Cue-weighting in native and non-native speech perception

This thesis investigates cue-weighting in native and non-native speech perception. Speech perception models (i.e. the Perceptual Assimilation Model (Best, McRoberts, & Sithole, 1988; Best & Strange, 1992; Best, 1995; Best, McRoberts, & Goodell, 2001) and the Native Language Magnet Model (Iverson & Kuhl, 1994)) predict that the degree of perceptual difficulty depends on the type of relation between the L1 and L2 phoneme inventory. In addition to these models, Broersma (2005) has shown that if native contrasts appear in non-native phonotactic contexts, non-native listeners use different cues but achieve similar accuracy levels as native listeners. By means of three separate experiments, the differences between L1 and L2 listeners were assessed in cue-weighting when multiple cues are relevant for phoneme identification. As a follow-up on Broersma (2005) reported on in this thesis, all experiments observed the use of fricative voicing and vowel duration cues to make phoneme identifications for /v/-/f/ and /z/-/s/ continua. In Experiment 1, fricative voicing was varied along a continuum while there were two vowel durations. In Experiment 2, vowel duration was varied along a continuum, while there were three fricatives. Finally, Experiment 3 evaluated the tolerance to contrastive cues between native and non-native listeners. The experiments showed that there are significant cue-weighting differences between native and non-native listeners, despite considerable experience in the second language