Introduction

Multimodal resources such as gestures, facial expressions, and body movements as well as their repetition serve relevant interactive functions in face-to-face interactions. They can help to acquire new knowledge in bodily instruction, as has been examined for dance instruction, for example (Ehmer, 2021; Keevallik, 2010). Moreover, they can establish and ensure mutual understanding, as can be seen in adult–child interactions in retelling a watched cartoon (Graziano et al., 2011) or in the contexts of second language learning (Majlesi, 2015, 2022). In this respect, adults or teachers make use of gesture repetitions after learners’ word searches while naming the searched lexeme and thereby display reference to the learner’s turn. In this way, the multimodal resources are “used practically both for maintaining mutual understanding and for creating teaching and learning opportunities” (Majlesi, 2015, p. 42). So far, however, the phenomenon of gesture repetitions has been largely studied from a cognitive psychological and psycholinguistic perspective (Alibali et al., 2017; Holler & Wilkin, 2011; Kimbara, 2006; McNeill, 1992, 2005). Thereby, most of the studies examine gesture repetitions in the context of descriptions, explanations, and (re-)narrations (Holler & Wilkin, 2011; Kimbara, 2006), and focus on repeated iconic or metaphoric gestures (McNeill, 1992) that visualize an absent or abstract object. As these types of gestures are held to be bound to a specific inner concept of meaning due to their form, cognitive psychological research assumes that the recipient’s repetition of iconic and metaphorical gestures provides insights into his/her process of (shared) understanding towards the previous speaker’s use of gesture (Holler & Wilkin, 2011). While this research focuses on the cognitive functions of gesture repetitions, little is known about the interactive functions of gesture repetitions.

The present study addresses this research gap and examines the interactive function of gesture repetition in peer argumentations of same-aged students. The data shows that children make use of repeating gestures as well as other bodily-visual resources of their co-participants when collaboratively co-constructing turns while strengthening and supporting the co-participants’ argument (Günthner, 2015; Kreuz, 2021). This is reminiscent of the practice of “format tying” that Goodwin and Goodwin (1987, p. 205) observed in children’s arguing: while making an opposition, interactants reuse and modify either the turn or turn components, thus demonstrating the reference of the utterances to each other.

The paper begins with a presentation and discussion of the current state of research on co-constructions in argumentation (“Collaborative co-constructions in argumentation” section), multimodal synchronization (“Multimodal synchronization in talk” section), and the repetition of gestures (“Gesture repetition” section). The following section introduces the examined data and the study design (“Method” section), before presenting three extracts of how children make use of gesture repetitions in conjunction with other resources in collaboratively co-constructed turns. Firstly, the analysis demonstrates that repetitions are not restricted to one resource, but entail multimodal gestalts, which are used to create coherence between parts of a co-constructed argument (“Multimodal repetition in collaborative reasoning” section). Secondly, it examines the sequential placement of multimodal repetitions and shows that collaboration may fail depending on its temporal embedding (“Failing collaboration in multimodal repetitions” section). Thirdly, the analysis demonstrates that co-constructions not only address the speaker who produced the first part of the co-constructed argument, but also third parties as mutual opponents (“Entering into coalition against a third party” section). A concluding discussion summarizes the findings (“Discussion” section).

Theoretical frame

Collaborative co-constructions in argumentation

As the present study examines collaborative co-constructions in the discursive practice of conversational argumentation, the concept of “conversational argumentation” must be clarified first before turning to the concept of co-constructions. The discourse activity of conversational argumentation is constituted by participants setting an emerging communicative problem or divergent claims relevant for further exploration or negotiation (Heller, 2012, 2021; Quasthoff et al., 2017; cf. for educational contexts Muller-Mirza et al., 2009; Baker et al., 2019). In this respect, acts of reasoning are made relevant by the participants as they point out the need to support their claims with reasons to each other (Arendt, 2019; Grundler, 2011; Heller, 2012; Quasthoff et al., 2017; Spranz-Fogasy, 2006). To this end, participants use various pragmatic devices as well as embodied resources to display opposition, objection, but also to agree with other participants (Couper-Kuhlen & Thompson, 2000; Gohl, 2006; Goodwin, 2000; Heller, 2021; Jacquin, 2017; Koshik, 2003; Morek, 2020). When participants negotiate divergent claims or explore a potential problem, the activity itself can be framed as either persuasive or collaborative (Ehlich, 2014; Heller, 2018). The present study is particularly interested in episodes in which the participants collaborate by co-constructing arguments.

