1 Introduction

Telling stories often requires holding the floor for consecutive turns (Sacks, 1974; Mandelbaum, 2013). To tell others something—whether it be a funny story, the description of a past event, or a trouble—also requires that what is being said is recognizable as such to recipients, and that recipients collaborate in appropriate ways. As such, the act of telling something to another, and for co-participants to allow for the telling to be told, requires a specific kind of interactional competence (IC). In most second language (L2) oral proficiency tests, interactional skills are assessed (cf. Sandlund et al., 2016), but the role of tellings and stories in the planning and interactional accomplishment of L2 test tasks is relatively underexplored.

Storytelling and its role in literacy development, socialization, and as a pedagogical tool has been studied extensively in educational research (see, e.g., Donald, 2001; Lucarevschi, 2016; Riley & Burrell, 2007; Vardi-Rath et al., 2014). While educational and developmental research generally view stories as situated in a social context, there is often a broader interest in cognitive and affective aspects, and in the story itself. Conversely, conversation analytic (CA) work distinguishes between the story itself (“the story”) and the activity of telling the story (the “telling”). The focus is on interactional practices deployed “to produce storytelling as a recognizable activity and through which they implement a variety of social actions” (Mandelbaum, 2013, p. 492). CA research has demonstrated how children gradually acquire the ability to participate in storytelling activities, and how co-participants collaborate in storytelling events by jointly accomplishing the narrative using linguistic and embodied resources (e.g. Burdelski, 2019; Filipi, 2017; Theobald, 2015). As such, it can be argued that participating in tellings in social interaction requires and displays interactional competence (IC) (Kasper & Ross, 2013; Kim, 2016; Salaberry & Kunitz, 2019; Sandlund & Greer, 2020; Sandlund & Sundqvist, 2019; Young & He, 1998), which develops through recurrent participation in social activities.

While conversation analytic work has demonstrated that IC is the “the ability to use the context-free interactional organizations (…) in a context-sensitive manner to participate in social activities” (Kasper & Ross, 2013, p. 24), the specific linguistic and embodied resources drawn upon to accomplish a particular social action may be language- and context-specific (cf. Doehler, 2018). Thus, accomplishing a telling in a second language relies not only on first language interactional skills, but also on linguistic and interactional competencies specific to the L2. For example, Huth (2014) demonstrates how telephone call openings are systematically different in German and American English. Understanding of such differences in particular activities across languages can be acquired over time, with repeated participation (and teaching, see Huth, 2014; Waring, 2018).

Research on the interactional accomplishment of oral tasks in L2 tests has shown that both test format (proficiency interviews, paired or small group tests) and sub-tasks (giving directions, describing a picture, or expressing opinions on given topics) provide the affordances for the types of contributions that a test-taker can appropriately make (Kasper, 2013; Sandlund & Sundqvist, 2011; van Compernolle, 2011). As a consequence, studies of how participants orient to and accomplish test tasks offer important information for the design of tasks, and for the teaching of IC (cf. Salaberry & Kunitz, 2019; Huth, 2020). Participants’ orientation to occasioned or invited tellings offers us insights into one aspect of a learner’s L2 IC. In addition, the role of tellings in the specific institutional interactional type of language testing, where multiple constraints operate for learners as well as for examiners, is relatively unexplored.

Even in assessment contexts where narrative skills are not specifically in focus, interaction around pre-set discussion topics frequently prompt test-takers to draw upon their own experiences (cf. Brown & Abeywickyama, 2010). The present work centres on a phenomenon which, from other perspectives, has been referred to as small stories (Bamberg & Georgakopoulou, 2008), which signals a distinction between extended narratives and more brief moments in everyday interaction. However, as Kasper and Prior (2015) rightly point out, such a distinction is researcher-based, and its unclear whether such a “researcher-stipulated distinction between big stories and small stories is relevant for storytellers and story recipients” (p. 228). Also, not all tellings lead to a multi-turn build-up of a “story”. This is perhaps especially true in institutional contexts where the overarching activity may be centred on completely different interactional goals. The present chapter, therefore, primarily focuses on brief narrative tellings that become part of the institutional activity of L2 oral proficiency testing. As such, the contribution of the chapter lies in the understanding of how such tellings may be invited, locally occasioned, accepted, or resisted in local interactional contexts where participants must balance multiple constraints on talk.

2 Storytelling and Educational Interaction

In folk linguistic terms, storytelling may bring to mind a long stretches of talk with several story components produced by one teller (e.g. Holt, 2017), but tellings can also refer to actions produced in relatively short sequences in interaction where they are deployed as part of activities such as complaining (Mandelbaum, 1991/1992; Holt, 2000), talk about troubles (Dooly & Tudini, this volume; Jefferson, 1988; Ta & Filipi, 2020), and joking (Holt, 2007; Sacks, 1974). In particular contexts, telling an institutional representative about some trouble is the core activity, such as in therapy sessions or calls to helplines (Peräkylä et al., 2008), but tellings are also a pervasive aspect of non-institutional talk (Jefferson, 2015).

