Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Situating the Neural Basis of Adjectival Modification Across Typologically Diverse Languages
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 E49 in Poster Session E, Thursday, October 26, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
This poster is part of the Sandbox Series.
Natalia Bekemeier1, Marianne Hundt1, Lena Jäger1, Fernando Zúñiga2, Xin Sennrich1, Jon Lapresa-Serrano1, James Magnuson3, Nicola Molinaro3, Alexis Hervais-Adelman1; 1University of Zurich, 2University of Bern, 3Basque Center on Cognition, Brain & Language
Analysis of word classes, even across typologically diverse languages, supports the existence of a number of parts of speech (PoS) that have different functional roles, for example: nouns (reference), verbs (predication) and adjectives (modification) (Bisang, 2010). The supposition that these PoS are functionally different has been investigated using the tools of cognitive neuroscience and multiple investigations have distinguished between neural representations of verbs and nouns, at least in prototypical cases of these (Kemmerer, 2014; Vigliocco et al., 2011). However, cross-linguistic comparisons indicate that existence of a distinct PoS Adjective may not be universally supported, or that the nature of adjectives is highly variable across languages (Croft, 2022). Compare, for example Basque, where the syntactic behaviour of adjectives approaches that of nouns, with Mandarin Chinese, where modification is often realised by means of verbs. Most cognitive neuroscience studies on compositional semantics of modifying expressions have been conducted on English (Zhang & Pylkkänen, 2018; Ziegler & Pylkkänen, 2016) or other Germanic languages (Norwegian Bokmål: Fritz & Baggio, 2022, Dutch: Kochari et al., 2021), and assume a priori the existence of a prototypical Adjective category. To determine neurobiological reality of PoS, it is necessary to probe the categorical construct of PoS across typologically distinct languages and to test the possibility of distribution of modifiers along a noun-verb continuum. The present project combines neurolinguistic and computational approaches, with a focus on the word-intrinsic and context-mediated attributes that are presumed to determine their meaning and function. It further seeks to determine whether the linguistic constructs of PoS are predictive of the way the brain processes linguistic stimuli. To advance our present understanding of the representation and organisation of adjectives we plan a cross-linguistic MEG investigation that will (i) elucidate the neurobiological signatures of the core PoS and combinatorial operations involved in phrase structure building in three typologically different languages (English, Basque and Mandarin Chinese) and (ii) generate a cross-linguistic MEG database from naturalistic linguistic stimuli in the auditory modality in these three languages. The study will employ naturalistic stimuli (excerpts from an audiobook). The effects of lexical semantics and local syntactic structure building will be controlled for by means of a Jabberwocky condition (content word roots will be substituted with Jabberwocky pseudowords, while preserving the original sentences’ syntactic structure) and scrambled sentences, respectively. Data will be analysed within and across languages employing: (i) mCCA to ensure optimal spatiotemporal alignment of the data across individuals and (ii) cross-validated regression of the source-localised signals (Arana et al., 2020; Heilbron et al., 2022; Huizeling et al., 2022). We will evaluate multiple encoding models that will include theoretical PoS predictors as well as metrics from language-specific embedding models. Across languages, the models that incorporate contextual predictors, such as predictability and surprisal, are expected to demonstrate the most consistent fit. The models that include syntactic structure vs. lexical root distribution predictors should fit differentially as a function of the linguistic structure of individual languages. Our results will promote insight into how PoS information contributes to the phrase structure building.
Topic Areas: Syntax and Combinatorial Semantics,