1.Introduction

In spoken interaction, participants often refer to their own thoughts and opinions with phrases such as I think, which take the grammatical form of a minimal clause and simultaneously express the speaker’s stance towards upcoming or preceding talk (see, e.g., Thompson and Mulac 1991Thompson, Sandra A., and Anthony Mulac 1991 “A Quantitative Perspective on the Grammaticization of Epistemic Parentheticals in English.” In Approaches to Grammaticalization: Volume II. Types of Grammatical Markers. ed. by Bernd Heine, and Elizabeth C. Traugott, 313–329. Amsterdam: John Benjamins Publishing Company. ; Aijmer 1997Aijmer, Karin 1997 “ I think – An English Modal Particle”. In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan, and Olaf Jansen Westwik, 2–44. Berlin, New York: Mouton de Gruyter. ; Thompson 2002Thompson, Sandra A. 2002 “Object Complements and Conversation: Towards a Realistic Account.” Studies in Language 26 (1): 125–163. ; Kärkkäinen 2003Kärkkäinen, Elise 2003 Epistemic Stance in English Conversation. A Description of Its Interactional Functions, with a Focus on I think. Amsterdam: John Benjamins. ; Keevallik 2010Keevallik, Leelo 2010 “Clauses Emerging as Epistemic Adverbs in Estonian Conversation.” Linguistica Uralica 46 (2): 81–101. ; Mullan and Karlsson 2012Mullan, Kerry, and Susanna Karlsson 2012 “Subjectivity in Contrast: A Cross-Linguistic Comparison of ‘I think’ in Australian English, French and Swedish.” In Subjectivity in Language and in Discourse, ed. by Nicole Baumgarten, Inke Du Bois, and Juliane House, 271–294. Brill.). For example, the English complement-taking predicate I think is routinely deployed as a stance-framing device before the speaker presents their point of view (Thompson 2002Thompson, Sandra A. 2002 “Object Complements and Conversation: Towards a Realistic Account.” Studies in Language 26 (1): 125–163. , 125; Kärkkäinen 2003Kärkkäinen, Elise 2003 Epistemic Stance in English Conversation. A Description of Its Interactional Functions, with a Focus on I think. Amsterdam: John Benjamins. , 35). However, stance-taking in conversation is not merely a case of quoting a speaker’s pre-formulated thoughts. Consequently, phrases such as I think may also be deployed retroactively to hedge the previous utterance or mark its completion due to, for instance, no uptake from interlocutors (Kärkkäinen 2003Kärkkäinen, Elise 2003 Epistemic Stance in English Conversation. A Description of Its Interactional Functions, with a Focus on I think. Amsterdam: John Benjamins. , 166; Auer and Lindström 2016Auer, Peter, and Jan Lindström 2016 “Left/right Asymmetries and the Grammar of Pre- vs. Post-positioning in German and Swedish Talk-in-Interaction”. Language Sciences 56: 68–92. , 86). Moreover, formatting upcoming talk as a ‘thought’ can be beneficial for managing the ongoing interaction. For example, in the context of joint decision-making, the construction of a proposal as a thought mitigates imposing on co-participants, which, in turn, allows the co-participants to reply in more open-ended ways (Stevanovic 2013Stevanovic, Melisa 2013 “Constructing a Proposal as a Thought: A Way to Manage Problems in the Initiation of Joint Decision-Making in Finnish Workplace Interaction.” Pragmatics 23 (3): 519–544., 521).

Swedish has three verbs that can be translated into the English verb to think, namely tänka ‘to think/cogitate’, tycka ‘to think/find’ and tro ‘to think/believe/guess’ (Aijmer 1997Aijmer, Karin 1997 “ I think – An English Modal Particle”. In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan, and Olaf Jansen Westwik, 2–44. Berlin, New York: Mouton de Gruyter. ; Goddard and Karlsson 2003Goddard, Cliff, and Susanna Karlsson 2003 “Re-thinking THINK: Contrastive Semantics of Swedish and English.” In Proceedings of the 2003 Conference of the Australian Language Society, ed. by Cristo Moskovsky. http://​www​.als​.asn​.au​/proceedings​/als2003​.html; Viberg 2005Viberg, Åke 2005 “The Lexical Typological Profile of Swedish Mental Verbs.” Languages in Contrast 5 (1): 121–157. ). Deployed in the present tense with a 1st person singular subject,11.The past form I thought (see, e.g., Kärkkäinen 2012Kärkkäinen, Elise 2012 “I Thought It Was Very Interesting: Conversational Formats for Taking a Stance.” Journal of Pragmatics 44 (15): 2194–2210. on English; Deppermann and Reineke 2017Deppermann, Arnulf, and Silke Reineke 2017 “Epistemische Praktiken und ihre feinen Unterschiede: Verwendungen von ich dachte in gesprochener Sprache.” In Verben im interaktiven Kontext. Bewegungsverben und mental Verben in gesprochenen Deutsch, ed. by Arnulf Deppermann, Nadine Proske, and Arne Zeschel, 337–375. Tübingen: Narr Francke Attempto. on German) has distinct functions and falls outside the scope of this study. these verbs may convey the speaker’s epistemic stance towards the rest of the utterance with slightly different meanings. In general, jag tycker ‘I think/find’ expresses an opinion and jag tror ‘I think/believe/guess’ a belief, while jag tänker ‘I think/cogitate’ can express the speaker’s direct reported thought, mental activity, or intention (Viberg 2005Viberg, Åke 2005 “The Lexical Typological Profile of Swedish Mental Verbs.” Languages in Contrast 5 (1): 121–157. , 144–151). However, jag tänker is not fully accepted as a stance-taking phrase in normative contexts (Kolu 2022Kolu, Jaana 2022 “Vad tänker du om det? Verbet tänka som epistemisk markör i tidningsspråk [What do you think about it? The verb tänka as an epistemic marker in newspapers].” In Svenskan i Finland 19, ed. by Siv Björklund, Bodil Haagensen, Marianne Nordman, and Anders Westerlund, 149–162. Vasa: Svensk-Österbottniska Samfundet., 150). The division of labour between the verbs also tends to cause difficulties for people learning Swedish as their second language (L2) (Viberg 2005Viberg, Åke 2005 “The Lexical Typological Profile of Swedish Mental Verbs.” Languages in Contrast 5 (1): 121–157. ; Toropainen and Lahtinen 2013Toropainen, Outi, and Sinikka Lahtinen 2013 “Argumentation på L2-svenska: inlärare skriver insändare [Argumentation in L2-Swedish: learners write readers’ letters].” In Svenskans beskrivning 33, ed. by Jan Lindström, Sofie Henricson, Anne Huhtala, Pirjo Kukkonen, Hanna Lehti-Eklund, and Camilla Lindholm. Helsingfors: Helsingfors Universitet., 485).

