‘ Go bolela , go a shikinya ’ – Shaking utterances in learning interactions

This article is an inquiry into how talking is used for learning. The focus is on utterances of significance where participants say something which brings some sense of surprise and cognitive dissonance, and the purpose is to develop an understanding of how such ‘shaking utterances’ contribute to learning. The study is conducted from a social interaction theory perspective and utilised conversation analysis methods to observe how such utterances come about, how they are sequentially organised, and how they contribute to learning. Findings indicate similarities in the origins and learning consequences of shaking interactions. The study demonstrates the value of conversation analysis research methods for the deepening of our understanding of the nature and learning benefits of talk in classroom settings.


Introduction
Interruptions in classroom interactions occur fairly regularly, and may include teachers changing course of actions for pedagogical purposes, or children requiring attention as result of a social or learning need.While educational interactions are ordinarily determined by the pedagogic intent and actions of the teacher, the flow may be interrupted by learners who would want to change the topic, steer the conversation in a new direction, attract attention to self/another, or want to solve a problem.
In this article we are concerned with interruptions in group interactions, the purposes they serve, and how they are organised socially.We take a conversation analysis perspective and consider utterances of significance in group interactions, i.e. utterances which typically break or change the flow and act as intervention with associated learning consequences.We want to explore how teachers and learners use talking as shaking utterances, how they prompt reaction and how such interruptions serve purposes of learning.
Shaking utterances are conceptualised here as Go bolela go a shikinya, which in Sesotho literally means "to speak, it shakes", implying that speaking makes things happen or moves things into action (M Monareng, pers. comm., 2012).These would be utterances such as those in everyday conversations, which indicate and create surprise, emphasis, confrontation, and made with the intention to change the flow/process of a conversation.This Sesotho interpretation is also universally relevant, in the sense that it draws the attention to the possibility of classroom utterances fulfilling functions of interruption and of movement.
We assume that in teacher and group interaction settings such utterances occur frequently, and that they are made by teachers and learners alike.For this study, we assume that such interventions have pedagogical and social intentions: they create dissonance and may or may not lead to learning gains in various forms.We also assume that it is characteristic of educational interactions that such talk is used to problematise, confront views, develop arguments, inspire, and, in terms of Piagetian (1977) theory, create cognitive dissonance.
Our purpose here is to identify examples of go bolela go a shikinya, as interventions in learning conversations, and to explore and describe how they contribute to learning.We focus on interaction patterns before, during and after such talk, and explore the associated learning gains using conversation analysis (CA) methods.We ask the question: what is the nature of shaking utterances in learning conversations?And: how are sequences of interaction structured/organised before, during and after shaking utterances, and how are they consequential for learning?We argue that a better understanding of these interactions will help teachers to support learning in the classroom and allow for deeper learning to take place.

