Search Abstracts | Symposia | Slide Sessions | Poster Sessions
A Merge-Based Approach to Early Syntactic Negativities
Poster Session C, Saturday, September 13, 11:00 am - 12:30 pm, Field House
Kevin Guzzo1, William Haddican1, Yasuaki Shinohara2, Valerie Shafer1; 1City University of New York, 2Waseda University
This research presents interdisciplinary experimental work at the intersection of neuroscience and syntax, aiming to illuminate language processing mechanisms through the integration of theoretical linguistics, behavioral data, and electrophysiological evidence. To achieve this, we focus on Early Syntactic Negativities (ESNs), which are rapid negative electrophysiological responses known to be elicited by phrase structure violations at the onset of unexpected lexical categories. Accounting for the unique properties of ESNs has inspired two distinct approaches. First, Friederici’s (1995) “Syntax First” model explains the early latency as a reflection of the automatic processing of lexical information in the first stage of its three-stage model. In contrast, Dikker’s (2010) Sensory Hypothesis attributes ESNs to ‘form-based’ prediction errors in the sensory cortex, rather than to syntactic violations. In this work, we develop a novel approach to ESNs. Our hypothesis posits a direct link between ESNs and violations of the operation Merge the sole structure-building operation in the minimalist program (Chomsky 1995). The approach thereby posits a direct electrophysiological correlate of a class of formal representational violations. A key component of our approach is the unification of two previously disparate experimental paradigms used to elicit ESNs in the literature, which we demonstrate are subsumed under a single syntactic distribution predicted by the Merge hypothesis. Furthermore, our hypothesis generates an additional behavioral prediction: a sentence with a Merge violation will be significantly less acceptable than a sentence containing other types of violations. In two experiments, we examine ESNs elicited from previously undocumented elliptical constructions, specifically sluicing and swiping, using event-related potentials (ERPs) and behavioral acceptability judgments. We recruited eighteen native speakers of North American English as participants, all of whom had no speech or language deficits and normal or corrected-to-normal vision. Each experimental sentence was visually presented in two parts. The first part consisted of the complete presentation of the initial conjunct containing the antecedent for ellipsis. Participants pressed the spacebar to proceed to the second part, containing one of three ellipsis conditions. Following a 300ms fixation cross, the second conjunct appeared word-by-word, each for 300ms. Upon completing the sentence, participants rated its acceptability on a seven-point Likert scale. Our experimental design contrasted violations of ellipsis constraints with Merge violations, systematically manipulating grammaticality within sluicing and swiping constructions. For example, the sentence “The receptionist received something from some organization,” was completed with one of the following: (a) “but I don’t know from what organization, honestly” (Grammatical Baseline); (b) “but I don’t know what organization from, honestly “(Ellipsis Violation); or (c). “but I don’t know what from organization, honestly” (Merge Violation). Our results revealed significant differences in the ESN effect for sentences containing a Merge violation compared to both grammatical and ellipsis violation conditions (p < .05). Additionally, we observed significantly lower acceptability ratings for sentences with Merge violations in comparison to those with ellipsis violations. Our findings suggest compatibility between the Merge Hypothesis and Dikker’s Sensory Hypothesis, as Merge violations were indexed by an ESN recorded at both inferior anterior and inferior posterior sites.
Topic Areas: Syntax and Combinatorial Semantics, Morphology