Co-constructions are hitherto commonly described as interactive processes in which co-participants jointly construct utterances (Günthner, 2015; Lerner, 2004; Mundwiler & Kreuz, 2018; Szczepek, 2000). According to the syntactic structure, two different forms can be distinguished (Ono & Thompson, 1996; Günthner, 2015; Kreuz, 2021; Kreuz & Luginbühl, this issue): completing and expanding the previous speaker’s turn. While in co-constructed completions a missing element is projected in the first speaker’s turn that is then complemented by the co-constructor, in expanding co-constructions, the co-constructor provides supplements to already finished turns of the first speaker. In each case, syntactic, semantic, and prosodic devices serve as markers for strong coherence between the co-constructed turns, such as syntactic junctions when completing another speaker’s turn (Lerner, 2004), or the repetition of turns and turn components to indicate semantic affiliation (Arendt & Zadunaisky Ehrlich, 2020; Kreuz, 2021), as well as prosodic features when adopting “melody, pause, tempo, loudness etc.” of the previous turn (Couper-Kuhlen, 1996, p. 367; cf. Szczepek, 2000; Bose & Hannken-Illjes, 2020).

However, co-constructions also concern the level of bodily-visual resources that exceed the verbal level. Uhmann (2015), for example, examines how participants co-construct deictic gestures in surgery. She demonstrates how participants manage to identify the located target when the speaker is unable to point to it. Co-participants then present potential targets by pointing in co-constructed expansions to achieve the common goal of action to thereby identify the intended object or point of interest. Furthermore, Heller (2021) demonstrates how, in decision-making processes, the speaker’s facial resources displaying of a “thinking face” (p. 9) serve to invite the co-participants to engage in the ongoing process of the argument development, and lead to mutually co-constructed reasons. In addition, Lerner (2002) examines the function of depictive gestures in co-constructions. He demonstrates how recipients’ depictive gestures that take place simultaneously with the speaker’s ongoing turn can expand and specify the present description. The depictive gesture in the co-construction also visualizes the recipient’s understanding of what the speaker has already exposed. The findings of the studies presented therefore reveal that bodily-visual resources need to be taken into account when investigating co-constructions, as the participants themselves make them relevant to the completion or expansion of the turns or actions of the co-participants.

Finally, by completing or expanding the turns of the previous speakers, several interactive functions can be achieved: participants can either use co-constructions collaboratively in the sense of displaying agreement (Günthner, 2015, p. 60; Kreuz & Luginbühl, 2020, p. 101), or, in a persuasive way, they only use the syntactic structure of the previous turn to introduce their own (discordant) argument (Günthner, 2015; Kreuz, 2021). As the paper focuses on the function of multimodal repetitions in collaborative co-constructions, it sheds light on how the resources of the co-constructor are synchronized with the speaker.

Multimodal synchronization in talk

This section discusses relevant findings on multimodal synchronization and, more generally, the coordination of speech and bodily resources. The coordination of various communicative resources has been investigated in various scientific fields, such as early ethnology (Efron, 1941/1972), kinetics (Birdwhistell, 1970; Scheflen, 1964), psychology (Condon & Ogston, 1966, 1971), anthropology and gesture studies (Kendon, 1970, 1972, 1990), and also in early conversation analytic studies (Heath, 1984, 1985; Schegloff, 1985). Thus, a variety of synonymously used terms for these coordination processes can be observed, such a synchrony/synchronization (Condon & Ogston, 1971; Ehmer, 2021; Hannken-Illjes & Bose, 2019; Kim, 2015; Krug, 2022; McNeill, 1992, 2005; Pfänder et al., 2017; Stevanovic, 2021; Szczepek Reed, 2011), coordination (Deppermann, 2014; Deppermann & Schmitt, 2007; Heath, 1985; Kendon, 1970; Mondada, 2014; Streeck & Hartge, 1992), resonance (Breyer & Pfänder, 2017; Warner-Garcia, 2013; Zima, 2017), and matching (Couper-Kuhlen, 1996; Lerner, 2002; Majlesi, 2015, 2022).

Although all these terms are linked in a broad sense, they differ in their specific foci. The concept of coordination can be understood as being an overarching concept that concerns the unfolding structure of interaction, including temporal, multimodal, and spatial aspects, as well as the finely tuned orientation of the participants towards each other (Deppermann & Schmitt, 2007, p. 23). The concept of resonance, however, emphasizes that the speaker’s action causes a specific response by the recipient (Breyer & Pfänder, 2017) and the term matching refers more to a recipient-oriented perspective which captures how his/her resources are aligned with the speaker (Couper-Kuhlen, 1996; Lerner, 2002). The last term poses difficulties, not least because it emphasizes the fit concerning the form (Couper-Kuhlen, 1996) and temporal-sequential unfolding (Lerner, 2002) between the first and the second speaker's utterance. The concept of synchronization, however, considers both the speaker and the recipient producing “corresponding actions” (Ehmer, 2021, p. 2). Typical of these actions is that they are closely linked due to their temporal configurations as well as their similarity in form, as shown by Ehmer (2021) for dance instructions. This paper examines the synchronization process in gesture repetitions in the corresponding action of co-constructions, in which the “gesture performed by the interlocutor has a subset of gestural features in common with the previous gesture of the speaker” (Kimbara, 2006, p. 41). Hence, it does not have to be mirrored in its whole gestalt (Ehmer, 2021) but in its “subcomponents of gesture such as handshape, movement, orientation” (Kimbara, 2006, p. 41).

