Language

THE DISCOURSE BASIS FOR LEXICAL CATEGORIES IN UNIVERSAL GRAMMAR Paul J. HopperSandra A. Thompson State University ofNew York,University of California, Los BinghamtonAngeles Most linguists who have investigated linguistic categories from a universal viewpoint have accepted the existence of two basic parts of speech, noun and verb. Other categories are found to be only inconsistently represented; thus adjective is manifested in many languages as a class of Stative verb. Furthermore, individual languages often have intermediate categories such as gerund, which cannot be unambiguously assigned to a single category. We suggest here that the basic categories N and V are to be viewed as universal lexicalizations of the prototypical discourse functions of 'discourse-manipulable participant' and 'reported event', respectively. We find that the grammars of languages tend to label the categories N and V with morpho-syntactic markers which are iconically characteristic of these categories to the degree that a given instance of N or V approaches its prototypical function. In other words, the closer a form is to signaling this prime function, the more the language tends to recognize its function through morphemes typical of the category—e.g. deictic markers for N, tense markers for V. We conclude by suggesting that categoriality itself is another fundamental property ofgrammars which may be directly derived from discourse function.* Introduction 1. The classification of the lexicon into categories or 'parts of speech' has been a concern of linguists since ancient times. This paper offers a new approach to the many problems raised when linguistic categories are considered from the aspect of universal grammar. 1.1. Categories. Over the centuries during which the Western grammatical tradition has evolved, the criteria for determining categories have shifted according to trends in linguistic thought as a whole. These criteria have included the following: (a)Morphological. This criterion takes the possibility of a form's combining with different types of inflectional morphemes as the basis for identifying categories.1 A noun is thus defined as a form which is 'declined' according to case, number, and gender (case being primary, indeed to some extent synonymous with 'noun'); a verb is a form which is 'conjugated' according to person, tense, and mood. This criterion clearly derives from a grammatical tradition within a classical Indo-European speech community; but in fact it extends remarkably well, as we shall see, to a wide range of non-IE languages. (b)Semantic. That the major categories are associated with fairly consistent semantic classes was recognized by the earliest grammarians. Thus Dionysius * We are grateful to several people whose comments have helped shape this study: Raimo Anttila, Joan Bybee, Wallace Chafe, Bernard Comrie, Jack Du Bois, John Haiman, Carol Lord, Francesca Merlan, Edith Moravcsik, Johanna Nichols, Michael Noonan, and Kenneth Whistler; and to Christian Matthiessen for help in preparing the manuscript. Authorship of this paper, as with all our joint work, is shared equally. 1 However, Lyons (1966:217) rejects the characterization 'morphological' for this approach. 703 704LANGUAGE, VOLUME 60, NUMBER 4 (1984) Thrax defined the noun (onoma) as 'a part of speech inflected for case, signifying a person or thing' (Robins 1979:33). The difficulty of applying semantic or 'notional' criteria consistently has been a perennial problem; yet the broad correlation is so obvious that linguists have repeatedly returned to it in one form or another. (c) Syntactic. Syntactic arguments for determining boundaries among word classes go back to earliest times; the distinction between 'subject' and 'predicate ' was well known to the Hindu grammarians, and was equated by them with the noun/verb distinction (Lyons 1968:19-20). Furthermore, the Greek term rhêma 'verb' appears originally to have referred to the predicate of the sentence (Robins 1979:26-7), rather to than the (conjugated) verb. Purely syntactic arguments are also used by modern structural linguistics; thus, in the transformational/generative paradigm, the criteria for assigning forms to the major categories are linked to the usefulness of the classification to the formulation of rules: 'The way in which we have arrived at the particular decisions [about word classes] ... is, in principle, irrelevant ... What matters is whether one classification rather than another enables the grammarian to formulate a set of rules which will include...