‘Yes? I have no idea’: teacher turns containing epistemic disclaimers in upper primary school whole-class discussions

ABSTRACT

Data from whole-class discussions in Dutch upper primary school show that teachers occasionally explicitly take downgraded epistemic stances through epistemic disclaimers such as ‘ik weet het niet’ (English: I don’t know (it)), which contrasts with their institutionally assigned epistemic authority. In the current study, we have collected turns in which such epistemic disclaimers occur, and analysed them using conversation analysis. In our analyses, we focused on the positions of the turns in which epistemic disclaimers occur, and on the varying ways in which these turns influence the subsequent course of interaction. We have found that teachers’ epistemic disclaimers occur in initiating turns, facilitating student participation, but also in responsive turns. The latter vary in the extent to which they facilitate participation, ranging from facilitating student participation in a similar way to the initiating turns, to blocking further student contributions altogether. This study furthermore demonstrates that teachers employ epistemic disclaimers to navigate two teacher roles, namely those of a teacher with epistemic authority, and of a facilitator of whole class-discussions.

KEYWORDS:

• Epistemic stance markers
• whole-class discussions
• conversation analysis
• teacher roles
• epistemic disclaimer

1. Introduction

Over the past decades, conversational approaches to teaching have received an increasing amount of both academic and societal attention. With the shift towards conversational approaches in the classroom, teachers and students engage in a different participation framework (Goffman 1981), where speakership is more evenly distributed and knowledge is not always simply provided by the teacher but also constructed by teacher and students collaboratively.

Within such a framework, the teacher moves beyond assessing students’ knowledge (Mehan 1979) and offering students new knowledge (Koole 2010; Macbeth 2003) as expected from the teacher having institutionally assigned epistemic authority (Heritage and Raymond 2005). Although teachers are the ones assumed to be more knowledgeable (K+), they are also found to offer students room for collaborative construction of knowledge or for using existing knowledge in classroom interaction (e.g. Barnes 2008). In doing so, teachers may be found to take a less knowledgeable (K-) epistemic stance (Heritage and Raymond 2012) by producing an epistemic disclaimer (Lindström, Maschler, and Pekarek Doehler 2016; Lindström and Karlsson 2016). However, specific research on teachers’ K- epistemic stance-taking has been scarce (but see Houen et al. 2019 for an analysis of teachers’ use of I wonder …). This is in contrast with an extensive body of work on epistemic disclaimers in other institutional interaction (e.g. Beach and Metzger 1997; Drew and Drew 1992; Potter 2004; Hutchby 2002; Jager et al. 2016; Lindström and Karlsson 2016) and ordinary interaction (e.g. Helmer, Reineke, and Deppermann 2016; Keevallik 2011; Pekarek Doehler 2016; Schegloff 1996; Tsui 1991). Since epistemic disclaimers are expected to contribute to a more conversational framework in classroom interaction, there is a need for studies investigating teachers’ downgraded epistemic stances in educational settings.

The current study presents epistemic disclaimers produced by teachers in a context where a participation framework of discussion and co-construction is being encouraged, and focuses on the different interactional contexts they occur in and the ways in which they influence the subsequent course of interaction and the participation framework along with that. As we will demonstrate in this paper, teachers’ epistemic disclaimers appear in both initiating and responsive turns and have encouraging as well as constraining effects on students’ contributions.

2. Theoretical background

2.1. Teacher practices in non-teacher-fronted classroom interaction

In recent decades, conversational approaches to teaching have received an increasing amount of attention. Within these approaches, which are often contrasted with more traditional teacher-fronted approaches, teachers are encouraged to step away from being at the centre of the interaction, and to invite students to participate and exchange knowledge (Durden and Dangel 2008; Howe and Abedin 2013; Massey 2004; Mercer and Dawes 2014). This extends students’ interactional role to using knowledge (Barnes 2008), rather than only displaying or gathering knowledge. Within non-teacher-fronted approaches, teachers employ varying practices that facilitate student participation.

Student participation is for instance facilitated when teachers expand the well-known initiation-response-evaluation (IRE) structure (Mehan 1979). By employing authentic questions or giving their own view concerning challenging issues and topics in initial position of the IRE structure, teachers might open up the floor for students to present their own ideas (Clark et al. 2003; Myhill 2006; Nystrand 1997; Nystrand and Gamoran 1991). In third position, elaborating on responses instead of closing the sequence might create opportunities for extended participation as well (Cazden 1988; Lee 2007; Nassaji and Wells 2000; Tabak and Baumgartner 2004; Wells 1993). By opening up the IRE structure and encouraging students to construct knowledge together, a Teacher-Student-Student-Student (T-S-S-S) turn-taking pattern is realized.