A growing body of CA studies have targeted storytelling practices in interactions with children, albeit often with a focus on first language (L1) storytelling among young children (e.g. Filipi, 2017; Theobald, 2015) rather than on adolescents and second language learners. Furthermore, as Mandelbaum (2013, p. 506) notes, more work on storytelling practices in institutional settings is needed, as the uniqueness of practices in such contexts likely contributes to the makeup of that institution. The present study responds to this call through its focus on a particular type of institutional interaction: L2 oral proficiency assessment. As few studies have specifically examined tellings in L2 testing contexts—perhaps because “authentic” troubles tellings are relatively rare in high-stakes language assessment—a brief examination of tellings and narrative actions in language learning contexts is warranted. These are the contexts in which language learners are taught and where they practise what is later assessed in tests.

2.1 Tellings and L2 Interactional Competence

Storytelling activities are relatively common in L2 classrooms. In a study of adult L2 learners of English, Waring (2013) examined how learners manage routine inquiries such as How are you, What’s up, and How was your weekend, which were not specific teaching objects, but still emerged as learning opportunities for L2 speakers. Through repeated participation in such routine sequences, learners have opportunities to develop their IC specifically in terms of understanding the sociopragmatics and pragmalinguistics of responding to such inquiries. As such, exchanges that are not part of the actual language teaching in the classroom may offer opportunities for the development of L2.

In foreign language classrooms, storytelling frequently serves to practise grammatical constructions, such as past tense forms required for narrating past events. Frantz (2021) examined a classroom activity which included repetition of a story with different recipients. The focal student examined was shown to develop over each telling in terms pronunciation, fluency, and autonomy and finally managed the telling without soliciting teacher support. The author argues for the usefulness of examining repetitious activities, in this case, storytelling, in describing the emergence of L2 learning.

As for storytelling and pedagogical objectives, Kasper and Monfaredi (2021) examine storytelling as an instructional practice in intermediate and advanced Persian language classrooms at two universities in North America. The analysis focuses on teachers’ impromptu, pedagogically purposed storytellings. The analysts show how teachers work to establish themselves as tellers, and how the teacher’s orientation to resuming the instructional activity is monitored by the students to jointly accomplish transitions between activities. The findings suggest that storytelling is used to make particular teaching objects comprehensible and to encourage student involvement.

Longitudinal studies of L2 learners’ changing participation in target language interactions are particularly important for knowledge on the developmental trajectories of L2 learners’ interactional competence. Berger and Doehler (2018) examine changes in storytelling practices of a second language speaker, Julie, an au pair, over the course of nine months. The findings indicate a substantial change in the position of the stories told. Initially, Julie’s stories were positioned as a response to other participants (i.e. in second position). However, gradually, stories told in first position increased. Other changes include the stories and their length and complexity, the interactional purposes that the stories accomplish, or the resources used to anticipate the climax of the story. Similarly, Greer (2019a) examined mealtime conversations in a homestay context in Brisbane, Australia, recorded over three weeks. Participants are Ryo—a Japanese undergraduate student and his homestay family. The study focuses on news-of-the-day tellings addressing Ryo by family members. Greer notes that there tends to be a ritualistic element involved in the accomplishment of stories connected to delivering “news of the day” in particular families, and for a newcomer to a family, such practices may call for negotiation. The analysis shows how all participants adapt their practices over time as the student (i.e. the novice) becomes more familiar with the evening routines and begins anticipating news-of-the-day tellings. Changes in the host family’s practices were also observed (Greer, 2019a, p. 161). Greer concludes that IC is not solely the responsibility of one speaker, but co-constructed (see also Greer, this volume; Wantanabe, this volume).

In sum, in connecting L2 IC to narrative activities, doing tellings in a second language requires practice and participation. Many features of tellings can be transferred from a speaker’s L1, whereas others can be learned (and taught). For the purpose of the present study, of particular relevance is the observation that teachers and learners frequently engage in tellings in regular and routine classroom activities, just as they do in their lives outside of education. Consequently, tellings and stories are familiar activities to learners. However, in assessment contexts, tellings are variously restricted by the institutional frame of testing, as well as by the conversation tasks assigned to test-takers.

3 Speaking Tasks and Their Accomplishment in L2 Oral Proficiency Tests

In designing speaking tests, the traditional oral proficiency interview (OPI) format with a candidate and a native speaker examiner has increasingly been replaced with paired or small group tests, where test-takers interact with one or more peers (Nitta & Nakatsuhara, 2014). An argument for peer-driven tests is that if interactional competencies are to be assessed, paired or small group tests yield talk more similar to naturally occurring conversation (Ducasse & Brown, 2009) and to peer activities common in language classrooms (Kasper, 2013, p. 259). While the present study focuses on paired or small group tests, studies on OPI interaction has contributed greatly to our understanding of the affordances for demonstrating IC in a testing context (see Kasper & Ross, 2013; Young & He, 1998). One study of relevance to the present analysis deals with the role of topics for conversation and examiners’ treatment of candidates’ reluctance to produce elaborate tellings in OPIs. Ross (1998) shows how an examiner can misinterpret a candidate’s reluctance in responding to a question as proficiency-related, and suggests that they may actually be due to different cultural orientations to questions and topics. Ross’ study emphasises the importance of viewing pragmatic competence in language proficiency interviews to include the candidate’s ability to decode a question through metapragmatic knowledge. Ross’ study points to the need to consider that displays of IC in a test are also about striking a balance between producing topical talk in interactionally appropriate ways in an L2, and monitoring the institutional frame for what is expected in terms of assessment.