2.Aim and research questions

This study investigates how L1 and L2 speakers of Swedish use the verbs tänka, tycka and tro, in the 1st person singular and the present tense, in the context of talk-in-interaction, with a specific focus on the previously uninvestigated phrase jag tänker ‘I think/cogitate’. Theoretically and methodologically, the study draws on Interactional Linguistics (IL, see Selting and Couper-Kuhlen 2001Selting, Margret, and Elizabeth Couper-Kuhlen (eds.) 2001 Studies in Interactional Linguistics. Amsterdam: John Benjamins Publishing Company. ; Couper-Kuhlen and Selting 2018Couper-Kuhlen, Elizabeth, and Margret Selting 2018 Interactional Linguistics. Studying Language in Social Interaction. Cambridge: Cambridge University Press.). IL advocates an interactional view on language and focuses on interactionally relevant categories to which participants orient in the real-time progression of talk in naturally occurring interactions (Couper-Kuhlen and Selting 2018Couper-Kuhlen, Elizabeth, and Margret Selting 2018 Interactional Linguistics. Studying Language in Social Interaction. Cambridge: Cambridge University Press., 18–26). Attention is also paid to how speakers utilize multimodal and material resources such as prosody, gestures, and artefacts in their interactions (Stivers and Sidnell 2005Stivers, Tanya, and Jack Sidnell 2005 “Introduction: Multimodal Interaction”. Semiotica 156–1/4: 1–20. ; Couper-Kuhlen and Selting 2018Couper-Kuhlen, Elizabeth, and Margret Selting 2018 Interactional Linguistics. Studying Language in Social Interaction. Cambridge: Cambridge University Press.; Streeck 2021Streeck, Jürgen 2021 “The Emancipation of Gestures.” Interactional Linguistics 1 (1): 90–122. ). Originally, the IL approach advocated a strictly Conversation Analytical study of language in interaction (Selting and Couper-Kuhlen 2001Selting, Margret, and Elizabeth Couper-Kuhlen (eds.) 2001 Studies in Interactional Linguistics. Amsterdam: John Benjamins Publishing Company. , 2–3; for Conversation Analysis [CA], see, e.g., Sacks, Schegloff and Jefferson 1974Sacks, Harvey, Emanuel Schegloff, and Gail Jefferson 1974 “A Simplest Systematics for the Organization of Turn-Taking for Conversation.” Language 50 (4): 696–735. ; Sidnell and Stivers 2013Sidnell, Jack, and Tanya Stivers 2013 The Handbook of Conversation Analysis. Chichester: Wiley-Blackwell.). However, the field has since broadened significantly and may include even other methods such as quantitative analyses of spoken language (Couper-Kuhlen and Selting 2018Couper-Kuhlen, Elizabeth, and Margret Selting 2018 Interactional Linguistics. Studying Language in Social Interaction. Cambridge: Cambridge University Press., 13).

The study presents a quantitative overview of the three phrases and a qualitative, interactionally informed examination of how jag tänker is deployed in its grammatical and sequential contexts. The data consist of task-based conversations, in which groups of either L1 or L2 speakers of Swedish discuss works of art. The research questions are:

1. How frequently do L1 and L2 speakers use the verbs tänka, tycka and tro, with a 1st person singular subject in the present tense, in the group conversations?

2. Given that its use in stance-taking contexts is descriptively and normatively less recognized, how do L1 and L2 speakers of Swedish deploy jag tänker ‘I think/cogitate’ in the group conversations?

Before presenting the empirical findings (Sections 6 and 7), I provide a brief outline of the Swedish verbs of thinking (Section 3), discuss L1 and L2 interaction (Section 4), and describe the data and methods of the study (Section 5).

3.Verbs of thinking in Swedish

The concept of ‘thinking’ includes aspects such as ‘cogitation’, ‘belief’, ‘opinion’ and ‘intention’ (Aijmer 1997Aijmer, Karin 1997 “ I think – An English Modal Particle”. In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan, and Olaf Jansen Westwik, 2–44. Berlin, New York: Mouton de Gruyter. , 12). In English, the verb to think conveys several of these meanings (Aijmer 1997Aijmer, Karin 1997 “ I think – An English Modal Particle”. In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan, and Olaf Jansen Westwik, 2–44. Berlin, New York: Mouton de Gruyter. ; Goddard and Karlsson 2003Goddard, Cliff, and Susanna Karlsson 2003 “Re-thinking THINK: Contrastive Semantics of Swedish and English.” In Proceedings of the 2003 Conference of the Australian Language Society, ed. by Cristo Moskovsky. http://​www​.als​.asn​.au​/proceedings​/als2003​.html). In contrast, both Swedish and German22.German is closely related to both Swedish and English. Moreover, Swedish and German are typologically similar; both are V2 (verb-second) languages and have several verbs of thinking. have several verbs that describe these meanings: in Swedish tänka ‘to think/cogitate’, tycka ‘to think/find’ and tro ‘to think/believe/guess’ (Aijmer 1997Aijmer, Karin 1997 “ I think – An English Modal Particle”. In Modality in Germanic Languages: Historical and Comparative Perspectives, ed. by Toril Swan, and Olaf Jansen Westwik, 2–44. Berlin, New York: Mouton de Gruyter. ; Viberg 2005Viberg, Åke 2005 “The Lexical Typological Profile of Swedish Mental Verbs.” Languages in Contrast 5 (1): 121–157. ), and in German denken ‘to think’, finden ‘to think/find’ and glauben ‘to think/believe’ (Auer and Lindström 2016Auer, Peter, and Jan Lindström 2016 “Left/right Asymmetries and the Grammar of Pre- vs. Post-positioning in German and Swedish Talk-in-Interaction”. Language Sciences 56: 68–92. ; Frommherz 2022Frommhertz, Yannick 2022 “Thinking Things in German versus Swedish. A Cross-linguistic Comparison of Verbs of Thinking in Two Genetically Close Languages.” Studia Linguistica 76: 464–506. ). However, there are differences between the Swedish and German verbs. For example, denken is heavily polysemous, while tänka is much more restricted in its usage (Frommherz 2022Frommhertz, Yannick 2022 “Thinking Things in German versus Swedish. A Cross-linguistic Comparison of Verbs of Thinking in Two Genetically Close Languages.” Studia Linguistica 76: 464–506. ). Denken is also commonly used to convey an affective stance, while tänka is not (Frommherz 2022Frommhertz, Yannick 2022 “Thinking Things in German versus Swedish. A Cross-linguistic Comparison of Verbs of Thinking in Two Genetically Close Languages.” Studia Linguistica 76: 464–506. ; Fiedler 2023Fiedler, Sophia 2023 “The Grammar-in-Use of Direct Reported Thought in French and German. An Interactional and Multimodal Analysis.” PhD dissertation, University of Neuchâtel/University of Hamburg., 326).

