A Computer Model of Child Language Acquisition

because generation cannot occur until comprehension learning adds words to the dictionary. Length of utterance Inoreases because the number of words available to express a oonoept inoreases during comprension learning. Knowledge of language meaning precedes knowledge of language syntax beoause syntax is indexed under word meanings and hence cannot be learned before the word meaning. Misunderstanding of utterances whose correct interpretation is not the semantically most probable occurs beoause children use their knowledge or probable meanings to augment gape in their understanding. Finally, misunderstanding of utterances whose syntax suggests an interpretation different from) the semantically most likely interpretation occurs beoause knowledge of syntax is learned after knowledge of meaning. This error is made when enough knowledge of aeaning has been acquired to produce an interpretation, but not enough syntax has been learned to produoe a correot interpretation. There are certainly aany other factors in child language acquisition which have not been considered here, but this paper offers support for the seaantically-indexed syntax hypothesis and the comprehension-driven generation hypothesis as components of a complete model of child language acquisition. Further research must address questions posed by the data of more advanced child language acquisition, which could support other methods of indexing syntactic knowledge and other possible relationships between comprehension and generation. ACKNOWLEDGEMENTS Dr. Roger Sohank's assistance in this work was invaluable. Dr. Richard Cullingford has made many valuable suggestions, and Dr. Katharine Nelson has provided many insights into problems of modelling child language learning. Peter Selfridge, Larry Birnbaum, Marie Bienkowskl and Jamie Callan have contributed both ideas and criticisms.