Classroom Discourse
ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/rcdi20
Applying conversation analysis to classroom
interactions: students’ ‘oh’
-prefaced utterances
and the interactional management of
explanations
Myrte N. Gosen, Annerose Willemsen & Frans Hiddink
To cite this article: Myrte N. Gosen, Annerose Willemsen & Frans Hiddink (21 Oct
2024): Applying conversation analysis to classroom interactions: students’ ‘oh’
-prefaced
utterances and the interactional management of explanations, Classroom Discourse, DOI:
10.1080/19463014.2024.2397127
To link to this article: https://doi.org/10.1080/19463014.2024.2397127
© 2024 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Group.
Published online: 21 Oct 2024.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=rcdi20
Applying conversation analysis to classroom interactions:
students’ ‘oh’-prefaced utterances and the interactional
management of explanations
Myrte N. Gosen a
, Annerose Willemsenb and Frans Hiddinkc
aCenter for Language and Cognition Groningen, University of Groningen, Groningen, The Netherlands;
bDepartment of Behavioural Sciences and Learning, Linköping University, Linkoping, Sweden;
cMultilingualism and Literacy, NHL Stenden University of Applied Science, Leeuwarden, The Netherlands
ABSTRACT
This study serves as an example of Conversation Analytic (CA)
research with a focus on the interactional management of learning
in classroom interactions, while simultaneously illustrating the several
steps and procedures used for conducting these analyses. In
this case, the method of CA is used to study students’ ‘oh’-prefaced
utterances in one-to-one classroom interactions centred around
explanations. The interactions are studied from the participants’
perspective by means of a turn-by-turn analysis of the selected
fragments on the basis of video recordings and detailed transcripts,
as is common practice in CA. This study aligns with previous CA
research focusing on the details of classroom interaction relating to
students’ learning processes. The results of the close analysis of the
data provide insight into the orientation teachers and students
show to interactionally exhibited knowledge displays in relation
to the entire explanation activity they are involved in. This detailed
analysis of one-to-one explanation interactions in secondary school
classrooms provides insight into students’ learning processes as
well as into the accompanying teacher practices.
ARTICLE HISTORY
Received 26 July 2023
Accepted 21 August 2024
KEYWORDS
Conversation analysis;
classroom interaction;
knowledge displays;
learning; explanations
- Introduction
This contribution showcases Conversation Analysis (CA) (Sidnell and Stivers 2013) as
a qualitative method for studying educational interactions. The general enterprise of CA
is to uncover how interactants accomplish their conversations as social projects.
Conversation analysts are generally interested in how participants make observable to
each other how they understand each other’s utterances (Heritage 1984b; Sacks,
Schegloff, and Jefferson 1974). This means, conversation analysts are only interested in
knowledge of interactants once this knowledge is interactionally exhibited (Heritage
2013). From this perspective, learning might only be observed in interaction by tracking
changes in the interactionally exhibited knowledge (Melander 2012).
CONTACT Myrte N. Gosen m.n.gosen@rug.nl Center for Language and Cognition Groningen, University of
Groningen, Oude Kijk in ‘t Jatstraat 26, Groningen 9712 EK, The Netherlands
CLASSROOM DISCOURSE
https://doi.org/10.1080/19463014.2024.2397127
© 2024 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/
licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or
with their consent.
In this paper, we illustrate the data-driven method of CA and its affordances by
examining the phenomenon of students’ ‘Oh’-prefaced utterances in classroom interactions.
‘Oh’ is described as a change-of-state token (Heritage 1984a) claiming a change in
understanding (Sacks and Jefferson 1992). This token is found to be used in response to
a previous action to express that something is heard as relevant, new or remembered
(Bolden 2006; Sacks and Jefferson 1992; Seuren, Huiskes, and Koole 2016).
We will show how analysing secondary school students’ ‘oh’-prefaced utterances
during explanation interactions with their teacher provides insight into students’ learning
processes as well as in teachers’ interactional conduct to facilitate learning (e.g. Koole
2015; Willemsen 2019). We moreover guide the reader through the different methodological
steps and decisions relevant for CA research: from recording data with a general
research interest (e.g. a specific task or learning outcome), via detailed transcription
enabling turn-by-turn analysis and the bottom-up discovery of recurring patterns, to
our presentation of the results in a conversation analytic publication.
- Conversation analysis in education
Originating from ethnomethodological theory (Garfinkel 1967), CA was established in the
1960s and 1970s by amongst others Sacks and Jefferson (1992) to uncover the organisation
of interaction. Presently, straightforward, but revolutionary in those days, was Sacks’s
idea to record authentic interactional data. To this day, the preservation and the analysis
of authentic data are still among CA’s crucial features. Collected data are not elicited (but
see Sert, Gynne and Larsson’s 2024) Discursive Timeline Analysis for a combination of CA
and other discourse data), recollected on the basis of memory, or derived from an
experimental setting, but instead recorded while naturally occurring.
