Functional Grammar


  • Martin Kay Stanford University and XEROX PARC



The term functional grammar has been used before, notably by Dik (1978). I risk adding to the number of its meanings here, and thus debasing its value, only because it is peculiarly apt for this new employment. I propose to outline a new grammatical formalism which, if it can be successfully developed, will be worthy of the name functional on three counts. First, it is required to function as part of a model of language production and comprehension. The formalism is interpretable by an abstract machine whose operation is intended to model tl1c syntactic processing of sentences by speakers and hearers indifferently. This is not to say that it is not also intended to represent a speaker's grammatical competence. Secondly, the formalism ascribes to every sentence, word, and phrase, a functional description which differs from the strnctural description of better known formalisms mainly by stressing the function mat a part plays in a whole rather than me position a part occupies in a sequence of omers. The names of grammatical categories, like S, NP, and VP will therefore play a secondary role to terms like subject, object, and modifier. Thirdly, properties that distinguish among logically equivalent sentences will have equal importance with properties that they share. In omer words, mis will be a functionalist view of grammar in which notions like topic and focus, given and new will have equal status with subject and predicate, positive and negative.