Phi-features reloaded: An event-related potential study on person and number agreement processing

Abstract

Linguistic analysis claims that verb agreement is composed of distinct phi-features such as person and number, but are these different phi-features processed distinctly or similarly? We used a sentence grammaticality task to explore the electrophysiological responses of Basque speakers when processing subject–verb person and number phi-feature agreement violations. We generated grammatical structures (grammatical control) and ungrammatical structures in which the verb disagreed with the subject in person (person violation), in number (number violation), or in both person and number features (person+number violation). Behavioral data revealed that, overall, participants were faster and more accurate detecting person and person+number violations than violations involving only number. Event-related potential responses revealed a N400–P600 pattern for all violation types. Person and person+number violations elicited larger P600 effects than number violations. These findings reveal different costs related to the processing of person and number phi-feature agreement and indicate that these features are distinct components of agreement computation.

Effects of subject-case marking on agreement processing: ERP evidence from Basque

“…Meanwhile, the observation of an early posterior negativity in the present study is also consistent with earlier findings in Basque (Zawiszewski & Friederici, 2009;Zawiszewski et al, 2016). Given its posterior distribution, this effect is likely distinct from the left-anterior negativity (LAN) that has been reported in earlier studies in other languages and might be more similar to the N400.…”

“…This positivity is sometimes found to be preceded by an increased negativity between 300-500 ms after stimulus onset. This negativity tends to have an anterior distribution and is at times found to be left-lateralised (a LAN; Caffarra, Siyanova-Chanturia, Pesciarelli, Vespignani, & Cacciari, 2015;De Vincenzi et al, 2003;Gunter, Friederici, & Schriefers, 2000;Münte, Heinze, & Mangun, 1993;Osterhout & Mobley, 1995;Silva-Pereyra & Carreiras, 2007), although in some cases it has also been found to be largest at central-posterior sites and has been considered an N400 effect instead (Barber & Carreiras, 2003;Guajardo & Wicha, 2014;Mancini et al, 2011;Zawiszewski & Friederici, 2009;Zawiszewski et al, 2016).…”

“…More recently, Zawiszewski et al (2016) examined participants' ERPs as they read transitive sentences in which the subject and the verb failed to agree in their (i) person feature, (ii) number feature, and (iii) person and number features. They found that all three types of S-V agreement violations elicited a posterior negativity followed by a P600 effect relative to the grammatical control.…”

“…Previous research has looked at different languages to examine and compare how the brain processes agreement that (i) involves different features, including number (De Vincenzi et al, 2003;Hagoort, Brown, & Groothusen, 1993;Kutas & Hillyard, 1983), gender (Barber & Carreiras, 2005;Hagoort & Brown, 1999), person (FrenckMestre, Osterhout, McLaughlin, & Foucart, 2008;Hinojosa, Martín-Loeches, Casado, Muñoz, & Rubia, 2003) as well as comparisons across features (Alemán Bañón, Fiorentino, & Gabriele, 2012;Alemán Bañón & Rothman, 2016;Barber & Carreiras, 2003, 2005Hagoort, 2003;Mancini, Molinaro, Rizzi, & Carreiras, 2011;Martín-Loeches, Nigbur, Casado, Hohlfeld, & Sommer, 2006;Nevins, Dillon, Malhotra, & Phillips, 2007;Silva-Pereyra & Carreiras, 2007;Zawiszewski, Santesteban, & Laka, 2016) and (ii) occurs between different constituents (Barber & Carreiras, 2005;Díaz, Sebastián-Gallés, Erdocia, Mueller, & Laka, 2011;Zawiszewski & Friederici, 2009).…”

Who are you talking to? The role of addressee identity in utterance comprehension

“…Two types of negative effects around 400 ms have been reported in response to a range of morphosyntactic violations (Molinaro, Barber, & Carreiras, ; see Caffarra, Mendoza, & Davidson, for a discussion on how distinguishable these two effects are): N400 (centroposteriorly distributed) and LAN (left anteriorly distributed) effects. Previous ERP studies on agreement violations similar to those examined here (person agreement violations in Basque) have consistently shown N400 effects followed by P600 effects (Mancini, Massol, Duñabeitia, Carreiras, & Molinaro, ; Zawiszewski, Santesteban, & Laka, ; see also Mancini, Molinaro, Rizzi, & Carreiras, for similar effects elicited by person anomalies in Spanish, and Molinaro et al, for a review), indicating an early stage of analysis possibly relying on information at the interface between morphosyntax and semantics/discourse followed by late controlled processes where higher‐level information is considered (Molinaro et al, ).…”