By means of the detailed method of CA, researchers create a clearer picture of common
interactional constellations in education, such as whole-class interactions, peer interactions
and one-to-one teacher–student interactions. One often starts with a general
research interest into a specific activity that can be considered to be the main reason to
collect recorded data. Once these data are collected, CA researchers take a data driven
approach. By repeatedly looking at the data, they arrive at more specific research interests
that can be studied in detail using transcripts of the interaction.
The recordings and accompanying transcripts enable the researchers to repeatedly
analyse the interaction as it unfolded in real-life. The recordings are therefore transcribed
in great detail including among others hesitations, pauses and overlaps (see for instance
Jefferson 2004). Multimodal transcripts furthermore include eye gaze, gestures, body
posture, etc. (see Mondada 2016). The transcripts however can never replace the original
recordings, but are an aid to put the recorded conversations on hold for analysis.
Questions initiating the detailed analyses on the basis of the recordings and transcripts
might for instance be: ‘How do teachers encourage student contributions in whole-class
discussions?’, ‘How do peers “reason together” during inquiry learning?’ or ‘What do
teachers’ explanations to individual students look like?’. These questions lead to
a subsequent close investigation resulting from the initial exploration of the data. The
initial question of how teachers encourage student contributions in whole-class discussions
for instance resulted in a study on teachers’ open invitations to their students
(Willemsen et al. 2018); the second question on peers reasoning together led to a study
2 M. N. GOSEN ET AL.
on how young children accomplish account sequences in problem-solving interactions
(Hiddink 2019).
When moving from a more general research interest to a particular interactional
practice, we thus move into the study of the details of interaction that enable us as
researchers to investigate how a particular interactional constellation is locally managed
from moment to moment in real time. While doing so, we focus on practices and actions
by teachers and learners in relation to each other. Herewith, we get a better picture of the
learning processes as they become apparent in interaction as well as of the teacher’s
conduct facilitating learning. It is precisely this detailed analysis that enables us to get
closer insights into the overarching educational interaction. At the same time, this type of
analysis comes close to educational practice as it is experienced by teachers and students
while participating. This illustrates the ‘emic’ perspective of CA: the analyst works with
what becomes observable in interaction, for the participants themselves as well as for the
researchers as analysing ‘overhearers’.
- Interactional organisation in education
A main driveway to study what becomes observable in interaction is described in the CA
literature as the next-turn proof procedure. By responding to a previous utterance,
a recipient displays how aspects of a prior turn were understood (Heritage 1984b).
Already by providing an answer, a recipient for instance indicates that the prior turn
was understood as a request for information. Schegloff (2007) uses the term ‘sequence
organization’ to refer to the fact that turns at talk are often organised in adjacent pairs
forming a coherent and orderly ‘sequence’ of actions (Schegloff 2007, 2) such as a request
for information and an answer. A particular action that is launched as a First Pair Part (FPP)
expects a fitting Second Pair Part (SPP). For instance, the request for information as an FPP
projects an answer providing the information as an SPP. This pair is furthermore often
followed by a third part in which the first speaker accepts the new information.
Concerning the organisation of one-to-one interactions between a teacher and one, or
sometimes two, student(s), previous research has focused on the interactional organisation
of these interactions in general (e.g. Breukelman et al. 2023) as well as on this
organisation in relation to more explicit displays of knowledge (Koole 2010, 2012).
These one-to-one interactions are regularly organised as assistance to the student(s)
completing individual exercises. Individual seatwork often follows whole-classroom
instructions by the teacher. When a student then runs into a problem completing an
exercise individually, the teacher may provide assistance in solving the problem. These
interactions are often initiated by a student’s request for assistance, followed by different
types of responses by the teachers in second sequential position (Breukelman et al. 2023).
The teacher responses are generally found to be either organised as discourse units
(Houtkoop and Mazeland 1985) i.e. longer turns explaining what to do next, or as the
launching of several sequences of questions and answers gradually leading students to an
understanding of how to proceed (Koole 2010). In general, these one-to-one interactions
are described in the literature as ‘explanation interactions’ (e.g. Breukelman et al. 2023;
Koole 2010, 2012).
As the starting point for these interactions concerns assistance-seeking in which ‘the
person producing the request for assistance is positioned as unknowing and the
CLASSROOM DISCOURSE 3
presumptive helper as knowing’ (Svahn and Melander Bowden 2021, 196), these interactions
are also studied from the perspective of learning, by focusing on aspects of knowing
and understanding. Displays of epistemic stances, for instance, are studied in relation to
these interactions (for example, Koole 2010, 2012). Within CA, ‘research into epistemics
focuses on the knowledge claims that interactants assert, contest and defend in and
through turns-at-talk and sequences of interaction’ (Heritage 2013, 370). Learning in this
sense is considered socially distributed (e.g. Markee and Kunitz 2013) and can only be
traced in interaction by tracking changes in the interactionally exhibited knowledge
claims (Melander 2012).
The particular interactional display of change in understanding (Sacks and Jefferson
1992) in the form ‘oh’ as a change-of-state-token has been elaborately described. This
token has been found to show a change in understanding in response to a previous action
(Heritage 1984a; Jefferson 1978; Schiffrin 1987). In several languages (e.g. English, Dutch)
it is found that interactants can use ‘oh’ to express that something is heard as relevant,
new or remembered (Bolden 2006; Sacks and Jefferson 1992; Seuren, Huiskes, and Koole
2016). The ‘“oh” is used to mark the receipt of the informing delivered in the preceding
turn or turns’ [italics in original] (Heritage 1984a, 301). Sacks and Jefferson (1992) already
distinguished between two modes of displaying change; claims and demonstrations.