Conceptualising Learning Conversations
Shaking utterances in learning conversations are part of a discursive process, where active participation involves the use of language and semiotic tools as instruments of learning and communication (Kumpulainen & Wray, 2002:35).It "… is an accountable, public and locally occasioned process which involves interaction as evidence of learning, and interaction as the place where learning is to be found" (Koschmann, 2013(Koschmann, :1039)).As such, the goal of learning through interaction is collective meaning making, shared understanding, and enculturation into practices, discourses and norms of the community (Kumpulainen & Wray, 2002:146;Magano, Mostert & Van der Westhuizen, 2010).
Learning conversations have in common that they pursue purposes of learning relevant to a particular topic, with the teacher being the participant with the 'epistemic authority,' and the institutionally determined plan/agenda to take the conversation to a particular, mostly predetermined, end (Mercer, 2010); see also Edwards (2006).Learning conversations unfold situationally, contextually, and discursively (Edwards & Mercer, 2012;Pike, 2010), which means they are authentic and unique.As such, learning conversations have pedagogical aims, but often lead to educational ends that are not always pre-determinable.
Learning conversations are intersubjective in nature (Pike, 2010:163-164;Stone, 1993).This means participants make inferences from what another speaker said in a prior turn, recreating the presuppositions underlying the utterance.When successful, this is an indication that learning has occurred (Stone, 1993).Pike (2010:165) noted however, that it is difficult using Conversation Analysis methods to claim "… that learning has 'actually' occurred; it can only ever seek to specify the conditions of talk in interaction that participants themselves orient to and treat as evidence for it." The intersubjective nature of learning conversations also needs to be understood as contextualised, joint activities where teachers and learners interact in what Mercer (2008) calls the interactional development zone where participants rely on presuppositions about the utterances made in the conversation (Mercer, 2000, quoted by Pike, 2010:164).Sequences of utterances constitute a display of intersubjectivity with regard the what-todo-tasks in the interaction, with specific repair actions as display of learning (Pike, 2010:178).For example, when participants talk about a text, their learning of new word meanings is evident in how they share understandings and offer repair of misunderstandings.
In school learning interactions, participants treat knowledge as a moral domain, and may, based on Stivers, Mondada and Steensig's (2011) distinction, be described in terms of three dimensions: epistemic access, primacy, and responsibility.The word epistemic comes from the Greek word epistēmē, and is the word for 'knowledge' and the ability to know or understand (Merriam-Webster.com, 2011).Epistemic access is defined in terms of a) knowing vs. not knowing, i.e. one participant knowing more or less than the other; b) the degree of certainty of knowledge expressed, i.e. a participant indicating what s/he knows about the topic of conversation, but doing so with some hesitation; c) access to knowledge sources, i.e. indicating where the knowledge was gained, and d) the directness of knowledge, i.e. the way of expressing knowledge.Epistemic primacy concerns the relative rights of participants to know, to claim knowledge, and to state their relative authority because of their knowledge.Epistemic responsibility concerns the way in which participants design their turns to exercise their responsibility to the other (Stivers et al., 2011:9; see also Heritage, 2012 andHeritage &Raymond, 2005).
The way knowledge is used in learning conversations is guided by institutional norms.These include roles and status, where the teacher guides the interaction and to the point that learning is displayed conversationally, and appropriated by the teacher.This involves some 'mutual stance' and conversational markers such as: 'that's right' to indicate shared understanding and aligning self with action(s)-in-progress (Barnes, 2011), based on epistemic access and rights.It is assumed therefore, that learning conversations are not static, and that 'topic transition sequences' (i.e., sequences consisting of utterances following one another) may also serve as indicators of (learning) progress, according to Drew and Holt (1988).Koole (2010) makes the distinction between displays of understanding and displays of knowing in interactions.He describes three conversational contextsa discourse unit understanding, which indicates understanding of what is said, a question sequence eliciting utterances that show 'having known,' and question sequences aimed at knowledge production.The latter contains utterances that show that a participant has acquired access to a correct answer 'here and now' (Koole, 2010(Koole, :207, 2012)).
Learning, from a conversation analysis perspective, is conceived of as happening in a micro-context, with knowledge demonstrated by means of a claim for a correct answer.Learning conversations therefore seem to include utterances/ sequences of knowledge appropriation where the teacher typically acknowledges statements as correct and relevant (see also Pea, 1993).For Paulus and Lester (2013), learning is observed in claims of change of state, ranging from extreme/explicit to denying a change of state and a neutral assessment of distancing self from a position taken.Wickman and Östman (2002) describe learning in interactions as discourse change, related to meaning.Meaning is constructed from differences and similarities in what is immediately intelligible when we act in an interaction (Wickman & Östman, 2002:603).Learning and knowledge are part of a dynamic process in human encounters with others and the world.In these encounters, a la Wittgenstein's notion of language games, meaning is evident in the rules of the interaction/language game (Wickman & Östman, 2002:604).
Finally, the Koschmann (2013:1039) review of literature on learning in interaction advances the notion that learning is conceived of as change in activity which requires going beyond judgements of regularity.Interactionally, "… learning is developmental change within our familiar and recurring activities," and embodied in the methods used by participants i.e., Garfinkel's (1967) notion of members' methods "… to detect and display changes in the ways their joint activity is organised" (Koschmann, 2013(Koschmann, :1039)).