Gesture repetition

As is known from early research by Goodwin and Goodwin (1987, p. 205), participants use the practice of repeating and modifying other participants’ turns, in the sense of “format tying,” especially in argumentative activities, to strengthen their own counter-position or to mark a source of trouble. In a later work, Goodwin and Goodwin (1992) examine the repetition of gestures in terms of aping the speaker while s/he continues narrating. By re-using verbal as well as multimodal elements of another speaker on the one hand, the turns as well as actions are linked to each other and the source of the upcoming turn is clearly recognizable. On the other hand, the initial meaning is changed by placing it in a new context (Goodwin, 2006). However, verbal and multimodal repetition in interaction can also serve functions that are different from showing disalignment. They can further indicate difficulties in understanding within a responding turn (Bolden, 2009; Schegloff, 2007), but also display agreement and affiliation (Pomerantz, 1984; Clark & Bernicot, 2008; Mundwiler & Kreuz, 2018; Arendt & Zadunaisky Ehrlich, 2020). Even though the functions differ, there appear to be some overlaps in the findings of these studies on verbal repetitions: they all state that speakers use repetitions to display a specific proximity or coherence (Tannen, 1989; Arendt & Zadunaisky Ehrlich, 2020) to the previous participant’s turn due to its temporal configuration and similarity in verbal or syntactical form. This is one way of showing that the participants are working mutually on a joint action (Schegloff, 2007). However, a similar picture emerges for the repetitions of multimodal resources as is shown for gesture repetitions, in which the coherence and reflexivity of turns become visually perceptible to all interactants (de Fornel, 1992; Koschmann & LeBaron, 2002; Lerner, 2002).

Whereas cognitive psychological and psycholinguistic studies focus on the question of what gesture repetition reveals about the inner, i.e., mental, processes of the current speaker (Holler & Wilkin, 2011; Kimbara, 2006; McNeill, 1992), studies inspired by multimodal conversation analysis are interested in the interactive function of gesture repetitions according to the contexts in which they appear, such as in dance instructions (Ehmer, 2021; Keevallik, 2010), collaborative brainstorming (Yasui, 2013), or (second) language learning contexts (Belhiah, 2012; Majlesi, 2015). Yasui (2013), for example, investigates group discussions in which participants work together to develop ideas for scene drafts in a jointly produced film. Here, the participants repeat depictive gestures (Kendon, 2004; Streeck, 2009) of the co-participants while co-constructing the previous speaker’s turn. However, differences in the form of gesture repetition occur. The co-constructors tend to repeat the previous speaker’s entire gesture when fully accepting the expressed proposal, and only partially repeat the source (Majlesi, 2015) or model gesture (Kimbara, 2006) when adding new aspects or displaying disagreement. In this context, the repetition of depictive gestures is considered a “social practice” (Yasui, 2013, p. 158) to show that the participants build on an established and shared knowledge base (ibid., p. 170).

Majlesi’s study (2015) reveals similar findings for gesture repetitions in language learning activities in Sweden. His data demonstrates that teachers repeat the learners’ depictive gestures in verbal corrections, for example. When learners had difficulties in finding a lexeme, teachers repeated the learners’ depictive gestures and named the missing lexeme to display their understanding as well as their “interactive co-engagement” (ibid., p. 42) in the student’s turn. Furthermore, by repeating the depictive gesture in a verbal correction, the correct lexeme and therefore the learnable source are highlighted. Thus, gesture repetitions can also serve the purpose of “creating teaching and learning opportunities” (ibid.; see for similar findings in the context of language cafés Majlesi, 2022). Even though the contexts differ, both studies pinpoint similar aspects: gesture repetition can serve to indicate that the speaker refers to the same knowledge base that is established by the previous speaker’s turn and within the ongoing interaction. Gesture repetition can also indicate mutual understanding in learning contexts, for example, as the repeated previous gesture ensures that both participants talk about the same target. Finally, the studies reveal that gesture repetitions also display the participants’ interactive co-participation in the joint action. In collaborative settings in particular, i.e., collaborative decision-making processes, the repetition of other participants’ gestures “can be one easily available resource for indicating acceptance of others’ ideas” (Yasui, 2013, p. 163).

However, still, little is known about gesture repetitions in collaborative argumentation (but cf. for confrontational contexts Goodwin & Goodwin, 1992), as previous studies have mainly dealt with its function in retellings, descriptions, and explanations or instructions. Considering other multimodal resources as well, this study addresses the overarching research question of which specific interactive functions multimodal repetitions serve in collaborative co-constructed turns in the discourse activity of children’s conversational argumentation. To this end, the study broadens the perspective by not only examining the repetition of depictive gestures, but also of pragmatic gestures (Kendon, 2004; Müller, 2004; Streeck, 2009). Pragmatic gestures assume relevant function in argumentative contexts, as they affect the understanding of an utterance and instruct the recipient how to interpret it, by “presenting” an abstract discursive object to another participant on a palm faced upwards gesture (Kendon, 2004; Müller, 2004), for example, or by highlighting aspects of speech by beat gestures (Kendon, 2004). Whereas depictive gestures concern the semantic level of speech, pragmatic gestures operate on the level of discourse (Müller, 2004). Hence, by investigating multimodal repetitions in peer argumentation, the study examines (a) how interactants synchronize their resources interpersonally during the process of co-constructing arguments and (b) which interactive consequences multimodal repetitions entail in argumentative contexts.