Producing just an ‘oh’ is described as a claim of understanding or recognition, while
delivering an ‘oh’ as a preface to a lengthier turn is described as a demonstration since
these turns reveal ‘some sort of analysis’ of B’s utterance and using ‘that analysis in
producing a next utterance’ (Sacks and Jefferson 1992, 253).
That the token ‘oh’ indicates a possible change enables us to come close to the
uncovering of interactants’ learning processes. In Dutch secondary classroom interaction
data, Koole (2010) already did a close analysis of displays of knowing and displays of
understanding in relation to students’ use of ‘oh’. He illustrated that different displays
occur in different sequential positions and have different functions within the overall
activity of explaining. For this paper, we follow-up on this research by focusing on ‘oh-
prefaced’ turns in particular. As the above-mentioned studies demonstrate the relevance
of the token, ‘oh’ is well suited as a focus in this paper showcasing CA as a method to
study the interactional details of learning processes. The following analysis therefore
illustrates how learning in a classroom setting can be studied from an interactional
perspective and enables us to show all the relevant steps and concepts related to CA
research in classroom interaction.
- Data set and object of study
For this study, data are used from a larger data set collected by Tim Mainhard, Astrid
Poorthuis and Janneke van de Pol (e.g. van Braak et al. 2021). In their project on
scaffolding in classroom interaction, they video recorded one-to-one interactions
during 150 lessons (virtually all lessons students attended in 2 weeks, ranging from
the subjects of Dutch language and literature to mathematics, taught by different
subject teachers) at four different lower secondary pre-vocational education schools
(students aged 11–13) in The Netherlands. The researchers’ aim was to get more
insight into scaffolding processes, such as teacher support and student independence,
during situations in which students worked on exercises independently. A recent
4 M. N. GOSEN ET AL.
publication for instance presented statistical analyses of coded behavioural engagement
and adaptive support in the interactions (van Braak et al. 2021). While their
quantitative findings are of great interest for classroom interaction research, one of the
researchers reached out to us as conversation analysts as she believed that CA
research could reveal deeper insights into the learning processes that became visible
in the interactions.
The data were made available to us as a large set of video recordings of individual one-
to-one interactions centred around explanations in which a teacher explains content
typically upon a student’s request for assistance to enable them to continue their work
independently. Two video cameras were installed in the classroom, teachers wore microphones,
and the students were recorded with audio recorders on their desks (one
recorder per two or four students). In line with the consent provided by teachers, students
and parents, the recordings were anonymised before our use of the data. The videos were
blurred and the voices were transformed. Sometimes, this complicated our transcribing of
the interactions, but this disadvantage was outweighed by the importance of protecting
the identities of the participants involved. The one-to-one interactions were transcribed
according to Jeffersonian conventions for verbal conduct and Mondadaian conventions
for multimodal transcription (Jefferson 2004; Mondada 2016, see Appendix).
The transcripts establish the possibility to analyse all features in the interactions.
Transcripts make it possible to ‘pause’ the interaction, where in real-time many features
come simultaneously, and at such a pace, that it is impossible to analyse them in detail.
Compared to methods of real-time observation and coding, the benefit of CA therefore is
that the transcripts enable us to dismantle all the different features of the interaction
while repeatedly looking at the data. In addition, transcripts enable us to take the
participant’s perspective by focusing on these features that are relevant on a turn-by-
turn basis for the interactants themselves. Without transcripts present for analysis, we
could only have taken note of our phenomenon to be happening and how often, without
drawing conclusions on the detailed workings of the phenomenon in the interactions.
As CA is a data-driven approach, one does not work with specific research questions
with accompanying hypotheses. As already stated in the introduction, one starts with
a general interest in a classroom activity. In this case, we adopted the overarching
project’s general interest in one-to-one explanation interactions in relation to learning
processes such as scaffolding (e.g. van de Pol and Elbers 2013; van de Pol et al. 2014, 2015;
van de Pol, Mercer, and Volman 2019). Therefore, we have worked with this data set to
gain a deeper insight into the learning processes as these become visible in these one-to-
one interactions in which a teacher explains content to a student. For this study, we have
selected the lessons on Dutch language and literature, resulting in 72 one-to-one
interactions.
While conducting observations on the basis of the (transcribed) data set, we encountered
a considerable number of ‘oh’-prefaced turns by the students. This caught our
interest because of CA’s general interest in ‘oh’ as a change-of-state-token (Heritage
1984a) and the more specific interest in displays of knowing and understanding in classroom
interactions in relation to students’ use of ‘oh’ (Koole 2010). We therefore decided to
focus on these ‘oh’-prefaced turns and attempt to find out what these turns reveal about
the learning processes and the accompanying teacher practices in these one-to-one
interactions. We then came to a more specific research question: ‘What do students do
CLASSROOM DISCOURSE 5
when deploying the practice of “oh”-prefaced utterances and how are these responded to
by teachers?’.
