| Summary: | The present thesis focuses on the claims of the PR-First Hypothesis (Grillo and Costa, 2014) about parsing. The PR-First Hypothesis intended to account for the relative clause attachment asymmetry. Its claim regarding attachment was that pseudorelatives created a methodological confound and, when accounting for pseudorelative availability, the same uniform locality driven parsing preferences should be observed across languages. Chapter 1 will present the relevant concepts of the thesis and establish its goals. In chapter 2, I discuss The PR-First hypothesis, the syntactic status and semantic preferences of pseudorelatives and a novel proposal for the availability of pseudorelatives in European Portuguese where pseudorelatives present a definiteness requirement. Chapter 3 presents an acquisition study in European Portuguese assessing the claim that the parser is uniform and continuous from childhood into adulthood. Results from a picture verification task show that children as early as 3 years of age display sensitivity to the same linguistic cues as adults when deciding attachment. Chapter 4 shows an ERP study in Italian looking at the electrophysiological response of ruling out the pseudorelative interpretation in favour of a relative clause interpretation. The goal was that the type of response to the violation would inform on the underlying principles that drive the pseudorelative preference. Results were inconclusive which created an asymmetry with the previous offline data conducted within the same hypothesis and which supported the PR-First Hypothesis. Chapter 5 presents two acceptability studies in Italian that try to account for the asymmetry in results between online and offline methods. I hypothesise that syntactic adaptation due to the uneven ratio between pseudorelatives and relative clauses can account for the asymmetry. The first acceptability study shows that the 1:3 ratio in favour of Relative clauses triggers an adaptation effect that throughout the experiment dissipates the initial gap in acceptability between pseudorelatives and Relatives. The second acceptability study shows that reducing the uneven ratio from 1:3 to 2:3 offsets the rate of the adaptation effect taking longer for the initial gap in acceptability between pseudorelatives and Relatives to dissipate. Results show that the parser is tracking the conditional probability of syntactic structures and trigger an adaptation effect that hides the parser’s natural preferences. Chapter 6 summarizes each chapter and discusses their conclusions.
|
|---|