The investigated Swedish verbs and phrases are presented in Table 1.

The Swedish tänka (see Table 1) reports “a thought either directly word by word as a quotation or indirectly in the form of a that-clause” (Viberg 2005Viberg, Åke 2005 “The Lexical Typological Profile of Swedish Mental Verbs.” Languages in Contrast 5 (1): 121–157. , 144). Grammatically, tänka projects either a clause (with or without the preposition på and the complementizer att) or a noun phrase (with the preposition på, cf. ‘think about’). It can take a propositional complement in the present tense only if the complement verbalises an actual, occurring thought that is comparable to a quotation, and thus, concerns literal ‘thinking’ (Goddard and Karlsson 2003Goddard, Cliff, and Susanna Karlsson 2003 “Re-thinking THINK: Contrastive Semantics of Swedish and English.” In Proceedings of the 2003 Conference of the Australian Language Society, ed. by Cristo Moskovsky. http://​www​.als​.asn​.au​/proceedings​/als2003​.html; Frommherz 2022Frommhertz, Yannick 2022 “Thinking Things in German versus Swedish. A Cross-linguistic Comparison of Verbs of Thinking in Two Genetically Close Languages.” Studia Linguistica 76: 464–506. ). Tänka is also used as an auxiliary verb to describe intention (cf. ‘will, to be going to’) (Viberg 2005Viberg, Åke 2005 “The Lexical Typological Profile of Swedish Mental Verbs.” Languages in Contrast 5 (1): 121–157. , 146). In turn, tycka and tro express the speaker’s stance towards the rest of the utterance. Tycka expresses a subjective opinion on the scale good–bad and does not require justification, whereas tro signals “a belief in the truth or probability of the utterance” on the scale true–false (Viberg 2005Viberg, Åke 2005 “The Lexical Typological Profile of Swedish Mental Verbs.” Languages in Contrast 5 (1): 121–157. , 146–147). In sum, a proposition presented with tänka is neither intersubjectively verifiable, as a proposition expressed with tro would be, nor a purely subjective opinion, as a proposition expressed with tycka would be (Goddard and Karlsson 2003Goddard, Cliff, and Susanna Karlsson 2003 “Re-thinking THINK: Contrastive Semantics of Swedish and English.” In Proceedings of the 2003 Conference of the Australian Language Society, ed. by Cristo Moskovsky. http://​www​.als​.asn​.au​/proceedings​/als2003​.html; Frommherz 2022Frommhertz, Yannick 2022 “Thinking Things in German versus Swedish. A Cross-linguistic Comparison of Verbs of Thinking in Two Genetically Close Languages.” Studia Linguistica 76: 464–506. ).

In talk-in-interaction, jag tycker ‘I think/find’ can signify that an utterance represents the speaker’s subjective perspective, signal an evaluation, or occur as an increment to a turn due to lack of uptake (Karlsson 2005Karlsson, Susanna 2005 Modalitet i interaktion. En studie av jag tycker och tycker jag [Modality in interaction. A study of jag tycker and tycker jag ]. In Samtal och grammatik. Studier i svenskt samtalsspråk, ed. by Jan Anward, and Bengt Nordberg, 119–138. Lund: Studentlitteratur.). Jag tror ‘I think/believe/guess’ is commonly deployed in interaction to signal that an utterance is based on restricted knowledge (Auer and Lindström 2016Auer, Peter, and Jan Lindström 2016 “Left/right Asymmetries and the Grammar of Pre- vs. Post-positioning in German and Swedish Talk-in-Interaction”. Language Sciences 56: 68–92. , 81). However, no previous studies have examined jag tänker ‘I think/cogitate’ in Swedish talk-in-interaction. In a recent study that concerns jag tänker in newspapers, Kolu (2022Kolu, Jaana 2022 “Vad tänker du om det? Verbet tänka som epistemisk markör i tidningsspråk [What do you think about it? The verb tänka as an epistemic marker in newspapers].” In Svenskan i Finland 19, ed. by Siv Björklund, Bodil Haagensen, Marianne Nordman, and Anders Westerlund, 149–162. Vasa: Svensk-Österbottniska Samfundet., 153–155) concludes that the phrase occurs in informal texts that report opinions or a person’s speech and is synonymous with jag resonerar som så att ‘I reason that’. Tänka can also suggest an ongoing process of thinking, while tycka and tro describe a more static, cognitive state (Blensenius 2014Blensenius, Kristian 2014 Jag tänker att… [“I think that…”]. Språktidningen 1/2015 https://​spraktidningen​.se​/2014​/12​/jag​-tanker​-att/; Kolu 2022Kolu, Jaana 2022 “Vad tänker du om det? Verbet tänka som epistemisk markör i tidningsspråk [What do you think about it? The verb tänka as an epistemic marker in newspapers].” In Svenskan i Finland 19, ed. by Siv Björklund, Bodil Haagensen, Marianne Nordman, and Anders Westerlund, 149–162. Vasa: Svensk-Österbottniska Samfundet., 157). Thus, in talk-in-interaction, jag tänker can be a useful resource, because it allows the speaker to present a proposition and explore it, without necessarily taking a more definite stance towards it.