Speaking tasks in oral proficiency tests are designed with the purpose of obtaining language output which matches the test construct, which means that tasks for assessment vary across educational levels, proficiency levels, and national contexts. Studying how tasks are accomplished in situated interaction, then, can inform the design of tasks. For example, Greer (2019b) analyzed paired EFL oral proficiency tests in Japan, and demonstrates how test-takers manage multiple constraints in accomplishing the test tasks while keeping within the time limit as set by a timer. Students who showed an excessive orientation to the institutional and technical constraints of the test were not as successful in closing their test talk in a natural fashion (p. 186). Similarly, Sandlund and Sundqvist (2011) examined task management in EFL oral proficiency tests, and contrasted observed learner strategies with teachers’ assessment of interactional skills and overall performance. Findings showed that test-takers who operated relatively freely from the task instructions and displayed willingness to “play the test-taking game” and “feign interest in topics and interlocutor contributions” (p. 115) had higher scores, while learners who displayed an excessive orientation to accomplishing the discussion task verbatim to the instructions scored lower. In terms of assessment, then, both Greer’s (2019b) and Sandlund and Sundqvist’s (2011) study indicate that successful test-takers display a moderate adherence to task guidelines, thus displaying a more natural-sounding interaction.

While the present chapter cannot do justice to studies on tasks in L2 test interaction, it seems clear from the studies reviewed that test-takers’ understandings of the institutional frame and the activity in which they are participating impacts the ways in which contributions are treated as appropriate or inappropriate. The present chapter addresses the role of task type in test interaction with a particular focus on tellings and stories as they are made relevant by participants in their management of speaking tasks assigned for the interaction.

4 The Study

The present study draws on data collected for a research project on interaction in L2 oral proficiency tests [Swedish Research Council, 2012–4129]. The tests collected form one part of the National Test (NT) of English, which is mandatory for all 9th graders (the final year of compulsory school) in the country.

4.1 The National English Speaking Test

The speaking part of the NT, the National English Speaking Test (NEST), constitutes one-third of the total test grade for English (see Borger, 2019; Swedish National Agency for Education, 2015).Footnote 1 The tests are administered and assessed by the learners’ own English teacher, who reports test grades to the authorities. Using teachers as examiners of high-stakes standardized tests has a long tradition in Sweden (Borger, 2019; Sundqvist et al., 2018).

The NEST construct targets learners’ free oral production and interaction (Swedish National Agency for Education, 2015), and this construct is grounded in the syllabus for English in compulsory school. The test is designed as a peer conversation, where test-takers are given a set of topics to discuss in pairs or small groups. The examining teacher is present, and allowed to help out when deemed necessary (Swedish National Agency for Education, 2015). Furthermore, the instructions state that the NEST can be taken in pairs or in small groups. Audio-recording the learners’ test interaction is recommended for assessment, but not required.

Typically the test consists of phases: a warm-up task, where students are asked to tell their co-participants something about themselves or describe a picture, they then move to more advanced interactional tasks where learners are given “topic cards” with statements or questions.Footnote 2 Learners then take turns drawing topic cards and initiating talk on each topic. While several topic cards are included with each year’s test, it is up to the examining teacher to decide whether to use all. Teachers (in all NT subjects) may use old national tests available at the test constructors’ website in the classroom in order to familiarize students with the types of tasks used (see, e.g. Nilsberth & Sandlund, 2021). The grading criteria are largely aligned with the descriptors of the communicative abilities in the Common European Framework of Reference for Languages (CEFR) (Council of Europe, 2001) and utilize a grading scale from A to E, where E is the lowest passing grade, and F is a failing grade (see Swedish National Agency for Education, Teacher information, 2014; Borger, 2019; Sandlund & Sundqvist, 2019).

4.2 Data and Participants

The data corpus consists of 71 audio-recorded paired and small group tests, encompassing a total of 161 9th grade learners of English as an additional language (EAL) in Sweden. English is not one of the official languages in Sweden and is taught as a foreign language from age 7. All 71 test-takers were students in the 9th grade of compulsory school (ages 15–16), from four schools in southwestern Sweden, which were approached by the researchers about participation the test context is high-stakes (Sundqvist et al., 2018) and audio-recordings were made by the examining teacher. Video recordings were not permitted given the high-stakes context, but audio-recordings were considered ethically acceptable and minimally disruptive. The project underwent ethical review at the host university, and all participants gave their written consent in alignment with national guidelines (Swedish Research Council, 2017). The test itself was under a secrecy ban between 2014 and 2020 as parts of tests are sometimes reused, and up to 2020, researchers agreed not to reproduce the tasks verbatim in publications.

The test instructions recommend a test duration of 15–25 min, and the recordings fall within this scope. For the picture prompts, learners are instructed to talk about one of several pictures and to “describe what you see. What do you think is going on? What are your feelings and thoughts?”. After the warm-up task, students move on to Part One (blue cards) and Part Two (yellow cards). The test package was themed “Who cares?” and the topic cards generally asked test-takers to imagine different hypothetical scenarios and express their opinions on moral dilemmas. Tasks also prompt students to speculate on what they would do in each situation, such as You see a dog locked up in a car on a hot day. How would you react? What could you do? What should other people do? Discuss with your friend! (Swedish National Agency for Education, 2014, Subtest A, English). Furthermore, there are topic cards with statements such as People care too little about how they use social media, followed by the discussion prompt Agree? Disagree? Explain why and give examples. Discuss with your friend! The test package includes eight photographs, twelve blue cards with hypothetical scenarios, and an additional twelve yellow cards with opinion statements. As both the task type and the pre-set topic interplay with the local interactional affordances for the emergence of tellings, sequences featuring all three task types were examined.