A T-S-S-S turn-taking pattern entails a different participation framework (Goffman 1981) than lecture-like teacher-fronted approaches. Such a framework has previously been described as a discussion framework (Gosen, Berenst, and De Glopper 2009, 2015) in which students participate through self-selection, respond to each other, and construct knowledge together. Using micro-analyses of interaction, several studies describe specific interactional practices that teachers use in offering opportunities for extended student participation (Gosen 2012; Willemsen et al. 2018, 2020a, 2020b, 2021). It is, for instance, described how different open invitations lead to different responses of students (Willemsen et al. 2018) and how avoidance of explicit evaluations leads to students reasoning in interaction (Gosen 2012). In addition to teachers’ interactional practices, attention has also been paid to the sequential details of moments in time in which students take initiative in classroom interactions (e.g. Duran and Sert 2021; Kardas Isler et al. 2019; Solem 2016). The current study aims to contribute to this field of research that sheds light on interactional practices that contribute to the realization of a discussion framework. It does so by focusing on one specific interactional practice: the use of epistemic disclaimers.

2.2. Epistemics and classroom interaction

Recently, there has been a growing interest in epistemics in classroom interaction (e.g. Heller 2017; Jakonen and Morton 2015; Kääntä 2014; Koole 2010, 2012; Rusk, Pörn, and Sahlström 2016; Sert 2013, 2015; Sert and Jacknick 2015). The field of epistemics makes a distinction between epistemic status and epistemic stance (Heritage 2012a, 2012b; Stevanovic and Peräkylä 2014). Epistemic status refers to the knowledge interlocutors hold, which can be positioned on an epistemic gradient that displays their epistemic statuses in relation to each other as being more knowledgeable (K+) or less knowledgeable (K-) (Heritage and Raymond 2012). However, knowledge positions are not fixed, but rather locally constructed and negotiated in social interaction. The knowledge position taken in interaction is the interlocutor’s epistemic stance, which is conveyed through ‘different grammatical realizations of propositional content’ (Heritage 2012a, 377), or the use of specific epistemic stance markers, such as the epistemic disclaimer I don’t know.

For everyday talk-in-interaction the use of these epistemic disclaimers has been studied before. These studies have mainly investigated their sequential environment and interactional functioning. Uses of I don’t know and its equivalents in other languages such as French (Pekarek Doehler 2016), German (Helmer, Reineke, and Deppermann 2016), Hebrew (Maschler 2017) or Estonian (Keevallik 2011, 2016) are often studied when occurring in second position, functioning as non-answers (Stivers and Robinson 2006) or prefacing trouble with the requested action (Beach and Metzger 1997; Helmer, Reineke, and Deppermann 2016; Keevallik 2011; Pekarek Doehler 2016; Schegloff 1996; Tsui 1991). Besides the epistemic functioning, several studies show the use of I don’t know and its equivalents as being more turn-organizational, for instance doing topic-closing (Aijmer 2009; Beach and Metzger 1997; Keevallik 2016; Pekarek Doehler 2016). Epistemic stance markers in first position have been studied less often but are found to be used in multiunit turns pointing forwards (Kärkkäinen 2003; Weatherall 2011). These markers therefore seem to work on speakers’ own turns and actions rather than on the sequential activities started by another speaker.

The use of epistemic disclaimers has also been investigated in different institutional settings, such as interviews (Potter 2004), counselling or therapeutic consultations (Hutchby 2002; Jager et al. 2016) and courtrooms (Beach and Metzger 1997; Drew 1992). These studies showed that in institutional interactions, responsive I don’t knows are also often used in utterances that reveal trouble with or resistance to the course of action that is deployed. Comparable findings are described for classroom interactions (e.g. Hosoda and Aline 2021; Ingram 2020; Sert 2013, 2015; Sert and Walsh 2013) in which epistemic disclaimers by students have been the main centre of attention.

The strong institutional epistemic asymmetry in classroom interactions makes that one might not primarily expect teachers to claim insufficient knowledge. Teachers are assumed to be more knowledgeable (K+) than students, which is also reflected in their epistemic stance-taking, through for example the often-employed known information questions (KIQs, cf. Mehan 1979). The third, position evaluations that follow the answers to a KIQ further evidence the teachers’ knowing epistemic status and function to re-establish the teacher’s epistemic authority. However, there are some studies that touch upon teachers’ K- epistemic stances. An example of this is a study by Skogmyr Marian, Henricson, and Nelson (2020) in which writing counsellors in academic counselling sessions claim insufficient knowledge in relation to their assessments of the student’s writing because of students’ primary access to the assignments and the subject area.

