A Typological Study on the Grammatical Encoding of the Causee in Causative Constructions

Abstract

The causative introduces an external agent into the clausal structure. When an additional agent is introduced into the event, and when it is construed as the agent ultimately responsible for the occurrence of the event, then this ultimate agent is encoded as the subject, and the one that is less responsible for the event is encoded as the direct or oblique object. Causative constructions can be classified according to two parameters; one is a formal (or morphological) parameter and the other is semantic. As for the semantic distinction, we have shown that the difference between distant and contact causation is reflected formally in many languages including Korean. In Korean, semantic differences between distant and contact causatives generally correspond to those between the analytic and the morphological types. The essential difference between distant and contact causatives is the degree of control exercised by the causee. The causee in distant causatives has some control over the action, whereas the causee in contact causatives has no control over the action. That is, the causer of contact causatives behaves like the agent of a typical transitive verb, in that he or she both initiates and directly carries out the action. One interesting aspect of this, with respect to the control parameter of the causee, is that distant causatives usually require animate causes. Where the causee is inanimate, the distant causative suggests that the causer might invoke some external force, such as "magical power", in view of the indication that there is an absence of physical contact between causer and causee. The other issue that we have investigated is the morphological encoding of the causee. We established the case hierarchy for the Korean causative construction as accusative > dative > nominative, which coincides with the degree of control exercised by the causee, from least to greatest. We also investigated the validity of Comrie's (1981) hierarchy account through a careful examination of data from Korean. We have shown that there are many languages that do not conform to this hierarchy, and that allow doubling on certain grammatical relations. Korean permits doubling on Direct Object, Indirect Object, and even on subject positions. Most importantly, Comrie's hierarchy account fails to explain why case-markings are used contrastively. For example, the case-marking contrast of the causee between the accusative and the dative/oblique indicates a semantic contrast. Basically, Comrie's hierarchy account is rooted in a purely syntactic perspective without considering the semantic function of case-markers. We have shown that there is a possible semantic contrast between different encodings of the causee in causative constructions. As the basic morphological encoding of the patient, the accusative typically refers to an entity (causee) with a very low degree of control. On the other hand, the oblique case (or whatever case that might be selected for passive agents) is frequently used for an entity with a high degree of control. As the typical exponent of the experiencer or recipient role, the dative occupies an intermediate position.