4.3 Analytic Approach

Conversation analysis (CA) has contributed greatly to the field of language testing and to the description and specification of IC for teaching and assessment (Kasper & Ross, 2013; Salaberry & Kunitz, 2019; Salaberry & Burch, 2021). The approach is uniquely suited for the understanding of how “proficiency” is displayed, how tasks for assessment are oriented to, and in providing empirical description of actions of relevance to IC. For the purpose of the present chapter, the analytic approach centers on the interactional resources participants draw on in producing or responding to tellings, and on participants’ displayed orientation to the speaking tasks-at-hand. Audio-recordings were transcribed in their entirety, and selected sequences were transcribed in detail using CA transcription conventions (Hepburn & Bolden, 2017).

As for phenomenon delimitations, “stories” identified in the corpus are generally not multi-turn sequences with clear story prefaces, a recognizable stepwise production of turns projecting a next part of the telling action, and closings (cf. Sidnell, 2010; Mandelbaum, 2013). Instead, the analysis centers on shorter tellings. As such, co-participants will have to rely on their knowledge of sequential organization and on features of delivery in the recognition of the talk as a telling, and respond to such tellings in ways that comply not only with interactional preferences, but also according to their understandings of the assessment activity. The interactions, while “naturally occurring” as L2 test interactions (cf. van Compernolle, 2011; Sandlund et al., 2016), are heavily constrained by the tasks themselves and the local testing context. It is worth pointing out that none of the pre-set tasks specifically invited storytellings. Rather, discussion topics specifically ask test-takers to speculate on possible scenarios and offer their moral assessment of them. As a result, a majority of discussion topics as they are managed by students remain on an abstract, non-specific level, not necessarily “oral versions of personal experience” (Schegloff, 1997: 98). The analysis centres on two aspects of tellings in tests: (1) the task management contexts in which tellings and stories are made relevant and (2) the interactional management of such occasionings. We will begin with two contrasting examples of stories in the picture prompt task.

5 Telling in a Test: An Examination of Tellings Across Task Types

5.1 Tellings in Warm-Up Tasks: Hypothetical Stories in Doing Description

The picture prompt tasks are usually accomplished through brief descriptions of people and surroundings in the picture, with just a few turns from each speaker and no questions from co-participants. However, on occasion, test-takers make the warm-up task highly interactional by turning descriptions into hypothetical stories, as in Excerpt 1 below.

Excerpt 1

[TT 11041051, 00:00:22-00:01:17] “Terrorist”

figure a

This interaction takes place at the very beginning of the test, and Sid (line 123) has just volunteered to talk about his selected photo prompt. In line 125, he begins describing the photograph with increasing specifications as he goes along. He describes soldiers, specifies this description with their affiliation being the United Nations, and also suggests a location for the event (“Brazil or something”, lines 126–127). He also describes seeing a baby being carried. So far, Sid has provided descriptive talk on the characters in the picture, but as he continues (131–134), he also offers a description of a possible event behind the photo—a terrorist raid on a village that the UN officers are trying to save.

A transition point is made available after line 138, and Eddie requests elaboration of the hypothetical scenario as to the events preceding the terrorist raid, where he also offers several candidate proposals (bad leadership, meanness, religion). With his question, Eddie prompts Sid to elaborate on his hypothetical story and displays to Sid his active listenership through fitting his question to the informing in Sid’s turn. Sid aligns with the question and produces an I think-prefaced response, and attributes the terrorist attack to leadership. Eddie confirms this (line 148), and their exchange on the issue of terrorism continues for another few rounds of turns (not shown) before Eddie brings the topic to a close and initiates talk on a different photograph.

As mentioned, most warm-up sequences stop at consecutive descriptions of characters in a picture, whereas Excerpt 1 shows two test-takers collaboratively treating the task as one of constructing a hypothetical story about the events leading up to what is depicted in the photograph. By treating the task this way, participants go beyond describing what they see and also accomplish on-task talk which gives them the opportunity to display their IC for assessment. It must be underscored that this is not a typical treatment of the warm-up task, but shows one way in which test-takers can deploy hypothetical stories to accomplish an interactional warm-up task, thereby turning a relatively monological task into a collaborative one.

5.2 Tellings as Second Stories for Task Accomplishment

Another slot in which tellings are occasioned in relation to picture prompts is when a description of something in a photograph produced by one speaker is treated as recognizable as an experience in their lives by a co-participant. In many ways, what test-takers produce in such instances works as second stories (cf. Wong & Waring, 2020) to topics initiated in general terms. Excerpt 2 constitutes an illustration, where Ellen and Mia are gazing at a picture of a man who appears to be donating blood. After Mia has described what she sees and both test-takers have agreed that donating blood is important (lines 70–83, not shown here), Ellen volunteers a telling about her dad who donates blood (line 80). This informing is met by Mia with a second informing (line 82) that her mom also donates blood:

Excerpt 2

[TT 35072042 00:05:21-00:05:40] “My mom does it too”

figure b

After this informing, she produces a telling reporting on a recent event that her mom had in fact donated blood as recently as the day before. Both girls’ tellings are produced with smile voices and interspersed laughter particles. Without elaborating on the specific potential story of yesterday’s events, Ellen returns to general matters, such as the side effects from donating blood (line 86), and this sequence continues on the topic of blood donations for another few minutes.

