Learning on the Run… or Finding Their Own Path? A Multimodal and Longitudinal Study of Motion Events in L2 Dutch

The typological differences between verb-framed (V-framed) and satellite-framed (S-framed) languages identified by Talmy (2000) have also been shown to be reflected in co-speech gestures (e.g., Kita & Özyürek 2003; McNeill & Duncan 2000; Stam 2006, 2010, 2018; Ünal, Mamus & Özyürek 2023). The Interface Hypothesis predicts that the semantic components encoded gesturally (e.g., manner vs. path) will mirror the way information is packaged in speech (Kita & Özyürek 2003). Other studies have found that path gestures align with different linguistic units in V-framed and S-framed languages (e.g., Stam 2006, 2010, 2018). Co-speech gestures should therefore be considered when examining speakers’ thinking-for-speaking patterns (e.g., Stam 2018).
Against this background, the present study investigates how motion events are expressed in speech and co-speech gestures by L1 French speakers, L1 Dutch speakers, and CLIL French-speaking learners of Dutch. The study also includes a longitudinal component: the L2 learners repeated the experiment two years later, allowing us to explore whether their thinking-for-speaking patterns changed as their proficiency increased.
An elicitation task was conducted in which 15 L1 French speakers, 14 L1 Dutch speakers, and 12 L2 Dutch speakers recounted scenes from Tweet Zoo (1957). Using established taxonomies (e.g., Cadierno & Ruiz 2006; Kopecka 2006; Özçalışkan & Slobin 1999), we identified the semantic components (manner and path) encoded in verbs and satellites. Gestures were categorized as iconic, beat, metaphoric, deictic, or pragmatic (Kendon 2004; McNeill 1992). Iconic and deictic gestures were further analyzed for the motion components they conveyed (e.g., manner, path, ground) and for viewpoint (character vs. observer; McNeill 1992). Finally, we examined combinations of semantic components across speech and gesture, and we analyzed gesture–speech synchronization following Stam (2006).
The results of this study show that even though Dutch-speaking participants focus on manner in their oral descriptions, they tend to produce path gestures only when describing self-propelled movements. These findings contradict the Interface Hypothesis, which predicted conflated gestures (manner and path together). French-speaking participants tend to focus on path both in speech and in gestures. The gestural differences between the two languages lie in the alignment between gestures and speech: French speakers tend to produce path gestures simultaneously with the verb, whereas Dutch speakers do so at the same time as the satellite. As for the learners, they seem to be aware of the "manner" character of Dutch, but it is difficult for them to express movements like L1 speakers: in fact, complex verbs are very rare in the data, and they do not tend to accumulate semantic components in their speech. At the gestural level, learners have their own behavior: they tend to produce more non-referential gestures (pragmatic and beats) and more referential gestures (here iconic and deictic) per utterance. Learners also produce more manner gestures than L1 speakers. We can thus speak of an intergesture just as learners have an interlanguage (Selinker 1972).