addressing personal information), and c) language learners may be more or less proficient in their language abilities to engage in leave taking. Drawing on an ethnographically informed conversation analytical approach (Deppermann 2000; Kallmeyer and Schwitalla 2014), we present and discuss typical examples of closing sequences from our corpus in light of their potential impact on relationship building and consider the implications for language learning in general.

Conversation analysis (for) second language acquisition

While drawing on ethnographic conversation analysis (CA) and CMC research, our research of online chats can also be placed within CA-SLA, i.e. the use of CA to investigate second language acquisition contexts (Kasper and Wagner 2011; Markee and Kunitz 2015), originally referred to as CA-for-SLA (Markee and Kasper 2004). Language acquisition in this tradition is viewed from a usage-based perspective (Tomasello 2003), which holds that the learning or acquisition of language and linguistic structures happens while individuals are engaging in interaction. Most importantly, the nature of any interaction is through back-and-forth engagement with each other, which points to the social nature of interactions. Hence, ‘CA can provide highly detailed insights into how opportunities for language learning and teaching are socially constructed by participants who observably orient to the practices of institutional varieties of talk’ (He 2004, 573). In orienting to each other’s practices, learners rely on the knowledge they have received through previous interactions, which includes ways of interacting that may be specific to particular contexts (e.g. online chats) and language groups (e.g. Canadian English vs. German). The goal of CA-SLA is two-fold: 1) to gain insight into the ways in which language is used and acquired in a variety of learning contexts and 2) to use this knowledge for teacher training (e.g. Betz and Huth 2014-2016). This article deals mainly with the first aspect but some discussion of the second aspect will follow in the implications.

Relationship building, positioning and intersubjectivity

Our interest in relationship building in interaction derives most generally from the notion of intersubjectivity, i.e. the need for negotiating mutual understanding with an orientation towards common ground, including an understanding of the interactional architecture (Sidnell 2014) and of cultural norms. Within the context of online discourse, Darhower (2002, 256) defines intersubjectivity ‘as a shared orientation or perspective on a communicative task.’ This includes the participants’ demonstrated understanding about their relationships and the task at hand, e.g. during leave taking. Within chat communication research, Darhower (2002, 2007) investigates leave-taking in the context of community-building in telecollaboration between learners of English and of Spanish. He further argues that extensive exchanges about positive experiences and talk about ‘futurity’ (i.e. keeping in touch in the future) indicate positive relationships and the building of a ‘camaraderie level’ (Brown 2001) of community, especially in one of the groups (Darhower 2007, 574). Also building on the notion of intersubjectivity, Enfield (2013) draws attention to the intricate ways in which interactants use linguistic and other resources to engage in relationship work. We draw on these perspectives by analyzing the kinds of relationships learners construct on an interactional level by positioning each other in certain ways, which includes the construction of professional and personal identities (Harré and van Langenhove 1991).

While this kind of relationship building through positioning occurs in any interaction, we are particularly interested in the ways in which our dyadic groups of language learners construct their relationships in online exchanges. Rather than analyzing groups of learners over the course of several weeks, however, as e.g. Darhower (2002, 2007) has done, we analyze leave taking as one particular site and focus of the interaction, where relationship negotiation and community-building becomes relevant (ibid).

Chat communication research

Chat has been defined as ‘synchronous computer-mediated discourse (SCMD) that takes the form of either private interaction (e.g. one-on-one communication via instant messaging) or public interaction (e.g. group communication in a chat room)’ (Blyth 2012). In other words, chat is text-based communication that takes place online, the common abbreviation for which is SCMC (synchronous computer-mediated communication). Most chat platforms that people are familiar with are quasi-synchronous, rather than synchronous, since the interlocutors can only see a particular message once the initiator hits ‘enter’ or ‘send,’ not while typing is happening. This is also the case for the chats we are analyzing.

Computer-mediated discourse analysis has been a growing field of research (e.g. Crystal 2006; Herring 1996). Interactional research on chat communication, particularly in the late 1990s and early 2000s, has centered on investigating its communicative structure (cf: Garcia and Jacobs 1999; Geers 1999), although much of this research has been related to the turn-taking system at work in SCMC, rather than on the negotiation of opening or closing sequences (Schoenfeldt and Golato 2003). Within computer-assisted language learning (CALL), the use of written SCMC has been widely researched, especially in foreign language learning and telecollaboration contexts, as Jenks (2014, 17–33) shows in his review of CMC studies. While some of that research has turned to the analysis of voice-based chat rooms (e.g. Jenks 2014), the investigation of text-based chat communication is still valuable for several reasons: 1) it allows for (diachronic) comparison of different forms of (CMC) communication, 2) it provides insights into the pedagogical value of text-based chats for language learners, and 3) it provides insights into transcultural aspects of communication that are important and transferable to different contexts. For the purpose of our paper, transcultural aspects of communication are viewed as those which encompass or extend across or beyond multiple languages and cultures (Kramsch 2012).

