Comprehension of sentences with reversible semantic roles is sensitive to phonological STM capacity.

Comprehension of sentences with reversible semantic roles (e.g., The boy is kissing the girl.)is difficult for many individuals with agrammatic aphasia (e.g., Schwartz, Saffran & Marin, 1980), especially in the context of non-canonical sentence structures (e.g., passive). Early accounts attributed this difficulty to a specific deficit in syntactic processing that affected both comprehension and production Caramazza & Zurif, 1976; Caramazza & Berndt, 1981). This account was challenged in subsequent studies reporting that impaired comprehension of ‘semantically reversible sentences was not present in all people with agrammatic aphasia (e.g., Miceli, Mazzucchi, Menn, & Goodglass 1983) and that it was present in some people with other aphasic syndromes (e.g., Caplan & Hildebrandt, 1988). Even in their seminal paper, Caramazza & Zurif (1976) reported the difficulty in comprehending semantically reversible sentences in conduction aphasia, but attributed this to an impairment of short-term memory (STM). In another seminal study, Linebarger, Saffran & Schwartz (1983) who demonstrated that impairment in comprehending sentences with reversible roles did not preclude the ability to judge grammaticality of sentences. From this, they proposed the “Mapping Hypothesis”: Difficulty with comprehending semantically reversible sentences lies in the mapping of grammatical roles specified in the syntactic representation onto the underlying thematic roles in the semantic representation of that utterance. These and other similar findings (see R. Martin, 2006 for review) led to an increased interest in the role of verbal STM (semantic and phonological) in sentence comprehension. In the context of the mapping hypothesis, that role would be related to a reduction in processing capacity needed to assign grammatical roles of a sentence’s surface structure onto the underlying thematic roles. In this study, we provide evidence that is consistent with this hypothesis. We examined the comprehension of five sentence structures with and without reversible semantic roles by people with aphasia under two response conditions. We compared performance on the two semantic role conditions (reversible vs. not reversible) and examined the contributions of aphasia severity and verbal STM deficits (WAB-R score, semantic STM and phonological STM) to detriments in performance on the reversible semantic role condition.