Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Fake Diamonds: an EMEG study bridging from semantic composition to semantic cognition
There is a Poster PDF for this presentation, but you must be a current member or registered to attend SNL 2023 to view it. Please go to your Account Home page to register.
Poster B30 in Poster Session B, Tuesday, October 24, 3:30 - 5:15 pm CEST, Espace Vieux-Port
Ryan Law1, Olaf Hauk1, Matt Lambon Ralph1; 1University of Cambridge
During comprehension, concepts are accessed and combined to build complex conceptual structures to support verbal and non-verbal behaviours. The anterior temporal lobes (ATLs) emerge as critical regions in semantic composition (Pylkkänen, 2019, 2020) and semantic cognition more broadly (Rogers et al., 2004; Patterson et al., 2007; Lambon Ralph et al., 2017). The current study aims to import insights across literatures, using conceptual variables like concreteness and specificity to study models of semantic composition and linguistic semantics to study models of semantic cognition. Within the hub-and-spokes (H&S) framework, conceptual content is represented partially by sensorimotor areas, while generalizable coherent concepts emerge when these sources of information are integrated across sources and time by the left and right ATLs (Rogers et al., 2004; Patterson et al., 2007; Lambon Ralph et al., 2017). This framework explains neuropsychological and imaging data. Concrete relative to abstract concepts, and less relative to more specific concepts, are more robust against ATL deterioration (Rogers & Patterson, 2007; Woollams et al., 2008; Jefferies et al., 2009; Hoffman & Lambon Ralph, 2011) and elicit stronger ATL responses (Gauthier et al., 1997; Grabowski et al., 2001; Bright et al., 2004; Tyler et al., 2004; Dhond et al., 2007; Hoffman et al., 2015; Farahibozorg et al., 2022). The neural bases of semantic composition has been studied using two-word phrases. When two concrete words form a phrase, left ATL activity increased at around 200-250 ms post second-word onset (Pylkkänen, 2019, 2020). Crucially, when modulated by conceptual specificity, composition effects were only elicited in phrases with a low-specificity (e.g., blue boat) relative to high-specificity noun (e.g., blue canoe) (Westerlund & Pylkkänen, 2014; Zhang & Pylkkänen, 2015). Recent work on composition varying denotational semantics (e.g., current president is a president, but fake president is not) by Fritz and Baggio (2020) showed EEG sensitivity: increased negativity around 500-800ms post second-word onset. We address three research questions here. First, we ask if the composition engine is engaged for phrases denoting concrete and abstract concepts (e.g., bad costume vs. bad success). If so, we predict ATL compositions effects for both. Second, we probe how neural representations of phrases are transformed across time as a function of denotational semantics (e.g., blue diamond vs. fake diamond) using spatiotemporal decoding (King & Dehaene, 2014). Third, we directly modulate specificity (e.g., bird vs. nocturnal bird vs. owl) and ask if the neural processing of compositional semantics mirrors that of activating more specific concepts. One prediction stemming from the H&S framework is ATL activity increases as a function of increasing specificity. We plan to record electro- and magnetoencephalographic (EEG and MEG) activity as participants read phrases that differ in concreteness (Q1), in denotational semantics (Q2), and conceptual specificity (Q3). Within a contrast, we matched stimuli on a range of psycholinguistic features. To encourage semantic processing, we opted for comprehension questions as a parallel task suitable for all our conditions and facilitate comparisons. We are planning to pre-register this study, recruiting 36 participants, with 100-140 trials per condition to power our study.
Topic Areas: Meaning: Lexical Semantics, Syntax and Combinatorial Semantics