How Talk Can Intervene -Shaking Utterances in Learning Conversations
'Shaking utterances' may be observed in social conversations in various forms, doing the work of announcing, requesting, claiming, correcting, and so on.They are perhaps mostly unexpected, and made for different purposes, i.e. to display curiosity, opposition, dissonance, or dissatisfaction.Such talk are also indicators of asymmetry between two speakers; as well as a sense of unevenness or irregularity (Sidnell, 2012).
In learning interactions, we would assume that talking that shakes is interventionist in nature, serving pedagogical purposes of cognitive dissonance, which, according to Piaget's theory, creates disequilibrium, leading to learning (Rogoff, 2008).Other pedagogical purposes include the creation of curiosity, and motivation to learn (Järvelä, Järvenoja & Veermans, 2008), and internal dialogue (Vygotsky, 2012).Such interventionist utterances also serve social purposes in contributing to stance taking (Kumpulainen & Mutanen, 1999), and the maintenance of a conversation (see Clark, 1996).In addition, shaking utterances may also serve the purpose of claiming authority.For example, epistemic authority is claimed when participants assess the "state of affairs' in the interaction and then 'index' their independent opinion, often as a 'first position assessment'" (Heritage & Raymond, 2005:16).
Because of the institutional norms governing learning conversations in a classroom setting, allowing the teacher to maintain authority, one would expect students/learners to make use of shaking or interventionist utterances as a way of shaping the conversation, and of claiming their authority.They may also be about gaining access, which involves the use of interactional resources.For example, where a participant poses a question, the questioner presupposes recipient access and willingness to answer (Stivers et al., 2011:10-11).
Shaking utterances furthermore display social norms of alignment and affiliationthey influence conversations, especially when interactants "show themselves to be accountable for what they know, their level of certainty, their relative authority, and the degree to which they exercise their rights and fulfil their responsibilities" (Stivers et al., 2011:9).They may also be a way of exercising a right to know and to claim what participants know (i.e.epistemic primacy) as indication of how participants orientate themselves to the asymmetry in the interaction.Participants exercise their epistemic rights depending on "relational closeness" and their own sense of having sufficient knowledge and authority (Stivers et al., 2011:14).

This Inquiry
The purpose of this inquiry is to analyse examples of shaking utterances in a group interaction session in the micro context of the sequences they occur in, in order to understand how their sequential organisation contributes to learning.We collected examples of go a shikinya type utterances in a learning conversation with adolescents on the role of gender in violence.A group of seven adolescent boys in Grade 12 participated voluntary in a learning conversation on gender-based violence.This session was part of the school programme aimed at addressing gender discrimination, deemed necessary by the teacher given the gender mixed composition of the school.The session was set up by a teacher, who was the school counsellor at a private school in the affluent northern suburbs of the city Johannesburg, as an educational session to help clarify views about violence (Bachrach, 2010).Approval for this research was provided by the relevant ethics committee as well as the participants in the study.
The session started with the group listening to the song "Kim" by Eminem, after which the facilitator invited them to talk about the question: "What role does gender play in violence?" In the analysis we describe examples of shaking utterances in the micro-context of sequences.We drew on CA analytic principles from the Epistemics in Interaction Framework by Stivers et al. (2011), which includes epistemic access, primacy, and the uptake of epistemic responsibility.
In particular, we looked at sequence organisation and the response preferences of participants building up to and following shaking utterances, with a view to understanding the turn organisation and what learning outcomes may be associated with such utterances.We identified shaking utterances according to the criteria of topic change (an utterance changing the direction and introducing a new topic), content novelty (an utterance which introduces a topic which departs from the topic at hand), and response preferences of surprise following the utterance, such as prolonged silence and/or gestures of uneasiness.

