A usage-based approach to language processing and intervention in aphasia

• Main Author: Bruns, Claudia
• Other Authors: Beeke, S. ; Varley, R. ; Zimmerer, V.
• Published: University College London (University of London) 2018
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.763334

Non-fluent aphasia (NFA) is characterized by grammatically impoverished language output. Yet there is evidence that a restricted set of multi-word utterances (e.g., "don't know") are retained. Analyses of connected speech often dismiss these as stereotypical, however, these high-frequency phrases are an interactional resource in both neurotypical and aphasic discourse. One approach that can account for these forms is usage-based grammar, where linguistic knowledge is thought of as an inventory of constructions, i.e., form-meaning pairings such as familiar collocations ("wait a minute") and semi-fixed phrases ("I want X"). This approach is used in language development and second language learning research, but its application to aphasiology is currently limited. This thesis applied a usage-based perspective to language processing and intervention in aphasia. Study 1 investigated use of word combinations in conversations of nine participants with Broca's aphasia (PWA) and their conversation partners (CPs), combining analysis of form (frequency-based approach) and function (interactional linguistics approach). In study 2, an on-line word monitoring task was used to examine whether individuals with aphasia and neurotypical controls showed sensitivity to collocation strength (degree of association between units of a word combination). Finally, the impact of a novel intervention involving loosening of slots in semi-fixed phrases was piloted with five participants with NFA. Study 1 revealed that PWA used stronger collocated word combinations compared to CPs, and familiar collocations are a resource adapted to the constraints of aphasia. Findings from study 2 indicated that words were recognised more rapidly when preceded by strongly collocated words in both neurotypical and aphasic listeners, although effects were stronger for controls. Study 3 resulted in improved connected speech for some participants. Future research is needed to refine outcome measures for connected speech interventions. This thesis suggests that usage-based grammar has potential to explain grammatical behaviour in aphasia, and to inform interventions.