Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Getting to the 'root' of semantic and syntactic processing of morphologically complex words: MEG evidence from Arabic
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 A37 in Poster Session A, Tuesday, October 24, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
Samantha Wray1,5, Suhail Matar2,5, Sherine Bou Dargham3, Linnaea Stockall4,5, Alec Marantz3,5; 1Dartmouth College, 2Basque Center on Cognition, Brain and Language, 3New York University Abu Dhabi, 4Queen Mary University of London, 5New York University
[Intro] Crosslinguistically, syntactic and semantic processing has been shown to occur across different time spans and using different neural bases. The dissociation is exemplified in single-word reading of nonwords which are ill-formed due to mismatch between an affix and a stem’s syntactic category (e.g. resend but *re-soft) and those which are ill-formed due to semantic incompatibility between stem and affix (e.g. *re-smile). The latter has been found to elicit greater activity in orbitofrontal cortex compared to the former during a time window of 350-500ms after word reading (see Stockall et al. (2019) for English and Neophytou et al. (2018) for Greek). Unlike languages such as English which primarily utilize concatenative morphology in which affixes are appended to cohesive, contiguous stems, Arabic and other Semitic languages are characterized by non-concatenative morphology: noncontiguous roots interleave with noncontiguous patterns (e.g. k-t-b ‘writing’ + ma--a- ‘place’ = ‘office’). Arabic roots are traditionally analyzed as being underspecified; a root may appear in verbs more often than nouns, but does not receive either its category or full semantic information until after interleaving with a pattern. This allows us to investigate if dissociable processing for semantics and syntactics in morphologically complex words is truly attributable to separability of semantics and syntactics, or if these distinct processing stages are consequences of more underspecified distributional properties. [METHOD] N=18 Arabic speakers participated in a visual lexical decision task with concurrent magnetoencephalography. In addition to reading grammatical, attested words, speakers read two types of nonwords: Syntactic Violation, in which roots only attested in tandem with nominal patterns were interleaved with a verbal pattern (e.g. ʕ-q-r-b ‘scorpion’ + ta-a--a- ‘passive/reflexive’ = ‘*scorpioned’); and Semantic Violation, with roots attested in verbs but interleaved with an unattested verbal pattern that indicates a passive or reflexive reading (e.g. z-ɣ-r-d ‘trill’ + ta-a--a- ‘passive/reflexive’ = ‘*was trilled’). [RESULTS] Spatiotemporal cluster-based regressions at 10,000 permutations were performed to determine significant clusters of activity in time and space within a window of 300-500ms after stimulus presentation. Three clusters were identified, all in orbitofrontal cortex: in the left hemisphere, Semantic Violation items elicited significantly more activity than Syntactic Violation items from 426-492ms (p < 0.01), and in the right hemisphere, Semantic Violation items elicited more activity than Syntactic Violation items from 421-478ms (p = 0.11) and from 479-500 (p = .15). These results suggest that the processing stages of morphologically complex words which have previously been claimed to be sensitive to syntactic category (Manouilidou & Stockall 2014; Schreuder & Baayen 1995) may actually be picking up on more distributed statistical properties of morphosyntax, such as which syntactic category is the root/stem likely to emerge in. These results thus not only inform us about how syntactic and semantic information is processed during word recognition but is also compatible with theoretical linguistic accounts that posit distribution of semantic and syntactic properties across the grammar such as Distributed Morphology (Halle & Marantz 1994), as opposed to the bundling of these properties within lexical items.
Topic Areas: Morphology, Syntax and Combinatorial Semantics