Abstract
Background: Agrammatic speech can manifest in different ways in the same speaker if task demands change. Individual variation is considered to reflect adaptation, driven by psycholinguistic factors such as underlying deficit. Recently, qualitative investigations have begun to show ways in which conversational interaction can influence the form of an agrammatic speaker's output.
Aims: To explore qualitative patterns of individual variation in the output of a single case, in order to address the following questions: (1) in what ways do grammatical structure, verbs and argument structure differ qualitatively between data samples?; and (2) what is the influence of interaction on the structure of utterances?
Methods & Procedures: A man with severe and chronic agrammatism is videotaped completing tests of spoken sentence construction, composite, and cartoon strip picture descriptions and a story telling task. In addition, he independently videotapes a sample of conversation with a family member at home. Analyses of these data draw on cognitive neuropsychological, linguistic and psycholinguistic methodologies as well as the data‐driven procedures of conversation analysis.
Outcomes & Results: The qualitative analysis uncovers considerable variation between data sets with respect to grammatical structure, verbs and argument structure. The agrammatic speaker appears more skilled at imposing order on his spontaneous speech than the results of sentence construction tests might predict. Interaction influences his output style. Adaptation is found to be a collaborative process that occurs between two speakers in the quest for mutual understanding, not something that takes place within an individual.
Conclusions: It is concluded that an interactional approach to agrammatism has the potential to yield important insights into the characteristics of telegraphic speech and individual variation. The failure of sentence‐level tests to capture this individual's skill in producing systematically structured utterances in conversation implies that the assessment of agrammatism could benefit from the additional sampling and analysis of conversational grammar.
Notes
1. All names of participants are pseudonyms, as are the names of any people or places referred to in the conversation data.