Finally, L2 speakers of a language tend to use stance-taking phrases in a manner that differs from L1 speakers, due to, for example, their L1 or other languages that they know (Baumgarten and House 2010Baumgarten, Nicole, and Juliane House 2010 “ I think and I don’t know in English as Lingua Franca and Native English Discourse.” Journal of Pragmatics 42 (5): 1184–1200. , 1185; Diaz et al. 2020Diaz, Maria Angela, Ken Lau, and Chia-Yen Lin 2020 “Pragmatic Functions of I think in Computer-mediated, Cross-cultural Communication between Taiwanese and Japanese Undergraduate Students”. Pragmatics 30 (4): 509–531. , 511). In general, the influence of English is strong in both Sweden, where Swedish is the majority language, and Finland, where Swedish is the first language of about 5% of the population and is widely studied as a second language (Saarela 2021Saarela, Jan 2021 Finlandssvenskarna 2021 – en statistisk rapport [Finland Swedes 2021 – A statistical report]. Helsingfors: Svenska Finlands Folkting.; Skjold Frøshaug and Stende 2021Skjold Frøshaug, Andrea, and Truls Stende 2021 Does the Nordic Language Community Exist? Analysis no. 01/2021. Copenhagen: Nordic Council of Ministers. , 13). In Sweden, concerns have been raised regarding the increasing use of jag tänker as an opinion-giving phrase, at least partially due to the English use of I think (Kolu 2022Kolu, Jaana 2022 “Vad tänker du om det? Verbet tänka som epistemisk markör i tidningsspråk [What do you think about it? The verb tänka as an epistemic marker in newspapers].” In Svenskan i Finland 19, ed. by Siv Björklund, Bodil Haagensen, Marianne Nordman, and Anders Westerlund, 149–162. Vasa: Svensk-Österbottniska Samfundet.). In Finland, research shows that English has a clear impact on the Swedish of Finnish-speaking Finns, who study both English and Swedish at school (see, e.g., Lahtinen 2010Lahtinen, Sinikka 2010 “Min kompis plejar drums. Engelskans inflytande på finska högstadieelevers svenska [My friend plays the drums. English influence on Finnish upper elementary school pupils’ Swedish].” In Svenskans beskrivning 30, ed. by Cecilia Falk, Andreas Nord, and Rune Palm, 177–186. Stockholm: Stockholms universitet.; Toropainen and Lahtinen 2013Toropainen, Outi, and Sinikka Lahtinen 2013 “Argumentation på L2-svenska: inlärare skriver insändare [Argumentation in L2-Swedish: learners write readers’ letters].” In Svenskans beskrivning 33, ed. by Jan Lindström, Sofie Henricson, Anne Huhtala, Pirjo Kukkonen, Hanna Lehti-Eklund, and Camilla Lindholm. Helsingfors: Helsingfors Universitet.). For instance, Finnish-speaking elementary school pupils commonly use tänka instead of tycka, most likely due to its phonological similarity to the English think (Lahtinen 2010Lahtinen, Sinikka 2010 “Min kompis plejar drums. Engelskans inflytande på finska högstadieelevers svenska [My friend plays the drums. English influence on Finnish upper elementary school pupils’ Swedish].” In Svenskans beskrivning 30, ed. by Cecilia Falk, Andreas Nord, and Rune Palm, 177–186. Stockholm: Stockholms universitet., 182). In argumentative texts, tycka is overall the most common verb for taking a stance, but the use of tänka is especially common among pupils at lower proficiency levels (Lahtinen 2010Lahtinen, Sinikka 2010 “Min kompis plejar drums. Engelskans inflytande på finska högstadieelevers svenska [My friend plays the drums. English influence on Finnish upper elementary school pupils’ Swedish].” In Svenskans beskrivning 30, ed. by Cecilia Falk, Andreas Nord, and Rune Palm, 177–186. Stockholm: Stockholms universitet., 182; Toropainen and Lahtinen 2013Toropainen, Outi, and Sinikka Lahtinen 2013 “Argumentation på L2-svenska: inlärare skriver insändare [Argumentation in L2-Swedish: learners write readers’ letters].” In Svenskans beskrivning 33, ed. by Jan Lindström, Sofie Henricson, Anne Huhtala, Pirjo Kukkonen, Hanna Lehti-Eklund, and Camilla Lindholm. Helsingfors: Helsingfors Universitet., 485–486). In sum, it is highly relevant to both examine the specific function(s) of jag tänker in talk-in-interaction and compare how L1 and L2 speakers use tänka, tycka and tro in the 1st person singular in their online production of talk.

4.Comparing L1 and L2 interaction

Both Interactional Linguistics and Conversation Analysis see the structure of conversation as orderly and regulated by a number of systems, such as turn-taking organization and preference organization (see Sacks, Schegloff and Jefferson 1974Sacks, Harvey, Emanuel Schegloff, and Gail Jefferson 1974 “A Simplest Systematics for the Organization of Turn-Taking for Conversation.” Language 50 (4): 696–735. ; Sidnell and Stivers 2013Sidnell, Jack, and Tanya Stivers 2013 The Handbook of Conversation Analysis. Chichester: Wiley-Blackwell.). A speaker’s turn at talk may include one or several turn-constructional units (TCUs), which consist of words, phrases, clauses, or sentences, and are recognizable as possibly complete (Sacks, Schegloff and Jefferson 1974Sacks, Harvey, Emanuel Schegloff, and Gail Jefferson 1974 “A Simplest Systematics for the Organization of Turn-Taking for Conversation.” Language 50 (4): 696–735. , 702; Schegloff 2007Schegloff, Emanuel 2007 Sequence Organization in Interaction. Cambridge: Cambridge University Press. , 3–7). Turns that are connected to each other form larger sequences, which enable speakers to perform courses of action in an orderly manner (Schegloff 2007Schegloff, Emanuel 2007 Sequence Organization in Interaction. Cambridge: Cambridge University Press. , 2). When speakers design and produce their turns, they take into account not only the preceding conversation, but also how they expect their co-participants to treat their upcoming talk, based on what they know about them (Drew 2013Drew, Paul 2013 “Turn Design.” In The Handbook of Conversation Analysis, ed. by Jack Sidnell, and Tanya Stivers, 131–149. Chichester: Wiley-Blackwell., 131).