Excerpt 2 shows an instance where a potential story could have been made relevant, such as a description of the mother’s experiencing of side effects the day before, but the potential story stops at the “she > did it < ↑↑YESterday£”, which shows uptake of and affiliation with Ellen’s informing and a reciprocation. This type of one-TCU tellings in the sequential slot of a second story is common in the data. They rarely move, however, beyond a two-turn mini-story. By expanding on the topic in this way, Mia displays her analysis of Ella’s talk and the topic of the task. Such second position tellings, then, are “built to show that they are touched off by and/or are picking up the point of the story to which they are responding” (Mandelbaum, 2013, p. 505). Moving on to the more advanced interactional tasks, we will now examine some ways in which tellings are volunteered and rejected.

5.3 Tellings in Topic Discussion Tasks: Volunteering Stories in Support of Claims

A common way in which stories are volunteered in the NEST data is in the sequential context of making a claim or taking a stance on a topic. As discussed earlier, the yellow topic cards use statements to which the test-takers are instructed to react by agreeing or disagreeing, and accounting for their expressed agreement/disagreement. In Excerpt 3a below, participants are test-takers Henrik and Fred, and their teacher (TEA in transcript).

The sequence is taken from one of the statement tasks. Henrik is prompted by the teacher TEA (line 124) to draw a yellow card, and Henrik shows his understanding of the request by initiating a reading aloud of the card statement and the discussion prompts printed (lines 127–131):

Excerpt 3a

[TT_File_32111231_08.02 – 09.07] “Marilyn Manson and COD”

figure c

Having finished reading the statement, Henrik immediately embarks on the first part of accomplishing the task instructions: that of agreeing or disagreeing with the statement (lines 134–135), where he signals disagreement (I don’t think they are caring too much). His turn shows orientation to the card formulation too much in his claim of disagreement, but subsequently produces a caveat (lines 136–139). The exception to his general opinion on the matter is that the unspecified category of “people” displays excessive “care” about teens playing computer games specifically in relation to violent games: the only thing i- in that case is VIolent games becuz they think (.) they (0.7) mess (0.5) the teens’ ↑head up > ‘ur somethin (.) like that so < we’re going to be like mass murderers or somethin’”. Here, Henrik takes a general stance on others’ involvement in teenagers’ gaming activities as relatively unproblematic—except for in the case of violent games. His reference to “they think” conveys an implicit assessment of adult attitudes on the matter as over-the-top by use of a membership category (e.g. Stokoe, 2012), mass murderers, to depict what adults may believe. The turn is produced with emphasis or volume on several key items (violent, mess, heads) and is hearable as a designedly exaggerated version of the mindset of adults. Up to this point, Henrik has engaged only with the agreement/disagreement part of the speaking task. The co-present teacher (line 141) produces an acknowledgement token with slightly rising intonation, projecting that more talk is expected. Now, pay close attention to Henrik’s next actions, lines 143–153, where Henrik delivers a telling of a story.

Excerpt 3b

[TT_File_32111231_08.02 – 09.07] “Marilyn Manson and COD”

figure d

In line 143, Henrik indicates the link between prior talk and projected talk with “kind of the same with”, followed by a lengthy silence. He restarts his formulation in line 145, again indicating trouble with delivering the gist of his message in the initial abandonment of a new turn beginning. While he does project an upcoming comparison of some kind, it is not a typical story preface (Sacks, 1974). Instead, the storytelling is initiated by presenting it as an example: “we jzt had an example when (therewz) a kid that wz playing like (0.5) Marilyn Manson an::d (hh). CO:(hh)D”. Co-participants can therefore hear Henrik’s turn as initiating an elaboration of his disagreement, and since the gist of his point is left unsaid at line 147, a continuation is projected. No verbal receipt from his co-participants is forthcoming, and Henrik (line 149) continues with the brief story. The connection to prior talk is displayed with the turn-initial and, and Henrik self-repairs his initial “he got-” before restarting with “he killed somebody and > theywr like < ”. So far, his story has two parts: there was a kid (the protagonist), he listened to music by rock singer Marilyn Manson and played the video game Call of Duty (COD) (a description of the protagonist’s lifestyle as background), and this kid killed someone (the main event of the story) (cf. Goodwin, 1984; Holt, 2017; Mandelbaum, 2013). However, the specific interactional role of the story is not yet revealed. At this point, Henrik projects an upcoming enactment of talk using the reported speech preface “and > theywr like < ” (Golato, 2000; Sandlund, 2014), which follows next. Note that the category of speakers referenced here remains underspecified (“they”). The enactment is animated with an initial and prosodically marked “↑O:H↓”, and the enactment takes the grammatical form of direct reported speech, where antagonists of video games are depicted as blaming the murder on Manson’s music and video games. As is often the case with enactments, they simultaneously report and convey an assessment of the talk reported (Holt, 2000; Sandlund, 2014). Henrik’s turns so far have already given away his assessment, as he enacts the talk reported with a prosodic shift and the depiction of the connection as an exaggerated simplification of cause and effect in the story (its’ only becuz). Talk following up on the climax of a story can be designed to facilitate recipient uptake, such as an upshot or an assessment (cf. Jefferson, 1978). Henrik’s turn (line 154) provides his stance on the position enacted (don’t th[ink THAT’s THE ↑CA::SE = ) and proceeds to provide a contrastive analysis: many people play video games, and presumably, not all end up murdering someone, although this point is only implied (for an analysis of subsequent talk, see Sandlund & Sundqvist, 2019). Henrik, then, re-orients to the first part of the task, the card statement, and summarizes the upshot of his prior talk: “bud◦ < I: don’t think they caring s: so much about it s: really”.

