Corpus-based transitivity biases in individuals with aphasia

ABSTRACT

Background: This study investigated whether individuals with aphasia (IWA) retain verb biases in expressive language. Verb biases refer to the likelihood that a given verb will occur in different sentence structures. We focused on the likelihood of verbs occurring in transitive and intransitive structures.

Aims: The main goal of this study was to determine whether IWA and controls show similar verb biases or whether IWA show a preference for transitive or intransitive structures that supersedes individual verb biases. We also investigated whether IWA show a preference for intransitively or transitively biased verbs, whether verb biases differ as a function of aphasia type, and how verb bias affects errors in IWA’s speech production.

Methods & Procedures: The current study analysed 236 transcribed interviews of IWA from AphasiaBank. All uses of 54 verbs were coded based on the sentence structure and the presence of errors. We report data from 11 transitively biased and 11 intransitively biased verbs.

Outcomes & Results: IWA’s transitivity biases were indistinguishable from controls’ biases. In addition, IWA produced more intransitively biased verbs than transitively biased verbs overall. In ungrammatical productions, IWA’s error rates were higher in sentence structures that conflicted with verb bias and highest when an intransitively biased verb was attempted in a transitive structure.

Conclusions: These findings indicate that IWA are sensitive to verb bias and verb complexity within expressive language. The effects are consistent with previous literature concerning IWA’s sensitivity to verb bias in receptive language tasks and to verb complexity in verb retrieval tasks.

KEYWORDS:

• Aphasia
• transitivity
• verb bias
• expressive language
• AphasiaBank

Acknowledgements

We would like to thank the AphasiaBank consortium for providing the database for this research. We would also like to thank members of the Undergraduate Biology Research Program (UBRP) and the Speech Language and Brain Lab for their support.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplemental data for this article can be accessed here.

Notes

1. The verb start was coded as an intransitively biased verb due to a calculation error, though it did not meet the criteria in Gahl et al. (2004) data. The current data were analysed with and without the verb start. The results did not change in the two sets of analyses. Results reported here include the verb start so as to maintain the balance between the sets.

2. We used the term “prepicle” to refer to a word before classifying it as preposition or particle.

Additional information

Funding

This work was supported by the HHMI under Grant 520006942 to the Undergraduate Biology Research Program at the University of Arizona and NIDCD Grant K23DC010808.