| Summary: | In the first 12 months of life, infants develop a robust knowledge of the acoustic dimensions in their phonetic inventory (Booth & Waxman, 2002; Kuhl, 2006; Yeung & Werker, 2009). The associative model (Apfelbaum & McMurray, 2011) predicts that infants may utilize perceptual cues such as talker variability to help discriminate between sounds. This variability may help infants establish acoustic boundaries of sounds, particularly if those sounds are acoustically similar. We find that 7.5 month old infants accurately discriminate between /p/ and /b/ in syllable onset position regardless of the presence of talker variability. We hypothesize that because /p/ and /b/ occur frequently in syllable onset position, infants may already have robust categorizations of the two sounds. We then examine whether talker variability is useful in discriminating between two more acoustically similar sounds, /n/ and /ŋ/. The sound /ŋ/ does not occur in syllable onset position in English; therefore, infants will likely have less robust acoustic boundaries for /ŋ/ and may have a harder time distinguishing /ŋ/ from /n/. We predict that 7.5 month old infants familiarized in a Multiple Talker condition will more accurately discriminate between /n/ and /ŋ/ than infants in a Single Talker condition.
|
|---|
