CLS Speaker Series - Grant Berry (Penn State University) The long and short of it: How short-term alignment and cognitive processing may influence sound change
Apr 22, 2016
from 09:00 AM to 10:30 AM
|Where||127 Moore Building|
|Add event to calendar||
Human beings are adept at processing variation in speech, and a wealth of research attests to individuals’ ability to quickly adapt perception to their input. Another, immediate consequence of exposure to variation may be modifications in the listener’s subsequent production (alignment/accommodation). Over the course of a conversation, interlocutors may align in their production of fine phonetic detail, including speech rate, pitch, spectral properties of vocalic production, and voice-onset time. However, interlocutors using the same language may also differ at the level of their phonological inventories (e.g., pen and pin are homophonous for Kansas City natives like me, but not for most Northeasterners), which affects both perception and production. Accommodation in production at the phonological level remains understudied, but may be essential to understanding how subtle, short-term variation in production is related to language change on a larger scale.
In this talk, I discuss two studies investigating phonological production in discourse. The first, resulting from collaboration with Mirjam Ernestus at Radboud University, investigates phonetic alignment in English as a lingua franca among Spanish-English participants and Dutch-English confederates. We examine dynamic changes to the production of two key phonological contrasts (/i/-/ɪ/ and /ɛ/-/æ/) in English, finding that Spanish participants align to the English of their Dutch interlocutors, which involves a merged /ɛ/-/æ/ category but a distinction of /i/ and /ɪ/, rather than more native-like English (which would require a four-way distinction). These results imply that phonological category production dynamically updates in response to one’s input, even after a single conversation. The second is a pilot study addressing how individual differences in processing variation may correlate to differences in the adoption of variable phonological rules over time. I collected personal narratives from Spanish-English bilinguals who are longstanding residents of Philadelphia, focusing on their production of three context-restricted sound changes-in-progress in that community (Ey Raising, where bait and beat become near homophonous; Canadian Raising, where the vowel in price raises and becomes distinct from the vowel in prize; and AE-tensing, where the vowel in ham raises and tenses and becomes distinct from the vowel in had) with distinct social valuations (non-salient, slightly salient, and socially salient, respectively). I then correlate adoption of these changes-in-progress with individual difference measures (proactive control, reactive control, and the Autism Spectrum Quotient). Notably, the effect of individual difference measures depends on the social value of the variable analyzed. Cognitive processing measures better describe changes-in-progress with low social awareness (Ey Raising, Canadian Raising) than they do salient changes-in-progress (AE-Tensing). This suggests that while the way an individual processes variation in his/her input may have implications for his/her adoption of changes present in the environment, social valuation can suppress these effects. I conclude this talk by outlining a working hypothesis regarding the importance of cognitive control and phonetic alignment in the actuation of sound change.