ABSTRACT
Classifier constructions in American Sign Language (ASL) were originally described as a type of mimetic depiction (Klima & Bellugi, 1979), rather than as morphologically complex signs. However, Supalla (1978, 1982) argued that classifier verbs were composed o f discrete morphemes that express object form or category through handshape morphemes, object location through a “base grid” o f discrete spatial loci, and movement through verbal roots. Similarly, Newport (1982) proposed that “ ‘mimetic’ forms within American Sign Language are not at all analogue in nature; rather, like morphologically complex forms in spoken language, they are constructed from a relatively small number o f discrete components, which mark familiar distinctions o f meaning and are combined in familiar ways” (p. 465). However, whether all meaningful components within a classifier construction are discrete morphemes has recently come into question (e.g., Liddell, 1995). In the series o f experiments reported here, we investigated whether aspects o f classifier constructions are treated as discrete categorical morphemes or as gradient, analogue representations.