As we have seen, a brief story is introduced as an elaboration of an expressed position on the topic card statement. The story was volunteered by the learner, and is told as part of the task accomplishment of explaining one’s reasons for agreeing or disagreeing. We can see Henrik’s successful accomplishment of the brief story using a hypothetical enactment of talk depicting an extreme position on the matter of video games, the production of connections to prior talk, and the upshot and expression of disagreement. A further observation is that co-participants, with the exception of the teacher’s acknowledgement token that displays her expectation of further talk from Henrik, do not verbally contribute to the story or provide assessments of it, but await further talk. This is rather typical of these test interactions, where participants orient to the test format as an interaction where test-takers often “exhaust their commentary on a topic before a turn shift is made relevant” (see Sandlund & Sundqvist, 2016, p. 128). As such, Fred’s (and the teacher’s) lack of contributions during Henrik’s telling may reveal this institutional orientation where Fred anticipates that he will offer his view on the topic separately, after Henrik has displayed completion of his talk. Some evidence for this interpretation comes next, when instead of, for example prompting Fred to engage with Henrik’s story talk, the teacher re-orients the interaction to the card statement and prompts Fred to begin with offering agreement or disagreement (line 163). In sum, the brief story in this piece of test interaction is fitted to the task at hand, but the story as such does not result in further elaboration or a second story.

5.4 Tellings in Topic Discussion Tasks: Collaborative Storytelling

While Excerpt 3 showed recruitment of the story format in accomplishing the opinion task, other stories are less tied to the topic card formulations. In Excerpt 4, Sally, Jill, and Sophie are midway through the yellow statement cards, and the task this time is a card formulation reading People care too little about equal rights for men and women, followed by the same instructions as in Excerpt 3 (i.e. Agree/disagree, explain, give examples). Prior to Excerpt 4, Sally, at first, displays non-understanding of the expression “equal rights”, and co-participants (including the teacher) exemplify the concept by way of salary differences between men and women. Having agreed that such differences are unfair (464–489), the teacher asks whether there are things women can do which men cannot. This question initiates new topical talk regarding femininity and the difficulties boys face when wishing to adopt “girly” attributes. Sally (line 532 below) makes a claim based on prior talk that it is easier for “us woman” to shop in a boys’ department (only implied as Jill and Sophie show immediate understanding), and in line 537, she provides the contrast: “but a boy cannot (.) sh[op,”. This is overlapped by Jill’s turn and Sally drops out of the competing talk. The analytic attention is on what Jill initiates in line 538, in overlap, which is the beginning of a story.

Excerpt 4a

[TT 33062212202 00:20:06-00:21:12] “Girls’ pants”

figure e

In line 538, Jill produces an emphatic “↑↑YE:AH” in response to Sally’s upshot and initiates a telling about what she has seen on Instagram from a famous Swedish male pop singer. Her telling is thus launched with a place reference, on Instagram, which is a common practice for story beginnings (Dingemanse et al., 2017). Sophie displays her recognition of the event in her empathetic affirmative token (line 540), which mirrors the animated tone of Jill’s story initiation. Sophie has thus indicated some epistemic access to what is to be told, but she does not produce further talk on the matter, displaying that she recognizes Jill’s turn as a storytelling action in which co-participants appropriately suspend further talk (Mandelbaum, 2013). Jill proceeds with the next part of her story—what pop star Saucedo had bought (“buyed”). Here, Sophie latches on and demonstrates her equal epistemic access to the event, and produces “buyed a ↑↑GI[RL pant.”. She completes Jill’s turn, but Jill persists in telling her version, and produces “[↑GIRL pants” in overlap with Sophie’s turn completion. While Sophie and Jill appear to be competing to tell the story, their collaborative building of the story also displays active listenership, which has been connected to the display of IC (Lerner, 2004; Sert, 2019) through their joint turn-building actions.

Jill proceeds with the story and conveys that the protagonist, Saucedo, wrote something (presumably on Instagram), but she is displaying interactional trouble in formulating the full story contribution (“he: s’- he:: (0.3) he ↑wrote that *e:::::* the::: = ”; line 543–544). For the second time, Sophie latches onto Jill’s incomplete TCU and begins completing it in the form of reported speech on Saucedo’s written comment (“they say they were girl pan[ts”, line 545). The remainder of the enactment of Saucedo’s words is produced chorally:Jill produces “but I bought them” and Sophie simultaneously produces the equivalent ``but I bu::y” (line 547). Both test-takers thus complete the telling chorally and collaboratively, and Jill displays some persistence in beating Sophie to the punchline by, in turn, completing Sophie’s delivery of the first part of the reported speech. Sophie drops out of the overlap (line 547), whether because Jill had initiated the story, or because she notices Jill’s embedded correction of the past tense form bought.