Analysis and Findings
The transcription in Appendix A is the full transcription of the interaction session observed, containing at least three examples of shaking utterances that meet the set criteria, to a greater or lesser extent.For the ease of reading, we insert in this analysis section three episodes which contain the shaking utterances, as tables 1, 2 and 3.
In Table 1, in example 1 (line 20), participant ((Laugh/breath)) [Well we live in a society that perceives (.) that perceives it to be a male dominated society (0.3) like consciously or subconsciously even a the females would would probably agree that it's a uh male dominated 9 F: =umm 11 12 P 1: Even people who would say they aren't sexist (.) probably um (.) are sexist without knowing it (.) just by maybe underestimating a woman or uh uh overest.. overestimating a man 13 14 15 16 17 18 P 2: It.it It.I think its integrated so deeply into our society that its ((Rubs nose)) that its almost its almost (.) its in our subconscious (.) even though people are aware of it but they're also not aware of um it's in everything we do (.) its in the jobs its in jobs (.) there's stereotypes um its in the way (.) its in way it's the way that people live their lives that um gender stereotypes influence the way they do things The utterance by P3 follows the Facilitator's solicitation question "What is the main perception about gender?," and P1 and P2 not answering the Facilitator's question directly.In response to the Facilitator's original question, in lines 2 and 3, P1 and P2 talk in general terms, referring to perceptions of society (lines 6 and 7), asserting that people don't know when they are being sexist (lines 11 and 12).When the Facilitator (F) repeats the question, the response by P3 is direct, and may be taken as a 'shaking utterance,' offering an assessment of views, and refocusing the conversation.The shaking done by this utterance seems evident in the silence in 21, the question by F in 22 wanting to confirm what P3 said, and the overlap/immediate response made at the same time in 23 by P1, P2 and P4.The latter is a response preference of repair, made together by a few participants, claiming that men are dominant and not violent, repeated in 24 and extended by P2 in 25.The sequences that follow in 29 to 38 involve F pursuing and extending the meaning of the original utterance in 20.
The interaction before, during, and after the shaking utterance clearly show causes and consequences.The prompting questions by the teacher seem to have had a pulling and challenging effect.P3 could not resist the invitation and preferred to make a strong statement.The latter, in conversation analysis terms, can be taken as an assessment of what is going on in the interaction around the teacher's questions.At the same time, it reflects an explicit claim of a point of view, which does the work of changing the tone of the talking.
The conversation segment in Table 2 is a direct follow-up of the segment in Table 1, and shows how the conversation continued.Here, in lines 46 and 47, another example of a shaking utterance is found.
The utterance is by P2 in lines 46 and 47, followed by extensions in 50 to 53.Here, P2 introduces the words 'wolf pack' to the conversation, challenging participants to consider a new line of conversation, and followed by some hesitance in responses.In response to F's question in 41, repeating the question about the role of gender in violence, P2 says that here, when someone feels dominant ( 42) arrogant (42), superior (43) and with "an aggressive attitude" (45), "it's like a mirror of what you see […] in the animal world […] the 'wolf pack."The giggle response in 48 and F's "Uhmm" in 49 is followed by an extended explanation of the behaviour of "the alpha dog."This is followed by the response preferences of P6, P5 and P1 supporting and confirming P2's shaking utterance.Learning seems to be reflected in the varied responses to the initial utterance, all leading to the Facilitator confirming in 79 that stereotyping needs to be challenged.And (.) and then there's the alpha dog and (.) the alpha dog if someone wants to eat before the alpha dog they know they can't and if they try the alpha dog will put the put the other the other wolf in place (.) alpha wolves so ahm so I think I think that its (.)I think that it could also (.) be linked with the effect (.) the effect that gender that gender has on uh on violence 54 P 6: [Women being subservient they would receive (.) they would receive violence.55 P 2: [Be put in place 56 P 6: Ya= The finding with regard to the second example of a shaking utterance is that teacher talk of prompting by means of a question (from line 40), is followed by one learner making two utterances (42 and 44), prompting P2 to introduce the idea that the occurrence of violence can be associated with the behaviours of a "wolf pack" in the animal world.This utterance was extended by P2 in line 50 onwards with further explanation of what the alpha dog does in a wolf pack.
In the segment directly following this one, the conversation is extended in a way that takes the impact of utterance in 46 even further.
In Table 3, the third example of what may be deemed a shaking utterance, is made by P5 in lines 66 and 67.Here, P5 makes a claim in response to exchanges of turns among three of the participants on the topic of violent acts of men and women.He says "… We see a man killing another man, not so much a woman killing another women," an utterance made after there have been some exchanges among the learner participants from 56 to 65, followed by the claim by P6 in 54 that "Women […] would receive violence", and in 57 and 58 "… not being able to take control […]."The shaking utterance can be seen as an intervention that confronted and challenged others to consider a new idea.The response preferences include utterances and gestures of uneasiness.Learning was appropriated by the facilitator in line 79 when she said that stereotyping needed to be challenged.
The finding of the analysis of shaking utterance 3 in 66 and 67 is similar to the previous examples.In this example the utterance, which does the work of prompting, takes the form of a claim by P5 in 58 about how gender roles are reversed.The shaking utterance itself was followed by responses to which P5 using turns to further explain her claim.The Facilitator used these explanations to prompt in turn 79 for final conclusions on the question of the conversation about gender-based violence.