Houen et al. (2019) are, to our knowledge, the first to study teachers’ use of epistemic disclaimers in classroom interaction. With their study of ‘I wonder … ’ formulations, the authors argue that this interactional device facilitates student participation, as it elicits children’s displays of knowledge through inviting them rather than expecting them to respond. Notwithstanding the interesting findings reported by Houen et al. (2019), it remains unknown how other, more explicit epistemic disclaimers are used by teachers. Our data of whole-class discussions around texts show many instances of explicit epistemic disclaiming by teachers. Studying this phenomenon builds our knowledge of the functions of teacher conduct in whole-class discussions and offers more insight into how teachers facilitate and manage a discussion framework, as well as how epistemic stance-taking is managed in whole-class discussions.

3. Data and method

In order to gain detailed insight into the occurrences and uses of teachers’ epistemic disclaimers in whole-class interaction, we applied Conversation Analysis as our method of research. Conversation Analysis (CA) focuses on studying actual practices as they occur in natural conversation, by analysing observable interactional conduct (Maynard 2012; ten Have 2007). In the context of the current study, CA allowed us to uncover the ways in which teachers’ epistemic disclaimers are shaped and influence the ensuing interaction.

3.1. Data

The data used for this study comprise part of a dataset gathered during an overarching research project on whole-class discussions around texts (Willemsen 2019). For the current study, the data consist of video recordings of 11 history lessons. These recordings were made in two different classrooms with different teachers in the sixth grade of Dutch primary school in the north of the Netherlands. The students were 9 to 10 years old. The third author of this paper was present at the lessons to make the recordings. The average length of the lessons was 48 minutes, totalling approximately 9 hours of recordings.

The history lessons in this dataset were set up as whole-class discussions around texts. This type of classroom interaction is not common in Dutch primary schools, where history lessons usually consist of reading texts in a textbook and answering questions in the accompanying exercise book afterwards. In order to be able to study teacher conduct during whole-class discussions and the effect their conduct has on the ensuing interaction, the teachers in this study were therefore explicitly instructed to organize their history lessons in another way. While still making use of the curricular texts, the teacher and students had conversations about the texts that revolved around a discussable question (e.g. what do you think it was like for parents that their children now went to school instead of working in the factory?). Such questions without a readily available right answer provided room for different views and ideas and hence for reasoning, to come to a joint answer at the end of the lesson.

The instructions given to the teachers were based on rather general recommendations in the literature (e.g. Cazden 1988; Myhill 2006; Soter et al. 2008): the teachers were encouraged to provide students with ample opportunities to take the floor and respond to each other instead of dominating the discussion themselves by acting as a primary respondent (see McHoul 1978). The teachers were not explicitly encouraged to produce epistemic disclaimers or any other specific practices. The lessons were unscripted, and the teachers were free to implement the instructions in their own manner, and indeed did so. The data can hence be considered as more or less naturally occurring, even though the whole-class discussion character was explicitly encouraged.

3.2. Transcription and analysis

The video recordings of the lessons have been transcribed according to Jeffersonian transcription conventions (Jefferson 1986) supplemented with transcription of bodily conduct (Mondada 2018, see appendix) when relevant for the analysis. The Dutch data have been translated into English, while staying as close to the original interactions as possible. Scrutiny of the dataset resulted in a collection of 25 instances of epistemic disclaimers produced by the teacher. In this collection, we included all occurrences of both complete and reduced instances of I don’t know (it) (Dutch: ‘ik weet (het) niet’) and I have no idea (Dutch: ‘ik heb geen idee’). This includes instances that contain modifiers, such as I don’t exactly know (Dutch: ‘ik weet het niet precies’). The collection also includes evident nonverbal equivalents such as clearly visible stand-alone shrugs. While all instances evidently convey a downgraded epistemic stance, the exact K- position differs among the instances. For instance, I don’t exactly know displays a more knowledgeable stance than I don’t know.

As the epistemic disclaimers in our collection are often embedded in an utterance rather than stand-alone, and since we are interested in how the interaction unfolds after teachers produce a turn with an epistemic disclaimer, the object of analysis in the current study is the whole teacher turn, and not merely the epistemic disclaimer or utterance it occurs in. As previous research has shown that epistemic disclaimers have different functions depending on where they occur in the turn (e.g. Kärkkäinen 2003; Weatherall 2011; Pekarek Doehler 2016), this is included in the analysis as well.

