Pronunciation is usually one of the biggest difficulties that second language (L2) learners meet. One of the helpful teaching strategies and resources that previous research has identified, is the use of gestures. However, exactly how and why such embodied training can be helpful is still unclear.

This study investigates the potential effects of embodied pronunciation training on the acquisition of the Swedish complementary length contrast (vila ≠ villa), a well-known difficulty for L2 learners of Swedish: In Swedish, a stressed syllable contains either a long vowel followed by a short or no consonant (V:(C)) or a long consonant following a short vowel (VC:). 

We investigate adult L2 learners’ perception and production of the above-mentioned contrast through a pre-/ post-/ delayed post-test design, that is, before and after receiving pronunciation instruction with or without gestures. The instruction phase consists of a training video on the given contrast in three different conditions (between subjects): no gestures (audiovisual speech only), and two gesture conditions, testing two different sets of gestures. Additionally, a fourth condition with no training (control group) is included.

Learners’ production is investigated through a word-reading and repetition task and is assessed through native-speaker ratings and acoustic analyses. Perception is measured by means of a visual-word eye tracking experiment, which constitutes a novelty in this field of research. Eye tracking data is analyzed using a growth curve analysis, where the time course of target looking will inform us of the learners’ processing abilities. Moreover, participants are asked to fill out a language background questionnaire and perform auxiliary tests, such as a memory task and a music ability test, to control over individual differences. Data collection is scheduled for spring 2025 and we aim to present the first results at the conference.