- Method
Based on the subset of 72 one-to-one interactions during Dutch language and literature
lessons, we localised all student turns that were ‘oh’-prefaced. This led to a collection of 21
explanation interactions containing at least one ‘oh’-prefaced turn by the student. As we
wanted to include as much information as possible about the learning processes of the
students, we did not include stand-alone ‘oh’s. Because of the emic perspective of CA, we
as analysts only base our analyses on what the participants make observable to each
other. The utterances following the ‘oh’ provide us with information on what exactly in
the teacher’s turn the ‘oh’ is responding to and how the ‘oh’ hence is to be interpreted in
the interaction.
Once we established this collection, we conducted an inductive analysis focusing on
the sequential position of the ‘oh’-prefaced turn in relation to the surrounding turns and
the action these turns were performing in the interaction. We noticed that in only 11 of
the fragments the students actually responded to the teacher’s explanation by means of
their ‘oh’-prefaced turn (other turns responded for example to something outside the
one-to-one interaction or to something unrelated to the explanation). The next step in our
data collection was then to focus on these 11 ‘oh’-prefaced student turns. This might
seem like a rather small collection of instances. However, for a more in-depth exploration
of the data, this can be considered sufficient in CA terms, especially since we focused on
the entire explanation interactions containing the ‘oh’-prefaced utterance and not just on
the immediately preceding and following turns surrounding this utterance. The analysis
thus provides us with an extensive image of the explanation interactions and a thorough
scrutiny of how the ‘oh’-prefaced turns functioned in the organisation of the explanation.
For each of the 11 fragments, we iteratively conducted a close sequential analysis of the
trajectory (from start to closing of the explanation interaction) and refined the transcripts
accordingly (Broth, Musk, and Persson 2020). This enabled us to draw conclusions on what
the ‘oh’-prefaced turns were doing in relation to the prior explanation as well as in relation
to the ensuing interaction.
- Results
While being in an explanation interaction with their teacher, students in our data set
regularly respond to the (ongoing) explanation by means of an ‘oh’-prefaced turn. As ‘oh’
is demonstrated to be a change-of-state token, this is of interest for the teacher involved.
The student signals a change, which can be considered the aim of an explanation
interaction. We will demonstrate that ‘oh’-prefaced turns may either claim or demonstrate
a change. We will also demonstrate that the teachers interactionally treat a claim as
insufficient. As we will show in our analyses, an ‘oh’-prefaced demonstration appears to be
conditional for the explanation interaction to be brought to a close. Both teacher and
students a) orient to a demonstration at the end of an explanation interaction as being
sufficient, or b) orient to a claim as insufficient by continuing the explanation interaction
with the elicitation of a demonstration, or c) orient to a claim as being sufficient in those
6 M. N. GOSEN ET AL.
cases where a demonstration has been provided earlier in the explanation interaction. As
is common in conversation analytic papers, the different patterns in the data will be
illustrated by presenting extracts that are representative for the 11 fragments analysed.
6.1. Orientation to a demonstration as being sufficient
Extract 1 shows a clear example of an ‘oh’-prefaced demonstration treated as sufficient for
closing the explanation interaction. As soon as the students in this extract demonstrate
change in an ‘oh’-prefaced turn, the teacher starts turning and walking away from their
desks.
In line 1, student 1 requests assistance by producing a request for information in
the form of a content question (Englert 2010) asking about the meaning of the
word ‘pendelen’ (shuttling or commuting) while stressing the second syllable. Using
the same pronunciation, student 2 (seated next to student 1) aligns with the
question (l. 2). In overlap with this, the teacher initiates a repair sequence addressing
the students’ pronunciation of the word by stressing the first syllable instead
of the second (l. 3). When no response is forthcoming, the teacher repeats the word
with the right pronunciation (l. 5). This time it is received by one of the students
demonstrating to now know or now remember (Seuren, Huiskes, and Koole 2016)
how to pronounce the word (l. 6) which closes the repair sequence. Then, the
teacher resumes the question-answer sequence initiated in line 1 and explains the
meaning of the word (l. 7–8) by invoking it in relation to an object that was used in
a previous explanation. It is the turns in lines 9 and 10 then that belong to our
collection of ‘oh’-prefaced turns after an explanation (line 6 of course also contains
1 STU1: wat is pendelen.=
what is shuttling
2 STU2: =ja pen[delen]
yes shuttling
3 TCH: [ pend]elen.
shuttling
4 (1.8)
5 TCH: pendelen.=
shuttling
6 STU1?: =⭡OH PENDELEN.
oh shuttling
7 TCH: PENdelen.=dat a- had ik uitgelegd gi- vorige keer.
shuttling I explained that yest- last time
8 met die over die bus:. die ergens tussendoor.=
with that about that bus that (drove) through something
9 STU1: =o[&:h de pendel&bus. ]
oh the shuttle bus
10 STU2: [&oh (de pendel&bus.)]
oh the shuttle bus
tch: &turns away—–&walks away–>>
Extract 1. The shuttle bus (B/5_int.1545DL)
CLASSROOM DISCOURSE 7
an ‘oh’-prefaced turn, but not in response to an explanation, so therefore this turn
is not included in our collection). Here, both students simultaneously demonstrate
change in response to the explanation in lines 7–8 by producing the compound
word ‘pendelbus’ (shuttle bus). The teacher turns away from the students early on
in their turns (lines 9–10) and starts walking away from their desks at the end. This
clearly shows the teacher’s orientation to the production of an ‘oh’-prefaced
demonstration as an indication of a successful explanation interaction ready to
be closed (in this case by moving away from the students, Broth and Mondada
2013).
