Abstract
This study investigates the phonetic implementation of Galician /epsilon-e/ and /o-o/ as produced by Galician-Spanish early bilinguals. It examines whether there is variation that can be explained by differences in participants' linguistic histories (as captured by their language dominance scores). Based on production data from one point in participants' lives, the study uses regression to predict phonetic variation from participants' language dominance at that point in life. Results reveal that, although participants produce a robust /epsilon/ vs /e/ contrast, the L1 specific category, /epsilon/ presented a more fronted position, more like /e/, as a function of lower dominance in Galician. However, this effect was not replicated for the back vowels. Given our results, we argue that differences in language dominance may trigger phonetic variation in bilingual speech production in some variables, but not others. This result is consistent with the current cognitive approach in bilingualism research claiming the plasticity of native phonetic domains throughout an individual's lifespan. Results also align with recent claims that language dominance does not equally affect multiple phonological processes for the same individual.