Spruiell, Bill
Dept. of English Language and Literature
Central Michigan University

Bill Spruiell
Dept. of English Language and Literature
Central Michigan University
Mt. Pleasant, MI

3lfyuji@cmuvm.csv.cmich.edu

Title: Wearing Multiple Hats: A Usefully Complex Theory of Participant Roles


All languages code distinctions among different types of participants in different types of events; whether a given general linguistic theory calls the categories thus marked "deep cases", as in Fillmore's (1968) work, "theta roles" or "thematic roles" as in several types of American formalist approaches, or "participant roles" as in Systemics and some other functionalist approaches, it will deal with them in some manner. The issues involved are in many ways central to linguistic theory, as they involve questions about the ways in which language both models and constitutes experience. Much debate about such roles in formal approaches centers around a pair of fundamental questions: (1) Are such roles primarily semantic, syntactic, or a mixture of the two? (cf. for example Jackendoff (1987)), and (2) Is there a fixed set of such roles universal to human language, or is there infinite variation? (cf. Dowty (1991); Beard (1990)). In functionalist approaches, such as Systemics, the answer to the first question is frequently "yes"; the second question occasions more debate, but most functionalists are suspicious of categorical universals, avoiding strong claims for fixed sets of roles. It is possible, however, to discern functionalist equivalents of those questions: (1') What is the relation between participant roles, as categories, and other types of linguistic (or extralinguistic) categories or processes?; and (2') How do we account for the similarities and differences among the role-inventories we use when describing disparate languages?

In this paper, I will propose an account of participant roles in which a given participant may be considered as existing in more than one role simultaneously, and argue that this relativizes the answers to the questions above. Some accounts (Van Valin (1993), Beard (1990)) allow participants to be assigned multiple roles, but those roles exist in separate functional domains; Beard makes a division between syntactic and semantic role categories, while Van Valin distinguishes roles from macro-roles in a way that appears to present them as fundamentally distinct. I will extend Van Valin's analysis and claim that roles may form a reticulated category structure and that a given participant may be seen as existing in multiple super- and sub-ordinate role categories, each of which has the potential to determine aspects of the emerging communicative event. I will further argue that the distinction between "levels" of subordinacy represent differences in degree rather than kind. Spruiell (1990) has argued that such a system is useful for accounting for the meanings of participant nominalizations, and I will further argue that it accords with some models of child language acquisition, that it provides a mechanism whereby role systems of languages may develop strong similarities without being formally identical, and that it bridges linguistic and non-linguistic experience. In addition, it provides for both extremely delicate ("every verb has its own specific set of roles") and highly reductionist ("there are four fundamental role-types in Language X") descriptions within the same functional domain.