Perception of novel phonetic contrasts in a second language has been studied extensively, but suprasegmentals have seen relatively little attention even though difficulties at this level can strongly impact comprehension. Available studies suggest that the perception of segmental and suprasegmental categories are subject to similar factors, but evidence is not entirely conclusive. Additionally, studies focusing on the perception of lexical prominence have suggested that perception across languages with different accent types might be particularly problematic, which means these effects in particular would be independent of the direction of language learning.
This dissertation explores both of these questions by studying the perception of lexical prominence by second-language learners in Spanish (a stress-accent language) and Japanese (a pitch-accent language). Following a bidirectional approach, it examines whether the perception of phonologically different types lexical prominence is subject to similar effects as those traditionally identified for cross-linguistic segmental perception, and how these relate to the direction of learning.
A first set of studies provides a comparative acoustic description of prominence in both languages, and presents the results of a identification task with natural words in different positions within a sentence. Using multiple speakers, these tests showed that the difficulties seen by both groups are different and related to features in their L1, and that despite phonological differences, contexts existed in which high performance was possible. A second set of studies explored the sensitivity of non-native listeners to secondary acoustic cues and the development of new accentual categories, and showed effects of learning for both groups and a strong sensitivity to duration for learners of Spanish. Learners of Japanese showed extremely poor category development for unaccented words in particular.
Overall results show that existing research on SLA is applicable to suprasegmental perception, and that the transfer effects affecting both groups have different domains and scope. The implications for theories of L2 acquisition are discussed.
Japanese (Temporarily offline)
Spanish (Temporarily offline)
For certain values of soon… ↩