Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Automatic morphosyntactic processing in the brain: ERP evidence
Poster Session C, Thursday, October 1, 10:45 am - 12:45 pm, Wangari Maathai
Marije Soto1, Leonardo Cabral1; 1Federal University of Rio de Janeiro (Brazil)
In this EEG study, we investigated whether morphosyntactic features are accessed and employed for combinatorial processing during the comprehension of phrases in Brazilian Portuguese. Also, we wanted to gauge the automaticity of morphosyntactic parsing. In order to do so we observed whether the syntactic mismatch negativity (sMMN) (Pulvermüller & Shtyrov, 2003, 2006) would be modulated in its amplitude according to an increase in number of grammatical violations in the phrases included in the experiment. Thus, we applied a multifeature odd-ball paradigm (Näätanen et al, 2004), comparing 3 levels of (a)grammaticality. 23 participants listened to a grammatical phrase (“eu adot-o”, I adopt.1p.sg), a phrase with 1 violation (“eu adota*”, I adopt.3p.sg) and a phrase with 2 violations (“eu adot-am”, I adopt.3p.pl), while watching a muted episode of a tv show. We hypothesized that the sMMN effect is proportionate to the number of violations, i.e. larger for the 2 violations condition than for the 1 violation, as a reflection of representation of morphosyntactic abstract features and their processing through parsing mechanisms. Results presented a complex picture: we found a cumulative effect (i.e., higher amplitude evoked by the 2 violations condition compared to the 1 violation condition) of negative amplitude at left anterior electrodes during the first interval, 150-250ms after the critical morpheme inflection. This cumulative negative response is sustained, also observed between 250-450ms, albeit less lateralized and more frontal. The contrast between violation and grammatical conditions presents a topographically widespread negative response. As topographical and latency characteristics of the response seem different from those associated to the MMN, and as the standard stimulus from blocks 3 and 4 (the phrase from the “1 violation condition”) evoked higher negativity than the grammatical phrase as a deviant, we consider the possibility that we actually found a directly morphosyntactic brain response that is automatic and independent of attention, even though we cannot rule out a contribution from the MMN effect. Syntactic processing in a non-oddball context of distraction is not unheard of, as studies have already reported such automatic response (e.g., Hasting & Kotz, 2008; Alekseeva et al., 2024). However, because these studies did not vary the amount of morphosyntactic feature violations in each phrase, the effects they reported may also be explained on the basis of probability violation: a verb or noun that disagrees with its preceding pronoun is both agrammatical and unexpected. Because the violation conditions in our study involve an illegal combination of pronoun and verb, both present verbs that have zero probability of occurring after their correspondent pronouns, with only an increase of morphosyntactic feature violation. Thus, our findings point to a potential sum of processes involving the comprehension of sentences that vary in number of violations, with left anterior activity suggesting that at least a part of these responses can be associated to grammatical processing. The data suggests that morphosyntactic information is accessed and employed in the course of sentence comprehension, with limitations on the study restricting a stronger claim and highlighting the need for replication with improvement of methodology.
Topic Areas: Morphology,