Subject-verb agreement in real timeactive feature maintenances as syntactic prediction

  1. Bojana Ristic
Supervised by:
  1. Nicola Molinaro Director
  2. Manuel Francisco Carreiras Valiña Director
  3. Simona Mancini Director

Defence university: Universidad del País Vasco = Euskal Herriko Unibertsitatea

Year of defence: 2020

  1. Patrick Sturt Chair
  2. Javier Ormazabal Zamakona Secretary
  3. Isabel Fraga Carou Committee member

Type: Thesis


The current dissertation tests whether the long-distance subject-verb establishment is maintained active over the course of the sentence, by maintaining morphosyntactic information such as syntactic category and number features. To this end, we looked at how the maintained representation affects the interpolated elements, focusing on two effects that the maintained features might generate: similarity-based interference and disambiguation. We performed four eye-tracking experiments (reading and visual world paradigm) and showed that subject-verb dependency establishment is characterized by active maintenance of the subject¿s category feature (English and Spanish experiments) and number feature (Basque experiments). Our effects, which occur prior to the integration site (the verb), can be ascribed to the top-down pre-activation mechanisms and thus syntactic prediction. Importantly, this implies that subject-verb agreement occurs in real-time sentence comprehension, i.e. it is psychologically real.