Advances in Cognition, Education, and Deafness

Deaf Readers' Comprehension of Complex Syntactic Structure

Deaf Readers' Comprehension of Complex Syntactic Structure Diane C. Lillo-Martin Vicki L. Hanson Suzanne T. Smith What differentiates adult deaf readers who are proficient at reading English from those who are less proficient? Are there aspects of English grammar that the good reader understands but the poor reader does not? Is there some cognitive factor in the process of reading that good readers are better at? In this paper we address these questions by examining good and poor deaf readers' understanding of some aspects of English grammar. In particular , we examine comprehension of one complex syntactic structure-relative clauses-using tests of the comprehension of relative clauses in written English, signed English, and American Sign Language (ASL). We also investigate how good and poor deaf readers' use of phonological coding in short-term working memory contributes to reading differences (Conrad 1979; Hanson 1982; Lichtenstein 1985). Studies with good and poor hearing readers have indicated that poor readers exhibit a variety of language-related impairments, including phonological segmentation difficulties, naming difficulties, working-memory deficits, and impaired performance on tests of the comprehension of complex syntax (for review see Liberman & Shankweiler 1987). However, despite these seemingly disparate imThis work was supported in part by National Institutes of Health NIDCD Grant #DC00183 (formerly NINCDS #NS18010) to Diane Lillo-Martin at Haskins Laboratories . We would like to thank Deborah Kuglitsch for running the subjects; Mike Karchmer, Ken Epstein, and Gallaudet University for assistance in gathering subjects; and the subjects who participated. 146 Deaf Readers' Comprehension of Complex Syntactic Structure 147 pairments, a unified account of poor readers' difficulties can be given by locating the source of reading impairment in the phonologically based working-memory system (Shankweiler & Crain 1986). Under this account, poor hearing readers have trouble using phonologically based coding in working memory, and this difficulty underlies their apparent lexical and syntactic problems. What makes phonological coding in working memory so important? In reading and listening, individual words of a sentence must be retained while the grammatical relations among words are determined. Evidence suggests that working memory is most efficient for verbal material (including written material) when the processing involves phonological coding. For readers suffering from impaired phonological coding in working memory, processing individual words and putting these words together into phrases and sentences can be computationally overloading , impairing overall reading performance. Is it even possible for deaf readers to use phonological coding in the absence of auditory input? We note that, despite common assumptions that a phonological code must be auditory, this assumption is not supportable. Phonological units relate not specifically to sounds, but rather to articulatory gestures and, as such, could also have motor and visual components (see Hanson 1989 for a discussion). Studies of good and poor deaf readers have indicated that use of a phonologically based working-memory code is characteristic of good deaf readers (Conrad 1979; Hanson, Liberman & Shankweiler 1984; Lichtenstein 1985). Thus it is possible that poor deaf readers, like their hearing counterparts, suffer from difficulties in using a phonologically based code. It is interesting to note, however, that tests of short-term memory for American Sign Language stimuli show evidence that deaf readers use a sign-based code (Bellugi, Klima & Siple 1975; Krakow & Hanson 1985; Poizner, Bellugi & Tweeney 1981). Since hearing children are exposed to spoken language in sufficient quantity for normal language acquisition to occur, it is perhaps not surprising that most hearing children with reading difficulties do not have a specifically syntactic deficit. However, for deaf readers, since much of the input for English syntax comes from written material, it is reasonable to ask whether syntactic deficits might remain, possibly in addition to deficits in phonologically based workingmemory processes. Evidence for a processing deficit over a syntactic deficit in deaf readers takes the following form: In tests of syntactic processing, subjects (adults and children, good and poor readers) will always make some errors, and certain sentence types might be more taxing than others. If good readers have syntactic competence, which poor readers lack, then poor readers should have an overall depressed level of performance on syntactic tasks, and we do not predict a particular pattern of errors across sentence types. However, if poor readers have the same syntactic competence as good readers but are suffering from a processing impairment of the type suggested by the unified-processing-deficit hypothesis, then we would expect to find a similar pattern of errors across sentence types for the two groups. That is, sentence types that are especially taxing for...