Research indicates that the basic mechanisms of social interaction are the same in L1 and L2 conversations (Pekarek Doehler and Pochon-Berger 2011Pekarek Doehler, Simona, and Evelyne Pochon-Berger 2011 “Developing ‘Methods’ for Interaction: A Cross-Sectional Study of Disagreement Sequences in French L2.” In L2 Interactional Competence and Development, ed. by Joan Kelly Hall, John Hellerman, and Simona Pekarek Doehler, 206–243. Bristol: Multilingual Matters. , 206; Skogmyr Marian 2020Skogmyr Marian, Klara 2020 The Development of Interactional Competence in a Second Language. A Multimodal Analysis of Complaining in French Interactions. Neuchâtel: Université de Neuchâtel., 25). Drawing on Conversation Analytic studies of L2 interaction (see, e.g., Kasper and Wagner 2011Kasper, Gabriele, and Johannes Wagner 2011 “A Conversation-Analytic Approach to Second Language Acquisition.” In Alternative Approaches to Second Language Acquisition, ed. by Dwight Atkinson, 117–142. London: Routledge.; Skogmyr Marian 2020Skogmyr Marian, Klara 2020 The Development of Interactional Competence in a Second Language. A Multimodal Analysis of Complaining in French Interactions. Neuchâtel: Université de Neuchâtel., 54), I regard L2 speakers as active and legitimate participants in these interactions, even though their L2 proficiency varies. In addition, the current study concerns L2 interaction “in the wild”, i.e., between peers or outside educational situations without the presence of ‘experts’ such as teachers or L1 speakers who could potentially affect the interaction (see Wagner 2015Wagner, Johannes 2015 “Designing for Language Learning in the Wild: Creating Social Infrastructures for Second Language Learning.” In Usage-Based Perspectives on Second Language Learning, ed. by Teresa Cadierno, and Søren Wind Eskildsen, 75–104. Berlin: De Gruyter Mouton. ). To engage in social interaction, speakers draw on their interactional competence, i.e., the socially grounded “set of routinized yet context-sensitive procedures with which we reason our way through the moment-to-moment unfoldings of our interactions” (Hall and Pekarek Doehler 2011Hall, Joan Kelly, and Simona Pekarek Doehler 2001 “L2 Interactional Competence and Development.” In L2 Interactional Competence and Development, ed. by Joan Kelly Hall, John Hellerman, and Simona Pekarek Doehler, 1–15. Bristol: Multilingual Matters., 2–3). Research on the development of L2 speakers’ interactional competence (see, e.g., Hellermann 2008Hellermann, John 2008 Social Actions for Classroom Language Learning. Clevedon: Multilingual Matters. ; Pekarek Doehler and Pochon-Berger 2011Pekarek Doehler, Simona, and Evelyne Pochon-Berger 2011 “Developing ‘Methods’ for Interaction: A Cross-Sectional Study of Disagreement Sequences in French L2.” In L2 Interactional Competence and Development, ed. by Joan Kelly Hall, John Hellerman, and Simona Pekarek Doehler, 206–243. Bristol: Multilingual Matters. , 2015Pekarek Doehler, Simona, and Evelyne Pochon-Berger 2015 “The Development of L2 Interactional Competence: Evidence from Turn-Taking Organization, Sequence Organization, Repair Organization and Preference Organization.” In Usage-Based Perspectives on Second Language Learning, ed. by Teresa Cadierno, and Søren Wind Eskildsen, 233–268. Berlin: De Gruyter Mouton. ; Skogmyr Marian 2020Skogmyr Marian, Klara 2020 The Development of Interactional Competence in a Second Language. A Multimodal Analysis of Complaining in French Interactions. Neuchâtel: Université de Neuchâtel.) indicates that interactional competence is not merely transferred from L1 to L2; it is also recalibrated, and over time, L2 speakers develop more diversified methods for accomplishing actions and acquire more diversified linguistic resources to do so. This diversification leads to a growing ability to produce talk that is designed to the recipients’ needs, to engage in context-sensitive conduct and to project upcoming actions, i.e., make the speaker’s upcoming actions recognisable to co-participants. Research also suggests that the interactional competence of L2 speakers develops to resemble the competence of L1 speakers. For instance, Pekarek Doehler and Pochon-Berger (2011Pekarek Doehler, Simona, and Evelyne Pochon-Berger 2011 “Developing ‘Methods’ for Interaction: A Cross-Sectional Study of Disagreement Sequences in French L2.” In L2 Interactional Competence and Development, ed. by Joan Kelly Hall, John Hellerman, and Simona Pekarek Doehler, 206–243. Bristol: Multilingual Matters. , 236–237) show that advanced L2 speakers of French accomplish disagreements in similar ways to L1 speakers of French.

In IL, comparisons commonly concern linguistic and interactional phenomena across different languages (see, e.g., Ono and Couper-Kuhlen 2007Ono, Tsuyoshi, and Elizabeth Couper-Kuhlen 2007 “Increments in Cross-Linguistic Perspective: Introductory Remarks.” Pragmatics 17:4: 505–512.; Laury, Etelämäki and Couper-Kuhlen 2014Laury, Ritva, Marja Etelämäki, and Elizabeth Couper-Kuhlen 2014 “Introduction.” Pragmatics 24:3: 435–452.; Lindström, Maschler and Pekarek Doehler 2016Lindström, Jan, Yael Maschler, and Simona Pekarek Doehler 2016 “A Cross-Linguistic Perspective on Grammar and Negative Epistemics in Talk-in-Interaction.” Journal of Pragmatics 106: 72–79. ). Similarly, comparing L1 and L2 interactions in the same language and in comparable contexts illustrates similarities and differences between the two (see, e.g., Lehti-Eklund 2006Lehti-Eklund, Hanna 2006 “Att planera och reparera. Skillnader mellan talare av svenska som förstaspråk och andraspråk [To plan and repair. Differences between speakers of Swedish as first and second language].” In Lek och lärt. Vänskrift till Jan Einarsson, ed. by Sofia Ask, Gunilla Byrman, Solveig Hammarbäck, Maria Lindgren, and Per Stille, 120–131. Växjö: Växjö University Press. for Swedish). For instance, Rönnqvist and Lindström (2021)Rönnqvist, Sara, and Jan Lindström 2021 “Turn Continuations and Gesture: “And Then”-Prefacing in Multi-Party Conversations.” Frontiers in Communication 6: 670173. , who study turn continuations with å sen ‘and then’ in Swedish group discussions, conclude that L2 speakers use the interactional practice in question in a broadly similar manner to L1 speakers. However, their deployment of grammar and gestures differs to some extent: for example, L2 speakers tend to use more depicting gestures than L1 speakers, and they may depart from the canonical word-order of Swedish.

5.Data and method

The empirical data (see Table 2) consist of eight experimentally set up, task-based discussions, during which groups of three to five university students33.Students of art were not included; the students were “lay” participants in the discussions. discuss eight laminated pictures of visual art (paintings, installations, or photographs). The discussions were recorded, transcribed and anonymised at the University of Helsinki in Finland in 2018–2019. Participation in the discussions was voluntary and the informants were chosen based on their own, reported language proficiency. Afterwards, the L2 speakers were estimated to be at CEFR level C1/C2 in Swedish.

The discussions were unmoderated, but the participants were provided with a list of questions, which invited them to describe, discuss, and evaluate the works of art. Although all the participants draw heavily on the pictures of art as material resources, the L2 speakers commonly use additional resources, such as other languages (Finnish, German etc.) and depicting (iconic) gestures, to elicit help with the interaction, while the L1 speakers do not do this as much (Rönnqvist and Lindström 2021Rönnqvist, Sara, and Jan Lindström 2021 “Turn Continuations and Gesture: “And Then”-Prefacing in Multi-Party Conversations.” Frontiers in Communication 6: 670173. ).