Three instances within the collection do not fall within the scope of this paper, because these instances do not seem to function as epistemic disclaimers, but rather as a trail-off. An example of these can be seen in Excerpt 1 below. In this instance, the teacher aborts the turn-in-progress through an I don’t know.

Excerpt 1 – Dye (S1.2 - 0:38:42.6)

4. Analysis

In this section, we will describe when teachers employ epistemic disclaimers and how the turns that these occur in influence the interaction that follows. Apart from instances occurring in initiating teacher turns which facilitate student participation (similar to the I wonder … practice described by Houen et al. 2019), we will demonstrate that the data also includes epistemic disclaimers occurring in responsive turns which influence the course of the interaction in multiple different ways.

We define initiating turns as turns through which teachers launch a new sequence, and responsive turns as turns through which teachers respond to students. These teacher turns have been labelled initiating and responsive, rather than in terms of their sequential position, as the responsive turns occur in both second and third position. As some sequences are student-initiated, and others are teacher-initiated, the student contributions that the teacher is responding to occur either in first or in second position.

4.1. Initiating teacher turns

Our collection contains multiple instances where teachers produce epistemic disclaimers in initiating turns. The initiating turns in our dataset encourage student contributions either by nominating a specific student to respond or by opening the floor to all students in the classroom. The latter are often manifested through or combined with gazing around the classroom. These teacher turns constitute requests for information that make relevant answers by the students. The epistemic disclaimers generally occur either in turn-initial or turn-final position. The excerpt analysed here shows an example of both.

In Excerpt 2, the teacher asks the class how fast horses can run, which is followed by a whole-class discussion about the answer. Prior to the start of this excerpt, the teacher produces the statement that horses, as a method of transportation, were not very fast compared to the newly upcoming train in the Industrial Revolution. Suzet then states that horses can run fast when they are galloping, to which the teacher responds in line 1.

Excerpt 2 – Horses (S2.1 - 0:13:45)

In line 2, the teacher initiates a new sequence, preceded by a turn-initial epistemic disclaimer in the form of ‘I dunno … ’ (Dutch: ‘kweenie’, derived from ‘ik weet niet’), thereby taking a K- stance in advance and momentarily stepping away from his institutionally assigned epistemic authority. In the same turn (l. 2–3), he proceeds to ask Suzet whether she knows how fast horses can go. Through the ‘toevallig’ (English: coincidentally, translated here as do you happen to know that), he identifies her as possibly having a K+ status, while leaving room for her not-knowing. While shaking his head he then downgrades his epistemic stance even further (‘ik heb geen idee hoor’: I have no idea; l. 4). This turn-final epistemic disclaimer occurs in a separate turn-constructional unit (TCU), as is the case in all turn-final instances in this category. As such, it seems to underline the authenticity of the question. Subsequently the teacher states that he is ‘just’ asking while directing his gaze at various other students and thereby potentially selecting them as next speakers (l. 4–5). In line 7, he rephrases his question to include all students, by asking ‘who knows how fast a horse can go?’. These teacher turns thus contain multiple elements that downgrade the teacher’s epistemic stance and thereby seem to amplify the eliciting function of the question in line 2–3, as the teacher’s epistemic disclaimer facilitates provision of a response by students who are in a K+ position.

Josco’s response to the teacher contains hedging and an uncertainty marker (‘ofzo’: or something), so while he takes a K+ stance by responding, he downgrades his certainty as well. It is also produced softly, and the teacher’s third position turn following it (l. 12–14) seems to be a repair initiation due to a problem of hearing, rather than an evaluation. In response to the teacher’s repetition of Josco’s contribution, Geeke, Manon and Nomar join the discussion by taking a K+ position (apart from the uncertainty marker produced by Nomar here; l. 16–18), thereby realizing a T-S-S-S turn-taking pattern. Thus, the teacher’s factual question combined with the epistemic disclaimer facilitates a whole-class discussion in which the students collaboratively try to construct the answer, although the initial question remains unresolved.

Other instances of teacher turns containing epistemic disclaimers generally do not project a factual response to the teacher’s question, but rather an opinion or assessment of a situation. However, all sequences in which teachers employ epistemic disclaimers in initiating turns progress in a similar manner to Excerpt 2. The initiating turns in our collection are all followed by multiple student contributions, sometimes constructed by multiple students collaboratively or simultaneously, and sometimes by a single student, after which the topic is extended through new sequences with other students. This shows that teachers’ initiating turns with epistemic disclaimers seem to contribute to facilitating student participation and discussion.