Having arrived at the story climax, Jill continues with an assessment of the event as such (line 551), and the pants themselves (line 555), but in line 554, she also adds to the story with how there was also a picture of the protagonist in his pants, after which an assessment is relevantly due: “and it wz really nice” (line 555) and that she liked them “very much” (line 557). Jill, then, has established herself as the primary teller of the story who has more to add. The sequence takes a new turn as Sally, who does not appear to have access to the event, asks for more information about the pants (line 558):

Excerpt 4b

[TT 33062212202 00:20:06-00:21:12] “Girls’ pants”

figure ffigure f

In lines 559–560, Jill again faces competition from Sophie on describing Danny Saucedo’s pants as Sophie beats her to the delivery of a colour descriptive term. Candidate formulations regarding the appearance of the pants follow from both girls (lines 561–562), with apparent trouble regarding whether the pants are described as partially “white” or “wide”, but in line 563, Jill turns towards a different type of description: “they were like *e:::* all kinds ofe:h (1.3) ofe:h (0.6) ye(hh)ah”, which is treated as an invitation to join in her laughter (lines 564–566). Following the subsiding of joint laughter, Jill restarts her attempt to describe not only the colour and fit, but the print (“i(hh)t wz like circles and it was no- I dunno”). The teacher orients to the repeated dysfluencies of Jill’s turn, and offers a candidate description: “different patterns’ur” (line 570). In line 573, Jill confirms this as correct and adds an additional assessment of the event: “an’ itwz- but it was really nice annit was good thadhe wrote that ↑too”, which shows orientation to how the story was occasioned—the topic of gender equality rather than fashion. She produces a final agreement and assessment “ = yeah. (0.2) I like that.” (line 579) before the teacher produces an upshot of the discussion-so-far.

In Excerpt 4, we have seen how test-takers diverge from the first topic card statement in launching a broader discussion on gender equality. The story itself can be viewed as sequentially positioned to do task accomplishment by providing examples of a viewpoint, but the sequence also differs from Excerpt 3 in that the story is collaboratively accomplished by all participants—two with displayed knowledge of the event told, and two who contribute to the elaboration (Sally) and upshot (the teacher) of the story. From the animated production of the sequence initiation, the storytelling shows something like genuine interactional engagement (cf. Sandlund & Greer, 2020) rather than explicit task orientation. While the task is of the same type as in Excerpt 3, the way in which test-takers manage it is highly collaborative. As a consequence, test-takers seem to temporarily suspend institutional norms of staying on task and instead allow the conversation to develop on a side-topic developed through a story initiation. The ways in which test-takers manage to accomplish such topic digressions through stories may be something to actively encourage in tasks, as it allows for the display of a wide range of competencies.

5.5 Tellings in Topic Discussion Tasks: Resisting Invitations to Tell

There are also sequences in which the initial topic card talk occasions an elaboration question from a co-participant, and where such invitations are rejected, or treated as an invitation to produce a more general, non-personalized telling. In Excerpt 5 below, the topic card at hand reads You find out that a friend is writing mean comments about another friend on Facebook. What could you do about it? Three test-takers—Kelli, Emma, and Annica and their teacher—are the participants, and the topic card is read by Emma in line 270 (not shown). Between lines 270 and 311, all three test-takers contribute to the topic by offering strong assessments of online bullying. In line 311, Kelli, using a polar question, hearable as an invitation to tell, asks her co-participants “ > have you ever seen something? (.) on facebook? (.) like (hh) someone is, = ”.

Excerpt 5

[TT 33182032102, 13:45-14:16] “Not that popular

figure g

In lines 311–312, Kelli’s question, which is restarted and repaired, indicates a laboured attempt at formulating the question and appears designed to elicit co-participants’ own experiences of witnessing social media bullying. Her incomplete TCU at the end of line 312 may be oriented towards some display from Emma in terms of her readiness to respond. Alternatively, it is deliberately designed as an invitation to the others to produce other-repair or contribute in other ways. Emma latches on to Kelli’s incomplete TCU with a turn that casts the question as unanswerable, (i.e. Facebook is an irrelevant platform). While Kelli’s question orients to the topic card formulation, Emma instead delivers a claim about other social media platforms (lines 315–316). As such, Emma disattends to any opportunity to produce a story or to respond on the basis of personal experiences, and instead reconfigures the context for when such a question would be relevant. Emma (line 319) continues with a general observation that she indeed “sometimes” sees pictures of “special” people on Instagram, as opposed to giving a personalized telling of a specific instance.

This type of sequence, where a co-participant offers a storytelling invitation but recipients either disattend the invitation, reject it completely, or opt to respond on a general topic level, is common in the dataset. This brief extract represents a collection of instances where test-takers orient to topical talk as general rather than personal in the context of the topical discussion tasks. In a case study of one specific test interaction where the two test-takers had been assessed very differently by three different raters, despite showing similar grammatical and lexical proficiency (Sandlund & Sundqvist, 2016), a close analysis revealed that one of the learners recurrently recruited his personal experiences in contributing to topical talk, while the other kept his contributions in general terms. The authors discuss whether the first test-takers’ claims of epistemic access to topics through displaying personal involvement and experience was one of the reasons behind his higher test grade (p. 128). However, most test-takers appear to treat personal perspectives as irrelevant and even inappropriate in task-based test interaction. Further research should examine closely whether test-takers systematically treat pre-set topic tasks in this particular way, and also, whether the tasks themselves could prompt non-sensitive storytelling in ways that allow test-takers to display more of their IC (cf. also Sandlund et al., 2015).

