ssfld! (Listen!): Do infant need vowels to spot words?

No Thumbnail Available
Journal Title
Journal ISSN
Volume Title
Crosslinguistic evidence suggests that the Possible Word Constraint (Norris et al., 1997) is followed by both adult and infant listeners to rule out non-viable lexical candidates. Listeners seem to spot words faster and with lower error rates when these were preceded by a segment that contained at least one vowel compared to words preceded by vowelless segments. An exception to this rule is adult speakers of some (e.g., Tarifit and Tashelhiyt) Berber languages where vowelless open class words are abundant. Berber listeners spot word equally well regardless of whether the previous segment contains a vowel or not. To better understand the origins and development of the Possible Word Constraint (PWC), this experiment explores Tashelhiyt Berber-learning infants’ word segmentation performance. In a Central Fixation (CF) procedure, 17 infants were familiarized with two words. They were then exposed to passages that contained either those familiar words or other novel words; both types of words were attached to a string of segments that either contained only consonants or minimally a vowel. Looking times per conditions were measured to test whether infant attended more to one condition than the other. Confirmatory statistical analyses revealed familiarity effects on word recognition, in which the majority of children (n = 15) looked longer at familiar target words. Descriptive statistics also revealed a trending PWC-like effect which diminished with age. These results, though preliminary, suggest that the PWC might be a language-independent bias which is based on a universal consonant/vowel dichotomy which is exploited by infants and is later modulated by language-specific lexical viability.
Faculteit der Letteren