6.2. Orientation to a claim as being insufficient
In the data, the students’ ‘oh’-prefaced turns do not always include a demonstration. In
these cases, the ‘oh’-prefaced turns are only claiming change. These claims are clearly
treated as being insufficient by the teacher, who pursues a demonstration as
a prerequisite for closing the explanation interaction. An example of this is Extract 2,
where the teacher explicitly asks for a demonstration when only a claim has been
produced by the student so far.
1 TCH: Ferry. sorry.
Ferry sorry
2 &(0.8) &(0.6)
tch: &comes to F&leans on desk–>
3 STU: een: wildwaterrivier is een soort wat.
a whitewater river is a kind of what
4 (0.3)
5 TCH: nou dat heb ik &net voorgedaan &daar.
well I just explained that there
tch: –>& &points to board–>
6 (.)
7 TCH: een huisvesting is een soort vesting,=een
a housing accommodation is a kind of accommodation
8 huisvrouw is een soort vrouw, .Hh huistelefoon is
a housewife is a kind of wife .Hh house telephone
9 een soort telefoon,
is a kind of telephone
10 (0.4)
11 TCH: een ziekenhuis is een soort hui[s. &]
a nursing home is a kind of home
12 STU: [o:h&] ja.
oh right
tch: –>&
13 (.)
14 TCH: ja?=& dus een wildwaterrivier is een soort_&
yes so a whitewater river is a kind of
tch: &leans over and points in book———&
15 STU: &°rivier.°&
river
tch: &trns away&walks away–>>
Extract 2. Whitewater river (G/17_int.1567DL)
8 M. N. GOSEN ET AL.
In this extract, the student (Ferry) initiates an explanation interaction by reading aloud an
exercise to the teacher: ‘a whitewater river is a kind of’ followed by the question word
‘what’ (l. 3). In this way, he indicates his trouble filling in the requested answer (the head of
the compound word: river) without producing an explicit request for assistance or
information. Nonetheless, the teacher treats it as such by starting an explanation after
the 0.3 second pause in line 4. In the explanation (lines 5–11), the teacher refers to
a previous classroom activity and explicitly models the procedure with multiple other
examples of compound words. In line 12, the student just says ’oh ja’ (oh right), thereby
claiming a change in response to the previous action of the teacher, but not (yet)
demonstrating it in any way. The teacher responds with a request for confirmation (yes?
- 14), but without awaiting an answer. Instead, she proceeds with ‘so’ and repeats the
exercise under discussion while withholding the last item, i.e. the answer (Elicitation
Completion Device/ECD, Margutti 2006 or Designedly Incomplete Utterance/DIU, Koshik
2002). In doing so, she indicates a relationship between the student’s claim and the ability
to complete the exercise now (by means of ‘so’) and invites him to demonstrate his
understanding. Indeed, responding to the ECD/DIU and hence formulating an answer to
the exercise, the student now demonstrates his understanding. Again here, already at the
beginning of this demonstration the teacher starts turning away and walks away at the
end of the demonstration, treating the student’s demonstration as sufficient to close the
explanation interaction.
6.3. Orientation to a claim as being sufficient after a prior demonstration
Contrary to Extract 2, Extract 3 shows an explanation interaction coming to a close after
a student’s ‘oh ja’ (oh right), i.e. a claim rather than a demonstration. In this case, however,
the student has already demonstrated change earlier in the interaction. We found multiple
similar examples of this in our data, showing that teachers do not pursue demonstrations
when a demonstration has already been provided earlier, before the ‘oh’-prefaced
claim. The student’s demonstration is however followed by an extended explanation by
the teacher eliciting a claim. In these extended explanations, the teacher typically refers to
the general rule to be applied to the specific type of assignments discussed, rather than
further explaining this particular exercise.