6 Conclusions and Implications

Tests are highly consequential for learners, and tasks in tests should allow test-takers to display a wide range of competencies without feeling “trapped” by topics on which they have little to say. Tasks for speaking assessment should be designed not only to capture the construct (i.e. aspects of IC and L2 proficiency), but also to match the test-taker targets in terms of maturity, proficiency level, interests, and life worlds (Green, 2014; Fulcher, 2003). The relationship between storytelling in language teaching and storytelling activities in language assessment is relatively under-researched in CA-for-SLA studies (with a few exceptions, e.g. Kim, 2016), although recent work in conversation analysis of L2 learning and interactional competence has argued that practices building up IC are indeed “teachable objects” (Waring, 2018, p. 65). As for affordances in test interactions, it has been argued that L2 speaking tests at best uncover a learner’s IC as a test-taker in the domain-specific context of participating in test interaction (e.g. van Compernolle, 2011). This claim may be even more salient for interview tests, but it would be difficult to claim that the test conversations are close matches of everyday conversations, or other types of institutional interactions in a learner’s future, as few dinner conversations or work meetings revolve around pre-set tasks and topics. In this regard, we can see that the stories that are occasioned in the dataset were, with one exception, less collaborative than what has been observed in previous CA research (e.g. Mandelbaum, 2013; Goodwin, 1984), and affiliation with tellers of the kind observed by others (Stivers, 2008) is relatively rare. We also know from educational research that storytelling activities in the classroom—across subjects—is considered fundamental for learning and for understanding oneself and the surrounding world (e.g. Nelson, 2005). This suggests two important questions for further research: (1) Do assessment contexts match the types of interactional activities that language learners become familiar with through classroom activities, and (2), is storytelling viewed as an important part of a learner’s IC?

As for the first question, if the telling and recipiency of stories at sequentially relevant junctures are to be considered an important IC, as it appears to be in language classrooms, learners should be given the opportunity to produce stories also in testing contexts. The speaking tasks examined did not specifically invite tellings or stories, but instead, were designed to capture more general L2 interactional abilities such as comprehensibility, fluency, accuracy, adaptation to purpose, situation, and recipient (Swedish National Agency for Education, 2014). It may be that since these learners are relatively fluent, describing and narrating events may be considered more appropriate for earlier educational stages. However, as demonstrated, tellings and “small stories” (Bamberg & Georgakopoulou, 2008) are occasioned by test-takers’ displayed understandings of what task accomplishment entails—in supporting claims, by some mention of an event or phenomenon reminiscent of participants’ prior experiences, or explicitly invited by others. Such stories may be volunteered, invited, or resisted by recipients, and in the dataset, participants tended to treat test topics as abstract rather than objects for personalization (cf. Sandlund et al., 2015). This feature of storytelling reveals something about the institutionality of speaking tests, where test-takers display their understanding of the assessment context by remaining within the boundaries of generality in their task accomplishment. This observation bears pedagogical implications: if the purpose of tasks is to give learners a variety of opportunities for displaying proficiency and IC, then tasks should perhaps be less opinion-based, which seems to invoke abstract talk, and consequently, make stories interactionally delicate. It can also be noted that discussions of hypothetical scenarios generate possible stories, but rarely become elaborated and collaboratively negotiated. However, when participants opt to diverge from strict task adherence, spontaneous and collaborative storying becomes possible, and a more natural-sounding interaction unfolds, and test-takers can display a wide range of competencies for assessment thus, for the second question on storytelling as IC, if storytelling is considered important for L2 learning, it should also be viewed as important in assessment.

7 Some Other Recommendations

Emanating from the present study include the importance of preparing students for tests (cf. Nilsberth & Sandlund, 2021). Considering the observations, it is crucial that students, in the testing context, know what becomes treated as relevant for assessment, and that strict task adherence is less important than interactional engagement, for example. Furthermore, if tellings of personal experiences can put learners in an uncomfortable position (cf. Ross, 1998), hypothetical or fictional stories could be encouraged. That is, not all contributions need to be “true” in order to function as IC/proficiency output. Also, if narrative oral tasks are common in the learning context, but absent in tasks for assessment, it is not surprising that expanded and collaborative tellings are relatively rare. However, as noted, learners are resourceful in drawing on first and second story initiations (hypothetical or experience-based) in managing task accomplishment. As such, the telling and receipt of small stories could be encouraged, and even included, in scoring rubrics for interaction.

Finally, it can be argued that all participants in second language tests are competent as interactants already—they are “masterful navigators in changing social environments and they know methods by which to engage in constructing and maintaining social order” (Eskildsen, 2018, p. 69). The question for storytelling research across contexts, and for practitioners of language teaching, then, is what language-specific methods and practices learners must become familiar with when it comes to practices of storytelling, which in turn can inform tasks for speaking assessment. In their research-based resource book for ESL teachers, Wong and Waring (2020) outline classroom exercises that involve practising initiating as well as responding to stories in a second language. Their recommendations are based on empirical work in CA, and illustrate hands-on practice and exercises to be used in the classroom. Likewise, studies on how test-takers manage tellings in high-stakes contexts can provide an impetus for the development of an expanded list of competencies for the overall IC construct, but also for the development of tasks that encourage and support collaborative tellings, as it is clear that tellings not only work as resources for task accomplishment, but also allow for a variation in what test-takers may appropriately display in a test.