Discussion
The three examples of shaking utterances seem to have some similarities: the utterances themselves are assessments of previous turns, and represent claims that prompt gestures such as giggles, and silence are followed by turns of surprise.Preceding these utterances, we have either the facilitator prompting, or an exchange of ideas by different participants.The utterances are followed by repair extensions and elaborations.These utterances are indicative of the 'movement' following shaking utterances.
In the micro-context of all three of the shaking utterances, the presence of the conversation facilitator is distinct.It was the Facilitator's questions that led to the episodes within which the shaking utterances occurred.The frequent reminder of the question as the purpose of the conversation indicates the epistemic authority of the Facilitator who, after her initial silence following the shaking utterance, exercised her authority by bringing the focus back to the conversation question.
The finding that such utterances follow exchanges that invite engagement and reaction was to be expected.The shaking utterances themselves are social actions of intervention, changing the topic and direction of the interaction.Such changes in topic orientation may be taken as a form of learning, as has been shown in international studies (Melander & Sahlström, 2009).They are also presented as interruptions, as defined in international literature by Sawyer and Berson (2004), which clearly invite response from other participants, and clearly offer assessments of the pre-ceding sequences, an action also serving the purpose of mobilising response (Stivers & Rossano, 2010) and enriching the conversation, as well as challenging perspectives by intervening in the flow.I mean the same situation could've happened and the man cheated on the woman (.) she wouldn't be able to >to take control of the situation in such a way that she would actually kill the man< (.)I mean the woman that the man ((Inaudible)) [I mean the woman the man cheated.I think thats so deeply ingrain::ned that men should be in control that the male is the dominant one (.) and that women that women have kind of (.) kind o::f accepted that they are to be ( The shaking utterances identified may furthermore be understood as claims of knowing made by participantsthat they have views that are different, or, in Heritage's (2012) terms, that they see themselves as being more knowledgeable.As such, they are displays of epistemics-in-action and used by participants to make claims and contribute to the flow of the conversation (see Heritage, 2012).The utterances also display epistemic access the way they show certainty and directness, uptake of responsibility, and participants exercising their relative right to tell and inform the group of the knowledge they have (see Stivers et al., 2011).
Interrupting utterances display participants' uptake of epistemic responsibility, adding their views to the conversation, and participating further in the conversations that follow.The teacher ignoring could serve the function of restoring asymmetry/epistemic authoritya finding which is only tentative.The level/extent of shaking utterances is determined by responsessometimes ignored, sometimes acknowledged.

Conclusion
The focus of this inquiry was on classroom talk that brings movement into an interaction.Our analysis allowed for some exploration of interactions around shaking utterances, and can be noted as limited in scope.The findings and discussion seem to highlight the use of shaking utterances as a distinct part of the classroom interaction pattern, with clear social and learning consequences, supporting the value of the 'Go Bolela' metaphor to this research filed internationally.At the same time, it emphasises the importance of learner actions in classroom learning of Life Orientation topics (Frantz, 2015).
It is reasonable to conclude that sequence organisation around shaking utterances displays some similarities which warrant further inquiry into forms of shaking utterances among learners/ students themselves in group learning conversations and cooperative learning settings, where participants play the role of facilitator.Further research into the pedagogic implications of shaking utterances may also be valuable.

Note
i. Published under a Creative Commons Attribution Licence.

Table 3
Shaking utterance 3: "… women not attacking …" Exactly↑ It's how it's how religion it is religion that what is the prime↑effect is the prime is the prime cause of war today 96 F: Okay I want to bring you back to gender how is gender↑ gender and religion possibly linked to violence?