Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Neural and Experiential Semantic Correlates of the Unergative-Unaccusative Verb Distinction
Poster B31 in Poster Session B, Tuesday, October 24, 3:30 - 5:15 pm CEST, Espace Vieux-Port
Songhee Kim1, Jeffrey Binder1, Jiaqing Tong1, Stephen Mazurchuk1, Joseph Heffernan1, Lisa Conant1; 1Medical College of Wisconsin
Introduction. Two types of intransitive verbs, unergative and unaccusative, are distinguished by their argument content and syntactic behavior. Unergatives (e.g., WALK) have an agentive subject argument that actively initiates the verb action (SHE WALKED). Unaccusatives (e.g., FALL) have a non-agentive grammatical subject that is underlyingly a direct object of the action (THE TREE FELL). Unaccusatives are thought to entail additional processing related to derivation of the grammatical subject from the underlying object. Prior fMRI studies have generally shown stronger activation for unaccusative verbs, though the regions implicated varied substantially (Shetreet et al., J Cogn Neurosci, 2010; Agnew et al., Lang Cognition Neurosci, 2014; Meltzer-Asscher et al., Brain Lang, 2015). This inconsistency could be due to the small number of verbs in each study, potentially increasing the influence of random stimulus variables. While observed differences are usually attributed to the additional syntactic movement required for unaccusatives, a substantial literature attests to differences in semantic content between unergatives and unaccusatives. In this study we measured activation differences between these verb types using a larger set of verbs than in prior studies. We hypothesized that some of these differences could be due to systematic differences in semantic content between the two verb classes, and tested this hypothesis by characterizing these semantic differences in terms of a set of neurobiologically-defined experiential features. Method. 86 English intransitive verbs (44 unergative, 42 unaccusative) were rated for 69 experiential features using crowdsourcing. The feature set (Binder et al., Cogn Neuropsychol, 2016) encompasses sensory, motor, temporal, spatial, affective, social, attentional, and event structure phenomena. Participants rated the degree to which they think each type of experience is relevant to the meaning of a given verb. The fMRI study was conducted with 18 healthy participants, who read the verbs 4 to 6 times in a fast event-related design, indicating how familiar they are with the event described by the verb. Activation differences between unergative and unaccusative verbs were contrasted at the group level with cluster-based FWE correction (alpha < .05) determined by permutation testing. Results. Increased activation for unergative relative to unaccusative verbs was found in left angular gyrus, posterior cingulate/precuneus, and medial prefrontal cortex. Left IFG showed greater activation for unaccusative verbs. The two categories differed significantly in their ratings on many experiential features. Unergatives were rated higher on 21 features related to i) human face or head movements, ii) body and limb movements, iii) intentionality and agency, iv) sound generation and perception, and v) communicative and mental phenomena. With these features included as covariates in the unergative-unaccusative contrast, no residual activation differences were observed, suggesting that they account for much of the observed differences in activation. Summary. The different direction and localization of the intransitive verb class effect in our results might be due to our larger stimulus set and use of a task focused on concept retrieval rather than syntactic processing. Our results suggest that the unergative vs. unaccusative distinction in the brain might be explained by their covarying experiential content.
Topic Areas: Meaning: Lexical Semantics, Syntax and Combinatorial Semantics