It is proposed to continue testing and elaborating a model of speech learning developed during the last two grant periods. The goal is to better understand the structure and content of phonetic categories, how and when they are established, and under what circumstances they may be modified when a second language (L2) is learned. Twenty experiments examining the production and perception of English have been planned. Many will compare native speakers to adult non-native speakers in three groups: Early L2 Learners who learned English as young children; nonproficient Late L2 Learners who learned English as adults and speak it with a strong accent; and proficient Late L2 Learners who pronounce English well even though they, too, learned it in adulthood. In addition to cross-sectional studies of English vowels and consonants, a longitudinal study examining the production and perception of English vowels will be undertaken. A wide range of differences has been noted among the Late L2 Learners examined in past grant periods to produce and perceive particular L2 sounds. This has led to the inclusion of experiments that aim to identify factors contributing to exceptionally good l2 pronunciation by Late L2 Learners. Another series of experiments will attempt to identify the basis for variations in degree of perceived foreign accent (FAO). Finally, it is hoped to continue developing improved methods for speech training. Illustrative training experiments designed to improve Late L2 Learners' perception -- and perhaps indirectly production -- of English vowels and consonants are presented. The proposed research will lead to a better understanding of how phonetic systems are organized, and provide insight into the processes and mechanisms through which they develop over the lifespan.