Method

The data for the present study stems from a larger corpus of 30 videotaped argumentative group discussions (Heller, 2018, 2021), from which 21 discussions were used for the present analyses. To study multimodal repetitions as an interactive practice rather than as an acquisition phenomenon, groups of discursively proficient children were selected. The groups were composed of three to six same-aged children (7.0 to 13.6 years), who are from different social milieus and attend the same class in North Rhine-Westphalia. They were asked to deal with a fictitious problem, i.e., a shipwreck scenario (Kreuz & Luginbühl, 2020, this issue), that requires mutual decision-making to jointly explore and solve a communicative problem (Quasthoff et al., 2017): the groups were asked to imagine that they were travelling on a ship and are suddenly caught in a storm that capsizes the ship. They now have to jointly discuss which three out of nine items they want to take with them on their escape to a desert island. Due to this setting, it was likely that their choices would diverge and the children would steer into the discursive practice of argumentation (Quasthoff et al., 2017).

For the present study, the 21 videotaped group interactions were annotated for co-constructions in which gesture repetition occurred. Even though pointing gestures also occur in multimodal repetitions, the study at hand focuses on depictive and pragmatic gestures (Kendon, 2004; Müller, 2004; Streeck, 2009), since the present data largely shows that pointing gestures are particularly used to establish reference and introduce an item. They do not tend to be used while reasoning, however. The collection comprises 13 cases of repetitions of depictive and pragmatic gestures. With respect to the research interest, the study focuses on expanding and completing collaborative co-constructions during argumentative negotiating.

After annotating the data, the respective sequences were transcribed according to GAT II (Selting et al., 2011) and expanded by the transcription of multimodal resources such as gesture (gs), gaze (gz), and body posture (bo), which are annotated with corresponding abbreviations under the English translation of the verbal transcript (Goodwin, 2000; Heller, 2021; Mondada, 2018). To protect the personal rights of the interactants, all names were pseudonymized and the stills were anonymized with an image-editing program.

Focusing on multimodality, the analytical approach is based on the multimodal interaction analysis (Goodwin, 2000). Therefore, the conversation analytic approach, which particularly focuses on the sequential organization of talk, is expanded by the microanalysis of multimodal resources. This analytic approach allows for the capturing of how various resources are temporally coordinated and functionally used in social interaction (Mondada, 2014, 2018).

Multimodal repetitions in collaborative co-constructions

As collaborative co-constructions serve as interactive practices in decision-making processes to strengthen and deepen other participant’s positions or reasons (“Collaborative co-constructions in argumentation” section), the repetition of gestures (“Gesture repetition” section) occurs repeatedly as a recurring element, which makes the collaboration with another party visually perceptual for the co-participants. Thereby, multimodal resources like pragmatic and depictive gestures emerge with other resources such as body posture and gaze as well as gestural movements that together form a well-orchestrated complex multimodal gestalt (Mondada, 2014).

Therefore, the first analysis (“Multimodal repetition in collaborative reasoning” section) demonstrates that not only single gestures but entire multimodal gestalts are repeated in co-constructed argumentative turns. In this way, a strong coherence between the turns is created that strengthens the mutual constructed argument. The second analysis (“Failing collaboration in multimodal repetitions” section) shows that co-constructions may fail when the sequential embedding of the co-constructed turn is missing. As a consequence, the co-constructed turn does not receive attention on the part of the previous speaker. The last analysis (“Entering into coalition against a third party” section) describes how the co-constructor’s bodily orientation changes when entering into a coalition against a third party, while the resources of the multimodal repetition remain adjusted to the previous speaker’s gestures.

Multimodal repetition in collaborative reasoning

The first analysis demonstrates how multimodal repetitions in their entire gestalt serve as markers of coherence between turns, and thereby strengthen the mutual position of the speakers in co-constructed completions. In the following excerpt, the four third graders (Faris, Sila, Sejla, and Damira) discuss the fictitious shipwreck scenario (“Method” section) by proposing and negotiating various items from which they can choose. In the beginning, two lines of argumentation are pursued simultaneously so that a schism (Egbert, 1997) is established: while Sila and Damira negotiate the item “knife,” Sejla repeatedly introduces the “flashlight,” which is agreed on within an expanding co-construction by Faris. After renewing the proposal a third time, Sila turns her attention to the other line of argumentation. She now co-constructs Sejla’s proposal by a completion and repetition of the pragmatic palm-up gesture with lateral movement that occurs as part of an entire multimodal gestalt.

figure a
figure b

In this excerpt, a series of three co-constructions can be observed. Although the focus is on the third instance in which Sila co-constructs Sejla’s and Faris’ argument, it is necessary to show how it builds on the first two co-constructions.

