Differences in the perception and time course of syntactic and semantic violations

Abstract

A reading time and an ERP experiment conducted in Italian investigated the parser's responses to a syntactic violation (subject-verb number agreement) and to a semantic violation (subject-verb selectional restriction), examining the time course of comprehension processes until sentence end. The reading-time data showed that the syntactic violation was detected earlier than the semantic one and that the two violations differed in the time-course. The ERP data fully supported the reading time data: Syntactic anomalies elicited a left anterior negativity (LAN) and a P600. Semantic anomalies elicited a N400 centred on the parietal sites which started 90 ms later (latency 430 ms) than the LAN. Furthermore, the N400 evoked by the words that followed the target word continued and increased until sentence end. The results are discussed with respect to the hypotheses that the parser constructs distinct syntactic and semantic analyses of a sentence and that this characteristic holds cross-linguistically. The appropriateness of different methodologies to the study of sentence processing is also evaluated.

The brain basis of syntactic processes: functional imaging and lesion studies

“…Processes of thematic role assignment which are based on morphosyntactic information signaling agreement between different phrases or between elements within a phrase are reflected by a left anterior negativity between 300 and 500 ms (LAN), a component which was not investigated in the present set of experiments but in a number of experiments in different languages including English (Coulson et al, 1998), Dutch (Gunter et al, 1997), German , Italian (Angrilli et al, 2002), and Hebrew (Deutsch and Bentin, 2001). Most recently it has been argued on the basis of ERP findings in Italian that the two processing streams indicated as taking place during phase 2 should be modulated with an inherent seriality of morphosyntactic and semantic processes (De Vincenzi et al, 2003). Thematic role assignment in noncanonical sentences requires the identification of long-distance relations between a moved element and its original position in a sentence and/or the manipulation of moved elements during online sentence processing.…”

Agreeing to Disagree: Deaf and Hearing Children's Awareness of Subject–Verb Number Agreement

“…We expected the slowing on agreement errors in the self-paced reading study to be different when the anomalies are based on grammatical versus semantic mismatches, as has been found for adults (Di Vincenzi et al 2003). If deaf children process agreement like hearing children, then we would expect that their response to agreement anomalies would also be different from their response to semantic anomalies.…”

“…Moreover, we anticipated that this grammatical effect would differ from their response to semantic anomalies, as shown in adults (De Vincenzi et al, 2003). It was less clear what level of awareness the RA and deaf children would have.…”

“…We also expected CA children to show sensitivity to agreement errors in the self-paced reading study in a similar manner to hearing adults -slowing down on the anomaly and the word after (De Vincenzi et al, 2003).…”

“…at around 250 -500ms post presentation onset of the verb, followed by enhanced late posterior-parietal positivity at around 600ms (P600, Coulson, King, & Kutas, 1998;De Vincenzi et al, 2003;Osterhout & Mobley, 1995), semantic anomalies are known to elicit central-posterior negativity at around 400ms post-onset, referred to as the N400 effect (De Vincenzi et al, 2003;Kutas & Hillyard, 1983). These effects have also been documented in deaf participants' responses to semantic and syntactic errors in American Sign Language (Capek et al, 2009).…”