1 STU: mevrouw u:hm &geef is toch het enige werkwoordelijk gezegde
miss uhm hand is the only verb phrase right
tch: &leans on STUs table—->
2 STU: (of) in de zin of- in deze zin (of eh)
(or) in the sentence or- in this sentence (or uh)
3 TCH: nou. kijk eens of er nog een stukjeh:
_
well just look whether there (is) another piece
4 (0.4)
5 TCH: en volgens &mij staat die zin zelfs @°op het& bord.
°
and I think that sentence is even on the board
tch: &looks at whiteboard—————&
stu: @looks at whiteboard—->
6 (0.7)
Extract 3. To Hand out (C/6_int.2505DL)
CLASSROOM DISCOURSE 9
In lines 1–2, the student produces a request for confirmation of her answer to the exercise,
while leaving room for alternative answers (‘or’, l. 2). The teacher encourages the student to
look at the sentence (l. 3) and then directs her attention to the whiteboard (l. 5), where this
sentence is presumably written down and parsed already. Looking at the whiteboard, the
student then produces an ‘oh’ and demonstrates her understanding by adding the preposition
‘out’ to her answer (l. 7). The teacher confirms this as the right answer and extends her
confirmation by providing the infinitive form of the verb: ‘to hand out’. Even though the
explanation could be complete at this point – as the student demonstrated to be able to
formulate the right answer – the teacher extends the explanation (l. 11–12). This serves as
a justification for the right answer as the teacher refers to a more general rule: a (finite) verb
phrase consists of all elements that (in Dutch) are written together as one word in the infinitive
form of the verb. This explanation is not only relevant for this particular exercise, but for all
other parsing assignments potentially. After an ‘okay’ by the student, the teacher requests
confirmation (‘do you see that’), answered by the student in line 17. The teacher however does
not await the student’s response, she expands her explanation once more further explicating
the general rule (l. 16). Only at the end of this expansion, she seems to consider the explanation
interaction to be complete as she stands up and walks away after the student’s claim of
understanding in line 18. Extracts like this one show us that teachers not only monitor their
students’ progression in the explanation of the exercise under discussion, but moreover
prepare them for individually completing similar exercises they may encounter next by
explicating the general rules even after students have demonstrated that their problem was
solved for this particular exercise.
An additional example of an explanation interaction coming to a close after a claim,
preceded by an earlier demonstration is displayed below. This extract is exceptional in our
7 STU: heh? (0.5) oh. geef uit,
huh oh hand out
8 (0.3)
9 TCH: jha:! uitgeven.
yes (to) hand out
10 (0.2)
11 TCH: .H het is @een werkwoord. wat je in twee stukjes kunt (0.3)
.H it is one verb which you can divide into two
stu: -–>@
12 verdelen.
parts
13 (0.3)
14 STU: o@ke.
okay.
stu: @sits up more straight–>
15 (0.4)
16 TCH: zie je dat?=[dus dat] zijn &sa:hmen &is dat het gezegde. H
do you see that so those are together that is the verb phrase H
17 STU: [(oh ja)]
oh right
tch: —->& &stands up straight-–>
18 STU: (°oja°)
o right
19 (0.5)@&(1.2)
tch: &walks away-–>>
stu: –>@leans forward and starts writing–>>
10 M. N. GOSEN ET AL.
1 ((STU standing at teacher’s table))
2 @ (3.7) @(2.0)
stu: @flips through book@
3 TCH: ja:,=je hebt een vraa[g. ]
yes you have a question
4 STU: [°ik] snap de (opdracht)
I don’t fully understand
5 niet @helemaal.°
the (exercise)
stu: @puts book down on tch’s desk and leans forward–>
6 .HH [@ °want j]e moet e:h (0.4) het
.HH because you have to uh the
7 TCH: [@(eh je moe-)]
uh you have to
stu: –>@stands slightly leaned over to table–>
8 STU: <werkwoord> in tegenwoordige tij[d° ]
verb in present tense
9 TCH: [nou]:,=je
well you
10 schrijft eerst,
first write
11 (0.2)
12 STU: °ja.°
yes
13 (.)
14 TCH: <IS> ut tegenwoordige tijd of verleden tijd.
is it present tense or past tense
15 (0.3)
16 TCH: dus .hh met een betraand gezicht, puntje puntje,
so .hh with a tear-stained face Irina dot dot
17 Irina gisteren naar de foto van alexander.=
at Alexander’s picture yesterday
18 =moet dat
should that (be)
19 (0.4)
20 STU: °verleden tijd.°
past tense
21 (0.3)
22 TCH: verleden tijd. ja.
past tense yes
23 .hh °en daar achter schrijf je dan het
.hh and behind that you then write the
24 werkwoord.=op de goe[ie man@ier.°]
verb in the right way
25 STU: [ oo:@h ] (.) oke.=
oh okay
stu: –>@stands upright–>
Extract 4. Past tense (J/20_int.4675DL)
CLASSROOM DISCOURSE 11
dataset, since this is the only example in which a student comes up to the teacher’s desk
with their request for assistance. In all the other extracts, it is the teacher visiting
a student’s desk and as such, being in control over the ending of the interaction (i.e. by
walking away from the student’s desk). Extract 4 will show that the teacher is not the only
one monitoring the interaction and determining when the explanation interaction can be
closed. In this extract, we see that the student also clearly orients to closing down the
interaction with both his body and words once the teacher seems to have produced
a complete explanation.