Excerpt 2 ended with another epistemic disclaimer (l. 21) which is different from the ones in line 2 and 4. This I don’t know namely occurs in response to a student turn, which will be explored in the following section.

4.2. Responsive teacher turns

We identified responsive turns containing epistemic disclaimers as turns in which the teacher claims insufficient knowledge in response to a student contribution. The format of these turns ranges from stand-alone epistemic disclaimers to more extensive turns, in which the epistemic disclaimer mostly occurs in turn-initial position. Furthermore, most of the instances are hearable as a separate TCU, and as such convey a lack of knowledge (Pekarek Doehler 2016). As we will show below, these responsive turns also vary with regard to their function in the interaction. Some of the turns are responsive while at the same time eliciting more student contributions, whereas others steer towards closure of the topic and limit or block further student contributions.

Excerpt 3 shows an example of a responsive turn containing an epistemic disclaimer that elicits further discussion. Such turns are what Willemsen et al. (2020b) describe as ‘pass-on turns’: turns by means of which teachers indicate that they will not respond while simultaneously inviting other students to do so. This excerpt begins with the teacher recalling a statement previously made by Pim, namely that work in the factory became lighter with the arrival of the steam engine, and asking the class in an uptake (Nystrand 1997; Collins 1982) whether they agree with Pim’s statement.

Excerpt 3 – Factory work (S1.2 - 0:06:47)

While the teacher recalls Pim’s previously made contribution, thereby marking it as relevant, he does not explicitly evaluate it in terms of correctness. Instead, he presents it to the other students, inviting them to state whether they agree or not (l. 1–7; cf. ‘uptake’ Nystrand 1997). Louis then takes the floor and produces an aligning response (l. 8–9). After a short pause, Louis resumes his turn with an interpretation regarding the implications this has for the boss of the factory (l. 10–12). Louis’ turn is followed by a 3.5 second pause at the transition relevance place (l. 13), in which nobody takes the next turn. During those 3.5 seconds, Louis’ gaze does not leave the teacher, while other students also direct their gaze at the teacher, indicating that they expect him to respond. At the same time, the teacher’s gaze zigzags across the room displaying a minimal invitation to respond (Willemsen et al. 2020b).

After the 3.5 second pause, Louis produces a request for confirmation (l. 13), thereby showing an orientation to the IRE structure and the teacher as an epistemic authority, as well as treating the nods in lines 10 and 12 as insufficient to be a convincingly positive evaluation of his prior contribution. Mid-telling nods are generally only treated as affiliative when there is a post-telling vocal affiliative response by the recipient of the telling as well (Stivers 2008), and the teacher does not give one here.

Louis’ question in line 13 both anticipates and makes interactional room for a negative evaluation. The teacher responds in overlap with a subtle shrug, which foreshadows the epistemic disclaimer it precedes (l. 15). With that claim, which stands alone both as a turn and TCU, the teacher ‘refute[s] the epistemic superiority conferred to [him]’ by the students’ turn allocation, and at the same time ‘de-select[s himself], which in turn allows [him] to avoid providing an answer’ (Pekarek Doehler 2016, 155).

In the silence that follows the epistemic disclaimer, the teacher once again looks around the circle, and Fons raises his hand to bid for a turn (l. 16). The teacher then provides a delayed evaluation of Louis’ contribution and, by uttering the word ‘but’ and pointing to Fons, projects a contribution by Fons (1. 17–18). With his response (l. 19), Fons continues the discussion on the same topic from another perspective. As Fons already raises his hand in line 16, it appears that the teacher’s epistemic disclaimer together with his gaze in line 15 works as a pass-on turn and facilitates student participation to the whole-class discussion. The teacher’s delayed evaluation of Louis’ contribution (l. 17) indicates that while facilitating a framework in which students collaboratively construct knowledge, he also remains in his role as a teacher having epistemic rights (and perhaps responsibilities) to evaluate student contributions.

The instances of these kinds of pass-on turns containing epistemic disclaimers in our collection vary in format, including one purely nonverbal ‘turn’ in which the teacher displays an evident shrug, and an instance in which the epistemic disclaimer is accompanied by an explicit invitation to respond. Nonetheless, the interactions all unfold similarly to Excerpt 3 and end with contributions by other students, and hence these pass-on turns all work to further facilitate student participation.

