A continuous and categorical approach to the L2 acquisition of Spanish voiced stop allophones

Previous work on the second language (L2) acquisition of Spanish voiced stop weakening by first language (L1) speakers of English has followed either a categorical or a continuous approach to allophonic production. This study combines both types of analyses to explore the phonetic development of intervocalic voiced stops. In addition, utterance initial voiced stops are examined to compare the development of different allophones, more precisely, approximants, voiced stops, and voiceless stops. Twenty-seven participants were recorded performing a reading task at the beginning and at the end of a semester during which they took a college-level Spanish phonetics course. Voiced stops in intervocalic and utterance initial position were analyzed for type of allophone production, and the consonant-vowel intensity ratio was used as a continuous measure of consonant weakening. Results show that learners develop their production of voiced stop weakening both at the segmental level, with an increase in approximants, and at the subsegmental level, with approximants being produced with increasingly less constriction. Moreover, an asymmetry between the development of intervocalic approximants and utterance initial voiced stops is observed, with the latter displaying a slower pace of acquisition. This finding is discussed within theories of L2 phonology that model the role of similarity in the acquisition process.