Adult perception of nonnative contrasts differing in assimilation to native phonological categories

In cross-language speech research, adults often have difficulty discriminating non-native phoneme contrasts, although this difficulty is not universal nor is it necessarily permanent. Best et al. [J. Exp. Psychol.: Human Percept. Perf. 14, 345–360 (1988)] recently proposed that phonologically mature listeners assimilate nonnative contrasts to native phoneme(s), by four possible patterns: (1) Both phones assimilate to a single category (SC) and are difficult to discriminate; (2) they differ in category goodness (CG) and are discriminated relatively well; (3) they assimilate to two categories (TC) and are discriminated near ceiling; or (4) they are nonassimilable (NA) and are discriminated relatively well as nonspeech sounds. Those authors found evidence supporting their prediction of good discrimination for nonnative NA contrasts. The present research extended to predictions that TC contrasts would be discriminated better than CG contrasts, which, in turn, would be discriminated better if the goodness of fit to the relevant native phone is marked (strong CG contrast) than if the difference is subtle (weak CG contrast). Results were consistent with predictions, thus providing further support for the phonological assimilation model. Additionally, a recency effect in auditory memory was found only for the weak CG contrast. [Work supported by NIH Grants DC-00403 and HD-01994.]