Pupillometry and filler gap dependencies

Fernandez, Leigh (2013) Pupillometry and filler gap dependencies. Doctoral thesis, Northumbria University.

Text (Doctoral Thesis)
fernandez.leigh1_mphil.pdf - Submitted Version

Download (1MB) | Preview


In this thesis, pupillometry was used to investigate how the parser comprehends filler gap dependencies. It has been shown that the parser automatically searches for and forms dependencies as soon as a gap is encountered. The parser utilizes verb information when positing upcoming gaps; however, in languages in which the verb occurs late in the sentence (e.g. Japanese) it has been shown that pre-verbal information is used to predict upcoming gaps. In this thesis a series of studies was conducted investigating filler gap dependencies. These dependences are typically unbounded in length; however there are some exceptions known as islands, which do not allow a filler and a gap to be associated. There are also exceptions to island constraints known as parasitic gaps. The first set of studies investigates processing costs involved with parasitic gap-like constructions in order to adjudicate between competing accounts of island constraints. The data suggests that the parser actively searches for gaps and forms dependencies as soon as possible (regardless of semantic fit). The second set of studies in this thesis investigates pre-verbal gap prediction in a verb-medial language (i.e. English). The results suggest that the parser consults transitivity information when positing upcoming gaps. These data support theories of grammar in which the parser forms the simplest (syntactic) interpretation in line with good-enough theories, in order to minimize processing costs when forming filler gap dependencies.

Item Type: Thesis (Doctoral)
Uncontrolled Keywords: island clauses, parasitic gaps, eye tracking
Subjects: C800 Psychology
Department: Faculties > Health and Life Sciences > Psychology
University Services > Graduate School > Master of Philosophy
Related URLs:
Depositing User: Ellen Cole
Date Deposited: 02 Jun 2014 08:32
Last Modified: 03 Oct 2022 15:30
URI: https://nrl.northumbria.ac.uk/id/eprint/16487

Actions (login required)

View Item View Item


Downloads per month over past year

View more statistics