Learning from professional conversation: A conversation analysis study

This article analyses conversation around classroom discipline to establish how teachers learn through professional conversation. The study was a qualitative study that originally adopted an ethno-methodological research design. Purposive sampling was used to select 6 primary school teachers from the East London Education District in the Eastern Cape. A video recorder was used to capture the conversation which lasted for 31 minutes 56 seconds after school hours. The recording was viewed repeatedly and transcribed verbatim. Three learning episodes were selected from the transcript for further transcription following Jefferson’s notations for conversation analysis purposes. Clayman and Gill’s (2004) conversation analysis levels were used to analyse selected episodes to establish how teachers learn through professional conversation. The findings show that teachers learn through requesting advice and testing ideas, and through the sharing of ideas. The findings also indicate that teachers use response preferences (response favourites both in agreement or disagreement during conversation), repairing or assisting one another through talking, nodding, and laughing as learning strategies. The study concludes that professional conversation is relevant for continuing teachers’ professional development. We recommend that teachers should embrace professional conversation for exchanging knowledge and experiences for learning purposes. We also encourage teachers to adopt conversational strategies highlighted in this study for professional learning purposes. Research experts on teacher learning should be involved in school workshops in order to further enhance teacher learning in specific areas.


Introduction
Professional conversation offers teachers a number of learning opportunities. This involves learning on one's own terms, learning from problem-posing by drawing on expertise and judgment within the group, and learning to address problems and questions of common interest (Wood, 2007:289). According to Msomi, Van der Westhuizen and Steenekamp (2014:800), professional learning ingredients such as deliberate reflection, inquiry, and sharing insights are imperative in the improvement of teacher learning. Research (Thurston, Van de Keere, Topping, Kosack, Gatt, Marchal, Mestdagh, Schmeinck, Sidor & Donnert, 2007) shows that one essential element of teacher professional learning is the quality of conversation that takes place among those teachers within a specific context. Thurston et al. (2007:488) further note that collaboration through conversations facilitates the reconstruction and elaboration of ideas through dialogue among teachers. Furthermore, a previous study (Pedder, James & MacBeath, 2005) submits that teacher learning is expanded through collaborative activities. Jackson and Bruegmann (2009:2) argue that experienced and skilled teachers increase the skills and knowledge of those with whom they interact; while Pedder et al. (2005:221) further suggest that teacher learning also occurs through talking about and valuing learning.

Learning from Professional Conversations
Professional conversation, according to Timperley (2015:6), refers to the intentionally organised formal and informal dialogue that occurs between education professionals including teachers, mentors, coaches, and school leaders, and is focused on educational matters. Horn (2007:39) suggests that teachers' colleagues can have a huge influence on teachers' approaches to classroom practice as well as providing them with the opportunities to shape their responses to challenging educational reforms. What is therefore fundamental about professional conversations in the context of teaching and learning, according to Rust (1999), entails teachers working together within a given setting to frame and solve education and classroom-based problems by creating their own powerful opportunities for learning. Van der Westhuizen (2015b:121) claims that professional conversations have been the focus of recent studies seeking to understand the discursive nature of professional preparation in teacher education. Van der Westhuizen (2015b:134) further shows how learning is facilitated by mentors' conversational moves of assessing, asserting, requesting, and accounting for views.
Conversation provides novice teachers with opportunities to learn how to improve their teaching practices from their experienced colleagues. New teachers, in particular, learn about strategies that are related to specific subject matter from peer conversations (Leonard, 2012;Timperley, 2015). They also learn about how to address individual differences relating to dealing with specific learner-related cases (Henderson & Petersen, 2008). Professional conversation provides opportunities for teachers to learn about the best and tested assessment practices from their experienced peers (Miller, 2008:78). According to Bhattacharjee (2015) and Giridharan (2012), professional conversation provides teachers with specialised situational learning in which they work together to co-construct knowledge through conversations about their everyday classroom experiences. In addition, professional conversation provides learning opportunities for teachers to exchange ideas and points of view. For instance, through peer conversations, principals learn new ways to interact with their staff, such that the conversations serve as an effective means of learning (Healy, Ehrich, Hansford & Stewart, 2001:339-341).

