Cortical Underpinnings of Conceptual Composition and Lexical Polysemy: The Many Meanings of English have
Poster D46 in Poster Session D with Social Hour, Friday, October 7, 5:30 - 7:15 pm EDT, Millennium Hall
Maria Mercedes Pinango1, Muye Zhang1, Cheryl Lacadie1; 1Yale University
Phenomenon: English speakers readily accept ambiguous sentences like "the maple tree has a car" with a locative meaning, if followed by facilitatory locative contexts like "there is a motorcycle under the pine tree". This indicates that English 'have' can express not only possession but also location without the support of a prepositional phrase, raising the question of have’s meaning encoding given its polysemy(1,2,3). Question: Is 'have''s meaning encoded in the brain, as measured through fMRI, as a unified conceptual structure or as two 'haves', one for possession and another for location, the latter expressed through syntactic cues? This leads to competing hypotheses (2): (I) Unified Meaning Hypothesis: Comprehension of 'have' involves the activation of a parametrized conceptual space determined by the degree of causality perceived from linguistic and contextual cues, and from which possession-location readings emerge. If so, comprehension of NP-have-NP sentences involves (a) lexical-semantic composition where the first NP is composed with the lexico-semantic conceptual structure of have and (b) conceptual contextualization when the complement is composed determining the specific relation (possession vs. location) between the two NP entities. (II) Two-Meanings Hypothesis: Comprehension of 'have' involves the implementation of a syntactic repair operation triggered by failure of possessive interpretation. Syntactic-repair inserts a locative-PP (overt or covert) to support the alternative, normally dispreferred, locative meaning. Methods: Thirty native English speakers (16 female, ages 18-29). 100 target NP-have-NP sentences followed either a facilitatory locative context (50_context+sentence) or a non-facilitatory possessive context (50_context+sentence) plus 50 fillers. Context+sentence pairs analyzed as two events: Event_1:context+subject of target sentence; Event_2:have+NP-complement. Subtraction: Event_2 "has a car that is red" after location_context minus Event_2 "has a car that is red" after possession_context. Event_2 is the segment where location interpretation is determined. Crucially, the Event_2 subtraction involves identical linguistic material. Accordingly, any neural activation differences observed are due to the contexts (location_v_possession). Contexts are captured, separately, in Event_1. The Unified Meaning hypothesis predicts preferential activation for Event_2 of the left posterior and of the prefrontal cortex previously reported as implicated in lexico-semantic composition and contextualization (4). In addition, given the high context dependence of have-interpretation, we expect that Autism Quotient (AQ) scores (5), our measure of variability in linguistic meaning context-sensitivity, will correlate with individual participants’ activation magnitudes in the Event_2 contextualization operation. The Two-Meanings Hypothesis predicts preferential activation of the LiF cortex reflecting morphosyntactic composition, and no interaction with context-sensitivity metrics, as syntactic repair is only dependent on presence/absence of a prepositional-phrase. Findings: The key subtraction Locative>Possessive associated with Event_2 revealed bilateral angular gyrus, left supplementary motor area, precuneus and right frontal cortex_(BA 8) activation consistent with lexico-semantic composition and contextualization processes (p<.001, corrected). AQ scores correlated in opposite ways for the composition and contextualization areas, suggesting that context matters for the interpretation of the target sentence. Conclusions: Brain activation patterns of meaning of English 'have' are consistent with the Unified Meaning hypothesis, suggesting a compositional role for conceptual structure independent of syntactic composition and grounding cases traditionally viewed as lexical polysemy. References: 1:Zhang et al. 2018;2:Zhang,2021;3:Zhang_et_al.2022;4:Lai_et_al.2017;5:Baron-Cohen_et_al.2001.
Topic Areas: Meaning: Lexical Semantics, Meaning: Combinatorial Semantics