The data also include epistemic disclaimers in responsive turns that do not elicit, but nonetheless allow subsequent student contributions. In these instances, a post-expansion with just one student and the teacher follows from a student contribution, rather than a discussion framework where multiple other students participate. While such a short one-on-one interaction can still be at the base of a multi-party participation framework revolving around a certain topic (an example of which can be found in Excerpt 2), in the cases represented by Excerpt 4 they are always proceeded by topic closure. In these instances, the teacher refrains from explicitly evaluating a student contribution in terms of correctness, and then steers away from the topic raised by the student. Prior to the start of Excerpt 4, another student has stated that he thinks working from home was better than working in the factory. Tristan responds to this, arguing that working in a factory six days a week would pay equally much money as working from home only once would.

Excerpt 4 – Working from home (S1.2 - 0:30:07)

In this excerpt, the teacher challenges Tristan’s argument from line 4 onwards. In line 1–3, Tristan summarizes the historical assertion he has been constructing. The question that forms the teacher’s response seems to imply that Tristan’s argument is incorrect (l. 4) and is followed by a pause in which the teacher shakes his head, which prefaces the teacher’s upcoming counterargument. Tristan seems to go along with this uncertainty by producing an epistemic disclaimer (l. 5) and the teacher continues with a third position closing turn. However, Tristan self-selects to respond again (l. 8), upgrading his epistemic stance slightly. Rather than explicitly evaluating, the teacher produces a confirmation check. This is immediately followed by an epistemic disclaimer, after which he phrases another counterargument that is marked with uncertainty (l. 9). While he still does not explicitly evaluate Tristan’s contribution, he averts his gaze and orients to the material (l. 11), which can be interpreted as a sign of sequence closure and moving forward (Rossano 2012). Tristan again takes the turn, producing another counterargument, but the teacher reverts to minimal responses (l. 14–16) and when Tristan produces a softly pronounced trail-off, he allocates the turn to another student, clearly marking off the transition prosodically (l. 18–19).

While the epistemic disclaimer occurs in a separate TCU, this excerpt shows how the turn it occurs in allows further contributions from the initiating student while the teacher seems to be working towards topic closure. Many of the teacher’s utterances are marked by uncertainty and he does not give a conclusive answer, but rather ends the sequence after a few turns by allocating the turn to another student and raising a new topic. At the same time, Tristan both challenges the teacher’s argument and employs uncertainty markers, indicating that he switches between taking a K+ position and aligning with the teacher’s uncertainty on his contribution. Excerpt 4 thus shows how teacher and student locally manage, establish, and contest their own and each other’s epistemic identities throughout the interaction (cf. Melander 2012).

Our collection also contains instances where the topic is not abandoned after a discussion between the teacher and a student, but immediately in a teacher’s third turn in response to a student contribution. In these instances, the teacher selects another student at the end of his turn containing the epistemic disclaimer, or moves on to another topic during his turn, leaving no room for post-expansion and thereby inhibiting any further contributions. An example of this can be found in Excerpt 5, in which the class is discussing a picture from the teaching material portraying houses from the Industrial Revolution.

Excerpt 5 – Chimneys (S1.5 - 00:17:55)

The excerpt begins with the teacher asking the students what they notice about the houses (l. 1). After a few turns in which the teacher selects a student who refuses to respond, the teacher allocates the turn to Jamiro (l. 10). Jamiro responds by comparing the chimneys on the houses in the picture to present-day chimneys (l. 11). The teacher treats this answer as surprising, producing the change of state token ‘oh’ (Heritage 1984; Bolden 2006). Moreover, he downgrades his own epistemic stance by stating that it ‘could be’ and that he does not ‘actually know that’ (l. 13), anaphorically referring to Jamiro’s assertion. By downgrading his epistemic stance, again in a separate TCU, he evades explicitly evaluating the response and indicates that he cannot give an affiliating response. This is reinforced by another incremental ‘yes could be’ (l. 14), which still does not convey agreement or disagreement. Potential topical irrelevance of Jamiro’s answer is marked by the emphasis on you in ‘you notice the chimneys’ (l. 16). At the subsequent transition relevance place the teacher immediately produces a new utterance allocating the turn to Tristan (l. 16) and thereby closing the topic raised by Jamiro. In doing so, he leaves no room for any of the students to continue the discussion, as is the case in Excerpt 4.