Swedish is a pluricentric language with two national varieties that share a single written norm; most differences between the varieties concern spoken language (Norrby et al. 2020Norrby, Catrin, Jan Lindström, Jenny Nilsson, and Camilla Wide 2020 “Pluricentric Languages.” In Handbook of Pragmatics 23, ed. by Jef Verschueren, and Jan-Ola Östman, 201–220. Amsterdam: John Benjamins Publishing Company. ). Sweden Swedish, which is the dominant variety, has a clear impact on Finland Swedish, which is spoken by a clear minority in Finland. Eight of the twelve L1 speakers originate from Finland, while only two originate from Sweden. Two informants have multilingual backgrounds, but studied in Swedish at the university and had skills comparable to L1 speakers. Consequently, the L1 results are mostly based on the Finland Swedish variety of Swedish. The L2 results are mostly based on L1 speakers of Finnish, as fifteen of the sixteen L2 informants had Finnish,44.In Finnish, epistemic stance is expressed with adpositional phrases (e.g., minun mielestäni ‘in my opinion’; see Viberg 2005Viberg, Åke 2005 “The Lexical Typological Profile of Swedish Mental Verbs.” Languages in Contrast 5 (1): 121–157. , 147) as well as verbs with 1st person subjects (see Helasvuo 2014Helasvuo, Marja-Liisa 2014 “Agreement or Crystallization: Patterns of 1st and 2nd Person Subjects and Verbs of Cognition in Finnish Conversational Interaction.” Journal of Pragmatics 63: 63–78. ). a Finno-Ugric language, as their L1.

The programme AntConc was used to identify instances of the phrases in the transcripts of the discussions. The phrases were then located in the video recordings of the data and analysed grammatically and sequentially. As Swedish is a V2 language, the searches covered both the subject–verb word orders jag tänker/tycker/tror and the verb–subject word orders tänker/tycker/tror jag, with their various spellings. The quantitative results include both literal and stance-taking uses of the phrases, but instances with a different meaning (e.g., tänka used as an auxiliary verb, the phrasal verb tycka om ‘to like something’) were removed from the collections. The following sections present the empirical findings.

6.Findings

01 SAB:  sen   e   den ju   no
         then  is  it  prt    prt
         ‘then it is’
02       ja tycker   denhä  e-  e       vacker
         I  tycka.prs  dem        be.prs   beautiful
         ‘I think this is beautiful’
03       eller liksom  dehär  lj [use   e      (mera)]
         or    like    dem      light.def  be.prs  (more)
         ‘or  like this light is more’
04  CEL:                         [jå  ja  (me)       ]
                                 ‘yes me too’
05       [dedä  lju]se   e            (  )
         dem      light.def be.prs
         ‘that light is’  
06  MAG: [jå       ]
         ‘yes’ 
07       (0.5) 
08  SAB: fint=
         ‘nice’
09  CEL: =de  e      bra    såhä  (stäm–)  .h
         it   be.prs good   dem     (athmos-) 
         ‘it is good like (athmos–)’
10       (1.1) 
11  SAB: å   no   e      de  ju:   
         and prt  be.prs it  prt 
         ‘and it really is’ 
12   →   ja tänker    sånhäna liksom
         I  tänka.prs    dem      like
         ‘I think these kinds of like’ 
13       (0.5)  
14       pt mor    å   barn   grejer     e      ganska  tidlösa=
            mother and child  thing.pl   be.prs  quite    timeless.pl
         ‘mother and child things are quite timeless’
15       =på ett       sätt  att int<
         on  indf.art  way  comp neg
         ‘in a way that not’
16       (0.6)  
17       om    man s-  titta     på:   på   liksom
         cond  one     look.prs prep  prep  like
         ‘if one looks at at like’
18       (0.5)  
19       denhär .h hur   dehär    e:      uppställt   denhär    bilden
         dem        how dem     be.prs  set.up.pctp  dem      picture.def
         ‘this how this is set up this picture’
20       (.)
21       så e        de ju    endast
         so be.prs  it prt   only
         ‘so it is only’
22       (.)
23       eh  liksom  me   kläder       
             like    with clothes.pl 
         ‘like with clothes’ 
24       å   så där   ser      man ju    att
         and so there see.prs  one prt   comp
         ‘and there one sees that’
25       dehär   e      int<  liksom  nytt
         dem    be.prs neg    like    new
         ‘this isn’t like new’ 
26       men .h  om     man sku    göra    me     samma,
         but     cond  one would  do.inf  with same
         ‘but if one would take with the same’ 
27       (0.4)
28       på  nå    sätt  set up
         in  some  way   set up ((English))
         ‘in some way set up’
29       (.) 
30       ehm:
31       (.) 
32       ett       foto   me    som    helt     vanliga     kläder      ida, 
         indf.art  photo with  like  totally  ordinary.pl  clothes.pl  today
         ‘a photo with like totally ordinary clothes today’
33       (0.5) 
34       så int   e      de ju  (.) int
         so neg  be.prs it prt      neg
         ‘so it is not’ 
35       här  finns   int  liksom  dendä   tidsaspekten
         here exists  neg    like     dem     time.aspect.def
         ‘the time aspect is not present here’ 
36       på de       sät [te kanske   s]å  [mycke hell]er
         on def.art  way.def   maybe      so       much  either 
         ‘in that way maybe so much either’ 
37  MAG:                [m:m          ] 
38  CEL:                                  [m:m       ] 
39       (4.3) 

01   SAB:  .h mm  nå    de  e      ju    ganska överraskande eller som
                  prt  it  be.prs prt  quite  surprising      or    prt
                 ‘well it  is  quite  surprising  or  like’
02         ja:  >inte vet     ja<  men fö   mej åtminstone
           yes   neg know.prs   I     but for  me  at least
           ‘yes  I don’t know but for me at least’ 
03         att       att:       tänka     att
           inf.mark  inf.mark   tänka.inf comp
           ‘to to think that’
04         de e       en         mörkhyad,
           it be.prs indf.art  darkskinned
           ‘it is a  person of colour’
05         [(0.3)]  som  e        avbildad
                    that be.prs  picture.pctp
                   ‘that is pictured’ 
06   CEL:  [mm,  ] 
07   SAB:  att   de e        ju     ändå      väldigt,=
           comp   it be.prs   prt    after all  very
           ‘that it is after all very’
08   SAB:  =.h västerländsk konst  e       ju    nog,
               western      art    be.prs prt    prt
              ‘Western art is’
09        (0.6) 
10        såhä   vit
          ‘like  white’
11        (0.6) 
          centr [erad  å   ga ]nska normativ  å   så=
          ‘centered    and quite    normative and so’
12   CEL:       [pt  jå,      ]   
                    ‘yes’
13   SAB: =att   de  e       redan    överraskande.=
          comp   it  be.prs  already  surprising
          ‘that that is already surprising’
14   MAG:  =°mm°  [.h ] 
15   CEL:         [°  ] å   ännu  fö  artonhundratal    liksom  [(   )]°
                 ‘and even for the eighteen hundreds like’
16   SAB:                                                       [m:m. ] 
17        (.)
18 → MAG: men ja tänker     just   att
          but I  tänka.prs   prt    comp
          ‘but I was just thinking that’
19        #eh:# är       de  nån     från Frankrike  så ä [r     de,]
                be.prs  it  someone from France     so be.prs   it 
          ‘if it is someone from France then it is’ 
20   SAB:                                                 [m:m:.    ] 
21        (.)
22   MAG: kolonialmakt    att   de
          colonial.power  comp   it
          ‘a colonial power so it’
23        [[kan   ju   va]     liksom     dehä ]ra (.)  [[exotiseringen ]] av,
            can   prt   be.inf  like    dem              exotization.def     of
           ‘can be like this exotization of’
24   CEL: [mm,           ]                           [°ja just de.°   ]        
                                                     ‘yes just that’
25   SAB:  [m̲m:                            ]         [°sant de°,      ]
                                                     ‘true that’ 
26        (1.2) 
27   MAG: av (de)  liksom att       visa upp      i  Europa  då.=
          of  it   like   inf.mark   display.inf in Europe  then
          ‘of that like to display in Europe then’
28   SAB: =[jå anta  ]gligen,=
          ‘yes probably’
29   CEL: =[mm,      ] 
30   MAG: =ja,
          ‘yes’ 
31        (1.5) 

