LANGUAGE LEARNABILITY AND LANGUAGE DEVELOPMENT

该项目试图了解儿童的语言发展，

The project attempts to understand language development in children, using regular (walk-walked) and irregular (come-came) inflection as a tractable, intensively-studied model system. The two kinds of inflection are thought to be examples of the two kinds of mental computation used in language. Irregular words are stored in the mental lexicon, an associative memory which occasionally generalizes patterns To new words based on their similarity to previously-learned words. Regular forms are created by symbolic rule-processing in the mental grammar, which manipulates abstract symbols (e.g., 'past = verb + ed'). Two empirical methods are used. The first is intensive analysis Of children's transcribed spontaneous speech-focusing on errors like comed, brang, and he walk and their correct counterparts. The errors are important data because they are examples of creative linguistic processing in the child's mind, and the studies will quantitatively document when, how often, and why children make them. The second method uses laboratory experiments where children produce or judge novel verbs with different properties. The studies attempt to show whether children are sensitive to a word's sound when computing how to inflect it, and how rule-created and memory-stored forms interact. The laboratory methods will also be applied to certain aged and developmentally impaired populations, including those with Specific Language Impairment, Alzheimer's Disease, Parkinson's Disease, and Williams Syndrome, because the methods are ideally suited to teasing apart their impairments in memory versus rule- processing. The ultimate goals are an extensive quantitative database on children's acquisition of a single rule, to be used as a well-documented test case in the empirical study-of language development, a test of the hypothesis that there are mentally-implemented linguistic rules, and a set of benchmarks against which the language of impaired populations can be assessed.