At the start of this extract, the student stands next to the teacher’s desk, flipping through the
pages of his book, and puts the book down on the teacher’s desk once invited to ask his
question (l. 2–5). Consequently, the student will have to pick up the book again and go back
to his own table when the explanation is completed. This is exactly what we see him orient
to when the teacher’s explanation is potentially complete in line 24. While producing an
‘oh’-prefaced claim (‘oh okay’, l. 25) he starts standing upright again (note that he already
demonstrated to be able to complete the exercise by providing an answer in line 20). When
the teacher then summarises her explanation of the exercise with, again, a reference to
a more general instruction to be applied to this specific type of assignment, rather than to
this particular exercise (just like in extract 3), he takes his book and starts walking away even
before the teacher has completed her turn (l. 26–28). In this way, this extract clearly shows
that students also orient to the closing of the explanation interaction once the explanation
seems to be complete and a demonstration has already been produced (and a claim hence
suffices here). The altered body posture in line 25 functions as a pre-closing element
projecting the end of the explanation interaction. The fact that the teacher nonetheless
continues with a concluding summary of the instructions for this (type of) exercise shows us
that student and teacher together negotiate the closing of their explanation interaction (just
like in extract 3) at the end of which one of both parties will walk away from the other’s desk.
- Discussion
The presented extracts have shown that a close analysis of the fragments containing an
‘oh’-prefaced turn following the teacher’s explanation provides more insight into how the
interactants organise one-to-one explanation interactions in relation to knowledge
26 TCH: =ja?=>dus je [zegt<] veetee streepje: (0.2) en
yes so you say pee tee (past tense) dash and
27 STU: [ja. ]
yes
28 TCH: dan: [@wat] het moet @zij[n.]
then what it should be
29 STU: [@ja ] [is] goe[d.]
yes all right
30 TCH: [ J]A?
yes
stu: –>@takes book—-@walks away–>>
31 STU: ja.
yes
12 M. N. GOSEN ET AL.
displays. The analysis reveals that the closing of the explanation interaction is negotiated
by both participants, verbally as well as bodily in relation to their displays of changes in
knowledge distribution. Additionally, the analysis has shown that teachers tend to move
beyond the exercise occasioning the explanation interaction. By referring to general rules
for parsing, for instance, teachers show to assist the students on a more general level as
well (see also Koole and Elbers 2014).
The close investigation of the explanation interactions in this paper has revealed more
insight into the process of conducting a conversation analytic study with a focus on
learning processes in education. Earlier, we described that one often starts with a general
research interest into a specific activity in education. Usually, this interest was the main
reason for collecting the data. Repeatedly looking at the data then results in more specific
research interests, followed by a close analysis of the data. This process of collecting,
transcribing, and iteratively analysing the data has been outlined in the foregoing and
reached its final stage in the results section of this paper. Hereby, the paper illustrates the
data-driven, qualitative approach of Conversation Analysis (CA).
By the close analysis of the turn-by-turn organisation of the interactions, CA highlights
the importance of the details that become visible by studying the video data and
transforming them into detailed transcripts. Although this method is considered to be
time consuming, we have shown in the results section that we could not have made the
empirical claims without a close investigation of the interactional details. It is in the
description of the close analyses in the results section of a CA paper that a reader should
get a grip on the real-time processes that participants themselves show an orientation to.
The study of sequential details of ‘oh’-prefaced turns contributes to earlier findings
revealing insights into one-to-one explanation interactions in education (such as
Breukelman et al. 2023; Koole 2010, 2012). By a close investigation of the ‘oh’-prefaced
turns as a particular practice for displaying change in response to the explanation
delivered in the preceding turn or turns, we have taken the emic participants’ perspective
as is commonplace in the CA tradition. Such an analysis enabled us as researchers to
understand what is happening from moment to moment in the classroom activity and
therefore to come close to the educational practices as they are experienced by teachers
and students while interacting.
By showing the analyses of the interactional organisation in the results section of the
paper, readers of a CA paper also get the opportunity to come close to the original data
and to check the analyses conducted by the researchers. This is different from papers
reporting statistical results on the basis of interactional data. For the reader, the actual
practices then become more abstracted codes that one cannot check, even when the
researchers of these papers have done ‘microanalytic’ coding from second to second (e.g.
van Braak et al. 2021). CA papers on the contrary show the second to second, moment by
moment, interactions as they happen in real time. This is also helpful in translating
findings to educational practice. We can immediately illustrate to practitioners what is
happening in interaction.
Nevertheless, one has to consider that CA papers typically only show representative
extracts of the complete data set. This enables the reader to check the analysis of the
presented extracts, but does not give them access to the entire collection (let alone the
video recordings), hence also missing the ability to double-check how representative the
selected excerpts actually are. A reader therefore still has to trust the analyses of the
CLASSROOM DISCOURSE 13
researchers who conducted the study. Their findings can be said to be validated by
a typical feature for CA: the findings get established by repeatedly looking at the data
and by working on the analyses with a team of researchers conducting data sessions and
discussing the patterns that occur in the data.
The analyses in the current study have shown that teachers as well as students are
oriented to the change-of-state token ‘oh’ as indicative of students’ learning processes.
Both parties treat ‘oh’-prefaced claims as insufficient for closing the explanation interaction
and a demonstration is pursued when the student has not provided one yet. Teachers
as well as students typically start walking away and seem to be satisfied with an ‘oh’-claim
when a demonstration has already been produced. Taking the participant perspective
that is central to CA research and hence basing our analyses on what the interactants
make visible to each other, we were able to uncover the prevalence of the orientation to,
and negotiation of, change in response to the previous actions for the organisation and
closing of the explanation interaction.
