The effect of L3 training on vowel reduction in L2 and L3

Longitudinal projects investigating Cross-Linguistic Influence (CLI) of phonological acquisition are gaining in popularity. Most of the studies to date have focused on segmental aspects of speech, such as rhotics production (Wrembel et al. 2019), vowel perception (Balas et al. 2019, Lewandowska 2019), or consonant discrimination (Kopeckova et al. 2018, 2019). The present study aims to investigate vowel reduction across the languages from the repertoire of multilingual speakers in order to provide insights into the suprasegmental level of L3 phonological acquisition. The participants of the study were 10 Polish university students, divided into two groups: Group 1 with L2 English and L3 German and Group 2 with L2 English and L3 French. They were enrolled into the first year of intensive courses of L3, starting from the beginner level and not pursuing their L2 training. Production data was collected through a semi-spontaneous task in which they had to make up sentences containing multi-syllabic token words presented in pictures. These words were later extracted from the participants’ recordings and analysed acoustically for the duration and intensity of unaccented vowels. The data was collected at two testing times: after 8 weeks of L3 learning and 6 months later. The languages under investigation were selected so as to provide matching (L2 English and L3 German) or non-matching (L2 English and L3 French) language pairs in terms of exhibiting the feature of vowel reduction. Assuming bi-directionality of CLI, this difference should be reflected in the longitudinal production data. We expect that the intense training in L3 German made speakers more prone to reduce unaccented vowels also in L2 English. Conversely, as the feature is not part of L3 French phonology, we expect no change in the L2 performance across the testing sessions for the L3 French group.