Subject relative clause preference in Basque: ERP evidence

• Published in: Brain and language Vol. 258; p. 105475
• Main Authors: Zimnukhova, Svetlana, Santesteban, Mikel, Zawiszewski, Adam
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier Inc 01.11.2024
• Subjects: Adult; Basque; Brain - physiology; Electroencephalography; Ergativity; Evoked Potentials - physiology; Female; Humans; Language; Male; Morphological markedness; Relative clauses; Subject-first preference; Word order; Young Adult; Basque; Subject-first preference; Word order; Morphological markedness; Relative clauses; Ergativity
• ISSN: 0093-934X; 1090-2155; 1090-2155
• DOI: 10.1016/j.bandl.2024.105475

•Basque prenominal and postnominal relative clauses are processed differently•Basque postnominal relative clauses processing supports Linear Distance account•ERP evidence on Basque postnominal relative clause processing shows the influence of ergative-marked subject-first preference•Absolutive case is not the preferred option in relative clause syntactic processing in Basque, an ergative language. Subject-object processing within relative clause (RC) attachments exhibits cross-linguistic asymmetries influenced by various factors, including filler-gap linear or structural distance, morphological case marking, and subject-first preferences (Lau & Tanaka, 2021). In the Basque language, filler-gap linear distance and morphological case marking have been posited as explanatory factors for the observed object relative clause (ORC) preference in prenominal RCs (Carreiras et al., 2010). However, recent studies by Yetano et al., (2019) have identified a behavioral preference for subject relative clause (SRC) constructions in Basque postnominal RCs. To ascertain the primary determinant impacting RC processing, we employed EEG signatures to scrutinize subject-object preferences in temporally ambiguous Basque postnominal RCs. Analysis of event-related potentials (ERPs) unveiled a SRC preference: ORCs elicited augmented negative (LAN: 200–400 ms) and positive (P600: 700–900 ms) components compared to SRCs. Our findings suggest that preferences in RC disambiguation are predominantly shaped by filler-gap linear distance and/or subject-first bias.