Parisot, A.-M. : "Do Plain Verbs Agree in Quebec Sign Language (LSQ)?", colloque TISLR, Amsterdam, juillet 2000. (syntaxe)

In ASL, according to Edge and Herrmann (1977), Fischer and Gough (1978), Friedman (1975), Janis (1995), Lillo-Martin (1986) and Padden (1983), there are verbs morphologically marked for agreement and verbs which are not. According to these studies, plain verbs (which do not agree with their arguments) assign case via structurally determined grammatical relations under SVO order. However, as shown by Bouchard et al. (1999), the primary data of LSQ indicate no canonical order, although SOV order seems to predominate for plain verbs in their corpus (67%). If there is no order and no morphological marking, how are cases assigned by plain verbs? In a corpus of 150 sentences with plain verbs produced by three native signers of LSQ, we noticed that the verb is often followed by one or two point signs. Without these point signs the semantic roles of arguments would be unclear. In addition, the phonological form of the point signs often assimilates to the form of the verb. We will first propose that these point signs are clitics and are interpreted as morphological marks of agreement. Even if simultaneous morphology is dominant in LSQ, this is not the only case of sequential morphology (Dubuisson et al., 1996; Vercaingne-Ménard and Dubuisson, 1998). Secondly, we postulate that there are two categories of plain verbs, one that requires morphological agreement marks (with clitic indexes) and one that does not. A verb like EAT does not need agreement marking because its lexical specification already includes information relative to the possible animate nature of the arguments (+animate agent, -animate patient) and there is no ambiguity. On the other hand, a verb of the second category such as LOVE requires agreement marking since there is ambiguity in the assignment of the theta role (both arguments are lexically specified as being +animate).