The first instance entails three components that, in consequence of their steady renewal, are also part of Sila’s co-construction in the third instance: a proposal for the flashlight, an accompanying palm-up gesture, and a co-constructed reason. Sejla proposes “the flashlight” (l. 073), pointing at the item and then transforming the pointing gesture into a short-lasting pragmatic palm-up gesture (Kendon, 2004; Müller, 2004). Her gaze remains on the sheet of paper; thus, she addresses no co-participant directly. However, Faris agrees with her by using an affirmation particle and a repetition of the proposal (l. 074: “yeah flashlight”). He then expands the turn by co-constructing the supporting reasoning, “so that one can see” (l. 076) while imitating the movement of swaying the flashlight. Even though he is addressing first Damira and then Sejla by gaze, both co-participants do not take up the proposal in the further course of interaction.

In the second instance, the three components, i.e., the proposal, the palm-up gesture, and the co-constructed reason, are renewed and thereby slightly modified. Sejla reintroduces her proposal (l. 084: “a flashlight”) now with an indefinite article as if it is established for the first time, while turn-initially and in overlap with Damira’s turn, pointing at the flashlight with one hand upwards and the other downwards. At the same time, she raises her eyebrows slightly, indicating a change of state (Dix & Groß, 2023) that, together with the pitch up-step as well as with the pitch accent, serves to add news value. She then transforms her pointing gesture (cf. #1) into a two-handed pragmatic palm-up gesture (Kendon, 2004) in the middle of the interactional space of the co-participants as if to present the discussed object on the table (Müller, 2004). Her hands retract shortly and then transform again into a palm-up gesture with a lateral movement and a shoulder shrug (cf. #2), which provide information about her epistemic stance and together mark her proposal as “obvious” (Kendon, 2004, p. 277; Debras, 2017). Here, an intensified use of multiple resources can be observed. However, Sejla’s gaze remains on the table and does not address a co-participant directly. Again, her proposal is only heard by Faris, who agrees and reintroduces the reason (l. 086: “yes so that one can see”) providing a pitch accent on the verb “see,” accompanied by a depictive gesture. At this point, it should be noted that Faris recycles the turn component of Sila (l. 085: “so that one”) and adds his previously brought in reason again. However, he now repeats Sila’s gesture, who had begun to depict the movements of thrust a second earlier. On the level of form, the turns seem to belong together because of the recycled turn component, the coordinated depictive gestures, and Faris’ gaze to Sila. Nevertheless, Sila is still challenging Damira and does not refer to Faris’ turn.

The third instance entails Sila’s co-construction of the (again) renewed proposal by finally repeating and completing the previous reason. Besides verbal repetitions, the collaborative co-construction comprises repetitions of several resources, such as prosody, pitch accent, gestures, and gestural movements that together form an entire multimodal gestalt. In the beginning, Sejla renews her proposal for the third time, and while the verbal resources remain stable on the lexical and prosodic level, her bodily resources, i.e., gaze and pointing gesture, are oriented to the sheet of paper, which marks her as the potential next speaker. She then reintroduces the item using the indefinite article “a” (l. 089) and again transforms the deictic gesture into a two-handed palm-up gesture. Simultaneously, Sejla shifts her body and turns her head in the direction of Sila, making her the recipient of the proposal through her bodily orientation and her gaze. In the following, Sila co-constructs and expands Sejla’s and Faris’ argument by also using multimodal repetitions, which create a strong coherence between the turns: Sila shifts her gaze and body to Sejla, and agrees using the particle “yeah” (l. 090) while strengthening her agreement with a nod. She now repeats the reason that Faris gave first on the verbal level (Faris in l. 076; Sila in l. 090: “so that one can see”) with the same prosody and pitch accent (cf. “Collaborative co-constructions in argumentation” section). She also adds a completion of the reason by saying “who is there” (l. 090). When repeating the reason, she also repeats Sejla’s both-handed pragmatic palm-up gesture as if to present the reason in the middle of the participants (Müller, 2004) while she is looking at Sejla. Her elbows are resting on the table while her hands are slightly raised, like Sejla’s (cf. #3, #4). At the same time, Sejla again makes a brief lateral movement with the held palm-up gesture, as she did in the second instance (l. 084), and then withdraws her gaze and gesture. Sila then transforms her palm-up gesture in a palm lateral gesture as well, contextualizing the completion of the reason (l. 90: “who is there”) as “obvious” (Kendon, 2004, p. 277). It therefore becomes apparent that a strong coherence between the turns is not only established on the verbal, i.e., turn-constructional, and multimodal level, but this also concerns the epistemic stance: in repeating both the palm-up gesture and its lateral movement, Sila displays the obviousness of the co-constructed argument as Sejla did before to highlight her proposal as obvious.

