Research project The grammar-body interface in second language interaction: a developmental perspective

This project brings together two research teams, one located at the University of Neuchâtel, Switzerland, and the other at Stockholm University, Sweden, to investigate how grammar and the body interface in second language (L2) development. Firmly grounded in multimodal CA but also drawing on insights from usage-based linguistics, we investigate how precise linguistic structures are bundled together with precise bodily conduct in L2 interaction, and how these bundles evolve over time as speakers gain increased L2 proficiency. We examine whether the use of particular bodily conduct precedes the development of certain linguistic constructions in L2 learning, and/or scaffolds such development. We also scrutinize to what extent L2 speakers develop grammar-body packages similar to what is known about L1 speakers. Taken together, these issues allow us to shed light on possible interactional motivations for the development of grammar and to enrich our understanding of multimodal (L2) interactional competence. The project is funded through the so-called International Co-Investigator Scheme (Swiss National Science Foundation & Swedish Research Council Formas) for the period 2026-2030.

Here is a more detailed description of the motivations, aims, methods and anticipated contributions of the project:

The fact that face-to-face communication is fundamentally multimodal in nature has for long been established in a variety of research paradigms: We humans use verbal-vocal means in concert with bodily-visual resources (gaze, posture, facial expressions, gestures, manipulation of objects) for both social and cognitive purposes, such as constructing and interpreting meaning, accomplishing and coordinating actions, and enhancing memorization and learning. Work on child language acquisition, for instance, has shown that vocal and bodily means precede the development of verbal communication; these represent precursors laying the (functional) grounds for the development of linguistic communication.

Surprisingly, however, the complex ways in which grammar and the body interface in L2 development have largely remained unexplored in research. Although many studies show that L2 learning is also a bodily process, these typically focus on observable gesture-language constellations ‘in the moment’; a longitudinal, developmental perspective is missing. Furthermore, given the predominant use of (semi-)experimental research designs and focus on manual gestures and their relation to lexicon and semantics in multimodal L2 acquisition research, we lack investigations on the body-grammar interface in L2 development as observable in people’s real-life interactions.
Bringing together insights from multimodal conversation analysis and usage-based linguistics, we seek to respond to these gaps by addressing four focal research questions:

• RQ1: Does the use of particular bodily conduct precede the development of certain linguistic constructions, and if so, how? That is: does bodily conduct fulfill functions / accomplish actions that later are accomplished by language? This would be a trajectory from gesture as a ‘prosthesis’ (possibly a functional template) for the later development of grammar.
• RQ2: Does bodily conduct ‘scaffold’ precise grammatical constructions, and if so, does such conduct recede along the developmental trajectory as the constructions become routinized? This would be a trajectory from gesture as a ‘crutch’—supporting (the development of) grammar—to grammar.
• RQ3: Do L2 speakers develop grammar-body assemblies similar to what we know about L1 speakers’ use of routinized ‘multimodal packages’? This would be a trajectory providing evidence for the functional role of bodily conduct as part of the development of a multimodal L2 interactional competence.
• RQ4: Are results related to (1) through (3) bound to a specific language, or can we observe convergences across two distinct (though typologically and culturally related) L2s?

In tackling these questions, we focus on two construction types: (a) complement-taking predicate constructions that are known to grammaticalize as interactional markers (e.g., ‘(I) dunno’, ‘you see’) and (b) ditransitive constructions (i.e., constructions with three arguments: ‘I gave Mary the book’). These two lines of investigation complete each other by allowing us to address a core facet of L2 development: diversification, over time, of action contexts and interactional uses of linguistic constructions, which cumulatively enable participants to deploy increased context-sensitive and recipient-designed conduct.

We address these issues based on two longitudinal corpora of video-recorded conversations among L2 speakers of French and of Swedish, respectively, and their fine-grained multimodal CA transcriptions. The data allow for (a) longitudinal tracking of individual speakers across up to 18 months and (b) pseudo-longitudinal comparison between speakers/proficiency levels. One of our methodological innovations pertains to the combination of qualitative multimodal CA and longitudinal CA with automatic extraction methods, the latter allowing us to rapidly detect possible candidate constructions, even in schematic forms (e.g. I Verb + Object + Object, for the ditransitive constructions); another originality lies in the use of both “forward-tracking”, following chronologically the emergence of focal phenomena, and “back-tracking”, starting from a given usage pattern in precise sequential contexts and rewinding to where it originates, or at least to its possible precursors. Forward-tracking informs on how body-grammar constellations for social-interactional purposes evolve over time. Back-tracking allows relating precise grammatical structures and the actions they accomplish (or: the sequential/action contexts in which they occur) to their possible bodily-visual precursors. Together, these two lines of investigation enable us to document the emergence of L2 constructions as they are used in the multimodal ecology of social interaction as well as their routinization and functional expansion for socio-interactional purposes.

In sum, the project ambitions to advance our understanding of fundamental yet under-researched questions in L2 acquisition and interaction pertaining to the role of bodily-visual conduct in L2 grammatical development, and ultimately in people’s ability to interact with others in their L2. The findings are anticipated to have implications for theory-building, providing empirical evidence for an interactional understanding of multimodally situated language use as a motor for grammatical development (i.e., how interaction drives grammar). Methodologically, it offers an exemplar of how qualitative CA can be combined with automated corpus-based extraction and how multimodal CA analysis can be brought to bear on authentic, mundane longitudinal interactional data, thereby setting precedents for future research.