Closing sequences in spoken interaction and text-based chat interaction

While spoken and written interaction are certainly not identical, the analysis of online text-based chats has profited from research about spoken interaction for its close connection to these forms of communication. In fact, Schoenfeldt and Golato (2003) among others (cf: Garcia and Jacobs 1999; Geers 1999) have demonstrated that despite the differences between face-to-face or telephone conversation and text-based chat, most particularly that one is an oral and one a written form of communication, participants orient very strongly to the same ‘rules and regulations’ regarding sequencing. Schoenfeldt and Golato (2003) write that ‘it has been noted that because of the simplified syntactic structures, colloquialisms, and spelling changes reflecting spoken language, chat logs often read almost like a transcript of spoken language’ (247). However, the rules for what makes up a turn in text-based chats are different than in spoken conversation (ibid.), since in quasi-synchronous chat, interlocutors tend to click ‘send’ or ‘enter’ only when their particular turn unit is completed, and the formulation process is only observable through screen capture technology, which was not used in the gathering of the data for this project, since analysis of chatlogs was not initially the intended focus (248). Schoenfeldt and Golato (2003) suggest that ‘the participants in chats treat each message as a turn and the posted message as a transition relevance place’ that might more effectively be called a ‘next-message respondable’ (248, based on Garcia and Jacobs 1999, 350). It is not uncommon, however, for a series of completed units to be fired off in quick succession, since there seems to be a preference for continued communication, rather than long waiting periods during which messages or units are being composed. On the other hand, one chat turn may also contain several meaningful units that may be more or less connected to each other, which has the effect that the first pair part of a terminal exchange may come as part of a longer turn.

As part of their seminal work on closing sequences in conversation, Schegloff and Sacks (1973) and Button (1987), who worked on examples from American English, found that closing sequences are governed by what these researchers call a ‘closing problem,’ which they describe as ‘how to organize the simultaneous arrival of the conversationalists at a point where one speaker’s completion will not occasion another speaker’s talk, and that will not be heard as some speaker’s silence’ (Schegloff and Sacks 1973, 294–295). This is to say, the ‘closing problem’ is the way in which interlocutors need to negotiate bringing the conversation to an end.

Schegloff and Sacks (1973) outline several steps that interlocutors take in order to achieve a successful closing sequence in conversation.

Closing sequence (from Schegloff and Sacks 1973, 307; simplified):

Table

1	Dorrinne:	Uh- you know, it’s just like bringing the blood up.
2	Theresa:	Yeah well, THINGS UH ALWAYS WORK OUT FOR THE [BEST
3	Dorrinne:	[Oh certainly. Alright tess
4	Theresa:	[Uh huh,
5		Okay,
6	Dorrinne:	Good-bye.
7	Theresa:	Goodnight,

To begin, closing sequences are usually initiated once the main topic, or ‘business’ of a particular conversation has been completed, which can be called a ‘closing relevant environment’ (Bolden 2017, 234). This is the case with Dorrinne’s turn in line 3. At this point, one of the interlocutors may then inquire whether there is anything else to discuss, i.e. inquire about ‘unmentioned mentionables’ (Schegloff and Sacks 1973). This can be done by using a possible pre-closing token (in English, this is often done through the use of ‘Okay?’) or other form of pre-closing such as Dorinne’s ‘Alright’ in line 3. The other interlocutor then either orients to the closing, as Theresa does in line 5, which usually results in a rapid end to the conversation through a terminal exchange of goodbyes (at least in English), or bring up an unmentioned mentionable, in which case the conversation may continue. Bolden (2017) states that ‘the general sequence of interactional tasks that leads to the termination of a conversation – a) establishing a closing relevant environment, b) checking for the presence of other “mentionables” via possible pre-closings, and then finally c) ending the conversation with a terminal exchange – has been documented in several languages aside from English’ (232).