These outcomes align with the findings concerning students’ use of ‘oh’ in one-to-one
classroom interactions (Koole 2010, 2012) and add new sequential insights into explanation
interactions in relation to displays of knowledge. While Koole mainly focused on the
difference between free-standing ‘oh’s and ‘oh’-prefaced demonstrations (Koole 2010),
our study zooms in on ‘oh’-prefaced turns and explicitly relates these to (the organisation
- of) the activity, the learning processes and teacher practices involved. The relationship
between previous findings and our results illustrates another feature of validity in
a qualitative sense, as we reason by analogy, describing similarities and differences between
the found patterns in the collection with those in other contexts (e.g. Smaling 2009).
Comparing our results with another strand of CA papers in classroom interaction
research, we would like to underline that our insights did not elucidate changing participation
in classroom interaction and its relationship with learning which has been the
centre of attention in other CA studies (as for instance summarised by Gardner 2019;
Gosen and Koole 2017). There is a line of CA studies investigating longitudinal development
of learners by focusing on their changes in participation (e.g. Melander and
Sahlström 2009; Slotte-Lüttge, Pörn, and Sahlström 2013). More specifically, there is
a strand of research called CA-for-SLA, for second language acquisition, focusing on the
changes in participation that become visible in interactions of learners engaged in
learning a new language (for an overview see for instance Kunitz, Markee, and Sert
2021). In a similar vein, there is a trend in CA research illustrating how certain sequential
moves contribute to an increased student participation, which is expected to contribute
to their development (e.g. van Balen et al. 2022; Willemsen 2019; Willemsen et al. 2020).
Taking these perspectives into account was outside the scope of this paper, but these
kinds of research from a CA perspective are for instance described in overviews of CA and
educational research, such as Gosen and Koole (2017), Koole (2015) and Gardner (2019).
This paper serves to illustrate the method of CA from beginning to end. We
have sketched the advantages of the tools CA has developed to describe what is
happening in interactions on a turn-by-turn basis from a participant perspective to
get a better grip on the workings of classroom interactions. A possible downside of
the use of CA is that the analytical procedure is quite demanding for novice
researchers. First of all, conducting line-by-line transcription of video data is
time-consuming. In addition, the iterative phases in the analysis require processing
14 M. N. GOSEN ET AL.
time and ask for consultation of other researchers to ensure validity of the findings.
This is achieved by having regular data sessions with other CA researchers
and by presenting work in progress during conferences to receive feedback on the
analyses. This is of course engaging and relevant, but also requires time and an
available network of CA researchers.
Another disadvantage of CA might be that it is a method one has to acquire by
practicing. One can intensively study conversation analytic handbooks and research
articles or practice it by training, still the mastery of CA also means ‘learning by doing
it’. Especially moving from a general research interest to phenomena, to specific
research questions that may be answered by a detailed study of certain practices,
can be hard to achieve for a new researcher in CA. One needs to be able to reason by
analogy on the basis of experience and knowledge of other CA driven research. The
focus on ‘oh’-prefaced utterances for instance came about by encountering these in
the data, having knowledge of the previous literature on ‘oh’ as a change-of-state-
token. In this paper, this previous research is summarised before the results are
presented. However, while conducting the study, the analyses and the literature
research are often done in the form of an iterative process.
For these reasons, for someone new to the field the analytical steps may take time (in
collaboration with more experienced researchers) and effort. This specific contribution
showcasing CA as a method with all its steps involved aims to assist researchers with
a new interest in CA to get a better (first) grip on educational data from this detailed
perspective.
Once the analyses have been done, the next step in our conversation analytic procedure
would be to think about how these findings might find their way back into education.
Quite recently, there has been a development of a successful method to use
conversation analytic findings in training sessions with practitioners: the Conversation
Analytic Role play Method (CARM, Stokoe 2014). With this method, researchers and
practitioners come together for a line by line ‘analysis’ of the data studied by the
researchers. By showing just one line of a transcript accompanied by the video recordings,
practitioners are encouraged to reflect on what they would do next if they were in that
particular conversation. On the basis of this, interesting discussions may emerge about
the interactional organisation that is displayed in the selected fragments. This is proven to
contribute greatly to the practitioners’ awareness of their interactional behaviour. For this
particular study, it would be interesting to pause the video after an ‘oh’-prefaced utterance
and discuss with teachers how they would proceed after this. They would then
become aware of the orientation that we have found as analysts, and a next step would
be to discuss with them whether this is desirable or not. It is not up to us as analysts to
make judgements about this; it is something that practitioners themselves can evaluate
once they have become aware of certain interactional patterns.
Our close sequential analyses of interactional data in the classroom can therefore
contribute to awareness of processes in the classroom that would otherwise not have
been discovered. This takes as a starting point the sometimes overlooked but nonetheless
great importance of interaction in the process of learning (and teaching). It is in interaction
with the teacher and with peers that students display their learning processes.
Therefore, the investigation of change as it becomes visible in interaction is of crucial