01   HEN:  [ja  mm                   ] 
02 → ONN:  [ja tänker    på   sådan  ]a saker
            I  tänka.prs    prep  dem         thing.pl
            ‘I think about  such things’
03          [(0.3) al-] 
04   HEN:   [jå:      ] 
            ‘yes’
05   ONN:   ja  liksom
            ‘yes like’
06          (0.6) 
07   HEN:   hela  tiden: [((laughs)) ∙nff             ] 
            whole time.def
            ‘all of the time’
08   ONN:                [hela  tiden    ut [an   att ] ]=
                          whole time.def    without  comp  
                         ‘all of the time without’               
09   EMM:                                   [mm         ] 
10   ONN:   =[ja] blir        påmint=
              I   become.prs  remind.pctp
             ‘being reminded’
11   HEN:    [mm]          =[ja  ]
                            ‘yes’
12   ONN:                   [(ja)]=
                             ‘yes’
13   EMM:   =mm (0.4) .h ja:
                        ‘yes’ 
14          (1.9)

01        (0.9) 
02  RII:  ä:h  denhä konsten   e        just  så
               dem    art.def   be.prs prt    so
              ‘this work of art is just so’ 
03        att   ja föstår-           ja   fattar   inte
          comp   I  understand.prs   I  get.prs  neg
          ‘that I don’t understand I don’t get it’
04 →      å   ja tror    att 
          and I  tro.prs  comp
          ‘and I think that’
05        okej om någon .h eh  säger    att
          okay if someone      say.prs  comp
          ‘okay if someone says that’
06        dehär  e      konst 
          dem     be.prs   art
          ‘this is art’
07        så kunde  ja också  vara (.)  en      konstnär
          so could  I  also   be.inf    indf.art   artist 
          ‘I could also be an artist’ 
08        (0.5) 
09  LAU:  [mm           ] 
10  JAA:  [[((chuckles))]] 
11  RII:   [eftersom:    ] ja kunde  också
          ‘because I could also’ 
12        ri[ta    nåt-   nånting   sånt ]
          draw.inf som-  something dem 
          ‘draw som- something like that’
13  LAU:    [mm (.) m:m                ] 
14        (1.0) 

15   LAU:   mt .h men  ja tycker    att
                  but  I  tycka.prs    comp
                 ‘but I think that’
16          denhär  är      ganska (0.4) cool 
            dem       be.prs  quite         cool ((English))
            ‘this is pretty cool’
17          (0.3) 
18   RII:   [.h (0.3) mm   ] 
19   LAU:   [på  något sätt] 
            ‘in some way’
20          (0.8) 
21 → RII:    .h nå   ja tror    att
                prt  I  tro.prs  comp
                ‘well I think that’ 
22          om    man (0.5) eh (0.4)  gör
            cond  one                do.prs
            ‘if one does’
23          nånting (.)  såhär  [.h  så  ]  
            something    dem           so
            ‘something like this’
24   LAU:                       [mm:     ] 
25   RII:   måste  den verkligen ha
            must   it  really    have.inf
            ‘it must really have’ 
26          (0.7) 
27          en        mening
            indf.art  meaning 
            ‘a meaning’
28          (0.5) 
29   JAA:   [j:å:¿ ]
            ‘yes’ 
30   RII:   [och   ] 
            ‘and’
31          (.)
32          en        djupare   mening
            indf.art  deeper   meaning
            ‘a deeper meaning’
33          (0.5) 
34   LAU:   men vi [vet]   inte  va    är    [de]
            but we  know   neg      what  be.prs   it ((non-canonical word order))
            ‘but we don’t know what it is’ 
35   RII:          [fö-]                      [mm]       
36          så (0.3)  därför    e      de konst
            so        therefore be.prs   it art
            ‘so that’s why it is art’ 
37   L:     mm (.) mm 
38          (0.5) 

01        (0.9) 
02  LAU:  ja skulle o̲ckså ta: (.)   dihä:r
          I  would  also  take.inf    dem
          ‘I would also take these’
03        å   sen (1.7) kanske  den hä:r
          and then      maybe   dem
          ‘and then maybe this one’ 
04        (0.5) 
05  JAA:  jå 
          ‘yes’
06        (.)
07  LAU:  men h 
          ‘but’
08        (2.0) 
09  →     °ja   vet  inte° .h  men ja tycker   att
           I    know neg         but  I  tycka.prs   comp
          ‘I don’t know but I think that’
10        n: om   vi hade  varit  (0.4) i  denhär
             cond  we had   be.pctp         in  dem  
            ‘if we had been at this’ 
11        (2.0) 
12  RII:  ja   där    [på  plats                  ] 
          yes  there   on  place
          ‘yes there on the spot’
13  LAU:              [⁎näyttely⁎               va ] [är      de ]
                       exhibition ((Finnish))  what  be.prs  it
                      ‘exhibition  what is it’
14  JAA:                                             [jå        ] 
                                                    ‘yes’ 
15        (.)
16  LAU:   >föreställ[ningen       ]<=
           exhibition.def
          ‘the exhibition’     
17  JAA:             [°föreställ-° ]
                     exhibit-                               
18  RII:  =[jo ] 
          ‘yes’
19  LAU:   [.h ] så: vi  skulle   minnas      den
                 so  we  would      remember.inf it
                 ‘we would remember it’
20        (0.6)
21        liksom  fö   resten    av
          like    for rest.def  of
          ‘like for the rest’  
22        ( 1.0) 
23        [vår ] liv=
          ‘our   life’
24        (.)
25  RII:  [ .h ]   
26  LAU:  [=>tycker ja< ]   för att   de är       så
          tycka.prs    I     because  it be.prs so
          ‘I think because it is so’
27  RII:  [m:m          ]
28        (.)
29  RII:  de kan ju  vara=
          it can prt  be.inf
          ‘it may be’
30  LAU:  =hemsk (0.3) å  (0.9) ovanlig också
          horrible     and      unusual also
          ‘horrible and unusual too’
31        (0.7) 
32  RII:  mm.
33        (0.9) 

