The hypothesis that memory for spoken sentences is facilitated by memory for sentence meaning was tested with 16 aphasic and 8 nonaphasic adults. Subjects were asked to make judgments of same or different on pairs of active and passive sentences separated in time. Sentence pairs were either identical in all respects or identical in just grammatical structure, subject-verb-object word order, or meaning. Nonaphasic subjects had higher sentence recognition scores, and larger percentages of meaning preserving responses than aphasic subjects. Aphasic subjects with the highest recognition scores made more meaning preserving responses than aphasic subjects with the lowest recognition scores. The results suggested that memory for spoken sentences is facilitated more by memory for sentence meaning than memory for structure of wording.