While this general pattern for a closing sequence is similar in different languages, what constitutes each of the steps and the overall length of the closing sequence varies. For American English, Schegloff and Sacks (1973) have noted that as long as no expansion takes place, an effective closing in a closing-relevant environment can be achieved in only four turns – a two-turn possible closing sequence (‘Okay?’ – ‘Okay’) followed by a two-turn terminal exchange (‘Bye’ – ‘Bye’). This suggests a preference, in American English, for an expedient closing. This preference is not the same for German. As Harren and Raitaniemi (2008) have suggested, there are differences in how closing sequences are negotiated in German as compared to English, something that also holds for other languages (e.g. Bolden 2017 for Russian; Bou-Franch 2011 for Spanish). In the case of German, Harren and Raitaniemi (2008) have demonstrated that the negotiation of a closing sequence is not as quick as Schegloff and Sacks (1973) and Button (1987) demonstrated for English. In fact, they have demonstrated that in German, there is a preference for a closing sequence that contains two separate parts consisting of an extended series of turns, which they call ‘negotiation sequences’ (206). Harren and Raitaniemi (2008) argue that unlike Button’s (1987) assertion that pre-closings are ‘movements out of closing,’ pre-closings are in fact ‘integral parts of the closing’ (Harren and Raitaniemi 2008, 206). Furthermore, as

opposed to the view of a single pre-closing pair and a terminal greeting pair as the essence of the closing, [Harren and Raitaniemi] recognized two more elaborate sequences in which the proceeding of the closing sequence is negotiated. First, participants negotiate whether they are ready to end the conversation, and then, they negotiate when to produce the terminal greetings. (206)

Harren and Raitaniemi (2008) suggest that the function of the two-part closing sequence in German is relationship negotiation, which is to say, a strengthening and re-establishment of positive social relationships. In American English, this kind of relationship work seems to be done during opening sequences that tend to have specific ‘extended’ sequencing (Schegloff 2007), while German tends towards a shorter opening sequence that moves quickly over into the ‘business’ portion of a given conversation (Selting 2007).

These observed differences in the organization of interactional structures such as closing sequences in different languages or communities are one important aspect underlying our analysis: if learners with different first languages come together in interactions, they bring their linguistic and sociocultural experiences and assumptions to the conversation (i.e. a certain ethnographic knowledge and background), which they use to negotiate with others in an interaction. This is especially important in learning environments that are less controlled than classroom discussions with the instructor, such as the text-based chat conversations in our data.

Materials and method

The data for this study were gathered between 2010 and 2012 as part of a project on transcultural learning based on a collaborative endeavour between Canadian learners of German and German learners of English. The groups watched a few episodes of two television programs – Türkisch für Anfänger (engl: Turkish for beginners) and Little Mosque on the Prairie – which encouraged direct engagement with and reflection on stereotypes. The students were paired (one Canadian student and one German student) and worked on a culminating Powerpoint presentation project addressing an aspect they found interesting about the shows, such as stereotypes, racism, or gender attitudes, through online collaboration using the university Learning Management System (LMS) chat system, discussion boards in the LMS, email, Facebook messaging, and Skype video and chat functions. Subsequently, students were required to comment on and ask questions about other students’ PowerPoint presentations via a discussion board in the LMS. The Canadian students participated in the project as part of an intermediate-advanced level German class, while the German students opted in outside of class, and only received a participation credit for an English class they were taking, rather than a grade like the Canadian students did.

The students were given minimal instructions for their conversations, although these instructions included encouragement to use English and German in about equal amounts. Students were given flexibility about how to achieve this, e.g. to alternate languages or to use English or German for some parts of their conversations. Most of the conversations happened during class time for the Canadian students. However, student pairs also chatted on their own time outside class. Overall, it was left up to the students how they organized their communication – some preferred to use Skype (i.e. the face-to-face video chat function) almost exclusively, while others used primarily email or Facebook chat, while others used the university-provided chat platform almost exclusively to complete the assigned tasks. The data corpus considered for this study consists of nearly 100 chat interactions on this university-provided platform between 21 learner pairs who participated in the project. For this article, 46 closing sequences negotiated by 14 student pairs were examined in more detail. Since we were interested in analyzing the construction of relationships, we eliminated those pairs that had fewer than three chat interactions. As such, pairs represented in the considered corpus had between 3–5 closing sequences. Excerpts for this paper were selected because they represent the complexity of language negotiation undertaken by the students. The excerpts discussed were selected primarily as illustrative of trends within the corpus and as representing common ‘types’ of closing sequences in the data (Deppermann 2000). While some representative excerpts are selected for the discussion below, the analysis has relied on all of these closing sequences.

Results and discussion

Professional vs. personal relationships

Despite being given the same objectives for their chat (and other) communication, the student pairs went about the project very differently. Some pairs were primarily concerned with finishing the assigned work (the PowerPoint presentation) as quickly and efficiently as possible and deviated very little from this topic. Six student pairings seemed to function in this way. This kind of task-orientation was reflected in the kinds of closing sequences these pairs negotiated. The following extract is an example of a typical closing sequence for those students who developed a primarily ‘business-like’ relationship through focusing on the task rather than engaging with personal information.

This is Pia and Gunnar’s second chat closing sequence, and it is typical for a number of reasons. First, the sequence is relatively short, consisting of only five turns, and there is very little transition time between the topic closing and the closing of the interaction. This sequence also closely resembles the sequencing outlined by Schegloff and Sacks (1973), including the possible pre-closing by Gunnar in line 164 which is ratified by Pia in lines 165–166, and then followed by the terminal exchanges in lines 167 and 168. It contains only slightly more than the minimal four turn closing Schegloff and Sacks suggest.

Schegloff and Sacks (1973) have noted that monotopicality is typical for U.S. American conversations between people who are not close familiars, which these students certainly were not at the beginning of this project. At the same time, however, as Bolden (2017) writes about Russian telephone conversations, ‘monotopicality is a feature of the call’s design, not its outcome, as designedly monotopical conversations may be extended beyond the first topic’ (240). This is also true of our chat data, since other pairs do engage in more than minimal closings. It is particularly these conversations, that were clearly extended beyond the main topic or purpose, which we are interested in examining in this article. We would argue that those pairs who did extend topics beyond the main topic or task (collaborating on the PowerPoint presentation) – who chatted about a variety of things outside of the academic project, including family life, hobbies and interests, as well as other more personal reflections on their transcultural experiences – also engaged in different kinds of closings. Extract 2 is a typical example of such students’ interaction.

Extract 1. Gunnar^a (Canadian) and Pia (German)^b.

161.	11:27 AM	Pia	Die Selbstbestimmung der Frauen .i think thats better[[trans.^c the autonomy of women]] .i think that’s better
162.	11:27 AM	Pia	Mit dem Titel versteht man es besser[[that title makes it easier to understand]]
163.	11:28 AM	Gunnar	Ok, das klappt.[[ok, that works.]]
164.	11:28 AM	Gunnar	Jetzt muss ich gehen. Bis nächste Woche?[[I have to go now. Until next week?]]
165.	11:28 AM	Pia	Ja, das ist gut[[yes, that’s good]]
166.	11:28 AM	Pia	See you next week
167.	11:29 AM	Gunnar	Tschüss[[bye]]
168.	11:29 AM	Pia	Byebye

^aAll names have been changed to protect the privacy of the students involved.

^bAll non-standard constructions of English are from the original transcripts and have not been changed.

^cEnglish translations for the German components of the chats are provided below the original surrounded by two square brackets. Non-standard spelling or grammar is indicated through [sic].

Extract 2. Paige (Canadian) and Lukas (German).

568.	11:22 AM	Paige	wir konnen dann zum naechsten mal alles finalisieren[[we can [sic.] finalize everything next time [sic]]]
569.	11:22 AM	Lukas	sounds wondeful:D
570.	11:23 AM	Paige	ok
571.	11:23 AM	Lukas	im going now
572.	11:23 AM	Lukas	need to get something to eat
573.	11:24 AM	Lukas	im starving
574.	11:24 AM	Paige	so between now and next week, don’t forget all your German.. as you will need to correct mine!
575.	11:24 AM	Paige	geh was essen:)[[go get something to eat]]
576.	11:24 AM	Paige	ich habe auch irgendwie hunger[[somehow i’m hungry too]]
577.	11:24 AM	Lukas	i will
578.	11:24 AM	Lukas	we both need to get some food now
579.	11:24 AM	Paige	vielleicht hole ich einen kaffee[[maybe i’ll get a coffee]]
580.	11:24 AM	Lukas	ich mag kein kaffee[[i don’t like coffee]]
581.	11:24 AM	Paige	‘the drink of the gods’ hat mein ehemaliger Kunstlehrer gesagt[[‘the drink of the gods’ my former art teacher called it]]
582.	11:24 AM	Lukas	i´ll get a bottle of water
583.	11:25 AM	Lukas	okay
584.	11:25 AM	Paige	lol i have a funny joke about that.. but there’s keine zeit[[lol i have a funny joke about that… but there’s no time]]
585.	11:25 AM	Paige	schoene wochenende lukas![[have a nice [sic] weekend lukas!]]
586.	11:25 AM	Paige	stay hydrated;)
587.	11:26 AM	Lukas	i will
588.	11:26 AM	Lukas	have a good one!!
589.	11:26 AM	Lukas	cya