Conversation Analysis
Conversation analysis (CA) explores how social interactions are structurally organised. CA analyses transcripts of video recordings in detail, examining actions such as turn-taking, length of pauses, inflections, and all other significant utterances (Hancock, Ockleford & Windridge, 2007:14). CA observes silences, and exclamations such as ums, errs, and overlapping speech when participants speak simultaneously (Moriarty, 2011:18). Apart from the "words-as-spoken, CA allows the researcher to highlight a range of production details concerning timing, intonation, and pace that have been proven important for the organisation of the interaction" (Ten Have, 2008:130). In CA transcripts, every single utterance, pause, overlap, change in volume, laughter, and non-verbal action is included (Niemi, 2016:33;Ten Have, 1999:79).
Conversations are organised sequentially in CA through turn-taking, response preferences, and repair actions (Flick, 2014;Ten Have, 1999;Van der Westhuizen, 2012a). Turn-taking involves getting in and out of conversation and can be achieved through turn allocation or self-selection (Koole, 2013;Ten Have, 1999). This means that utterances are produced in succession, giving each speaker the space to participate. Learning, according to Gardner (2008), depends on giving recipients the opportunity to actively respond to conversation stimuli. Thus, response preferences are ways of responding to previous utterances such as answering questions, giving feedback, suggesting, supporting, agreeing, or disagreeing with a speaker (Van der Westhuizen, 2012b). According to Wu (2013:89), participants in conversation give some feedback to a speaker to show that they are interested in what the speaker is talking about. This observation concurs with Melander (2007, cited in Van der Westhuizen, 2015a:14) who claims that during conversation, participants contribute towards keeping the conversation going.
On the other hand, repair actions (self-and other-repair) are used to replace, insert, rephrase, or correct talk during interactions (Van der Westhuizen, 2012a). Repair is not necessarily correcting errors but could be used to assist a speaker having trouble searching for a word that captures the intended message (Hutchby & Wooffitt, 2008;Koschmann, 2013). According to Van der Westhuizen (2012b:147), "self-repair and other-repairs serve the purposes of clarifying understanding, developing shared understanding and knowledge." Hence, conversation analysis was adopted in this study to show how participating teachers constructed their own learning through professional conversation.

Research Question
The research question that informed this study was: how do teachers learn in professional conversation?

Research Design and Methodology
This study originally adopted an ethnomethodological research design. Given that the conversation was organised by the researchers, autoethnography becomes the suitable design. Auto-ethnography focusses on the dialectics of subjectivity and culture, and in general, entails the detailed analysis of oneself qua member of a social group or category (Allen Collinson, 2006:19). According to Dyll (2018:149), autoethnographers are expected to be involved in the construction of meaning and values in the social worlds they investigate.
Six primary school teachers were purposively selected from the East London Education District in the Eastern Cape, South Africa. The teachers were aged between 25 and 58 years, with a minimum qualification of a Bachelor of Education degree. Among them, the teachers shared teaching experience of between one and 35 years. Five of the teachers taught in the intermediate phase, while one taught in the foundation phase.
A video recorder was used to capture the conversation among the teachers. Video recording is a qualitative research instrument that captures moving images with or without sound to study the visual detail of interaction and behaviour (Gibson, 2008:918). The choice of video recording allowed the researchers to capture both vocal and non-vocal behaviour (Clayman & Gill, 2004:592) of the teachers for conversation analysis purposes.
The procedure used to capture the data involved video recording teachers while discussing issues around classroom discipline. The recording, which lasted for 31 minutes and 56 seconds, took place in the school staff room. The teachers agreed that the conversation and video recording could take place after official school hours for purposes of privacy and non-distraction. The six teachers sat in three rows facing each other during the conversation while one of them initiated the conversation. Another teacher from the school (computer teacher) recorded the conversation on behalf of the researchers in order to reduce the researchers' intrusion.
The video recording was viewed several times and transcribed verbatim. Three learning episodes, represented in Tables 1 to 3 below, were selected from the transcript and transcribed again using Jef-ferson's notations symbols (Jefferson, 2004:24) for conversation analysis purposes (see Appendix A for the Jefferson notations symbols explanations). Clayman and Gill's (2004:596) conversation analysis levels were used to analyse the three learning episodes identified as: requesting advice, testing ideas, and sharing ideas. The levels include a macroscopic level (what is happening in the conversation); sequence organisation (participants' ordering of talk/actions); singular actions (how the participants individually or in the group participate in the conversation); and a microscopic level (participants' choice of words, intonation, and non-vocal behaviours). Pseudonyms (T1, T2, T3, T4, T5, T6 and AT) were used to represent the teachers' names. T meaning Teacher and AT meaning All Teachers.