To summarize, it is to be noted that repetitions within co-constructions are not restricted to single layers of multimodality, but entail entire multimodal gestalts (Mondada, 2014). In this respect, they are highly synchronized with the co-participants’ resources, and due to the synchronization and co-orientation to each other, the collaboration of specific participants becomes publicly visible for all interactants. In the course of this, multimodal repetitions serve to create strong coherence between the turns. Furthermore, the analysis shows that coherence can also be observed in terms of the epistemic stance: by repeating the palm-up gesture with a lateral movement, Sila aligns with Sejla’s epistemic stance that the choice is obvious. However, she slightly changes the point of reference of the gesture: whereas Sejla had temporally coordinated the gesture with mentioning “flashlight” (l. 084), Sila coordinates the repetition of the enriched gesture with a reason (l. 090: “so that one (can see)”) which complements the argument begun by Sejla. In this way, the reason (and not the proposal itself) is marked as obvious.

Failing collaboration in multimodal repetitions

The second analysis shows that the success of multimodal repetitions in collaboratively co-constructed turns, i.e., the approval of a co-constructed turn by the previous speaker, depends to a considerable degree on its sequential embedding: when a co-construction is realized although the current speaker displays a continuation of his own turn, the collaboration fails, even though the co-construction supports the first speaker’s argument. In the following, the four-graders Jannik, Aidan, and Dustin are involved in the decision-making process. In the beginning, Jannik makes his first list of proposals and expands the use of the knife in particular by illustrating it both verbally and by performing chopping and cutting depictive gestures (Streeck, 2009) at the bodily level. This is co-constructed by Aidan, who expands the scenario and repeats Jannik’s depictive gesture of cutting. Although Aidan’s turn supports Jannik’s proposal, Jannik shows only little acknowledgement with a short reception signal, and continues his previous turn by addressing the third co-participant, Dustin, again. The analysis thus demonstrates how collaborative co-constructions may fail due to the lack of timing in the sequential embedding of the turn.

figure c
figure d

In this instance, the focus lies on Aidan’s (failing) expanding co-construction that he provides as Jannik builds up a scenario in which the item “knife” might be useful (l. 041: “the knife in contrast can cut down trees” / l. 044: “also (um) coconuts”). The analysis therefore sheds light on the lack of its sequential embedding.

At the very start, Jannik proposes a list of three items (i.e., knife, lighter, first aid kit), by naming and pointing at them (ll. 035–37), so that the currently discussed items become visible for all interactants. After finishing the list, Jannik shifts his gaze to Dustin (l. 037), but a response remains absent (l. 038). Jannik then reinforces his turn (l. 039: “think about it”) by continuing to gaze at Dustin. Instead of Dustin, Aidan responds with an elliptical concession, in which he shows weak agreement (Pomerantz, 1984) with Jannik (l. 040: “yeah they need also”) and then proposes a new item by pointing at the flashlight. However, Jannik continues his previous turn in an overlap with Aidan and illustrates what his proposed item “knife” might be used for. This leads Aidan to break off his turn.

In the following, a scenario is established including a depictive gesture of chopping that becomes part of the later co-construction: gazing at Dustin Jannik creates a scenario, in which he illustrates that the knife might be used to cut down trees (l. 041), which is also visualized using a depictive gesture (cf. #2). This is followed by a short pause on the verbal level, but Jannik further depicts the motion of chopping and attaches the conjunction “and” (l. 042) with steady intonation. Both the movement of gesture and Jannik’s turn end simultaneously. Note that due to the sequential position of the gesture in the transition space between two turns and without a corresponding lexeme, the depictive chopping gesture seems to be taking place during a word search (Auer & Zima, 2021; Goodwin & Goodwin, 1986). However, Jannik does not elicit a response from the other participants as he now gazes at the paper, but continues his turn (ll. 042–43). Nevertheless, Aidan subsequently engages in a multimodal co-construction and expands Jannik’s turn (l. 041: “the knife in contrast can cut trees”) by adding another object that can be hewn with the knife (l. 044: “also (um) coconuts”), and thereby repeats the depictive gesture of chopping in its form and gestural movement (cf. #3). Within the expanding co-construction, linking the related turns with the lexeme “also” (l. 044), Aidan orients towards Jannik by shifting his gaze and turning his body in the direction of Jannik. Even though he does not offer a candidate item (Pomerantz, 1988) for Jannik’s chopping gesture as the verb is also missing in his turn, Aidan provides another object that can be chopped: the “coconuts” (l. 044). Within a short hesitation, Aidan repeats Jannik’s depictive gesture of chopping (cf. #3). By expanding the turn, Aidan supports Jannik’s proposal as he provides another scenario in which the knife might be useful. Furthermore, the multimodal repetition in the expanding co-construction also demonstrates a mutual understanding of the established scenario (Deppermann & Schmitt, 2008).

Considering the first speaker’s reaction, the lack of sequential embedding of Aidan’s co-construction becomes apparent: Jannik does not orient to Aidan visually (cf. in contrast Sila’s and Sejla’s orientation to each other, excerpt 1), but the overlap with Aidan’s turn shows his shifted orientation to the next item “lighter” on which he is tapping (l. 043). Due to his bodily resources, i.e., his gaze on the paper and the tapping gesture on the item of the lighter, Jannik projects a perceptible continuation of his own turn and further reasoning for the two items left on the list (l. 043). Jannik only acknowledges Aidan’s co-constructed expansion with a minimal reception signal (l. 045: “yes”). He immediately continues his turn, in which he does not refer to Aidan’s scenario but addresses Dustin again while reasoning the next item “lighter” (l. 047: “and lighter can light such a wood there”). This minimal reception signal in conjunction with the immediate turn continuation indicates that Aidan’s collaborative expanding co-construction fails.

