Macquarie University
Browse

Agents, events, causers, and states

Download (16.48 MB)
thesis
posted on 2025-01-23, 04:35 authored by Margaret Kathleen Ryan

This thesis examines the neuropsychological realisation of syntactic theory pertaining to the comprehension of sentence syntax, functional morphology, and sentence aspect and event structure, in typically- and atypically-ordered transitive experiencer-verb and agentive active and passive sentences without animacy, length, voice, or contextual coercion confounds. Importantly, this thesis allowed for the discovery of variation in thematic roles and sentence aspects and events due to the inclusion of a large number of active and passive experiencer verbs (N=480). In the first experiment, crowdsourced American-English speaking participants rated the degree of intent in the sentence event of subjects and objects of subject-experiencer (SE) and object-experiencer (OE) active and passive sentences, on a 5-point Likert scale. A second group rated the cause. Intentionality ratings demonstrate that both SE and OE sentences comprise stative and eventive subtypes, due to an alternation of the theme or experiencer thematic roles with an intentional agent. The SE result was unique, the OE consistent with syntactic theory. However, the interpretation of an argument as a cause(r) appears to be a sentence aspect rather than a thematic role, increasing the likelihood that a low-prominence thematic role is a sentence subject, since both low intent and high intent subjects were attributed causality. Additionally, despite variations in theoretical predictions for experiencer passives, these showed equivalent antipodean thematic roles and aspects on their ‘by’-phrase objects as on the (antipodean) subjects of their corresponding actives. This study is a unique test for OE subtypes as well as for the often-assumed causal aspect or causer thematic role.

The second experiment pairs the incremental processing speed of each sentence with replication of the identification of its thematic roles through ratings of intent. Australian-English speaking undergraduate students read agentive and OE actives and passives at their own pace phrase-by-phrase, then rated the intent of their subjects and objects. This experiment uniquely reveals that OE sentences are labile, a number able to be interpreted as either subtype. This second experiment also examines the relevance of the identity of the thematic role for sentence processing, with the instantaneous interpretation of the roles affecting the processing speed of the sentences that contain them: A causer, that theory and Experiment 1 ratings would attribute to active agentive and OE subjects, was processed like an internal thematic role in active voice when not judged to be an agent, indicative of a low-prominence theme or an aspect that can accompany a low-prominence theme. Additionally, stative passives were processed as adjectival passives, with external subjects, further supporting the low thematic prominence of their corresponding active subject.

This second experiment provides support for the postulation of a light verb projection (little v/CAUS) and associated event-structure morphemes in English, such that English (like other languages) needs multitiered morphology to assign the event characteristics of sentences even if the morphology is not pronounced. A stative morpheme may project an external experiencer or internal theme/ causer, independently of voice. An eventive or causative morpheme may not project an external causer thematic role unless it is an agent. This study provides unique evidence that the relative prominence of broader definitions of thematic roles, such as agent, experiencer and theme are relevant to processing speed. Finer detail in thematic roles, such as that which differentiates agent and causer or causer and theme, or theme and patient, is not relevant to processing speed and instead may be captured by lexical semantics.

A third experiment tests the possibility that we can diagnose event type through naturalness ratings, since theory suggests stative passives are less natural than eventives. Crowdsourced American-English-speaking participants rated SE and OE sentence naturalness on a 5-point Likert scale. Prototypical events that followed a cause-effect event progression were rated more natural, but hierarchical prominence ordering and event type also affected perceptions of naturalness. This experiment provides further evidence of SE and OE stative and eventive subtypes, as well as for the possible variations in sentence event types. It provides unique decontextualised judgements of naturalness that are likely based on the event structure of the sentence verbs themselves. The implications for theories of event types are discussed.

A final experiment supports a syntactic complexity account of the processing slowdown when listeners encounter thematically non-prominent sentence subjects. Australian-English speaking undergraduates completed a Cross-Modal Lexical Decision Task, listening to auditory sentences and reacting to orthographic (written) targets, followed by orthographic comprehension questions. Participants also completed a demographic questionnaire, two memory span tasks and a Grammaticality Judgement Task. When successfully comprehended, sentences more likely to have a non-prominent sentence subject, were more likely to show priming or inhibition in reaction time to the orthographic word target when it was semantically related to the subject, at a site consistent with the original position of a moved or copied internal argument. Furthermore, individual differences in processing were discovered: Degree and timing of reactivation coincided with processing speed, comprehension ability, language background, and musical ability. Because sentences in this task were auditory, as they are in a dynamic speech environment, the patterns of errors suggest that many slower processors are not comprehending all of the sentences they hear in communication environments, predicting comprehension difficulties for many skilled and impaired language users.

This thesis demonstrates that to comprehend a sentence, skilled-language users apply the same common, hierarchical syntactic frame, as soon as it begins to unfold. If the verb informs the comprehender that the thematic roles do not align with the applied template, the elements are copied or reactivated to fit the template structure, increasing post-verbal processing speed and comprehension difficulty.

History

Table of Contents

Chapter 1 Introduction: Why worry? -- i. Preamble to Chapter 2: Who's responsible? -- Chapter 2 The Causer's Intent -- ii. Preamble to Chapter 3: The Processing Race -- The 'Agent' and the 'Causer' in the Processing Race -- Chapter 3 The Contest of the 'Causer' Contender and the 'Agent' Defender -- iii. Preamble to Chapter 4: Preferences -- Chapter 4 Who Causes Emotional States? -- iv. Preamble to Chapter 5: Relationships -- Chapter 5 Trials Are Just Beginning -- v. Preamble to Chapter 6: Testing Relationships -- Chapter 6 Inadvertent Closeness -- vi. Preamble to Chapter 7: We Are All the Same and We Differ -- Chapter 7 Internalising the Experience and its Cause -- vii. Preamble to Discussion: The Key to Understanding Sentences -- Chapter 8 Discussion: Was it the Instrument or the Causer Who Made Me Surprised? -- Chapter 9 Conclusion: Consequences -- References -- Appendices

Notes

Thesis by publication

Awarding Institution

Macquarie University

Degree Type

Thesis PhD

Degree

Doctor of Philosophy

Department, Centre or School

Department of Cognitive Science

Year of Award

2024

Principal Supervisor

Paul Sowman

Additional Supervisor 1

Linda Cupples

Additional Supervisor 2

Iain Giblin

Rights

Copyright: The Author Copyright disclaimer: https://www.mq.edu.au/copyright-disclaimer

Language

English

Extent

421 pages

Former Identifiers

AMIS ID: 356599

Usage metrics

    Macquarie University Theses

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC