Getting Students to Talk: A Practice-Based Study on the Design and Implementation of Problem-Solving Tasks in the EFL Classroom

This study addresses a pedagogical practice-based issue, that is, difficulties with eliciting student–student co-constructed oral interaction in the EFL classroom. The study was conducted with a bottom-up approach to pedagogical research through the close collaboration of teachers and researchers who were equal partners in the research team. It was observed that students often engage in parallel monologues or unauthentic question–response sequences when accomplishing oral activities; thus, the research team aimed to design tasks providing opportunities for meaningful, co-constructed talk. The research design involved an iteration of task design and classroom testing in three cycles, and the student–student interaction was analyzed using conversation analysis. Findings show that the divergent problem-based task designed in this process did elicit purposeful and collaborative oral interaction, as the students engaged in co-constructed talk by visibly attending to each other’s turns-at-talk and by formulating fitting turns that fostered the progressivity of the activity. The task also included artifacts (i.e., material objects), the manipulation of which played an important role in the emerging collaborative interaction. These findings suggest that the implementation of open-ended problem-based tasks can develop students’ interactional competence, while the use of artifacts can help students make their reasoning tangible and visually accessible.


Introduction
The present practice-based, collaborative study focuses on the implementation of problem-based tasks in the EFL classroom in lower secondary schools in Sweden. Specifically, the analysis presented here and conducted with a conversation analytic (CA; see, Sidnell and Stivers 2013) lens shows how students in grade 7 engage in co-constructed interaction as they accomplish problem-based tasks that were designed following Ellis' (2003) framework. At the same time, the findings discussed here highlight "the relationship between participants, learning materials, and (inter)action" (Mathieu et al. 2021, p. 3) that is currently gaining increasing attention within approaches to the study of second language (L2) teaching and learning that see the material world as integral to such processes (Guerrettaz et al. 2021).
Our study is part of a larger project called From monologues to dialogues, financed by Swedish school organizers and universities. It is important to point out that our work is rooted in a practical, pedagogical problem concerning difficulties with organizing oral activities that elicit co-constructed talk in the language classroom. This problem was collaboratively addressed by schoolteachers and researchers with the aid of two different scientific frameworks; namely, Ellis' (2003) take on task-based instruction in general and task design in particular, and conversation analysis (CA; see Sidnell and Stivers 2013) with its focus on the details of interaction (see below). The novelty of our work lies precisely in the fact that it is grounded in a bottom-up approach to research: a problem observed during classroom practice was addressed in and through classroom practice, with a researchinformed pedagogical intervention that was collaboratively designed and subsequently analyzed by a research team of schoolteachers and researchers. This approach is similar to action research in that it takes a pedagogical problem as a starting point (Burns 2010;Elliot 1991); however, action research is commonly understood as teacher professional development (e.g., see: Banegas et al. 2013;Edwards 2021) and is focused on the production of local knowledge (Eriksson 2018). In our case, however, the collaboration entailed that the team members were on equal footing as they engaged side by side in addressing the same research object (Carlgren 2012;Eriksson 2018). Specifically, throughout the process, teachers were "included in the research as interpretative professionals making professional sense of particular educational events" (Carlgren 2012, p. 126). This collaborative way of working with teachers and researchers as equally knowledgeable partners produces "sustainable" (ibidem) results that are more credible and more applicable in the classroom. In this article, we thus describe the collaborative research process at the root of our work and present the findings of such work with a CA lens.
The process of designing tasks that promote collaborative interaction and of analyzing the implementation of these tasks led the research team to reflect on what interaction entails and to eventually adopt CA as the analytical framework for the study. CA views interaction as the primordial site of human sociality (Schegloff 1987) and L2 learning as an inherently social (and socially situated) process occurring in and through interaction (e.g., see: Kasper and Wagner 2011;Sahlström 2011), which in turn is guided by a constant "effort to achieve and maintain mutual understanding" (Kim 2019, p. 323). As evidence from longitudinal studies is cumulatively suggesting (see Pekarek Doehler et al. 2018 and some of the studies in Hellermann et al. 2019), repeated participation in locally contextualized, situated interaction fosters the development of an increasingly diversified repertoire of semiotic resources that allow participants to engage in context-sensitive and recipientdesigned conduct and thereby to show a higher degree of interactional competence (or IC; though see Hall 2018 for the distinction between IC as an underlying universal apparatus versus interactional repertoires as the variable set of semiotic resources for taking action that constitute the object of learning). In other words, language learners need to engage in interaction with the resources at their disposal in order to further develop and diversify their semiotic toolkit. Simply put, L2 learning is inseparable from L2 use (see Eskildsen andMarkee 2018 andKim 2019, among others) as language is both the object and the tool of learning.
A logical corollary of the view of L2 learning as rooted in interaction is that the classroom needs to provide opportunities for meaningful communication in order for students to practice their interactional skills and to acquire the language-specific semiotic repertoires that are essential to use the target language proficiently. This is also suggested in the integrative (i.e., aimed to reconcile cognitivist and social views) framework proposed by the Douglas Fir Group (2016). According to this group of scholars, at the core of the L2 learning process is the "micro level of social activity" (p. 25) represented by "individuals engaging with others" (ibidem) through the use of a variety of semiotic resources (ranging from linguistic to artifactual). At the same time, their interactions are "situated within and shaped by" (ibidem) social institutions, including schools. It is thus clear that schools play an important role in terms of the affordances for learning as social interaction that they provide.
The Swedish syllabus for English in compulsory school, prepared by the Swedish National Agency for Education (Skolverket in Swedish), has a clear communicative standpoint, as evidenced by the statement that the students should "develop all-round communicative skills" (Skolverket 2018, p. 34). Furthermore, it should also be noted that the syllabus is influenced by the Common European Framework of Reference for Languages (CEFR), which adopts an action-based, communicative approach to language education (Council of Europe 2020). In other words, the policy documents that are referred to in language education in Swedish schools have a clear communicative focus. Now, while Swedish students' language proficiency is generally high (Skolverket 2012) and many teenagers use English outside the school for communication purposes (Sundqvist 2009), some students have expressed that they do not get the opportunity to show their communicative skills in English in school (Myndigheten för Skolutveckling 2008).
Similar issues had been observed by the schoolteachers participating in this study. Specifically, these teachers perceived that the student-student interaction that is typically enacted in the L2 classroom resembles either a series of monologues where a change of speakership is marked by fixed expressions such as what do you think? or do you agree?, or unauthentic question-response sequences where the exchanged information is often known beforehand. Put another way, student-student interaction in the L2 classroom does not seem to entail mutual and collaborative engagement; that is, the kind of engagement that is deeply rooted in the attempt to make sense of the coparticipants' actions-in-interaction while establishing and maintaining intersubjectivity.
These issues are similar to issues analyzed in the literature on language testing, which has highlighted that test-takers often produce parallel monologues (Galaczi 2008(Galaczi , 2014 and "monological accounts" (Sandlund and Sundqvist 2016, p. 123); this kind of interaction essentially shows students "talking at rather than talking to each other" (May 2011, p. 137). Overall, then, there seems to be a discrepancy between the institutional aim for the development of communicative skills set by the syllabus and the skills that are actually practiced in the classroom and manifested in oral classroom activities and national exams.
Against this background, the research team hypothesized that the root of the problem might lie with the oral activities that are typically implemented in the classroom. Based on the participating teachers' experience, two kinds of activities seemed to stand out in the L2 classroom: (a) activities where the aim is to practice the use of specific grammatical features, words, or phrases, and (b) opinion-based tasks inspired by the Swedish national standardized tests (see below). The former can be likened with exercises (Ellis 2003(Ellis , 2016 focusing on form, often in the shape of dialogues, whereas in the opinion-based tasks, the participants are instructed to discuss a statement or a question by first expressing their own view and then asking for their friend's opinion (see below). Neither kind of activity provides opportunities for meaningful conversation or promotes collaborative and co-constructed talk. Consequently, the aim of the study was to design tasks that promote collaborative oral interaction in the lower-secondary EFL classroom, in line with the research on taskbased language teaching that emphasizes the importance of "purposeful and functional language use" (Ellis 2009, p. 222). To accomplish this goal, the teachers' knowledge of classroom practices and their vision of the desired interactional outcome were crucial for producing "sustainable" (Carlgren 2012, p. 126) results that could be implemented in the classroom. At the same time, the active and collaborative participation of teachers and researchers in the socially situated process of designing and testing tasks and of analyzing their outcomes allowed for the achievement of professional, pedagogical knowledge that the use of ready-made tasks would not have made possible.
Before we engage in the analysis and discussion of students' task-oriented interactions implemented in the L2 classroom, it is relevant to address the institutional framework within which Swedish teachers work in their local context and the theoretical and methodological framework that we intend to adopt in our study. Therefore, in what follows, we detail the conceptualization of oral proficiency that emerges from Swedish policy documents and from the national exams that inspire the teachers' work in the EFL classroom. We then discuss findings from previous studies on EFL oral interaction in Swedish schools and finally present a conceptualization of oral proficiency as IC that is rooted in CA-SLA work on L2 interactions and in CA's view of language as a tool for social action. Oral proficiency in the Swedish syllabus comprises oral production and oral interaction and is strongly linked to the ability to adapt language use to purpose, interlocutor, and context and to the use of strategies (Skolverket 2018). In relation to oral interaction, these strategies are intended to enable the students to "contribute to and actively participate in conversations by taking the initiative in interaction, giving confirmation, putting follow-up questions, taking the initiative to raise new issues and also concluding conversations" (Skolverket 2018, p. 37). The commentary adds that body language and gestures, as well as speech sounds and intonation, are important aspects of students' interactional skills (Skolverket 2017).
To further conceptualize oral proficiency, Swedish EFL teachers rely on the mandatory national standardized tests, which are carried out in grades 6 and 9 in compulsory school. While the main purpose of standardized testing in Sweden is to improve comparability and equity across schools, these tests are also used "implicitly" in order to "clarify and exemplify the view of knowledge and language expressed in the national curricula and syllabuses" (Erickson 2020, p. 1). In practice, this means that the construct of oral proficiency targeted by the national tests and related assessment criteria shapes the general conceptualization and understanding of oral proficiency in Swedish EFL classrooms by providing some guidance on the interpretation of the syllabi (see also Erickson and Börjesson 2001). For this reason, and for the extensive use of opinion-based tasks inspired by the national tests in the EFL classroom, we deem it important to describe the format of the national exams and the kind of interaction that is typically observed during such exams.
The oral proficiency test, called "Speaking", assesses both oral production and oral interaction and is, in recent years, carried out in pairs, without teacher participation. Normally, the test comprises a warm-up activity followed by opinion-based tasks related to a general theme, such as "The world around us". These tasks are implemented using conversation cards with a statement or a question followed by a short instruction asking the students to, for instance, agree or disagree and explain why. The students are also encouraged to engage in conversation by asking for their friend's viewpoints and to "BE ACTIVE AND SPEAK ENGLISH ALL THE TIME" (Skolverket 2020; capital letters in the original). It is important to note that several topics are covered during a test and that the students are not informed of the topics before the test. In the assessment guidelines, it is articulated that the assessment should be based on the presupposition that the student is willing and able to express and develop content, individually and in collaboration with others (Skolverket 2020). The assessment foci are (i) content and (ii) language and the ability to express oneself. Criteria for content involve intelligibility and clarity, richness and variation, context and structure, and the ability to adapt to purpose, interlocutor, and situation. The criteria for language and ability to express oneself comprise the use of communicative strategies, fluency and ease, range, variation and clarity, confidence and adaptation to purpose, interlocutor, and situation (Skolverket 2020). Turn-taking is not mentioned explicitly in the criteria but can be implied from the description of the communicative strategies (Borger 2019; Council of Europe 2020). To clarify the assessment factors and facilitate grading, teachers sometimes make their own scoring rubrics (Byman Frisén et al. 2021). In many respects, these rubrics resemble the official assessment guidelines, but with the addition of engagement/initiative and interactional skills. It is possible that these aspects are included to simplify the individual grading of a co-constructed conversation (Byman Frisén et al. 2021).
Studies also show that it is common to use sample tasks provided by Skolverket to prepare the students for the test (Sandlund and Sundqvist 2011;Sundqvist et al. 2013) and, as mentioned above, these opinion-based tasks also inspire teachers in the design of their own prompts for oral interaction practice in the classroom. Accustoming students to a certain test format by, for example, discussing the instructions and grading criteria can help the students avoid test-related trouble Sundqvist 2011, 2013). As indicated above, however, it is not just a matter of familiarizing students with a specific test format, in that these national tests also influence EFL teaching by affecting the conceptualization of oral proficiency that language teachers use in school and the opportunities for oral practice that they provide for their students. In the EFL teacher community, it is, for example, common to share lists of "useful phrases" accompanied by conversation cards similar to the ones used in the national tests.
In line with the problem that forms the background for this study, Sandlund and Sundqvist (2016) observe that, even if the national exam is considered a proficiency test (and therefore an objective assessment of what the students can do with language), the task-based interaction enacted during the national exam is very different from everyday conversation. For example, the task design used in the national exams does not encourage the pairs to collaborate and engage with each other's turns-at-talk; instead, they take turns to "exhaust their commentary on a topic" (Sandlund and Sundqvist 2016, p. 128), which means that they engage in relatively long turns with limited overlaps. It is also considered important that the students share the floor (Borger 2019;Sandlund and Sundqvist 2016); that is, that each of them speaks for roughly the same amount of time. Moreover, it is likely that the students' prior knowledge and personal experience of the topics affect their possibilities to initiate and elaborate on the topics to be discussed in the exams.
As mentioned above, for the purposes of this study, we rely on CA methodological affordances for the analysis of interaction, and, more specifically, we adopt CA's definition of IC. The concept of IC as it is used today builds on Hymes' (1972) view of communicative competence as the knowledge of how to use language appropriately in a social context. Such a view was then imported into the field of language education by Canale and Swain (1980) and later framed Bachman's (1990) model for testing communicative language ability. From a CA perspective (e.g., see Kasper and Ross 2013), in these models, components of communicative competence are considered statically, as cognitive properties of the individuals and anchored in a rationalist approach to pragmatics that is grounded on intention-based accounts of social actions, with an almost exclusive focus on the speaker. CA, instead, adopts an action-based account of interaction, which is seen as inherently co-constructed by all the participants, and it thus falls in the field of discursive pragmatics (for a discussion of the differences between rationalist and discursive pragmatics, see Kasper 2006). In CA terms, then, IC is the ability to accomplish recognizable social actions through the production of timely turns that are well-fitting with respect to prior talk and its praxeological (i.e., action-related) import. Being interactionally competent involves the ability to understand and respond to the local, emergent circumstances in which social actions are accomplished through talk-in-interaction (Eskildsen 2018). Therefore, "competence is displayed ( . . . ) in the ways in which the members act on their discourse within the sequential order of their interaction" (Lee 2006, p. 368). Studying IC, then, means studying "the interpretive work of understanding that participants display" (Lee and Hellermann 2014, p. 769) with each turn they produce, with each action they accomplish. Such ability is based on the appropriate use of the mechanisms organizing talk-in-interaction (such as turn-taking, repair, sequence organization, and preference organization; for an overview of these terms, see: Sidnell and Stivers 2013), coupled with the use of various linguistic and embodied resources (Hall 2018;Markee 2008;. We thus see a clear connection between this conceptualization of IC and the "all-round communicative skills" (Skolverket 2018, p. 34) that the Swedish syllabus for English aims for.
Overall, then, our study emerges from the practical need-as perceived by the participating teachers-for oral activities in which the students engage in collaborative interaction where their communicative skills can be clearly observable. As research has shown (e.g., see: Hellermann and Lee 2020; Mori 2002), it is not enough to instruct students to "have a conversation", since this prompt seems to elicit an interview type of interaction organized around a series of question-answer sequences that somewhat straightjackets affordances for topic development and mutual engagement, while reproducing a rather rigid system of turn-taking and turn allocation. Therefore, the research team was faced with the issue of designing classroom tasks that would be different from the task format typically adopted in the L2 classroom, and that would go beyond the "have a conversation" prompt. After a few cycles of design, implementation, and revision of different types of tasks, the research team eventually settled on open-ended problem-based tasks, in which students work with specific artifacts (e.g., cut out figures, actual material objects). While the task design phase was informed by Ellis' (2003) framework and the idea of providing opportunities for meaningful and functional language use, there remained the empirical issue of examining whether this task type might lead to the kind of engaged and collaborative interaction that the teachers were aiming for. The analysis presented here shows how students in grade 7 implemented the task-as-activity (versus the task-as-workplan; see Breen 1989;Coughlan and Duff 1994;Seedhouse 2005) and the kind of interactional features and semiotic repertoires the students used during their task-based interactions. As the analysis will show, the artifacts that were included in the task design played a major role in making visible and tangible the interactional co-construction of imaginative narratives in which the students engaged. Our practice-based study, therefore, contributes not only to CA-SLA research on L2 IC (see above) and on task implementation (e.g.,: Kunitz and Marian 2017; Lee and Burch 2017;Pochon-Berger 2011;Hellermann and Doehler 2010;Seedhouse 2005), but also to the recent strand of research focusing on the use of language learning and teaching materials (Guerrettaz et al. 2021) and on their impact on classroom discourse (Mathieu et al. 2021).

Materials and Methods
As mentioned above, the work presented here is framed within the project From monologues to dialogues, a classroom study carried out by a research team in which researchers collaborated with several primary and secondary school teachers in English and Modern languages (Berggren et al. 2019); the research team designed and implemented oral tasks in their classes. The research design involved a number of subprojects, each comprising an iteration of task design and classroom testing in three cycles (cf. design-based research, Anderson and Shattuck 2012). To facilitate comparability across the subprojects, Ellis' (2003) framework of design features informed the task design in all subprojects. Moreover, findings from previous subprojects contributed to subsequent designs.
The collaboration with schoolteachers ensured the ecological validity of the project, that is, "the degree of similarity between a research study and the authentic context that the study is purportedly investigating" (Loewen and Plonsky 2015, p. 56). As detailed in the introduction, the project was anchored in the teachers' practical problem of implementing oral tasks that would engage students in co-constructed, collaborative interaction for meaningful communicative purposes (and therefore in line with the syllabus for English in Sweden and with the CEFR). To further enhance the ecological validity of the study, data collection was carried out during regular class instruction; this meant that the students' regular English teacher was in charge of the students' implementation of the tasks.
The present paper builds on data from the first iteration of the task What happened to Kim? This task was designed based on earlier findings in the project indicating that problem-based tasks presented with short instructions and relying on the use of artifacts foster students' engaged and collaborative interaction. The instructions were displayed on a screen (see Figure 1) and presented orally by the teachers at the beginning of the class.
The students, who worked in groups of three, were asked to find a solution to the problem faced by Kim, a twelve-year-old who had left home early in the morning, never arrived in school, and was found later in the evening at a petrol station with a paper bag containing eight artifacts (see Figure 1). Each group of students received a paper bag with a mobile phone, a note saying "I love you", a roller coaster ticket, a torn photo, a receipt, a key, some coins, and a cuddly toy which was different for every group (in the data presented here, Group A received a teddy bear and Group B received a fish). The students were specifically instructed to select five artifacts that they thought could help Kim remember what had happened during the day. Essentially, the question the students were supposed to answer was What happened to Kim? The task was divergent in that there was no correct solution to the problem. The intended outcome was that the students would engage in a conversation about Kim's day and select five items to help Kim remember what had happened. a mobile phone, a note saying "I love you", a roller coaster ticket, a torn photo, a receipt, a key, some coins, and a cuddly toy which was different for every group (in the data presented here, Group A received a teddy bear and Group B received a fish). The students were specifically instructed to select five artifacts that they thought could help Kim remember what had happened during the day. Essentially, the question the students were supposed to answer was What happened to Kim? The task was divergent in that there was no correct solution to the problem. The intended outcome was that the students would engage in a conversation about Kim's day and select five items to help Kim remember what had happened. The task was implemented in two classes during an ordinary English lesson: one in grade 7 and one in grade 8. All the students present completed the task, but video recordings were only collected from the students who had consented to participate in the study; the recruitment procedure followed the ethical guidelines issued by the Swedish Research Council (Vetenskapsrådet 2017).
The analysis of the students' task-oriented interaction was completed in two stages: a first stage based on empirically derived categories and a second stage using CA. Both stages were accomplished collaboratively by the members of the research team (teachers and researchers) and intended to pinpoint characteristics of "good" interaction. For stage one, word-only transcripts, including a description of artifact manipulation, were prepared, and the students' utterances were categorized based on their function in the conversation, such as confirmations, new ideas, and elaborations. These categories were discussed among the members of the research team and emerged from their observations of the data and from what they deemed to be instances of "good" interaction. The purpose of this initial analysis was to provide an overview of the co-constructed interaction; however, it was clear that a more fine-grained analysis was needed.
It was at this stage in the project that it was decided to resort to CA and its theoretical and methodological affordances. CA is a "naturalistic, observational discipline" (Schegloff and Sacks 1973, p. 289) that is interested in how participants in interaction make sense of each other's actions-actions which are accomplished through the lamination of various semiotic resources (Goodwin 2013), including talk, prosody, and embodiment. The task was implemented in two classes during an ordinary English lesson: one in grade 7 and one in grade 8. All the students present completed the task, but video recordings were only collected from the students who had consented to participate in the study; the recruitment procedure followed the ethical guidelines issued by the Swedish Research Council (Vetenskapsrådet 2017).
The analysis of the students' task-oriented interaction was completed in two stages: a first stage based on empirically derived categories and a second stage using CA. Both stages were accomplished collaboratively by the members of the research team (teachers and researchers) and intended to pinpoint characteristics of "good" interaction. For stage one, word-only transcripts, including a description of artifact manipulation, were prepared, and the students' utterances were categorized based on their function in the conversation, such as confirmations, new ideas, and elaborations. These categories were discussed among the members of the research team and emerged from their observations of the data and from what they deemed to be instances of "good" interaction. The purpose of this initial analysis was to provide an overview of the co-constructed interaction; however, it was clear that a more fine-grained analysis was needed.
It was at this stage in the project that it was decided to resort to CA and its theoretical and methodological affordances. CA is a "naturalistic, observational discipline" (Schegloff and Sacks 1973, p. 289) that is interested in how participants in interaction make sense of each other's actions-actions which are accomplished through the lamination of various semiotic resources (Goodwin 2013), including talk, prosody, and embodiment. CA was deemed particularly suited to describe the students' task-oriented interaction precisely because of its focus on interaction and, in particular, on L2 IC (see above for definitions of L2 IC in the CA-SLA literature). At the theoretical level, CA adopts an emic (i.e., participantrelevant) approach to data analysis, which is empirically driven and grounded in the participants' observable behaviors. At the methodological level, CA relies on audio and/or video recordings of naturally occurring interactions, which are then transcribed in great detail. A CA analysis is rooted in an action-based view of language and interaction, in that CA is interested in "talk-that-does" (Schegloff 1990, p. 52) rather than "talk-about" (ibid., p. 52); therefore, the focus is on the participants' actions-in-interaction rather than on the topics of conversation.
For the present paper, we chose to work with two groups from grade 7 whose video recording was of good quality in terms of the audibility of the students' talk and visibility of the group's embodied actions. The recording of Group A lasts approximately 21 min, while the recording of Group B lasts approximately 19 min. As mentioned below, the analysis focuses on two stretches of talk that represent a specific phase of the task-oriented interaction, namely, the moment in which students, after manipulating and discussing the artifacts one by one, start formulating the first hypothesis about what happened to Kim by bringing together two or more artifacts (see below). Each excerpt presented in the analysis has been transcribed following Jefferson's (2004) convention for the transcription of talk. Relevant embodied actions are also described; their simultaneous occurrence with talk is marked with a plus sign (+) or with an asterisk (*). Frame grabs have also been included in order to give the reader visual access to some particularly significant moments in the interactions; in the frame grabs, the participants' faces have been blurred for anonymity purposes.

Results
The analysis illustrates excerpts from the two groups of students as they are engaged in a similar phase of their task accomplishment, that is, the co-construction of the first imaginative narrative in which different artifacts are brought together in the attempt to develop an organic story about what happened to Kim. This phase occurs after the initial artifact-manipulation phase, during which the students extract the artifacts from the paper bag and start identifying each one of them. For each group, we analyze the emergence of the first narrative based on multiple artifacts, the additions, agreements, and disagreements through which such narrative is shaped, and, finally, the ad interim summary produced by each group, bringing together the artifacts that have so far been taken as relevant for determining what happened to Kim. This summary also represents the participants' first attempt at providing a relevant task outcome, in that the task instructions (Figure 1) required students to select five artifacts that could help Kim remember what had happened during the day.
Excerpt 1 (divided in Excerpts 1a, 1b, and 1c) reproduces the task-oriented interaction accomplished by group A (Iman, Sam, and Walt). With Excerpt 1a, we join the students as Iman proposes the first multiartifact narrative, bringing together the roller coaster ticket, the teddy bear, and the I-miss-you note.
In line 1, as she is looking at the roller coaster ticket, Iman produces a high-pitched ↑ma:ybe and then adds that Kim got the bear and the note (lines 2-4) while putting the note on the bear. Her first narrative, formulated both embodiedly (through her gazing at the ticket in line 1) and linguistically (lines 2-4), seems to suggest that Kim won the bear with the note at a Luna park. The use of the epistemic hedge maybe (line 1) and the embodied action of showing the note and the bear to Sam (line 4, Frame Grab 1_1) make relevant a response from Sam, who indeed produces the acknowledgment token yeah in line 6. What is relevant to note here is that the holding of the bear and the note in Sam's direction provides material and visual access to the reasoning so far. That is, the manipulation of artifacts seems to have a crucial role in the interactional emergence and unfolding of the narrative.
Sam's turn is followed by a 0.4 s pause (line 7), during which Sam reaches for the photo. Sam then continues Iman's turn with the increment from (.) her (line 8), delivered with emphasis. Here the embodied action of reaching for the photo mobilizes the coparticipants' attention to a new, potentially relevant artifact while disambiguating the referent of her. Sam's actions provide an example of co-constructed interaction, in that-with her acknowledgment of Iman's proposal-so-far (line 6)-Sam displays the understanding that a response is expected from her. At the same time, with the increment that recompletes (Couper-Kuhlen and Ono 2007;Lerner 1991;Schegloff 2016) Iman's previous turn (↑ma:ybe he: got (0.2) this::., lines 1-4), Sam displays the rather sophisticated ability to monitor the coparticipant's turn by adding to it in a fitting way, in terms of action, content, and syntax. Iman displays strong agreement (line 10), as indicated by the loud volume on YEAH and by the use of the intensifier exactly, accompanied by the embodied action of picking the photo from the desk.
After Iman's agreement, Sam continues the developing narrative with and he went to a tivoli, as she reaches for the roller coaster ticket, lifts it, and moves it toward Iman (line 12). By putting her hand on the ticket (line 13), Iman shows receipt of Sam's emerging narrative and continues Sam's turn with the increment WIth (0.3) her (line 15) as she raises her hand holding the bear and the note. Iman's turn is produced in partial overlap with Walt, who agrees (I ↑think • so • , line 14) with the developing narrative. These lines too show the highly collaborative work accomplished in and through interaction, with the participants displaying agreement (Walt, line 14) with the emerging narrative and building on each other's ideas and turns-at-talk (Iman, line 15) while relying on the available artifacts to make visible their emerging reasoning.
At this point, after a 0.9 s pause (line 16), Iman attempts a continuation of the narrative by adding that Kim bought something (line 19). As she says buy, Iman takes one of the receipts from Sam's hand (by mistake, the paper bag received by group A contained two receipts). During the ensuing 1.5 s silence (line 20), both Iman and Sam look at the two receipts. Sam further displays acceptance of Iman's proposal that Kim bought something with and buyed (line 21). Her linguistic choice realizes an embedded correction (Jefferson 1987) of Iman's turn (and buy:, line 19) through the use of the past tense form; though buyed is not grammatically accurate, the replacement of buy with buyed in a turn that is otherwise identical to the one delivered by Iman in line 19 displays Sam's orientation to a narrative set in the past (see Iman's use of got in line 2 and Sam's went in line 12). Finally, it is Walt who completes the turn-so-far by producing apples in line 25. After a further look at the receipts (line 26), Sam confirms by repeating apples and adding yeah (line 27).
Overall, Excerpt 1a has shown how the participants closely monitor each other's talk in terms of propositional content, praxeological import, and linguistic formulation. The three students are able to elaborate on each other's ideas in a topically and syntactically fitting way (e.g., see the increments in lines 8, 15, and 25) while also displaying alignment and agreement. It is also apparent that the artifacts are instrumental as mediating tools that structure the students' reasoning in the interaction while also making it more tangible and available for the coparticipants' scrutiny.
In Excerpt 1b, which occurs soon after Excerpt 1a, we see both Iman and Sam (lines 33-34) observably wondering about the role of the apple purchase for the emergent narrative.
Languages 2021, 6, x FOR PEER REVIEW 11 of 23 that structure the students' reasoning in the interaction while also making it more tangible and available for the coparticipants' scrutiny. In Excerpt 1b, which occurs soon after Excerpt 1a, we see both Iman and Sam (lines 33-34) observably wondering about the role of the apple purchase for the emergent narrative.
After a 1.2 second pause (line 35), during which Sam further inspects the receipt, Walt formulates his own hypothesis regarding the role of apples: at the Luna park (line 38), Kim bought apples with the coins in the bag (line 40) in order to feed the bear (line 36). Note how his narrative is developed increment by increment, as Walt touches the relevant artifacts, one by one, in line with what is being brought into relevance in the talk. Specifically, Walt touches the bear and the ticket as he mentions these items (lines 36 and 38) and, as he refers to the money (line 40), he puts the coins on the desk, next to the bear and the ticket (Frame Grab 1_2), so that all relevant artifacts in Walt's narrative are grouped and displayed in the same area on the desk.
Walt's embodied action of putting the coins on a specific area of the desk mobilizes Iman's attention to that area; in fact, she gazes down in that direction (line 40) and starts frowning (line 41). Soon afterward, Iman explicitly rejects Walt's narrative with a turninitial no., followed by an open-class repair initiator (what?, line 42) that challenges the narrative proposed by Walt. In the following line, Sam too displays her rejection of Walt's proposal with a straightforward no:. (line 43). Sam's turn, however, is more articulate in that she questions Walt's proposal with why would he feed the be(h)ar(h)¿ (0.5) with apples, After a 1.2 s pause (line 35), during which Sam further inspects the receipt, Walt formulates his own hypothesis regarding the role of apples: at the Luna park (line 38), Kim bought apples with the coins in the bag (line 40) in order to feed the bear (line 36). Note how his narrative is developed increment by increment, as Walt touches the relevant artifacts, one by one, in line with what is being brought into relevance in the talk. Specifically, Walt touches the bear and the ticket as he mentions these items (lines 36 and 38) and, as he refers to the money (line 40), he puts the coins on the desk, next to the bear and the ticket (Frame Grab 1_2), so that all relevant artifacts in Walt's narrative are grouped and displayed in the same area on the desk.
Walt's embodied action of putting the coins on a specific area of the desk mobilizes Iman's attention to that area; in fact, she gazes down in that direction (line 40) and starts frowning (line 41). Soon afterward, Iman explicitly rejects Walt's narrative with a turn-initial no., followed by an open-class repair initiator (what?, line 42) that challenges the narrative proposed by Walt. In the following line, Sam too displays her rejection of Walt's proposal with a straightforward no:. (line 43). Sam's turn, however, is more articulate in that she questions Walt's proposal with why would he feed the be(h)ar(h)¿ (0.5) with apples, lines 43 and 46). Note that Sam is here reproducing Walt's narrative in detail. That is, this is not a simple (and generic) rejection of what had been said before (as is the case for Iman's turn in line 42); instead, Sam is explicit about what exactly she is rejecting by reproducing it almost verbatim. Furthermore, the interspersed laughter in the delivery of her turn shows that she is orienting to Walt's idea as a laughable. Walt responds by joining Sam's laughter (line 44) but without any further (or revised) elaboration of his own proposal. In the meantime, Iman backs Sam's rejection by shaking her head (line 45) and producing multiple no's (line 47; on multiple sayings as an interactional resource, see Stivers 2004). With her actions, Iman explicitly and strongly rejects Walt's course of action and its implications for task accomplishment.
Excerpt 1b could therefore be taken as an example of collaborative disagreement, in which we see two participants (Iman and Sam) siding with each other against a narrative proposed by the third coparticipant (Walt). At the same time, Sam's rejection-in its explicitness-shows that she has carefully attended to Walt's proposal.
In subsequent lines (not reported here), Iman and Sam discuss the role of the phone for the emerging narrative. Since they had received a broken phone, they develop a narrative according to which Kim must have fallen, broken his phone, and lost his memory. Excerpt 1c picks up the talk as Sam shows Iman the note (line 1). At this point, in seeming competition with each other, both Walt (line 4) and Iman (line 5) start formulating a summary of the artifacts that have been talked into relevance so far, and that might play a role in determining what happened to Kim. Specifically, Walt grabs the photo and points at Iman's hand holding the note, as he says I ↑think that ↑there, (line 4), thereby establishing a connection between the photo and the note. He then gives the photo to Iman (line 6), who has just grabbed the note from Sam's hands (line 2) and touched the roller coaster ticket and the bear (line 5). With these embodied actions of touching, grabbing, passing the artifacts, the participants seem to accomplish a material summary of the narrative so far.
Iman further brings these artifacts into relevance with minimal talk accompanied by observable embodied actions. That is, she says because THAT as she visibly puts the note and the photo in her right hand (line 8), and she says and then, as she grabs the bear (line 10). Iman's talk in these lines is produced with listing intonation, projecting continuation; her coparticipants do not take the floor but follow Iman's hand movements with their gaze (lines 10 and 12), thereby displaying that they are attending to Iman's summary in progress. Then, as Iman says and the (line 12), both Iman and Sam orient to the ticket as the next item in the summary, as indicated by Iman's action of touching it (and eventually lifting it in her left hand) and by Sam's action of pointing at it (see Goodwin 2007 on environmentally coupled gestures and Mondada 2014 on pointing). The summary so far is visible in Frame Grab 1_3, showing Iman holding the bear and the note in her right hand and touching the ticket with her left hand while Sam points to it. After a slight hesitation (see the cutoff and the pause in line 12), Iman uses the Swedish formulation bio (literally, "cinema", line 14) to refer to the roller coaster ticket. At this point, during a 0.5 s pause (line 13), Sam moves her hand toward the receipt and suggests it as a potentially relevant item for the summary (and maybe the::, line 14; see Lilja and Piirainen-Marsh 2019 on the role of gestures for action ascription). With these actions, Sam directs the coparticipants' attention to another artifact that they have discussed so far. Note that the students were instructed to select five artifacts that could help explain what happened to Kim; the receipt would therefore constitute the fifth artifact. However, this suggestion is discarded by Iman, who announces that she does not get the apple thing (line 16). Her turn is formulated with an initial but that projects disagreement. Sam carefully parses Iman's turn and actually produces • apples • in collaborative overlap with Iman (line 17) before explicitly aligning with her by saying no. (line 19). This final exchange then excludes the receipt as a relevant artifact for task completion.
Languages 2021, 6, x FOR PEER REVIEW 13 of 23 Overall, in this short stretch of talk that concludes the first cycle of imaginative narration of what might have happened to Kim, Iman is co-constructed as the student in charge of summarizing the discussion so far by bringing further into relevance the selected artifacts. While Walt and Sam initially seem to compete for such a role (lines 4-5), Walt eventually gives in by handing the photo to Iman (line 6). We have then seen how Sam collaborates in completing the list of five artifacts (lines 12-14) and how she accepts Iman's disagreement (line 19): the receipt is related to the purchase of apples and the role of apples in the narrative is not quite clear (lines 16-17).
In sum, Excerpt 1 shows the high level of co-construction and cooperation achieved by the students in group A. Specifically, the analysis of Excerpt 1a has illustrated how the students build on each other's turns-at-talk as they collaboratively shape the emerging narrative, with displays of agreement and further additions that are syntactically and thematically consistent. On the other hand, Excerpt 1b has provided an example of coconstructed rejection, with an initial rejection formulated in generic terms (line 42) being upgraded with a more detailed rejection (lines 43 and 46) that displays attentive listening of the original proposal (lines 36-40). Finally, Excerpt 1c has shown the collaborative coconstruction of a list of potentially relevant artifacts (see task instructions), which also has the function of summarizing the discussion so far and of providing a first, tentative task outcome. Crucial in all these excerpts is the presence of artifacts that are oriented to and manipulated by the coparticipants as they present, discuss, and summarize their ideas for the emerging narrative. Such use of the artifacts makes visible and tangible the developing narrative.
Overall, in this short stretch of talk that concludes the first cycle of imaginative narration of what might have happened to Kim, Iman is co-constructed as the student in charge of summarizing the discussion so far by bringing further into relevance the selected artifacts. While Walt and Sam initially seem to compete for such a role (lines 4-5), Walt eventually gives in by handing the photo to Iman (line 6). We have then seen how Sam collaborates in completing the list of five artifacts (lines 12-14) and how she accepts Iman's disagreement (line 19): the receipt is related to the purchase of apples and the role of apples in the narrative is not quite clear (lines 16-17).
In sum, Excerpt 1 shows the high level of co-construction and cooperation achieved by the students in group A. Specifically, the analysis of Excerpt 1a has illustrated how the students build on each other's turns-at-talk as they collaboratively shape the emerging narrative, with displays of agreement and further additions that are syntactically and thematically consistent. On the other hand, Excerpt 1b has provided an example of coconstructed rejection, with an initial rejection formulated in generic terms (line 42) being upgraded with a more detailed rejection (lines 43 and 46) that displays attentive listening of the original proposal (lines 36-40). Finally, Excerpt 1c has shown the collaborative co-construction of a list of potentially relevant artifacts (see task instructions), which also has the function of summarizing the discussion so far and of providing a first, tentative task outcome. Crucial in all these excerpts is the presence of artifacts that are oriented to and manipulated by the coparticipants as they present, discuss, and summarize their ideas for the emerging narrative. Such use of the artifacts makes visible and tangible the developing narrative.
Excerpt 2 (divided in Excerpts 2a, 2b, and 2c) illustrates the emergence of the first multiartifact narrative in Group B. The group members are Calle, Lina, and Marie. Lina maintains a bystander role in that she is mostly silent, but through her embodied actions (such as eye-gaze behaviors and body posture), she displays at least an orientation to the task-based interaction conducted by her two classmates.
Excerpt 2a picks up the talk as Marie connects two artifacts, that is, the receipt and the coins. With a hedged epistemic stance (maybe:, line 1), she suggests that Kim bought something and got some coins in return (lines 1-2). Marie's turn is formulated with deictic pronouns (this in lines 1 and 2), accompanied by embodied actions (i.e., grabbing the receipt and pointing at the coins). Calle agrees with a series of yeah (line 4) as he touches the coins and then provides the expression the return of the money (line 7) to verbally refer to the coins. With this action, Calle orients to the relevance of providing specific lexical items to name the referents that are brought into relevance in the interaction.
Subsequently, while Calle's turn in line 10 (↑YEAH¿ that-that seems) projects an assessment of the emerging narrative involving the receipt and the coins, Marie orients to the roller coaster ticket (line 12). Once again, Marie uses a deictic pronoun (this, line 12) as she touches the relevant artifact (see also Frame Grab 2_1, where Marie is seen holding the ticket). At this point, as shown in line 13, Calle is attending to two aspects of the narrative: on the one hand, he is making tangible the connection between the receipt and the coins by putting the coins on the receipt (Frame Grab 2_1), thereby seemingly materializing and finalizing the narrative-so-far; on the other hand, he is also following Marie's emerging idea, as indicated by his provision of the lexical item (the coupon, line 13) referring to the ticket.
Calle then further aligns with Marie in shifting the focus of attention to the roller coaster ticket, the role of which he problematizes with why why's:: (0.4) • this thing • . (lines 14 and 16). He reaches for the ticket (line 16) that Marie had been holding in front of herself (lines 13-14) and slightly moves it in his direction. At this point, all participants (including Lina) are looking at the ticket (line 17) and are quietly reading from it (lines 18-21). Marie then formulates the first hypothesis concerning the ticket's role in Kim's story by suggesting that Kim was possibly going to that place (line 23) where the ticket could be used (line 25). After competing with Marie for the floor (line 24), as he attempts to come up with his own hypothesis concerning the ticket, Calle produces a stretched yea::::::h. (line 26) with which he accepts Marie's proposal so-far (maybe he was <go::ing to that place>, line 23). This indicates that, while initially attempting to produce his own hypothesis regarding the role of the ticket for Kim's story (line 24), Calle quickly drops his attempt as he parses Marie's turn-at-talk (line 23) and agrees with the narrative she is proposing.
Then, Calle verbalizes what the ticket (previously referred to as the coupon, line 13) is for (this:: is like a ↑roller coaster or something¿, line 28). At this point, Marie reformulates her previous turn (lines 23 and 25) in more specific terms: maybe he was going to like an amusement park or something (lines 30-31). With this reformulation, she orients to the relevance of replacing an ostensibly vague location reference (that place, line 23) with a lexical item that identifies a specific location (amusement park, line 31; note that this was referred to as Tivoli by the participants in Excerpt 1). Once her turn is projectably complete, Calle takes the floor by mentioning what is written on the ticket: good for one ride (line 32).
Overall, Excerpt 2a shows various ways in which the coparticipants collaborate in task accomplishment through co-constructed interaction. At first, we have seen Calle offering unsolicited vocabulary assistance (lines 7 and 13), an action which might show his orientation to the accomplishment of a task targeting language practice (note that a similar orientation is displayed later by Marie when she replaces a vague reference with a more specific one; see line 31). Calle's action also displays his alignment with Marie: he is attending to her turns-at-talk, agreeing with what she proposes (lines 4, 7, 10), shifting focus when she mobilizes their attention to a new artifact (lines 11-14), while also providing specific lexical items that verbally label the focal artifacts. After all the participants engage in reading what is written on the ticket, both Marie and Calle start formulating hypotheses about the role of the ticket, with Calle quickly abandoning his turn (line 24) and aligning with Marie's emerging narrative (line 26).
Excerpt 2b takes place shortly after Excerpt 2a. Here, Calle further reformulates Marie's turn in lines 30-31 by suggesting: <↑maybe (0.3) he went to the (1.1) roller coaster park.> (lines 36-37; Marie's amusement park in Excerpt 2a, line 31, here becomes roller coaster park), as he holds up the fish (line 36) and points at the ticket (line 37). Note that Calle prefaces and concludes his turn with hedges (↑maybe in line 36 and the claim of insufficient knowledge i don't know in line 37; for the latter, see Sert and Walsh 2013) that weaken his epistemic stance.
After Marie displays alignment with mh mh, (line 38), Calle elaborates on the ongoing, emergent narrative by suggesting that, at the roller coaster park, Kim got the roller coaster ticket (line 39) and won the fish (line 40). At this point, Marie starts smiling (line 42) and produces a laughter token (line 43); the import of these actions is not clear. In subsequent lines, however, Marie talks into relevance the receipt (lines 45 and 47) and seems to suggest that the role that the receipt might have in the narrative contrasts with Calle's emerging narrative about Kim going on a roller coaster and winning a fish toy at a roller coaster park. Indeed, even though the formulation of her turn initially projects continuation (a:::::nd, line 45), the subsequent unfolding of her turn with but ( Note that, at this moment, Calle is further developing the narrative about the roller coaster park, as he adds that Kim won the fish like a:: hunt. (lines 44 and 46). As he says so, Calle enacts a throwing gesture (line 44, Frame Grab 2_2), which seems to suggest that Kim played some sort of game with which he won the toy fish. At this point, Marie is already orienting to the receipt, as indicated by her index finger touching the receipt (line 14, Frame Grab 2_2). That is, the divergence in the participants' orientation is also visibly manifest in their embodied actions, with Marie touching the receipt (lines 44-45) and Calle holding up the fish (line 44). After receiving no response to his subsequent confirmation request (yes¿, line 46), Calle shifts his attention to the receipt (line 47), in response to Marie's inquiry about its role, an inquiry that is specifically addressing Calle (see the receipt held in Calle's direction in line 47). Calle replies with °>I think.<° this was: (.) before.
(line 49), as he lifts the receipt to his eye level in order to inspect it closely.
Overall, Excerpt 2b shows that the participants attend to each other's talk, whether they align with it (line 38) or signal potential inconsistencies in the evolving narrative (lines 45 and 47). Note also that the inconsistency pointed out by Marie is recipient-Note that, at this moment, Calle is further developing the narrative about the roller coaster park, as he adds that Kim won the fish like a:: hunt. (lines 44 and 46). As he says so, Calle enacts a throwing gesture (line 44, Frame Grab 2_2), which seems to suggest that Kim played some sort of game with which he won the toy fish. At this point, Marie is already orienting to the receipt, as indicated by her index finger touching the receipt (line 44, Frame Grab 2_2). That is, the divergence in the participants' orientation is also visibly manifest in their embodied actions, with Marie touching the receipt (lines 44-45) and Calle holding up the fish (line 44). After receiving no response to his subsequent confirmation request (yes¿, line 46), Calle shifts his attention to the receipt (line 47), in response to Marie's inquiry about its role, an inquiry that is specifically addressing Calle (see the receipt held in Calle's direction in line 47). Calle replies with • >I think.< • this was: (.) before. (line 49), as he lifts the receipt to his eye level in order to inspect it closely.
Overall, Excerpt 2b shows that the participants attend to each other's talk, whether they align with it (line 38) or signal potential inconsistencies in the evolving narrative (lines 45 and 47). Note also that the inconsistency pointed out by Marie is recipient-designed in a contrastive way with respect to prior talk, which observably indicates that the role of the receipt in the story might not fit with the narrative developed by Calle.
In lines not reported here, the participants try to figure out the timeline of the story by checking the instructions (which say that Kim left home at 7:30 a.m. and was found at 11 p.m.) and the time specified on the receipt. After determining that at 1 p.m. Kim was at the store where he got change for one dollar, thereby reinstating the connection between the receipt and the coins (see Excerpt 2a), the participants start orienting to different artifacts. We join the discussion in Excerpt 2c with Calle mentioning the change (lines 1-2) as he touches the coins (line 1) and looks at the receipt (lines 2-3), while Marie grabs the note (line 2).
Languages 2021, 6, x FOR PEER REVIEW 17 of 23 designed in a contrastive way with respect to prior talk, which observably indicates that the role of the receipt in the story might not fit with the narrative developed by Calle.
In lines not reported here, the participants try to figure out the timeline of the story by checking the instructions (which say that Kim left home at 7:30 a.m. and was found at 11 p.m.) and the time specified on the receipt. After determining that at 1 p.m. Kim was at the store where he got change for one dollar, thereby reinstating the connection between the receipt and the coins (see Excerpt 2a), the participants start orienting to different artifacts. We join the discussion in Excerpt 2c with Calle mentioning the change (lines 1-2) as he touches the coins (line 1) and looks at the receipt (lines 2-3), while Marie grabs the note (line 2). Languages 2021, 6, x FOR PEER REVIEW 18 of 23 Calle then reaches for the phone (line 10) and assesses its potential role in the evolving narrative as making z:ero sense (line 11) since it's just a pho:ne (line 12), not charged (line 16). In these lines, Calle is engaging with the camera by turning the phone towards it (lines 12 and 15); the camera then seems to be invoked as a fourth participant to which evidence needs to be provided as a way to back one's assessments. Note that, in this stretch of talk, Marie smiles as Calle says that the mobile is just a phone (line 12) and then takes the floor in line 14, possibly in the attempt to develop a hypothesis concerning this artifact. However, she rejects her own idea (no:h., line 14) before verbalizing it.
At this point, after Calle's negative assessment of the role of the phone and after Marie's abandoned attempt at producing a potential narrative that includes the phone, Calle takes the floor again with an extended hesitation token (u:::hm::, line 18). After a 1.6 second pause (line 18), Calle starts listing the artifacts that might be relevant to develop the narrative about Kim's day. As he lists the artifacts, Calle moves them next to each other, thereby materializing the list: the phone (line 19), the coins (line 21), the ticket (line 23), and the fish (line 25) are brought together to occupy a specific area on the desk (see Frame Grab 2_3, where the phone, the coins, and the fish are visible). Calle then projects the continuation of the list with and uh as he grabs the keys and puts them aside (line 27). That is, Calle seems to be discarding the keys as a relevant fifth artifact.
However, Marie looks at the keys (line 28), points at them (line 29), and then, after raising a potential objection (but the key::s, line 29), grabs them, lifts them at eye level, and carefully inspects them (line 31). In the meantime, Calle seems to be considering other artifacts, as he touches the photo (line 30) and the note (line 32), while rejecting the potential relevance of the keys with no:::. (line 32). Marie finally produces the assessment that those keys are not the keys to a house (line 34). Calle confirms (no::., line 36) and further reinstates his rejection of the keys as the relevant fifth artifact (it doesn't help us. °like° at all., line 36). Marie, however, in partial overlap with Calle, starts developing a narrative concerning the role of the keys in Kim's day (maybe he was like, lines 37-38). At this point, Marie and Calle collaboratively complete the turn (lines 38-39). While Marie's talk here is unfortunately not audible, Calle distinctly completes Marie's turn-so-far with going to the school (line 39), with slightly rising intonation projecting continuation. Indeed, after Marie's acceptance of his emerging hypothesis (line 41), Calle completes the narrative with and something happened (line 42), with which Marie quickly aligns (line 43). We can then say that this is a case of joint turn construction, which is defined as "a practice whereby a participant in conversation completes a grammatical unit-in-progress initiated by another participant" (Hayashi 2014, p. 225).
Calle then reaches for the phone (line 10) and assesses its potential role in the evolving narrative as making z:ero sense (line 11) since it's just a pho:ne (line 12), not charged (line 16). In these lines, Calle is engaging with the camera by turning the phone towards it (lines 12 and 15); the camera then seems to be invoked as a fourth participant to which evidence needs to be provided as a way to back one's assessments. Note that, in this stretch of talk, Marie smiles as Calle says that the mobile is just a phone (line 12) and then takes the floor in line 14, possibly in the attempt to develop a hypothesis concerning this artifact. However, she rejects her own idea (↓no:h., line 14) before verbalizing it.
At this point, after Calle's negative assessment of the role of the phone and after Marie's abandoned attempt at producing a potential narrative that includes the phone, Calle takes the floor again with an extended hesitation token (u:::hm::, line 18). After a 1.6 s pause (line 18), Calle starts listing the artifacts that might be relevant to develop the narrative about Kim's day. As he lists the artifacts, Calle moves them next to each other, thereby materializing the list: the phone (line 19), the coins (line 21), the ticket (line 23), and the fish (line 25) are brought together to occupy a specific area on the desk (see Frame Grab 2_3, where the phone, the coins, and the fish are visible). Calle then projects the continuation of the list with and uh as he grabs the keys and puts them aside (line 27). That is, Calle seems to be discarding the keys as a relevant fifth artifact.
However, Marie looks at the keys (line 28), points at them (line 29), and then, after raising a potential objection (but the key::s, line 29), grabs them, lifts them at eye level, and carefully inspects them (line 31). In the meantime, Calle seems to be considering other artifacts, as he touches the photo (line 30) and the note (line 32), while rejecting the potential relevance of the keys with no:::. (line 32). Marie finally produces the assessment that those keys are not the keys to a house (line 34). Calle confirms (no::., line 36) and further reinstates his rejection of the keys as the relevant fifth artifact (it doesn't help us. • like • at all., line 36). Marie, however, in partial overlap with Calle, starts developing a narrative concerning the role of the keys in Kim's day (maybe he was like, lines 37-38). At this point, Marie and Calle collaboratively complete the turn (lines 38-39). While Marie's talk here is unfortunately not audible, Calle distinctly completes Marie's turn-so-far with going to the school (line 39), with slightly rising intonation projecting continuation. Indeed, after Marie's acceptance of his emerging hypothesis (line 41), Calle completes the narrative with and something happened (line 42), with which Marie quickly aligns (line 43). We can then say that this is a case of joint turn construction, which is defined as "a practice whereby a participant in conversation completes a grammatical unit-in-progress initiated by another participant" (Hayashi 2014, p. 225).
Overall, Excerpt 2c illustrates the first ad interim summary co-constructed by the students in Group B. In this excerpt, it is Calle who self-selects (line 18) and starts mention-ing the items that might go together while also physically moving them one next to the other. The summary emerges after the participants initially orient to different artifacts (lines 1-10) and once the mention of the phone does not lead to any hypothesis concerning that artifact (lines 11-16). Put another way, in a moment when the participants seem to have reached an impasse, Calle's action of listing potentially relevant artifacts while grouping them together on the desk summarizes the points of their discussion so far. As soon as he orients to an artifact not previously mentioned, the keys (line 23), Marie reopens the discussion as she formulates a narrative around the role of the keys, supported by a close inspection of the artifact. While initially discarding the keys as an item that does not help with task accomplishment (line 36), Calle eventually collaborates in the completion of Marie's emerging hypothesis.
In sum, the analysis of Excerpt 2 has shown the co-constructed, task-oriented interaction achieved by the students in Group B. In the three excerpts examined here, we have seen how the participants attend to each other's talk, whether they agree (Excerpt 2a) or disagree (Excerpts 2b and 2c) with the emerging narratives and the artifacts they orient to. Their turns-at-talk are recipient-designed to target prior talk as they align with it, indicate potential inconsistencies, mobilize their attention towards a shared focus, and, in some cases, offer unsolicited vocabulary assistance. As they accomplish all these actions, the two focal participants, Marie and Calle, collaboratively orient to the progressivity of the task.

Discussion
This practice-based study is the result of a collaborative research effort conducted by schoolteachers and researchers. The novelty of our work lies in its rootedness in practice and the involvement of teachers throughout the entire research process: from the identification of a practical problem observed in EFL classrooms in Sweden to the design of tasks that might help to solve the problem and, finally, to the analysis of the collected data. This kind of collaborative research produces knowledge that is directly relevant and actionable for the teaching profession (Eriksson 2018; see also Carlgren 2012) and does not need to be "translated" to be applicable in the classroom (Hultman 2015).
The point of departure for the larger project in which this study is framed stems from the participating teachers' observation that student-student interaction in the EFL classroom in Sweden does not seem to be characterized by students' collaborative engagement with each other's talk. The issue was attributed to the kinds of activities for oral practice that are typically used in the classroom and that are usually inspired by the task format adopted in the Swedish national exams. Indeed, as research has increasingly shown (e.g., Sandlund and Sundqvist 2011, 2013Sundqvist et al. 2013), the opinion-based tasks used in the national exams and for the purpose of classroom practice tend to elicit monologues rather than dialogues. As pointed out by Sandlund and Sundqvist (2016, p. 128), the speech exchange system enacted in these tasks is rather unnatural, as it is characterized by "extended, less co-participant-oriented turns with few overlaps and interjections". Arguably, this kind of talk is not illustrative of the students' communicative skills, which nevertheless are the main target of L2 education as envisioned by the Swedish syllabus for English (Skolverket 2018) and by the CEFR (Council of Europe 2020). The research team thus designed a problem-based task in order to verify whether this kind of task might be more conducive to the co-constructed interaction that the teachers saw as an indication of their students' communicative abilities. The analysis of the students' interaction was conducted with a CA lens (Sidnell and Stivers 2013), which we deemed particularly suited to describe in detail how students participate in and contribute to task-oriented interaction. In addition, we paid particular attention to the students' orientation to and manipulation of artifacts in the interaction, in line with recent calls (Guerrettaz et al. 2021;Mathieu et al. 2021) for an increased focus on materiality as central for learning and teaching research.
Our analysis, focusing on two groups of students in grade 7, has shown that the problem-based task designed by the research team did elicit co-constructed, collaborative talk that is very different from the parallel monologues seen in the national exam (Sandlund and Sundqvist 2016). More specifically, the students observably engaged with each other's turns-at-talk, not only by aligning with and elaborating on emerging narratives but also by challenging and disagreeing with proposals issued in prior talk. As the students collaboratively accomplished the task in and through the unfolding co-constructed interaction, they manifested their sophisticated interactional skills, such as the ability of parsing previous turns, of understanding their praxeological import, and of building on them in timely and fitting ways. In the interactions analyzed here, the turns are relatively short, and overlaps are common; there are also instances of incremental turn recompletion (Couper-Kuhlen and Ono 2007;Lerner 1991;Schegloff 2016) and of joint turn construction (Hayashi 2014), which show the students' ability to closely monitor the linguistic formulation of emerging turns. In other words, in finding out what happened to Kim, the students were speaking to each other (rather than at each other; cf. May 2011) and, thus, engaged in meaningful and functional language use (Ellis 2003(Ellis , 2009). In doing so, they visibly displayed their interactional competence (see : Eskildsen 2018;Hall 2018;. At the same time, it should be noted that, in their task-based interactions, the students relied heavily on the use of the available artifacts, the manipulation of which clarified the import of their turns and made tangible and visually accessible their reasoning as they engaged in the co-construction of imaginative narratives. In other words, one could say that the use of artifacts was part and parcel of the meaning-making process accomplished by the students as they were engaging with each other's ideas while maintaining and establishing intersubjectivity. Clearly, the task What happened to Kim? is different from the oral activities usually implemented in the EFL classroom in Sweden (not least for its inclusion of artifacts) and, therefore, from the opinion-based tasks used in Swedish national exams. On the basis of the detailed analysis of the students' interaction presented here and of our observations on the implementation of this specific task and its blueprint by various student groups at different levels (from grade 6 to upper secondary school), we maintain that this kind of problem-based task has proved to be a valuable opportunity for students to practice the "all-round communicative skills" that the Swedish syllabus for English aims for (Skolverket 2018, p. 34), while also giving students the chance to show their interactional skills in school (cf. Myndigheten för Skolutveckling 2008). Therefore, we believe that, in order to follow Skolverket's recommendations for the development of communicative skills, it would be relevant to include problem-based, open-ended tasks in students' regular class instruction. At the same time, at the theoretical level, we suggest that the CA-based, action-oriented notion of IC can meaningfully contribute to the conceptualization of oral interaction that is currently held in Swedish schools. In sum, our findings have implications for the (potential) diffusion of curricular innovations (Markee 1997) and testify to the relevance of bottom-up collaborative approaches to research where teachers and researchers are collaboratively involved in a joint investigation of a pedagogical problem that is explored and addressed through the combination of theory and practice.  Institutional Review Board Statement: Ethical review and approval were waived for this study due to the fact that sensitive information about the informants was neither collected nor used.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: Data sharing not applicable.