Ethical Considerations
Ethical clearance and permission to conduct this study were obtained from the University of Johannesburg (2016-057), the Eastern Cape Department of Education, and the East London District Office. A letter granting permission was received from the principal while letters of consent, which also permitted video recording, were obtained from the participants. The researchers fully disclosed the nature and purpose of the study to the teachers and participation was entirely voluntary. The participants' identities and responses were protected through pseudonyms, as indicated above.

Requesting Advice
The episode presented in Table 1 was selected to show how teachers learn through advice. T1 used his role as facilitator and requested advice from T4 on school discipline. and also have our frustrations that we feel (.) that we tend to take easy way out therefore sit down emotions should be left outside of any situation ((gestures with both hands)) 170 T1 Uhm 171 172

T4
You shouldn't think using your emotions but rather look at the problem first (.) then assess the problem 173 T1 Uhm ((nodding)) 174 T4 Find a way which will work both for you and the child in the classroom 175 T1 ( (nodding) In lines 157-160 in the episode above, T1, responding to T4's previous statement, requested his advice about teachers who complain that their hands are tied in terms of hitting learners because of the policy reform which forbids corporal punishment with overlap utterancecorporal punishment by T3 in line 161. T4, in lines 162-174, advised the teachers to leave emotions out when disciplining learners, and assess the problem to find ways that would benefit all (both the teacher and learners) in resolving the issue. The other teachers responded in agreement with this comment (lines 163, 165-166, 170, 173 and 175). In line 176, T4 acknowledged an earlier statement by T1 that discipline is democraticmeaning it is negotiable between the teacher and learners. The suggestion from T4 appears to have received an agreement response from T1 in line 178. T4, (lines 179-185), in self-repair, explained how teachers and learners should negotiate discipline in the classroom, with agreement responses from others in lines 181, 184, 186 and 187. After a pause in line 188, the advice was appropriated by T1, in a laughing mood with high tone -YOH WE ARE FULL (line 189)meaning that the advice was helpful, while the other teachers (AT) were laughing in agreement (line 190).
The episode in Table 1 shows how advice is requested and given during professional conversation. T1 requested advice from T4 on the policy reform that forbids corporal punishment at school. T4, in response, advised teachers to avoid emotions when disciplining learners but to rather follow a democratic method namely, negotiation (as was mentioned by T1), to establish a disciplinary approach suitable for all involved, with response preferences from other teachers. The advice was appropriated by the facilitator in a laughing mood and high tone, with other teachers laughing in agreement. The sequence in this episode comprised request/response, repair action and response preferences. Self-repair is evident in line 179. T4 rephrased his utterances as he tendered the advice. Response preferences from others ranged from uhm, yes, yeah, nodding, and laughing. These were support and agreement tokens used to assist the speaker. Nodding and laughing here display agreement and excitement.

Testing Ideas
The episode presented in Table 2 was selected to show how ideas are tested during professional conversation. Ehm: it is a challenge really to:: only verbally (.) discipline the learners (.) because some learners they take verbal discipline as a way of just warning them (0.1) ehm: only (.) when ((inaudible)) so↑ i would say↑ that verbal discipline for me (.) i can take it as a warning as well like the children perceive it (.) so that ultimately means there can be no corporal punishment ehm:: strategies that can be used to discipline children. 230 T1 ((nodding)) The episode presented in Table 2 was in the form of a question-answer-response sequence. T1, in agreement with the previous statement from T5, tried to test his idea on what the reason could be for teachers not showing support for the policy, which is against corporal punishment (lines 221-223). In lines 224 to 226, T5 responded that it was challenging to discipline learners verbally as the learners simply regarded such as just a warning. In lines 226 to 229, T5 further acknowledged that verbal discipline, perceived as a warning by the learners, does not carry any serious consequences. This idea was followed by a nodding response from T1 in line 230. T1 opened the sequence with an additional question seeking to understand the reason for teachers' lack of support of alternatives to corporal punishment. T5, in response, claimed that it was difficult to achieve discipline among learners without corporal punishment as they perceive verbal discipline as a simple warning. This response describes his own experiences, which tended to concur with those of the learners. T1, in response, acknowledged the idea with nodding, which served as a token of agreement.

Sharing Ideas
The final episode selected (Table 3) was to show how teachers shared ideas in professional conversation. Okay just to add on that (0.1) as teachers we need to follow our learners (.) just to make that thing:: we used to call it home visits because that's where you can get ahm:: the background of the child 352 T1 ((nodding)) T1 (lines 284-286) invited his colleagues to add to the discussion about classroom discipline. T5 (lines 287-288) noted that he could not verbally discipline a child over the same kind of … offence (repaired by T2 in line 289). This repair action (otherrepair) overlapped with T5's search of a word to complete his statement. T5, in line 290, acknowledged repair by repeating offence after he had mentioned behaviour, with a nodding response from T1 in line 291. He further suggested using another form of discipline other than corporal punishment once verbal discipline has proven to be ineffective and the misbehaviour continued. He, however, advised that teachers should first start with verbal discipline. As is clear from lines 292 to 301, other teachers appeared to agree with the suggestion. T1, in agreement, suggested more consequences as teachers cannot rely on only one method, but that verbal discipline should be the starting point. This elicited agreement responses from others (lines 302-307). T2 (lines 308-311) added that in instances where learners are disruptive, the parents are invited to a meeting to discuss the matter. Should the disruptive behaviour continue after that, the learner is asked to stay away from school for a week or so, but not longer than a week. In an overlapped repair, T1 assisted T2 with the correct term (suspension) for telling a learner to stay away from school for one week (line 312), which T2 acknowledged (line 313). In agreement, T5 added that parents needed to be actively involved during the suspension, as some learners might take advantage of suspension; instead of perceiving the suspension as a punitive measure, the suspended learner rather enjoys staying at home without adult supervision (lines 318-320).
In response to T5's comment, T2 added that learners were suspended for more serious offences, with T1 and T5 nodding in agreement (lines 324-325). T3, (lines 327-328), suggested that it was crucial to understand a learner's background before applying any form of disciplinary measure, which was subsequently followed by yes and nodding from T5 and T6 (lines 329-330). T3 shared a classroom experience of how, after a home visit, she stopped shouting at a learner who was in the habit of arriving at school very late (lines 331-335). She (T3) attributed most behavioural problems to the learner's background and reaffirmed her initial suggestion (lines 338-343), with ensuing agreement responses from other teachers (lines 336-337, 339, 341-342, 345-348). In addition, T6 added that teachers needed to follow up on their learners through home visits because such undertakings provided insight into the child's background (lines 349-351), with which T1 agreed by nodding (line 352).
In the above episode, T1 invited his colleagues to make some concluding remarks. In line with the invitation, T5 noted that he could not verbally discipline a child over the same kind of offence. He then suggested that other forms of discipline other than corporal punishment may be necessary, should the learner persist to misbehave after the verbal warning. T5 advised that the first option in disciplining a learner should be verbal discipline, which evoked agreement responses from other teachers. T1, in agreement, suggested more consequences as teachers could not rely on a singular method, but recommended verbal discipline as a starting point, with agreement responses from others. Sequence organisation took the form of invitation/responses, repair actions and response preferences. Response preferences in this episode were mainly agreement tokens (yes, yeah, exactly, uhm and nodding) in support of the others' ideas. Two other repair actions (lines 289 and 312) were noted. The first assisted the speaker in search of a word to complete his statement, while the second helped in providing the actual name for asking a learner to stay away from school for a week suspension. Both nodding and gestures in this episode served as non-verbal and paralinguistic actions.

Discussion
Based on the conversation analyses of the selected episodes, the findings indicate that teachers learn from requesting advice. One of the teachers, responding to this learning strategy, advised fellow teachers to avoid emotions when disciplining learners. However, negotiating with the learners on the best way to handle a particular misconduct seems to offer a more plausible solution to such learner misconduct. Requesting/giving advice in this context is seen as a way of learning among teachers with regard to classroom discipline. It is obvious that colleagues could learn more from conversation through requesting advice for work-related challenges. This shows the importance of professional conversation and how an action could initiate advice and/or learning. This finding is in agreement with Van der Westhuizen (2015b:134) who claims that learning is facilitated by conversational moves of assessing, asserting, requesting, and accounting for views. The finding also confirms an earlier claim by Wood (2007:289) that "professional conversation offers teachers a number of learning opportunities … to draw on expertise … in addressing problems and questions of common interest." The finding shows that testing ideas is a learning strategy in professional conversation. One of the teachers was tested on the policy shift from corporal punishment to alternatives to corporal punishment. The teacher, in response, expressed difficulties achieving discipline among learners after the abolition of corporal punishment. As a practising teacher living with the challenges, he viewed verbal discipline as just a warning that does not carry any consequences. Ideas are tested in educational settings mainly to receive opinions regarding policy reform, proposals, events, or experiences. This conversation provided teachers with the opportunity to meet and hear about colleagues' experiences and perceptions about discipline policy. Miller (2008:78) argues that professional conversation provides opportunities for teachers to learn about the best and tested practices from their experienced peers. A study by Horn (2007) demonstrates that collegial talk can be a huge source of influence on teachers' approaches to classroom practice and opportunities that shape their responses to challenging educational reforms.
The findings from this study show that teachers learn from sharing ideas. The conversation in which teachers discussed disciplinary challenges in South African schools offered teachers with a learning opportunity. All participating teachers in the study acknowledged that learner discipline was a huge problem in their school. They initiated and shared numerous ideas and suggested ways of improving their disciplinary practices. Some of the important ideas shared by the teachers include knowing learners' backgrounds, visiting learners' homes, making verbal discipline the starting point before adopting other forms of discipline, and avoiding emotions when disciplining learners. These findings confirm those by Wood (2007:290) who claims that teachers' learning involves shouldering of responsibilities by teachers themselves to carefully inquire into their present practices, reflect on what they have learned from experience, and engage in conversations with one another. The findings also relate to Msomi et al. (2014:800) who conclude that professional learning ingredients include deliberate reflection, inquiry, and sharing insights.
The findings show that teachers respond to requests, questions, tests, and invitation. They use utterances such as yeah, yes, uhm, and exactly to support others' opinions. These talking/learning strategies are called response preferences in the conversation analysis. According to Wu (2013:89) participants involved in conversation have to give some feedback to the speaker to show that they are interested in what the speaker is talking about. Alluding to this, view Melander (2007, cited in Van der Westhuizen, 2015a:14) maintains that during conversation, participants contribute towards keeping the conversation going.
Findings in the current study show that teachers assisted each other through repair actions. Both self and other-repairs were noted in the conversation. Self-repair was performed by speakers, while other-repair was undertaken by the listeners. T4 rephrased his utterances as he spoke (see Table 1). T2 used other-repair to assist T5 who was struggling to complete his statement, helping to locate the appropriate wordoffence (see Table 3). T1 employing other-repair, assisted T2 to name the process of telling a learner to stay away from school for one week, namely suspension (see Table 3). Repair is not always necessarily correcting errors; it could be assist-ing a speaker to recall a specific term (Hutchby & Wooffitt, 2008;Koschmann, 2013). Repair is also used to replace, insert, rephrase, or correct talk during interactions (Van der Westhuizen, 2012a).
Similarly, nodding and laughing were used repeatedly by the teachers in this study as back-up to the conversationways of expressing one's mind without being vocal. Both nodding and laughing display agreement and excitement during a conversation, which Clayman and Gill (2004:596) classify as non-vocal behaviour. Nodding is used globally, according to Veldhuis (2006), to express understanding, agreement and approval, while laughter is perceived as positive feedback that shows joy, acceptance, and agreement (Petridis & Pantic, 2011:217).

Conclusion
The study concludes that teachers should engage in professional conversations to solve numerous discipline-related issues in schools. The findings show that the study created opportunity for teachers to come together to review discipline challenges and their teaching practices. It is obvious that teamwork in the workplace and group participation is necessary in solving classroom problems. By requesting advice, testing ideas, and sharing ideas in the workplace, colleagues learn from each otherespecially from the more experienced ones.

Recommendations
Based on the findings of this study, the following recommendations are made: 1) We recommend that teachers embrace professional conversation for exchanging knowledge and experiences for learning purposes. This calls for teachers to make professional conversation part of their continuing professional development. 2) We encourage teachers to adopt aspects of the conversational strategies highlighted in this study for professional learning purposes. 3) Research experts on teacher learning should be involved in workshops at schools to further enhance teacher learning in specific areas. This calls for effective policy formulation that ensures the need for partnership between teachers and education experts in solving problems, such as managing school discipline, through continuous teacher learning in practice with assistance from relevant experts.