This thesis is concerned with the phonological representations of words used in lexical storage and retrieval, during speech perception by second language speakers. There is evidence of categorical perception of certain phonological features of words by native speakers of particular languages. This can be constrained by language- particular phonological properties, such as lexical tone, or by distinctions between certain types of consonant. If native language perceptual strategies were used in second language word retrieval, then this would mean there were differences in word storage for second language speakers. This would be reflected in differing patterns of word retrieval for second language speakers of a language with different phonological properties from their own. In order to test this possibility, Hong Kong Cantonese speakers with English as their second language were required to retrieve English words from their word store. Their native language is tonal, unlike English, and they have been found to perceive tones in their native language categorically. Subjects were presented aurally with English sentences which each contained a malapropism for the last word, and were asked to produce the correct word. The malapropisms were systematically varied in their phonological similarity to the target. The phonological variation was determined from evidence drawn from speech error analyses in production and from an analysis of a high frequency sample of words conducted as part of the thesis. Native speakers of English were used as controls in the experiments. Results showed similarities and differences in retrieval between the two groups of subjects. Both groups made use of a number of phonological properties in retrieval. The differences were associated with perceptual strategies involving a suprasegmental phonological property of English, that of lexical stress. Correct words could be retrieved by the Cantonese speakers when word stress was the only shared phonological property of error and target. Native speakers only made use of word stress when other phonological properties were shared by error and target. The use of a number of phonological properties by both sets of speakers during word retrieval is consistent with recent generative linguistic accounts of enriched phonological structure in phonological representations. It is possible that the mind takes account of such constituent structure during speech perception to disanibiguate phonetic stimuli. However, the phonological organisation of lexical representations may vary from one language to another, with information from the same sound signal being used differently by second language speakers of a given language from native speakers
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:308699 |
| Date | January 1991 |
| Creators | Partington, Ann |
| Publisher | London School of Economics and Political Science (University of London) |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0015 seconds