There are a number of differences between Extract 1 and 2, even at first glance. Regarding length of the sequence alone, Extract 2 is considerably longer than Extract 1, which has to do with the fact that so many additional aspects (‘unmentioned mentionables’) were dealt with before the interaction closed. This was Paige and Lukas’ third chat interaction, and while they did stay on-task and get work done during the rest of the chat conversation, their closing sequence is reflective of the remainder of their conversation in that the ‘business’ part is interspersed with personal anecdotes and jokes. At the same time as the closing sequence is typical for the interaction between Paige and Lukas, it also contains typical elements of other student pairs who positioned themselves as orienting towards personal relationships. Bolden (2008) suggests that ‘there are interactional issues involved in building and maintaining a particular kind of social relationship – a relationship in which inquiries into each other’s state of being are expected on a particular interactional occasion’ (116). By joking and discussing likes and dislikes, Paige and Lukas position themselves as familiars, invoking personal elements to establish and maintain their relationship. This is in keeping with Darhower’s (2002) observed differences among the closings in his online chat learner data: some groups used more elaborate closings than others, continuing ‘the type of social cohesion that they established and maintained throughout the entire chat session’ (267). Precisely this kind of ‘social cohesion’ is evident in Extract 2 above as well.

While the reference to their next interaction (line 574) is common in all interactions, whether the pair is more oriented towards the task-at-hand or exchanging personal information, the additional commentary and joking around (especially line 574- Paige’s good-natured self-deprecating comment about her German language skills), as well as the volunteering of personal information (line 580- Lukas does not like coffee) are not typical of such task-focused interactions, and instead serve to underscore and re-establish a personal relationship between the interlocutors that is typical of the more personal-oriented interactions.

Both Extract 1 and 2 can be considered successful closing sequences, in that they were clearly collaboratively negotiated, and both parties orient to bringing the chat conversations to a close. This can be seen by the fact that the possible pre-closing, which in both cases comes at a closing-relevant point, is oriented to as such by the interlocutor. Following this, in both cases, the participants arrive at the terminal exchanges after some minor expansion. Extract 1 ends in a typical “Tschüss“ – ‘Byebye’ sequence. Extract 2‘s terminal exchange is more complex, since the exchange contains more than a simple terminal pair. In the course of ending their closing sequence, Paige and Lukas cluster a variety of information into their turns adding personal information to their exchange. For Paige this includes a reference to a joke that she doesn’t have time to tell (line 584), a terminal pair part (line 585), and a well-wishing to ‘stay hydrated’ in reference to Lukas’ wanting to get water (586). These parts come in quick succession, i.e. in typical chat format, as discussed above with reference to Schoenfeldt and Golato (2008). Taken together, they are understood and taken up by Lukas as a first pair part to a terminal exchange. In his second pair part, Lukas starts with a response to Paige’s instruction to stay hydrated (line 587), followed by two second pair parts to a terminal exchange (line 588 and 589). In this way, he attends to the personal comment made before providing the prototypical second pair part to close the conversation, to which there is no need for Paige to respond again.

Personal relationships and the use of ‘then’ in closings

In our chat corpus, a wide variety of terminal exchanges are employed. As Extract 1 and 2 have shown, there certainly are examples of minimal terminal exchanges that function as outlined by Schegloff and Sacks (1973), with a simple ‘Bye – Bye’ or equivalent closing the interaction. In addition, Harren and Raitaniemi (2008) found that towards the end of a spoken interaction, the utterances become increasingly formulaic, something which holds true both for German and for English. However, although there were instances in the corpus when this was the case, the dominating tendency was for the terminal exchanges to look very little like the closing sequences outlined by Schegloff and Sacks (1973) or Harren and Raitaniemi (2008). Extract 3 is an example of such a sequence, where learners use a variety of strategies to mitigate the leave-taking and thus position themselves as orienting towards their interpersonal relationship.

Extract 3. Kieran (Canadian) and Nicole (German).

250.	11:10 AM	Kieran	hahah yeah and i’ll snd you my german ones
251.	11:11 AM	Kieran	is there anything else we should dicuss today, since we have ten minutes left ?
252.	11:11 AM	Nicole	i was just going to ask you the same. But I don’t think so, you?
253.	11:13 AM	Nicole	oh, one thing! are we supposed to ‘bewerten’ the other pp’s as a group or individually?[[oh, one thing! are we supposed to ‘evaluate’ the other pp’s as a group or individually?]]
254.	11:13 AM	Kieran	yeah i think we have all our info covered, would you mind making the PP look nice lol, im not very artistic or good at PP]
255.	11:14 AM	Kieran	umm i think we respond to the other PP’s separately, i have to respond to group 8 and 9s PP
256.	11:15 AM	Nicole	and you suppose i am??? lol, no it’s fine, I’ll do my best. and like i said, if you’re not happy with it, just tell me. ah ok, yeah I’m supposed to do the same groups
257.	11:16 AM	Kieran	lol thanks
258.	11:17 AM	Nicole	sure. well then, you’ll hear from me saturday at the latest
259.	11:17 AM	Kieran	excellent:)
260.	11:18 AM	Kieran	ok Nicky well then i’m going to head out now
261.	11:19 AM	Kieran	have a good thurs/fri and i’ll hear from you on sat
262.	11:19 AM	Nicole	allright! you too! and thanks for your ideas!
263.	11:19 AM	Kieran	no problem