As demonstrated by the excerpts above, responsive turns containing an epistemic disclaimer have different uses in the interaction. Some pass on turns to other students, some allow for a student to elaborate on their contribution and others block any further student contributions. Our analysis of representative points in our data shows that responsive teacher turns with epistemic disclaimers can vary in whether or not (and to what extent) they encourage participation, while initiating teacher turns with epistemic disclaimers all show to encourage participation.

5. Conclusion and discussion

This study has shown that teacher turns containing epistemic disclaimers can occur in different positions, and from those positions also influence the participation framework in different ways.

Initiating turns with epistemic disclaimers elicit student contributions and thereby facilitate student participation. In these instances, the teacher steps away from his role as epistemic authority while inviting the students to respond, which seems to support the establishment of a participation framework where students come up with suggestions and hypotheses (cf. Willemsen et al. 2018) and respond to each other. Epistemic disclaimers in initiating turns are either turn-initial or turn-final. In the first case, they seem to preface the subsequent teacher question. This is somewhat similar to what Weatherall (2011) found with regard to first-position turn-initial epistemic disclaimers in ordinary interactions, which display limited commitment aligning with a possible disclaim of knowledge authority (Weatherall 2011, 335). However, together with other eliciting markers such as gazing around the classroom (cf. lighthouse gaze, Cekaite and Björk-Willén 2018) and requesting more information these teacher turns extend Weatherall (2011) findings that the epistemic disclaimers work on speakers’ own turns and actions mainly. Turn-final epistemic disclaimers in initiating utterances appear to underline the authenticity of the prepositioned teacher question. To our knowledge, these turn-final epistemic disclaimers that form a stand-alone TCU have not been described in ordinary interaction.

When epistemic disclaimers occur in responsive turns, they sometimes similarly elicit further student contributions by passing on the turn, or do not explicitly elicit, but still allow student responses. Especially in the cases of passing on the turn, turns with epistemic disclaimers seem to function to open up the T-S-T-S turn-taking pattern and encourage students to respond to each other. In contrast to these turns, we have also shown that teachers use epistemic disclaimers in turns that block student contributions by closing the topic, in which case they do not function to support the discussion framework but rather put or leave the teacher in charge of the turnallocation as well as the topics under discussion. The responsive turns that are either followed by a short post-expansion and then topic and sequence closure or by immediate topic and sequence closure seemingly occur when the student contribution is not treated as correct or not treated as relevant. It thus seems to be functioning as a tool to contest student contributions without explicitly evaluating them. This topic-closing function of these utterances is also found in ordinary interaction (e.g. Aijmer 2009; Beach and Metzger 1997; Keevallik 2016; Pekarek Doehler 2016). However, in contrast with previous findings, these epistemic disclaimers in responsive turns are not to be seen as non-answers (Stivers and Robinson 2006) or prefacing trouble with the requested action (Beach and Metzger 1997; Helmer, Reineke, and Deppermann 2016; Keevallik 2011; Pekarek Doehler 2016; Schegloff 1996; Tsui 1991).

Epistemic disclaimers in responsive turns are mostly turn-initial or stand-alone, and all are hearable as a separate TCU. They thus seem to convey a lack of knowledge in response to a student contribution or question, and as such either pass on the turn, or occupy the evaluation slot in a third position response to a student contribution. While epistemic disclaimers in response to questions have been described extensively in previous research in both ordinary and institutional interaction (e.g. Stivers and Robinson 2006; Pekarek Doehler 2016; Beach and Metzger 1997; Keevallik 2016; Jager et al. 2016), the third position occurrences in the evaluation slot have, to our knowledge, not yet been described in other types of interaction than teacher-student interaction.

Our findings that turns with epistemic disclaimers do not always function to facilitate student participation are somewhat contrastive to the findings reported by Houen et al. (2019). Especially the instances in which the epistemic disclaimer is immediately followed by topic and sequence closure indicate that K- teacher turns sometimes also block any further student contributions. A possible explanation for this difference with the observations by Houen et al. (2019) lies in the format of the epistemic disclaimer under analysis. While ‘I wonder … ’ formulations seem to inherently invite interlocutors to answer the questions at hand, the variations of ‘ik weet het niet’ (I don’t know (it)) and ‘ik heb geen idee’ (I have no idea) that are central to the current study allow some things to be left in the open because there is simply no answer, which is also reflected in their use in both initiating as well as responsive turns.