7.Discussion and concluding remarks

In conclusion, the L1 speakers of Swedish discussing works of art deploy the verbs tänka, tycka and tro in the 1st person singular and the present tense to project different types of stance-taking. A stance put forth with jag tycker is a subjective opinion and does not call for justification, whereas jag tror expresses a lack of actual knowledge of some state of affairs that is intersubjectively verifiable. In contrast, jag tänker describes the speaker’s actual thoughts and may do the work of stance-taking in a more covert manner, for instance, when there is a contrast with previous talk. Jag tänker has the potential to project more talk than a single complement clause, i.e., a multi-unit turn that may consist of conditional clauses and conclusions. When speakers project more talk about a topic with jag tänker, they account for the relevance of their ‘thought’ in the unfolding interaction, i.e., make their thought process or line of reasoning available to their interlocutors and, simultaneously, portray themselves as rational, accountable beings (see Fiedler 2023Fiedler, Sophia 2023 “The Grammar-in-Use of Direct Reported Thought in French and German. An Interactional and Multimodal Analysis.” PhD dissertation, University of Neuchâtel/University of Hamburg., 83). This also mitigates imposing on the co-participants and reduces their obligation to (dis)agree with the actual point, only the reasoning behind it. Thus, the way in which the speakers frame their claims and interpretations of art as ‘thoughts’ with jag tänker serves the collaborative nature of the group discussions. Notably, Fiedler (2023Fiedler, Sophia 2023 “The Grammar-in-Use of Direct Reported Thought in French and German. An Interactional and Multimodal Analysis.” PhD dissertation, University of Neuchâtel/University of Hamburg., 370–371) shows that similar actions, i.e. expressing emergent reasonings that seek out understandings, can also be accomplished by the German verb denken (‘to think’, especially in its past forms) and the French verb se dire (‘to say to oneself’) in everyday conversation. Thus, it could be fruitful to pursue systematic cross-linguistic investigations of the verb tänka and its counterparts in talk-in-interaction.

In terms of the certainty of the expressed stance, Kolu (2022Kolu, Jaana 2022 “Vad tänker du om det? Verbet tänka som epistemisk markör i tidningsspråk [What do you think about it? The verb tänka as an epistemic marker in newspapers].” In Svenskan i Finland 19, ed. by Siv Björklund, Bodil Haagensen, Marianne Nordman, and Anders Westerlund, 149–162. Vasa: Svensk-Österbottniska Samfundet., 160) places jag tänker between jag tycker, which implies a high degree of certainty, and jag tror, which expresses a lack of knowledge. In the investigated data, however, I argue that jag tänker does not mainly concern the certainty of a proposition. Instead, it constitutes an interactional device to project more talk without clearly specifying the speaker’s stance towards it, for example, when there is a contrast with a previous speaker’s explicit stance. As a result, talk projected with jag tänker does not occasion stance-taking from the co-participants in the way that an opinion projected with jag tycker would. In conclusion, jag tänker clearly has its own functional niche in the L1 data, where it cannot be replaced with either jag tycker or jag tror. In this regard, the results of this study do not support the notion that jag tänker is used as a purely opinion-giving phrase in Swedish talk-in-interaction. However, the L1 results are mainly based on Finland-Swedish speakers. As previous research suggests that the verb tänka is more common in Sweden-Swedish conversations (see Nelson and Henricson 2019Nelson, Marie, and Sofie Henricson 2019 “Kognitionsverb i sverigesvenska och finlandssvenska handledningssamtal [Cognition verbs in Sweden-Swedish and Finland-Swedish supervision meetings].” Puhe ja kieli 39:1: 45–68. ), it would be highly relevant to investigate the interactional workings of this verb in more varied conversational data on L1 speakers from Sweden.

In the L2 data, jag tänker is rare and only used as a lexical resource for concrete ‘thinking’, not as a pragmatic, stance-framing device. However, L2 speakers accomplish stance-taking actions that express the speaker’s reasoning using other means, such as jag tycker (att) and jag tror (att), which may be deployed as fixed expressions, at the same time blurring the normative distinctions between the verbs. While the specific stance-taking function of jag tänker is not fully recognized and thus not explicitly taught in Swedish lessons in Finland, the meanings of tänka, tycka and tro are often contrasted with each other in textbooks (see Kolu 2022Kolu, Jaana 2022 “Vad tänker du om det? Verbet tänka som epistemisk markör i tidningsspråk [What do you think about it? The verb tänka as an epistemic marker in newspapers].” In Svenskan i Finland 19, ed. by Siv Björklund, Bodil Haagensen, Marianne Nordman, and Anders Westerlund, 149–162. Vasa: Svensk-Österbottniska Samfundet., 149). This can lead to L2 speakers regarding these verbs as similar. As the qualitative results show, an L2 speaker may use one of the three phrases as a general stance-taking device, instead of diversifying between the different phrases. Nevertheless, the L2 speakers can format turns that account for their reasoning in a relatively similar manner to the L1 speakers, as regards the initial positioning of the stance-taking phrase and the deployment of conditionals and conclusions. In other words, they recognize that such phrases have a function in the organization of the interaction, i.e. the potential to project a longer stretch of talk.

To conclude, it is extremely useful to compare how L1 and L2 speakers of a language interact in a similar context. Firstly, the investigation of how an interactional practice is deployed by L1 and L2 speakers sheds light on the structure of the language in question and its interactional practices in more general terms (see, e.g., Rönnqvist and Lindström 2021Rönnqvist, Sara, and Jan Lindström 2021 “Turn Continuations and Gesture: “And Then”-Prefacing in Multi-Party Conversations.” Frontiers in Communication 6: 670173. ). In this case, the resources for epistemic stance-taking in Swedish include the verbs tänka, tycka and tro, which have their own, diversified uses in with 1st person subjects in interaction. Secondly, comparisons between L1 and L2 usage are highly beneficial for educational and teaching contexts. For instance, the function of jag tänker ‘I think/cogitate’ and the overall division of labour between tänka, tycka and tro could be further discussed in advanced L2 studies of Swedish. Finally, comparisons between L1 and L2 interactions increase our understanding of how L2 speakers accomplish actions and take on interactional projects as active, legitimate participants in their interactions, even though the linguistic resources they put to use might differ from those deployed by L1 speakers.