Slide Slam A14
Indeterminate sentences in aphasia: investigating coercion and the nature of compositionality
Caitlyn Antal1, Alexa Ray Falcone1, Laura Pissani1, Kyan Salehi1, Roberto G. de Almeida1; 1Concordia University
INTRODUCTION Sentences such as “Mary began the book” are called indeterminate because they do not make explicit what the subject (Mary) began doing with the object (the book). These types of sentences have generated much interest because they represent a case study for a central issue in language representation and processing: compositionality. Specifically, (1) Is semantic composition simple (classical) or is it enriched with intended or implicit constituents? And (2) what is the nature of the linguistic and cognitive resources involved in the interpretation of the event that the sentence conveys? There have been at least two proposals for how the meaning of an indeterminate sentence is attained. One assumes classical compositionality, with much of the interpretation of the sentence being the product of pragmatic inferences (e.g., Fodor & Lepore, 2001; de Almeida & Riven, 2021; de Almeida & Lepore, 2018) triggered by a syntactic gap ([began [v [the book]]; de Almeida & Dwivedi, 2008). An alternative view assumes that some form of local semantic enrichment takes place—often via what is called “coercion” or “type-shifting” (e.g., Pustejovsky, 1995, 2011; Asher, 2015). Coercion relies on internal analyses of the noun complement yielding an enriched form of compositionality (viz., [begin the book]→[begin reading the book]). Thus far, the only study investigating this phenomenon in aphasia supported semantic coercion based on greater difficulty by “Wernicke’s patients” in understanding indeterminate sentences (Piñango & Zurif, 2001). We investigated the coercion hypothesis in a group of 14 individuals with aphasia from different etiologies, and with lesions in either the left or right hemisphere. METHOD Our sample included 5 non-fluent [NF], 4 fluent [FL], 3 mixed but predominantly non-fluent [MN], 2 with mixed aphasia [MX], and 41 healthy controls. Participants completed a sentence-picture matching task whereby a sentence was aurally presented, immediately followed by two pictures on a computer screen. Their task was to choose the picture that best represented the sentence they heard. Sentences were: (a) indeterminate (The academic began the research), (b) fully determinate (“preferred”: …conducted the research), (c) figurative (viz., in need of pragmatic enrichment: …dumped the research), or (d) determined but non-preferred (…abandoned the research). Only one picture was the correct choice for indeterminate and fully determinate sentences. The competing picture was the correct choice for the figurative and non-preferred sentences. RESULTS Repeated-measures ANOVAs revealed a main effect of group, sentence type, and an interaction. Overall, group analyses showed that, when compared to controls, NF individuals had significantly more difficulty choosing the correct picture when presented with indeterminate sentences. Further, pairwise comparisons revealed that FL, NF, and MX individuals performed worse with indeterminate sentences than fully determinate sentences. Results from case-series analyses will be presented. DISCUSSION We propose that indeterminate sentences may be resolved by a syntactic-gap detection and by pragmatic inferences. We take the difficulty with indeterminate sentences shown by the NF group (compared to controls) to suggest that they have problems computing the syntactic gap that may serve to trigger a search for an appropriate event during semantic composition.