Professor Catherine (Cathi) Best is the Chair in Psycholinguistic Research at the MARCS Institute of Brain, Behaviour and Development at Western Sydney University, Australia. Her research focuses on how adults’ and infants’ experience with their native language shapes their perception and production of the phonological elements of spoken words, including consonants, vowels, lexical tones and prosodic patterns. She has applied this theme broadly, investigating perception of unfamiliar non-native speech contrasts by naïve listeners, second language learners and bilinguals, in infants and toddlers. She has expanded that core interest in perceptual effects of experience to sign language and culture-specific musical structures.

Her most significant theoretical contribution is her model of the effects of language experience on perception: the Perceptual Assimilation Model (PAM: e.g., Best, 1984, 1994a, 1994b, 1995; Best & Tyler, 2007). Best's model and empirical work offered insights into why many non-native phonetic contrasts are difficult for adults and older infants to discriminate, but others remain much easier. Throughout her work, Best has taken an ecological, or direct realist, philosophical perspective, founded on James and Eleanor Gibson's ecological theory of perception. Prior to arriving in Australia, she was awarded an NIH Research Career Development grant, during which she gained several years of advanced training in phonetics and phonology, deepening her interest in articulatory information as a viable ecological basis for speech perception. That interest has been fundamental to the continued development of the PAM model, which has also provided the core motivation for her more recent line of research on the effects of regional accent differences in native language word recognition by infants, toddlers and adults.

Abstract:
Four decades of research has confirmed that during their first year infants perceptually attune to native consonant and vowel contrasts, as well as suprasegmental patterns. It is often claimed that discrimination is initially good for both native and non-native contrasts but declines for non-native contrasts by mid-year for vowels and by ~ 8-10 months for consonants, whereas discrimination of native speech contrasts remains good or improves. But these claims are over-simplified; perceptual data on a wide variety of both native and non-native speech contrasts is more nuanced, with some native ones remaining difficult into the preschool years and some non-native ones continuing to be discriminated well (e.g., Best et al. 1988, 1995). The Perceptual Assimilation Model (PAM: e.g., Best, 1993, 1995) posits that perceptual ease or difficulty for specific non-native contrasts is shaped by developmental changes in detecting articulatory similarities and differences from native phonemes and contrasts, based on increasing native speech experience. PAM has been supported by numerous studies, and accounts for the observed deviations from the “simple” developmental story (Best et al, 2016).

But how does first year perceptual attunement provide the grounding for word learning and recognition in the second year? Much is known about toddlers’ ability to detect critical consonant or vowel mispronunciations of words, which signals a grasp of the principle of phonological distinctiveness (Best, 2015). However, recognizing words despite non-distinctive phonetic variation in regional accents, or phonological constancy, is also crucial for lexical acquisition. We have conducted a number of studies on phonological constancy in toddlers’ word recognition across English accents, and how it may relate to their vocabulary size. More recently, we have examined how young children adapt to accent variation. Implications for theories of early spoken word recognition and episodic memory in lexical retrieval will be discussed. Relevance to adult second language speech and word learning will also be considered.