Colloquium: What language development in infants tells us about sound systems and change

February 2, 3pm, Social Sciences 122
Chandan Narayan, University of Toronto

The connections between first language acquisition, phonological typology, and sound change have been questioned in the past (e.g., Baudouin de Courtenay, 1895; Grammont, 1933; Jakobson & Waugh, 1979) mainly from the standpoint of the emergence and shape of children’s productive phonology. In this talk I take a different perspective, approaching possible links between acquisition, typology, and sound change as a function of infants’ innate perceptual biases and the acoustic nature of the primary input to infants, infant directed speech (IDS). The first part of my talk present cases in infant speech perception where infants do not perceive a phonetic contrast in a language-general fashion (i.e., their discrimination performance is relatively poor). I argue that (1) the contrasts that infants fail to discriminate are precisely those that are rare in the world’s sound systems, and (2) this weak perception is mediated by the acoustic salience of the contrast. The second part of my talk looks to the developing acoustic nature IDS as a potential “seed” for misperception-based sound change (Hombert & Ohala, 1979). I present new work on obstruent voicing in American English IDS that suggests that the primary acoustic cue, voice-onset time, is more variable than in adult-directed speech, leaving a secondary acoustic feature, post-stop f0, as a potentially more reliable cue. Taken together, infant speech perception and the acoustics of the caregiver-infant interaction conspire in potentially affecting the observed shapes of sound systems and the directions of sound change.