In addition, in Harren and Raitaniemi’s (2008) findings, the lexical item ‘dann’ (then) played a vital role in the negotiation of the German language closing track, with the closing sequence becoming increasingly formulaic as it continued (206). While there were several instances of both Canadians and Germans using ‘bis dann’ (until then/until later) in our data as part of a terminal exchange, ‘dann’ did not appear as frequently as a lexical item as one might have expected it to, given the importance it has in the Harren and Raitaniemi (ibid.) data. At the same time, when it was used outside of the more formulaic context, participants used the ‘dann’-construction in creative ways, as we will see in Extract 3.

As has been mentioned, Harren and Raitaniemi (2008) have suggested that the use of ‘dann’ in the German closing track serves an important interpersonal function. In Extract 3, the way ‘dann,’ or in this case, ‘then,’ is used serves a similar function, in connection with a number of other elements that point to the relationship-orientation of the interlocutors. This is the third chat between Nicole and Kieran. The first pre-closing comes in line 251 during which Kieran indicates that he will need to leave soon and to determine whether there are any unmentioned mentionables that need to be dealt with before their allotted time is up. Bolden (2017) outlines the difference between tacit and explicit pre-closings, demonstrating that explicit pre-closings were potentially more face-threatening and were therefore consistently embedded in other actions, such as floor offerings, or other mitigating efforts in order to maintain positive social relationships between interlocutors. In line with this observation by Bolden (ibid), Kieran’s explicit pre-closing here, referencing the time they have left, is common for most closing sequences in the corpus, since the chat sequences took place during class time, and as such, students knew that their interaction had a specific time limit. The vast majority of first possible pre-closings took this form, although the fact that it comes well in advance of the end of the conversation is less common in the corpus. Generally, the explicit pre-closing came much closer to the terminal exchange, or even included the terminal exchange first pair part within it. In Extract 3, however, there is no typical terminal exchange at all, as will be discussed below.

Both Nicole and Kieran position themselves as attending to their personal relationship by mitigating explicit pre-closings in a number of ways. In line 258, Nicole uses the kind of ‘dann’ construction that Harren and Raitaniemi (2008) outline, but translates it into English, prefacing her statement with the discourse marker ‘well’ signalling pre-closing (Schiffrin 1987). This ‘dann’ construction is important to the closing track in German because it is a pre-closing that addresses the interlocutor’s readiness to end the conversation, especially since it points to a future meeting during which the conversation will be continued. Interestingly, although this has not been identified as a common pattern in English conversations, there are a number of instances in the corpus where the ‘dann’-structures introduced by the German student are repeated by the Canadian student in a subsequent turn, as Kieran does in line 260. Scissors, Gill, and Gergle (2008) have demonstrated that linguistic mimicry shows trust and bonding in online communication, and this kind of recycling appears more frequently in the chatlogs of students who position themselves as more oriented towards interpersonal relationships. Besides the ‘then,’ Kieran recycles a number of elements from Nicole’s turn in his turn in lines 260–261, including both the discourse marker ‘well,’ which he further prefaces with the use of ‘okay,’ a discourse marker indexing pre-closing (Levinson 1983) – the fact that he needs to leave and end the interaction – as well as the confirmation of the day of their next interaction. By beginning with the discourse marker ‘okay,’ using his interlocutor’s name, and finishing with the reiteration of their plans to connect again, he softens the urgency of the assertion that he is ‘going to head out now,’ positioning himself as attending to the social relationship by mitigating and accounting for leave-taking.

The difference between the short vs. elaborated closing sequences in Extract 1 as compared to Extract 2 and 3 is in part a result of the construction of the student relationships (cf. Darhower 2007). At the same time, we would also contend that the difference between the less vs. more elaborate leave takings may arise from sociocultural aspects, i.e. ways in which learners are used to engaging in particular tasks according to particular patterns. Relevant here is Harren and Raitaniemi’s (2008) argument about the commonly brief American English vs. the elaborate German patterns of leave-taking. These sociocultural differences, which learners bring to an interaction, can be observed in how learners engage in leave takings in creative and transcultural ways, such as in Excerpt 3, especially in the final part of the closing sequence.

