Generalization of complex sentence production in agrammatism

The efficacy of Linguistic - Specific Treatment (LST) was investigated. LST focuses on developing metalinguistic knowledge of thematic roles in complex sentence structures and provides direct and repeated practice in production of these structures. Four sentence types were targeted in treatment: object clefts, matrix questions, embedded questions, and embedded actives. The first three are noncanonical and contain wh-movement, the fourth is canonical with no movement. All sentences contain overt material in the complementizer phrase (CP). Experimental single subject multiple baseline designs across subjects and behaviors were used to investigate treatment and generalization effects within and across sentences in five individuals with chronic agrammatic Broca's aphasia. Three subjects demonstrated generalization from object clefts to matrix questions and partial generalization to embedded questions. This partial generalization involved improved ability to produce the targeted syntactic structure but continuing inability to express overt material in CP. The remaining two subjects' behavior was best characterized by overgeneralization. No subjects demonstrated generalization between embedded questions and embedded actives. No clear pattern of generalization of treatment effects to morpho-syntactic variables in a narrative language task or performance in formal language testing was observed. However, four subjects demonstrated increases in measures of informativeness and efficiency in narratives and three subjects were judged by naive listeners to improve in content, coherence, and fluency of production. Results are discussed in relation to their contribution to the development of (a) the LST protocol for treatment of sentence production in agrammatism and (b) important criteria for subject inclusion. Responses of individual subjects to the treatment protocol are related to linguistic, capacity-constraint, and adaptation theories of agrammatism. Directions are provided for further developing a theoretical account of the sentence production impairment in agrammatism. Areas for future research are identified for applying this treatment approach to a broader range of subjects and developing more sensitive measures of generalization of treatment effects.