Advisor(s)

Neal J. Pearlmutter (1967-)

Contributor(s)

Kathryn Bock, Stephen G. Harkins (1948-), Joanne L. Miller

Date of Award

2011

Date Accepted

7-2011

Degree Grantor

Northeastern University

Degree Level

Ph.D.

Degree Name

Doctor of Philosophy

Department or Academic Unit

Bouve College of Health Sciences, Department of Counseling and Applied Educational Psychology

Keywords

imageability, language production, subject-verb agreement, semantic integration

Subject Categories

Psycholinguistics

Disciplines

Cognitive Psychology | Psychology

Abstract

The study of language production is concerned with how speakers translate non-verbal thoughts into meaningful, grammatical utterances. In all languages, information in one part of an utterance may depend on information that appears in another part of the utterance, and the distance between these parts of the utterance may span an arbitrary amount of intervening material. Subject-verb agreement is an example of a long-distance syntactic dependency in that the number information encoded in the subject and the verb of a sentence must agree. Agreement errors are likely to occur when a singular head noun (e.g., key) of a subject noun phrase (NP) is followed by a plural local noun (e.g., cabinets) in a modifier (e.g., the key to the cabinets *ARE), and this pattern is referred to as the mismatch effect (Bock & Miller, 1991). This dissertation follows the tradition of studies examining subject-verb agreement errors as a way to gain insight into the syntactic planning process.

Most theories of agreement production have suggested that syntactic structure is particularly important for computing agreement relations (e.g., Bock & Cutting, 1992; Eberhard, Cutting, & Bock, 2005; Franck, Vigliocco, & Nicol, 2002). The scope of planning account of agreement (Gillespie & Pearlmutter, 2011) does not rely on syntactic structure and instead suggests that it is the timing of planning of local nouns relative to their head that is im- portant for correct agreement computation. This dissertation examines two major themes. Experiments 1-5 investigated whether syntactic structure constrains agreement computation independent of other lexical properties, and how one particular conceptual property, image- ability, modulates the strength of the mismatch effect. Experiments 6 and 7 examined how another conceptual property, semantic integration (Solomon & Pearlmutter, 2004b), affects the timing of planning of elements within complex NPs, and how differences in timing of planning affect agreement computation.

Earlier studies supporting the scope of planning account did not test the influence of clausal structure on agreement, and there is evidence that subject-verb agreement errors are more likely to follow a complex NP with a prepositional phrase (PP) modifier (e.g., the editor of the history books) than complex NPs with relative clause (RC) modifiers (e.g., the editor who rejected the books; Bock & Cutting, 1992; Solomon & Pearlmutter, 2004b). Experiment 1 examined whether hierarchical syntactic structure constrains subject-verb agreement computation, by contrasting subject NPs with PP and RC modifiers that were matched in many properties that varied in previous experiments (e.g., overall meaning). The mismatch effect was replicated; however, there was no interaction of local noun number and structure. These results suggest that structure is not involved in subject-verb agreement computation during production.

Experiments 2 and 3 ruled out two word-level properties (frequency and transitivity of the embedded verb in RCs) as sources of the difference between Experiment 1 and earlier studies (Bock & Cutting, 1992; Solomon & Pearlmutter, 2004b) that had found support for structural effects. Experiment 4 examined the effects of imageability and syntactic structure to determine whether imageability was a factor contributing to previously observed structural effects. Agreement errors were more common when stimuli were less imageable, but structure had no influence on agreement error rates, replicating Experiment 1. Experiment 5 was a regression meta-analysis of Experiments 1-4 and Solomon and Pearlmutter (2004b; Exp. 5) which showed that error rates increased with decreasing imageability.

Semantic integration, the degree of conceptual linkage between elements within an utterance, has been hypothesized to influence the timing of planning of elements within a phrase, such that highly semantically integrated elements are planned with more overlap than less integrated elements. But the evidence that integration affects timing during production is indirect, as it largely comes from speech error data (Gillespie & Pearlmutter, 2011; Pearlmutter & Solomon, 2007; Solomon & Pearlmutter, 2004b). Experiment 6 examined the effect of semantic integration on temporal separation of words in recordings of stimuli from a subject- verb agreement elicitation study in Solomon and Pearlmutter (2004b). Linear mixed-effect regression modeling showed that temporal separation between the head and local noun de- creased with increasing semantic integration, consistent with and providing the first direct evidence for Solomon and Pearlmutter's claim that more integrated elements are planned with more overlap.

Experiment 7 examined how semantic integration affects utterance planning and subject- verb agreement error rates. Participants described picture displays with NP PP subject noun phrases and then completed the descriptions as full sentences. Semantic integration (Solomon & Pearlmutter, 2004b) was manipulated, and speech onset times and agreement error rates were recorded. Speakers were faster to initiate speech when the head and local noun were integrated than when they were unintegrated. Agreement errors were more likely when the local noun was plural than when it was singular. Supporting the scope of planning account of agreement (Gillespie & Pearlmutter, 2011), speakers who were slower to initiate speech produced more agreement errors, suggesting that when speakers do more advance planning they are more likely to experience interference during agreement computation.

Document Type

Dissertation

Rights Information

Copyright 2011

Rights Holder

Maureen Gillespie


View Full Text

Click button above to open, or right-click to save.

Share

COinS