We propose to account for what masquerade as preserved syntactic abilities by patients commonly characterized as """"""""agrammatic"""""""" (literally without grammar) and to highlight those aspects of their linguistic behavior that warrant special attention in developing successful diagnostic tools and therapies that can maximize their communicative abilities. Observations of spared syntactic abilities preserved in agrammatics might seem to disconfirm linguistic theories that assume autonomy between the syntax and the lexicon. When language breakdown fails to pattern along lines predicted by a grammatical theory, one is led to assume that the theory has nothing to contribute to our knowledge of language breakdown and its remediations. Furthermore, we might also be forced to conclude that data from aphasia make no contribution to the development or refinement of linguistic theory. Fortunately, the premise that data from agrammatics is incompatible with linguistic theory is false since such a conclusion would offer little hope for the development of fine grained diagnostic tools or linguistically informed remediation methods. We argue that the spared """"""""syntactic abilities"""""""" that have been attributed to agrammatics, making this label somewhat of an oyxmoron, are actually spared lexical abilities. Recent linguistic and psycholinguistic research, including our own, indicates that the lexical representation of verbs includes enough information about the type and number of arguments (noun phrases or sentential complements, as well as any other obligatory lexical material such as particles, measure complements, and associated idiomatic material) required by individual verbs to be equivalent to the phrase structure representation of its verb phrase in the absence of the effects of any movement rules required by the syntax. Crucially, in our view the structural configuration that characterizes the grammatical relations among the elements within a sentence is not part of the syntax, but is part of the lexicon. In fact, it is part of the lexical representation of a verb. Agrammatics retain and can use this lexical information. What they cannot do is unravel the effects of syntactic operations on this basic configuration. They only exhibit what has been interpreted as syntactic comprehension in those sentences where they can decode information on the basis of lexical knowledge. From the lexical entry of a verb, we can predict which sentences will be problematic. Specifically, we will use our research on argument structure and verb class organization to better characterize the sentence production and comprehension abilities of agrammatic speakers with a view toward contributing to the development therapies that can utilize the phrase structure information present in lexical entries to facilitate the production and processing of certain syntactic constructions while indicating those constructions that are likely to be less accessible. To verify that we have identified the appropriate and the specific grammatical deficit underlying agrammatic linguistic behavior, we propose to conduct an extensive series of sentence production and comprehension experiments, testing their on-line performance as well as their performance on more traditional secondary tasks. Agrammatics are a logical starting point for this type of study both because their pattern of language breakdown is compatible with the enriched theory of the lexicon we espouse and also because we have access to a large body of preexisting production data we can use as a testing ground for our model. However, we intend to extend our scope to include a representative range of aphasic subjects because we believe our model makes specific and verifiable predictions about systematic and distinct patterns of sparing and loss in other population of aphasics.
| Kegl, J (1995) Levels of representation and units of access relevant to agrammatism. Brain Lang 50:151-200 |