The presented example of collaborative expansion demonstrates finely tuned processes of multimodal synchronization by repeating the co-participant’s bodily-visual resources while supporting his reasoned proposals and thereby further demonstrating mutual understanding. However, the microanalysis shows how a lack of timing, i.e., when the current speaker displays a continuation of the own turn, the collaboration is put at risk, as the collaborative turn needs to be produced at a specific, sequentially, and temporally appropriate position in interaction.

Entering into coalition against a third party

As the formation of coalitions is common in conversational peer argumentation (Goodwin & Kyratzis, 2011; Goodwin et al., 2012), the third analysis demonstrates that participants make also use of co-constructions while addressing a third party as the mutual opponent by bodily orientation. In doing so, the co-constructor’s synchronization processes are attuned to both the first speaker and the mutual opponent. Hence, the co-constructed counter-argument is considerably strengthened. The following sequence is from a peer argumentation of the five four-graders Zarif, Yeliz, Deana, Rafik, and Bruno, in which Deana suggests taking the knife that is questioned by Zarif and further reasoned by Yeliz, who previously proposed this item as well. Deana, in turn, supports Yeliz’s reason by expanding it collaboratively and repeating her pragmatic palm-up as well as the subsequent depictive gesture of cutting while orienting to Zarif.

figure e
figure f

In this instance, the analysis focuses on collaboratively co-constructed arguments of Deana and Yeliz that address a third speaker, Zarif, as a mutual opponent who challenged a previous proposal. Deana and Yeliz mutually co-construct arguments for the item of the “knife,” and thereby make use of multimodal repetitions including also pragmatic palm-up and depictive gestures of “cutting.” In terms of their configuration, these gestures are understood as part of one multimodal gestalt.

At the beginning of this sequence, Deana proposes to take the “knife,” which is called into question by Zarif (l. 112, l. 114), who displays a critical stance towards the proposal by contracting his eyebrows and squinting the eyes (Author et al., 2023), which solicits further reasons for this suggestion. Although he addresses Deana by gaze, Yeliz, as the only participant who looks at Zarif, responds.

In the following, a scenario is established including a pragmatic palm-up gesture as well as a depictive gesture of cutting that becomes part of Deana’s later co-construction: Yeliz brings in the scenario (l. 116: “if um if they need something to cook or so”), and in so doing, turns both of her hand upwards as if to present the condition on her open palms (cf. #1, Müller, 2004). As soon as she formulates the consequence (l. 117: “they can cut it with the knife”), Yeliz adds a depictive gesture of cutting (cf. #2). It can be assumed that due to her bodily orientation (cf. #2), Deana perceives both gestures. She further offers a candidate understanding (Pomerantz, 1988) as Yeliz gazes at her by proposing the lexeme “cut” in overlap with Yeliz’s recently produced depictive gesture. This occurs rather early whereas the verbal counterpart “to cut with the knife” (l. 116) is uttered in the final turn position.

After a pause of two seconds, Deana shows agreement with Yeliz’s reasoning by initiating her turn with the particle “yes” and takes over the turn (l. 120: “yes they they go PTCL (down) with the ship”). She refers to the shipwreck scenario verbally and with a deictic gesture, and establishes the scenario as a prerequisite for her subsequent and co-constructed claim, stressed by the particle (l. 120: “ja,” Reineke, 2016) that serves as a marker for mutual knowledge. In the following, Deana’s supportive co-construction of Yeliz’s turn becomes apparent: she claims “they need (it)” (l. 121) and thereby recycles (Arendt & Zadunaisky Ehrlich, 2020) the lexeme “need” of Yeliz’s turn. This creates at least minimal coherence on the verbal level. However, the coherence of Deana’s and Yeliz’s turn becomes more apparent when considering the bodily resources: Deana’s claim is accompanied by a palm-up open hand gesture, which, even though it is one-handed, shows similarities to Yeliz’s two-handed palm-up gesture (Kendon, 2004) within her scenario (l. 116: “if um if they need something to cook or so”). As both participants orient their right-hand palm facing upwards in the direction of Zarif (cf. #1, #3), they seem to be synchronized with each other. Furthermore, Deana makes Zarif the addressee of her turn by orienting herself to him, just like Yeliz, through her gaze and posture.

Yeliz now expands her own turn (l. 122: “or to cut animals”) in overlap with Deana’s further co-construction of the scenario (l. 123: “or for example (if) pirates come”). Both turns are tied together due to Deana’s recycling of the conjunction “or” (l. 122, 123). Deana then adds a further exemplary scenario (ll. 124–126: “and afterwards they have a dispute / afterwards just (he freaks (out))”) by contextualizing it as comical due to her own smile, which later turns into laughing. Within this scenario, Deana repeats elements of Yeliz’s depictive gesture of cutting (cf. #4) still orienting towards Zarif. As the cutting gesture now visualizes a scenario of arguing or fighting, the gesture movement of the repetitive gestures is also slightly modified (Yasui, 2013) and is done with light sweeping movements a little like fencing. However, several aspects remain similar to Yeliz’s gesture, since it is produced with the right elbow placed on the table, the fingers curled to a fist, and the movements performed from a loose wrist.

