The Acquisition Of Syntactic Categories In Chinese On The Basis Of Distributional Information

Chinese is a language with impoverished morphology. And the identification of syntactic categories would depend primarily on syntactic distribution of words, which is one of key criteria. The distribution of words thus plays an important role in the acquisition of syntactic categories by Mandarin-speaking children. The present study investigates the possibility that distributional information found in parental speech directed to Mandarin-speaking children can provide reliable syntactic category information for language acquisition. The distribution of words given in terms of frequent frames is studied to evaluate the effect of the distributional information on the Mandarin-speaking children's syntactic category acquisition. Here, a frame is defined as a sequence of two jointly occurring words separated by an intervening word.The data used for the study come from the parental input (22137 utterances)to two children whose development was observed longitudinally from 01;02 to 02;00, as well as the spontaneous production of the children from 01; 09 to 02; 00 (2491 utterances). The data consist of 40 sessions of hourly audio- and audio-visual recordings.The study consists of two parts.In the distributional analysis based on 701 frequent frames from the parental input,the main finding is that frequent frames are an accurate and efficient source of information for categorizing words. Firstly, the distributional information of categories provided by frequent frames is robust, since the syntactic category information of words that constituted about two-thirds of each corpus was provided by the frequent frames that took up less than 3% of each corpus. Next, frequent frames in the parental speech can provide the reliable syntactic category information, as proved by high accuracy scores and prominence scores. Thirdly, the distributional properties of both lexical categories such as nouns, verbs, and adjectives and functional categories like adverbs and sentence-final particles were picked out by frequent frames.Furthermore, the frequent frames and their intervening words found in the parental speech and those observed in the children's production were compared to evaluate the effect of distributional information on the children's syntactic category acquisition. There are 494 overlapping frequent frames in the comparison. About 80% frequent frames and 60% intervening words from these frequent frames in children's production also appeared in the parental speech. The frequency of overlapping frequent frames in the parental input was very...