Abstract
Abstract Children overgeneralize intransitive verbs for causative verbs. This phenomenon is called the intransitive-to-causative bias. However, Lin & Tsay (2008) argue that the overgeneralization pattern in Taiwan Southern Min shows a resultative-to-causative bias, which is different from the bias mentioned above. On the other hand, Murasugi & Hashimoto (2004) follow the analysis of Larson’s VP shell hypothesis to account for certain phenomena in early Japanese, and argue that there exist four stages when Japanese-speaking children are acquiring verbs and morphological –(s)ase causative constructions. Previous studies did not discuss the relevant phenomena in Mandarin Chinese. Thus, the main purpose of this paper is to figure out whether Mandarin Chinese has the intransitiveto-causative bias or the resultative-to-causative bias. This paper also uses VP shell hypothesis to explain the development of the causative constructions in Mandarin Chinese. We observe the usages of element gei in early Mandarin Chinese for this purpose. After examing data of four children from Child Corpus developed by Hsu, we find that Mandarin-speaking children show the intransitive-to-causative bias, and that there exist three stages in the acquisition of Mandarin-speaking children. The first stage is when the small v is phonetically null. The second stage is is when the small v can be realized as gei. The last stage is the acquisition of morpho-syntactic causative constructions. The acquisition stages in Mandarin Chinese are different from those in Japanese. We argue that inflections cause the divergence. The evidence is the difference of tense morphology between Japanese and Mandarin Chinese.