Abstract
A new characterization of agrammatism is suggested, based on new data from Hebrew speaking agrammatic aphasics, and a reexamination of data from Russian and Italian. This characterization is formed in relation to linguistic levels of representation. First, the description of agrammatism as omission of closed-class items is challenged on the basis of the data, and a new description is suggested-viewing agrammatism as mis-selection of items + default: in English the default procedure may always be used, but in the other languages discussed, the patient is forced, for structural reasons, to unconscious guessing that results, in many instances, in syntactically aberrant sentences in which each lexical item is well formed. Second, after discussing issues concerning the proper relation between linguistic theories and processing models, a condition on a syntactic level (S-structure) in linguistic theory (Chomsky, 1981) is proposed, to account for agrammatic data from all the languages considered. It is then shown that agrammatic performance in a variety of tasks (including comprehension) is explained naturally as a consequence of this condition. Finally, several related processing issues are discussed. In particular, the relationship between the proposed structural account and the model offered by Bradley et al. (1980).
| Original language | English |
|---|---|
| Pages (from-to) | 99-120 |
| Number of pages | 22 |
| Journal | Cognition |
| Volume | 16 |
| Issue number | 2 |
| DOIs | |
| State | Published - Mar 1984 |
| Externally published | Yes |
Bibliographical note
Funding Information:*This work was supported by NIH grants l&M6 and 11408. I wish to thank Dr. S. Dame1 and Mr. Y. Schechter from Lowenstein Rehabilitation Center, Ra’anana, Israel for enabling me to interview their patients, and Fania Andelman for her help with the Russian text. Edgar Zurif and David Caplan commented on various drafts and kept encouraging me, and I wish to thank them for that. I am also grateful to Noam Chomsky, Merrill Garrett and Lisa Travis for their helpful comments. Reprint requests should be sent to: Yosef Grodzinsky, Department of Psychology, Brandeis University. Waltham, MA 02254, U.S.A.