| This thesis examines English speakers' perception and production of a non-native segmental stop contrast in Czech: alveolar /t, d/ versus palatal /c, j/. Brown's (1997, 1998, 2000) model of L2 phoneme acquisition argues that it is the L1 feature inventory (rather than individual segments) which define the boundaries within which novel phonemes are perceived. Specifically, L1 speakers can perceive novel L2 contrasts if that contrast is characterised by a feature present in their L1 grammar; conversely, if a particular feature is lacking in the L1 feature inventory, then perception of the novel phonemic contrast should be precluded. I argue that the contrasting feature between Czech alveolar /t, d/ and palatal /c, j/ is [posterior]. English requires the feature [posterior] to contrast two fricative segments: alveolar /s, z/ versus alveo-palatal / &smallint; , 3/. English speakers thus have the necessary building block for acquisition of the novel Czech contrast. |
