Abstract
This work makes several inquiries about Chinese durative phrases and the structure of aspect. Lin (2003) makes a distinction between process-related (P-related) and result state-related (RS-related) interpretations of Chinese duratives. Several questions remain, however. Semantically, there are two questions. First, it is well-known that English telic expressions are marginally compatible with the durative adverbs, as in (1a), but they can be modified by terminative adverbs, as in (1b) (cf. Verkuyl 1993, Krifka 1998, and Kratzer 2003). On the other hand, Chinese durative phrases seem to trespass this telicity constraint, unless a completive marker is present. See (2). This is the issue of telicity: (1) a. ??John built a house for three days. (2) a. Zhangsan gai-le fangzi san-tian. ZS build-LE house three-day ‘ZS built the house for three days.’ b. John built a house in three days. b. *Zhangsan gai-wan fanzi san-tian. ZS build-finish house three-day Second, According to Piñón (1999), the result state (RS-related) durative phrases can only occur with the event which denotes an explicit result state. See (3). However, Chinese durative phrases seem more unrestricted with respect to the result state. See (4). This is the issue of resultativity: (3) a. John opened the window for three hours. (RS-related) (4) a. ZS dakai chuanghu henjiu. ‘ZS opened the window for long.’ (RS) b. John built the house for three hours. (*RS-related) b. ZS gai-le fangzi san-tian le. ‘ZS built the house for three days.’ (RS?) Under close scrutiny, the presence of le’s (verb-le and sentence-le) is likely to introduce different interpretations. The ‘atelicity’ is allowed by the verb-le, while the relatively free result state by sentence-le. How should this be? I suggest that the answer lies in the nature of the two le’s. Following Liu (1988), the verb-le actually denotes a ‘realization’ aspect, which marks the SI (initial point of a situation). Therefore, the P-related interpretation is ensured by the presence of the verb¬-le. On the other hand, the sentence-le provides a reference time (R) in the Reichanbachian framework. Therefore, it provides an R for the measurement of the duration. Therefore, I argue that a fine-grained categorization should include the P-relate, the RS-related, and the RT-related. This solves the issue of resultativity since the post-event states are classified into two independent categories. In terms of Klein’s theory, the P-related modify the only phase in 1-phase lexical content (or the source phase 2-phase content), the RS-related modify the target state in the sense of Parson (1990), and the RT-related modify the resultant state, or the former two states with respect to a reference point. This analysis also solves the telicity problem. By virtue of the realization aspect denoted by the verb-le, the event is not necessarily a closed one; the SF (final point) appears only by implicature, which is cancelable in nature. Therefore, it is the time span from the initial to an arbitrary final point that is estimated in the P-related in Chinese. The same condition holds in English when an arbitrary final point is provided in context. Syntacitcally, the three interpretations of the duratives are treated as adjuncts licensed, respectively, by the three levels of the aspectual projections, as proposed by Tenny (2000). Therefore, the RS-related are licensed by the root verb (lower aspect), the P-related by verb-le (middle aspect), and the RT-related by the sentence-le (higher aspect). Conclusively, the model suggests an isomorphic mapping between semantics to syntax.