We argue that teachers’ use of epistemic disclaimers in whole-class discussions indicates that there is a precarious tension between different teacher roles in this type of conversational teaching. On the one hand, the teacher is instructed to take on a facilitating role, stimulating students to share ideas and knowledge together and have discussions with each other. On the other hand, the teacher also remains in his role of epistemic authority, following a certain agenda and teaching relevant and correct knowledge to students (cf. Schuitema et al. 2018). In our analyses, this tension between the two teacher roles comes forward for example in the use of epistemic disclaimers in the third turn to contest student contributions without explicitly evaluating them (see Excerpts 3 and 5), and in sequences in which the teacher initially takes a K- stance, which is then followed by a (e.g. more or less explicit) evaluation (see Excerpt 3). Furthermore, Excerpt 2 shows that factual questions in classroom interaction create a need for an epistemic authority who gives a conclusive answer, and that student guesses are not sufficient. In that case the epistemic authority is sought outside the classroom.

While teachers in whole-class discussions no longer have a central information transmitting role, they thus remain the leader of the interaction. This is also reflected in the epistemic stances of students in the subsequent turns. While we see some instances where students take a K+ position (e.g. Geeke and Manon’s response in Excerpt 2, l. 16–17, and Tristan’s continuations in Excerpt 4), we also see students using many uncertainty markers (e.g. Tristan in Excerpt 4) and explicit orientations to the teacher as an epistemic authority (Louis, Excerpt 3). It thus seems that there is no systematic way in which students alter their epistemic stance after the teacher downgrades his.

As has been previously mentioned, the turns under analysis in the current study do not all display the exact same epistemic stance. Furthermore, epistemic stance is displayed in different ways in differently composed turns. Nonverbal conduct, as well as prosody, also play a role in stance-taking, and influence the stance captured in the verbal utterance. Nonverbal conduct can also foreshadow the upcoming teacher stance, as is for example visible in Excerpt 3.

Lastly, there are some individual differences between the two teachers in this study. While the instances analysed in this paper are representative of both teachers, one of the teachers produced more epistemic disclaimers than the other. Nonetheless, it is noteworthy that both teachers used disclaiming practices and did so in similar ways. While typical history lessons in the Netherlands hardly involve whole-class discussions, the setting of this study (i.e. the fact that the teachers were instructed to have whole-class discussions and change their interactional conduct along with it) evidently led teachers to not only open up the participation framework but even openly reduce their epistemic authority, affording them to engage their students in a different type of classroom interaction. Future research could discover how widespread the practice of epistemic disclaiming is among teachers.

This study has shown that teacher turns containing epistemic disclaimers are used in multiple ways and can be a tool to balance their simultaneous roles of teacher and facilitator in conversational approaches in the classroom on a micro-level. Our results contribute to a growing body of research on epistemics in institutional interaction, and specifically to research on participation in classroom interaction. Furthermore, the obtained knowledge of epistemic disclaimers as a type of teacher conduct can be used in teacher training, to make teachers aware of the different types of facilitating conduct they may already demonstrate, even without explicit instructions to do so. When teachers are aware of this, they can purposely decide when to use it to facilitate non-teacher-fronted classroom interaction.

Acknowledgements

We are grateful to the teachers and students for allowing us into their classrooms to observe and record the lessons. We would also like to thank the anonymous reviewers for their valuable comments and suggestions.

Disclosure statement

There is no potential conflict of interest to report.

Additional information

Notes on contributors

Sofie van der Meij

Sofie van der Meij is a research master student in Linguistics at the University of Groningen. Her research interests include using conversation analysis to study institutional interaction and epistemics.

Myrte Gosen

Myrte Gosen is an assistant professor at Communication and Information Studies of the University of Groningen, the Netherlands. She has a particular interest in classroom interactions in relation to knowledge and she uses conversation analysis to identify the fundamental structures and practices in interaction that are related to knowledge construction.

Annerose Willemsen

Annerose Willemsen is a postdoctoral researcher at the Department of Behavioural Sciences and Learning at Linköping University, Sweden. Using multimodal conversation analysis and discursive psychology, she focuses on interaction in educational settings, such as whole-class discussions in primary school and traffic and mealtime socialization in Swedish preschools.

Appendix

Multimodal transcription conventions (following Mondada 2018)

* *Descriptions of embodied actions are delimited between

+ +two identical symbols (one symbol per participant and per type of action)

∆ ∆that are synchronised with correspondent stretches of talk or time indications.

*—>The action described continues across subsequent lines

—->*until the same symbol is reached.

>>The action described begins before the excerpt’s beginning.

—>>The action described continues after the excerpt’s end.

….. Action’s preparation.

----Action’s apex is reached and maintained.

,,,,, Action’s retraction.

ricParticipant doing the embodied action is identified in small caps in the margin.