Specifically, this particular closing sequence contains no terminal exchange at all in the canonical sense. However, an examination of the corpus suggests that Kieran’s utterance in line 261 seems to function as a terminal pair part. There are a number of similar instances by Canadian participants. Most of the closing sequences that did not finish with a more canonical terminal exchange (such as ‘Bye’ ‘Bye’ or similar) still seem to have been considered complete by the Canadian students when they typed something like ‘talk to you on Saturday’ or ‘talk soon,’ much like Kieran has done here. This suggests the importance of the preference structures students are used to – in particular the expediency of the closing track in English – in making sense of how closings unfold in a transcultural context. At the same time, we see further evidence of the transcultural context by the way in which this kind of terminal pair part is not oriented to as ‘terminal’ by Nicole, who instead thanks Kieran for his ideas in line 262. Interestingly, Kieran also orients to Nicole’s utterance in the final turn of the exchange, with ‘no problem’ (line 263) being the final utterance of this closing sequence.

Conclusions and implications

Learner chat closing sequences are highly significant for negotiating how to close a conversation effectively and for maintaining ongoing relationships. In the learner chat sequences examined in this paper, a number of trends were observed that can arguably be traced back to participants positioning themselves as attending to their interpersonal relationship, and thus foregrounding personal, rather than professional relationships. In Extract 1, we analyzed a closing track in which participants foregrounded professional identities, in which the closing sequence moved quickly from the ‘business’ portion of the conversation to the end of the interaction. In Extracts 2 and 3, however, we highlighted how participants negotiated the closing track in non-canonical ways that demonstrated orientation towards personal relationships, despite potentially differing expectations about precisely how to sequence a closing track. Harren and Raitaniemi (2008) have suggested that closing sequences in German have a series of additional integral discrete units as compared to North American English closing sequences. In the learner chat data, there was a tendency towards extended closings among 15 of 21 analyzed learner pairs. Given the importance of the lexical item ‘dann’ (then) in their conclusions, we would have expected to find a much higher frequency of ‘dann’-formulations (or ‘then’- formulations) in the corpus than we did.

Similarly, although there was evidence that both Canadian and German students seemed to be following patterns they are used to from the sequences most typical in their first languages – extending closing sequences in the German case, and closing expediently in the Canadian case, the variety of different closing sequences that are represented in the corpus points to an alternative conclusion. We would like to suggest that in the transcultural context in which learners in our data were operating, with the maintenance of social relationships central to their interactions (and their task outcome), a different set of ‘rules’ seemed to emerge. First, both Canadian and German students used a variety of strategies to mitigate explicit possible pre-closings, including producing discourse markers such as ‘okay’ and ‘well’ preceding the pre-closing, as well as referring specifically and explicitly to the next projected conversation within the same turn as the pre-closing. Especially Canadian students also employed lexical repetition in their subsequent pre-closings, which can be interpreted as emphasizing solidarity and trust (cf. Scissors, Gill, and Gergle 2008). Few of the closing sequences contained a canonical terminal exchange (‘Bye’ – ‘Bye’), but rather took a completely different form, where one pair part or even both of the terminal exchanges were completely missing in favour of other more personal utterances such as ‘i’ll hear from you on sat’ (see Extract 3) or ‘schoene wochenende lukas’ (have a nice weekend Lukas) (see Extract 2). Furthermore, these different ways in which conversational ‘slots’ (e.g. pre-closings, closings) are taken up by individuals index the constructions of (personal and institutional) identities and the creation of different kinds of interpersonal relationships (cf. He 2004). Our analysis has shown that learners attend to each other in the construction of personal relationships. They have done so in creative ways, especially when negotiating language-specific patterns such as the use of ‘then’. While we were interested in the ways in which these relationships were constructed and these patterns were negotiated, we have not attended to the questions of how far these constructions and negotiations affect further interactions between the same people or how they may lead to certain kinds of relationships over time. These latter kinds of questions could be addressed through a longitudinal study in the future.

CALL research, and CALL-CMC in particular, has widely contributed to a better understanding of language teaching and learning issues (Jenks 2014, 8). Based on our analysis, there are a number of insights gained that can be applied to classroom interaction and subsequently to teacher training. First, the acquisition of (context and culture specific) sequential organization is an important aspect of language acquisition. In particular, Jenks (2014) argues for considering more strongly the links between learners’ interactions outside and within the classroom, namely ‘what second language users do in informal learning settings should inform what is taught in the classroom’ (123). Engagement in interaction (whether online or in person) requires individuals (including learners) to attend to and act on others’ linguistic choices. These choices may not always be the ones individuals deem intuitively most appropriate, considering that they are used to doing things a particular way in particular contexts and cultures. Hence, adjusting to less intuitive ways presents learning opportunities and requires flexibility and comfort with ambiguity. The analysis on which the current article is based has revealed some of the ways individuals use to bring a conversation to a close in such situations and the skills they need to do so.

