Active Assignment of Quantifier Scope Guides Language Processing

Orth, Wesley James

doi:https://doi.org/10.21985/n2-p51t-nv93

Work

Active Assignment of Quantifier Scope Guides Language Processing

Public

Download PDF

Download All Files (.zip)

This dissertation investigates how scope relations are constructed and evaluated during real-time human sentence processing. Theoretical approaches to processing scope relations exist in a multi-dimensional space where trade-offs are made around how quickly scope relations can be computed, how many mistakes are made in computing scope relations, and how many different scope relations can be produced. Existing approaches are organized around the idea of minimizing mistakes in computing scope relations, and achieve this by sacrificing incrementality or by avoiding the calculation of scope relations. In this dissertation, I propose a new approach called the active scope model. In the active scope model, the processing of scope is organized around ensuring the parser can always produce a possible scope relation and can do so quickly in a real-time language processing context. The active scope model is also directly rooted in the grammatical concepts of Quantifier Raising and Scope Economy, and for this reason makes different predictions about processing behavior than other approaches. By investigating the processing of sentences with multiple quantifiers and competing constraints on scope, this dissertation determines that the active scope model appears to most accurately predict the observed processing profile. A major series of investigations also reveals that the distributional properties of the negative polarity item illusion are also predicted by the active scope model. The active scope model also predicts a novel illusion effect for positive polarity items, which is observed for the first time in this dissertation. The data presented in this dissertation constitutes a major expansion of the empirical landscape of polarity illusions, and supports an analysis of a single generalized polarity illusion which is the result of scope calculation during online sentence processing. With this supporting evidence, this dissertation argues that the real-time processing of scope relations is best captured by the active scope model proposed in this dissertation. As the active scope model is rooted in the grammatical concepts of Quantifier Raising and Scope Economy, this dissertation also proposes that these are necessarily components of the syntactic portion of grammar and that scope relations are calculated in the syntactic module of grammar rather than a later interpretive module.

Creator