JOURNAL OF MEMORY AND LANGUAGE 26, 505-526 (1987) Resolution of Syntactic Category Ambiguities: Eye Movements in Parsing Lexically Ambiguous Sentences LYN FRAZIER AND KEITH RAYNER University of Massachusetts Three experiments explored the effects of the interaction of lexical and syntactic pro- cesses during language comprehension. Subjects read sentences containing lexical items that resulted in syntactic category ambiguities (e.g., desert trains, where desert can be a noun or an adjective and trains can be a verb or a noun). During reading, eye movements were monitored as a reflection of on-line parsing activities. The experiments tested alterna- tive hypotheses about how the processor resolves syntactic category ambiguities. All ex- periments supported a delay strategy in which the processor delays assigning an analysis to a categorially ambiguous string until it receives disambiguating information dictating the correct analysis of the string. The implications of the results for a general theory of sentence comprehension are discussed. o 1987 Academic PKSS. 1~. Understanding the manner in which the sentences of a language are syntactically parsed is a central concern of any theory of language processing. In several prior in- vestigations we have examined the prin- ciples governing the analysis of syntacti- cally ambiguous sentences (Frazier & Fodor, 1978; Frazier, Clifton, & Randall, 1983; Frazier & Rayner, 1982; Rayner, Carlson, & Frazier, 1983). We have argued that the processor rapidly adopts the first syntactic analysis available to it, e.g., the analysis requiring the postulation of the fewest syntactic nodes. Syntactically more complex analyses of a sentence are identi- fied later, if and when disambiguating or biasing information is encountered. These This research was supported by Grant HD17246 from the National Institute of Child Health and Human Development and Grant BNS-8510177 from the National Science Foundation. We thank Patrick Carroll, Mary-Ann Palmieri, and Dolores Shank for their assistance in collecting and analyzing the data. We also thank Patricia Carpenter, Chuck Clifton, Susan Duffy, Marcel Just, Barbara Malt, Sara Sereno, Rosemary Stevenson, and an anonymous reviewer for their comments on an earlier version of the paper. Re- quests for reprints should be addressed to Lyn Fra- zier, Department of Linguistics, University of Massa- chusetts, Amherst, MA 01003. earlier investigations dealt with the anal- ysis of categorially unambiguous inputs and examined the syntactic structuring of input items that could unambiguously be categorized as nouns, verbs, etc. For ex- ample, we were interested in whether the adverb attaches to the lower or higher clause in a structurally ambiguous sentence like Tom said Sue left yesterday. The ex- periments to be reported here were de- signed to investigate whether the principle of adopting the first available analysis of an input extends to the analysis of categorially ambiguous items. Consider the ambiguous string in (1): (1) The warehouse fires. . . . The word warehouse occurs most often as a noun, though it has a derivative usage as an adjective, as in the (b) form of Sentences (2) and (3).’ i The correct linguistic analysis of items like desert in the noun phrase The desert train is not entirely clear. Assuming that English does not permit novel compounds to be generated without compound stress (stress on the left-hand member of the compound), words like desert must be analyzed as derivative ad- jectives, since compound stress is not obligatory on these phrases. Clearly, the syntactic rules of English 505 0749-596X187 $3.00 Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.