Second, this analysis demonstrates the importance of transcultural awareness in language teacher training. We view this as having two components. On the one hand, communication must be seen as intersubjective – constructing meaning (which includes in our analysis understanding and pursuing the closure of an interaction) is a social rather than an individual practice. On the other hand, given that meaning construction is collaborative and intersubjective, raising language awareness, specifically for ways of doing things differently, rather than incorrectly, is equally important. In the context of effective transcultural communication, this means suspending judgement and working towards adjusting and negotiating communication and relationships. Teachers who create spaces for students who develop translingual and transcultural awareness as laid out in the MLA report (2007), create spaces for learners to negotiate and use language in meaningful, collaborative ways.

A last consideration is that these findings attest to what adding a social dimension to learners’ transcultural encounters can do – namely, create a situation in which they must navigate their own norms and those of an interlocutor with potentially different norms for how a particular interaction must be negotiated. There is pedagogical value for language learners in written chats as much as other forms of CMC, since “[s]econd language chat rooms are rich with authentic input and allow students to engage in communication that is socially meaningful (Jenks 2014, 123). It has been suggested that pragmatic competence can be taught using data from real conversations, using a conversation analysis framework (cf: Huth 2010; Betz and Huth 2014-2016), and lesson plans have been developed for teaching closing sequences in German specifically (Kampen Robinson 2014). As such, it would no doubt be fruitful, especially given the particular topic the students involved in this particular project were working on – namely transculturality and stereotypes – to examine the how of their communication in much more detail, moving from the general tendencies or emotional reactions they had to particular utterances or exchanges and moving to a fine-grained analysis of both monolingual interaction between Germans and Canadians, as well as their own multilingual, transcultural interactions. Working with actual data extracts in the classroom could provide teachers-in-training as well as learners with important insights in the pragmatics of language in use. In a transcultural context, learners can struggle with closings, not only because they are culturally-specific (Bardovi-Harlig et al. 1991), but also because they may differ much from the (artificial) examples in textbooks (Zhang 2014). Chat is a particularly useful medium for analyzing learner interactions and for teaching students about communication. Students are already familiar with the chat medium, and it is an opportunity for them to engage with others in real time, focusing on language in a transcultural context.

Acknowledgements

We would like to thank Dr. Robert Simunek and his students in Germany as well as the students in Canada for participating in the course and for letting us collect data. Ethics clearance was received under ORE #: 16706. See also Simunek (2007) for an additional discourse-analytic take on the data. We would like to acknowledge the help of the following research assistants in the data collection and data management process: Stephanie Cooper, Fabienne Hargreaves, Christine Kampen Robinson, Katharina Schröder, and Gerlinde Weimer-Stuckmann. This work was supported by the University of Waterloo’s Learning Initiatives Fund under Grant number #202926. We also appreciate the detailed feedback and very helpful advice by Dr. Emma Betz on an earlier version of this paper. Lastly, we want to thank our reviewers for reshaping and strengthening our argument with their insightful comments.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the University of Waterloo Learning Initiatives Fund [202926].

Notes on contributors

Christine Kampen Robinson

Christine Kampen Robinson An interactional sociolinguist, Christine Kampen Robinson examines questions related to language, literacy, and identity in a variety of contexts. One of her most recent projects analyzed the co-told narratives of Low German-speaking Mennonite migrants from Mexico to Canada, with specific attention to the narratives they tell about their experiences learning and home language attitudes. In her research on classroom contexts, previous projects have included the language use of children in a German bilingual program, and most recently she has concentrated on how post-secondary students’ transcultural learning experiences are impacted through collaborative projects and study abroad.

Grit Liebscher

Grit Liebscher’s research in Applied Linguistics addresses questions of language and identity within and outside of the classroom context. In her book Language, Space and Identity in Migration (2013, co-authored with Jennifer Dailey-O’Cain), she discusses how German- Canadians use their multiple languages as resources, how forms of address can become markers of group identities, and how meaning is created through language that is saturated with historical contexts and ideologies. An ongoing project relates to language attitudes, and how we create them discursively through the use of language. In her research on classroom contexts, she has focused on the use of the first language in the second-language classroom, the use of ‘okay’ to mark transitions in peer learner interactions, and trans-Atlantic telecollaborations. Utilizing tools of CA-SLA, she is currently investigating the use of laughter and ‘okay’ in a German language classroom in Japan.