To summarize, the third example showed that collaborative co-constructions can also be addressed to a third party. Here, Zarif’s critical question occasions Deana’s and Yeliz’s argumentative collaboration, who in turn address Zarif as the mutual opponent. The co-constructors’ bodily resources are highly orchestrated, i.e., the deployed multimodal resources—in their entire gestalt—are synchronized with the previous speaker’s gestures, as the co-participants (re)use pragmatic palm-up gestures as if to present a scenario respectively a claim and then picture a specific scenario by using depictive gestures of cutting. Likewise, the bodily orientation of the collaboration party itself is directed to the mutual opponent. In addition, the third example shows that the gesture repetitions studied can also encompass more than one gesture or form of gesture in the entire multimodal gestalt of repetitions.

Discussion

The present study, based on a corpus of 21 group discussions among children, reveals that the use of multimodal repetitions within co-constructing arguments in decision-making processes appears to be a recurrent communicative practice (Schegloff, 1997) that serves several interactive functions. In this way, the analyses shed light on the synchronization process of multimodal resources within collaborative co-constructions by examining both how interactants synchronize their resources during the process of co-constructing arguments and what interactive consequences multimodal repetitions entail in argumentative contexts. The findings are threefold:

Firstly, in examining gesture repetitions, the analyses show that interactants create strong coherence between collaboratively co-constructed arguments not merely on the verbal, but also on the level of multimodal resources. It is thereby demonstrated that gestures are not repeated in isolation, but form an entire gestalt (Mondada, 2014) of multimodal repetition comprising several resources, such as speech, prosody, and gestures that are repeated as part of the collaborative co-constructed completion or expansion (cf. excerpt 1, 3). This study therefore extends previous research that focused primarily on the resource of hand gestures in repetitions (e.g., Holler & Wilkin, 2011; McNeill, 1992, 2005). Further, the analyses worked out that displays concerning the epistemic stance of the first speaker are repeated as well: in excerpt 1, Sila shows alignment with Sejla’s epistemic stance and repeats the displays that mark a proposal as obvious. Even though the original meaning of the gesture receives a subtle shift due to its placement in the new context of the co-constructed turn (Goodwin, 2006), the coherence between the speakers’ epistemic status becomes apparent.

Secondly, whether or not a co-construction is accepted by first speakers depends on its sequential embedding. As excerpt 2 demonstrates, collaborative co-constructions may fail depending on the lack of temporal-sequential embedding, such as when a turn-transitional point has not yet been reached and the current speaker displays a continuation of the turn.

Thirdly, co-constructors not only address the speaker they are supporting, but also third parties as mutual opponents (cf. excerpt 3). In doing so, the co-constructors show a bodily orientation towards the third party and address the third party with their gaze. Despite the change of bodily orientation, the synchronization of the resources such as pragmatic palm-up as well as depictive gestures remains adjusted to the previous speaker.

Overall, the present study reveals and expands previous findings that largely focused on the role of repeated depictive gestures (cf. “Gesture repetition” section; Belhiah, 2012; Yasui, 2013; Majlesi, 2015, 2022): the study reveals that both depictive and pragmatic gestures are part of multimodal repetitions. More specifically, in eight of 13 instances, pragmatic palm-up gestures, e.g., to present a discursive object or to mark something as obvious (Kendon, 2004; Müller, 2004), are repeated. Furthermore, the present study takes into account the discursive practice of (peer-)argumentation. In this context, where the mode can oscillate between collaborative and persuasive (Ehlich, 2014), the repetition of depictive and particularly pragmatic gestures not only has a coherence-creating function, but is also a visible device for all the interactants of (mutual) collaborations by certain participants. Hence, bodily-visible resources establish temporary intra-group collaborations, and particular positions and arguments therefore receive more weight in the ongoing interaction. This can also affect the social group order, as the participants (visibly) form collaborations and coalitions against opponents (cf. social order in peer group interaction Goodwin & Kyratzis, 2007). Based on the review of the corpus, it also becomes apparent that pragmatic gestures providing information about the intended speech handling (Streeck, 2009) are more frequently part of the multimodal repetition than depictive gestures. From an interaction-oriented perspective, it would appear fruitful to include data from other contexts in the study of multimodal repetitions, such as narratives or explanations in informal peer talk, to gain insights into the question of whether and to what extent the use of depictive and pragmatic gestures in multimodal repetition varies depending on the communicative context. As the present study focused primarily on the first speaker’s and the recipient’s synchronization of multimodal repetitions (but cf. excerpt 2), it might also be relevant to consider the uptake of the co-constructed and strengthened argument in the subsequent discussion by the first speaker or other participants within these multiparty interactions.