1 Classifier Predicates Reanalyzed with Special Reference to Taiwan Sign Language*

In this paper we reanalyze the so-called classifier predicates with special reference to Taiwan Sign Language, arguing that the so-called classifier morpheme is a superordinate and should be analyzed as a proform because of its primary anaphoric function. In addition, we develop an approach combining Talmy’s (1985) motion events with the proform analysis to account for how Figure and Ground are represented concurrently with spatial verbs in spatial constructions and how the proforms of Figure and Ground are formed to satisfy such purpose. It is shown that in Taiwan Sign Language the proforms of Ground are more predictable than those of Figure, because the Ground usually takes the handshape by the non-moving hand as its proform, while the Figure usually adopts a new handshape which is not part of the given sign to serve as its proform. Key words: Taiwan Sign Language, classifier predicates, proforms of Figure and Ground 1