Linking verbs to syntax

Lin, Yi

Permalink

https://hdl.handle.net/2142/109327 Copy

Description

Title

Linking verbs to syntax

Author(s)

Lin, Yi

Issue Date

2020-08-31

Director of Research (if dissertation) or Advisor (if thesis)

Fisher, Cynthia L

Doctoral Committee Chair(s)

Fisher, Cynthia L

Committee Member(s)

• Baillargeon, Renée
• Dell, Gary S
• Federmeier, Kara D
• Montag, Jessica L

Department of Study

Psychology

Discipline

Psychology

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• language acquisition
• verb learning
• syntax learning
• distributional learning
• multi-level learning
• language production
• language comprehension
• language processing
• verb bias

Abstract

Language learning requires learners to balance the productive power of language with its many constraints (e.g., Bowerman, 1988). One of these constraints is verb bias. Verb bias refers to the likelihood of a verb to appear in one of the multiple structures that the verb is allowed to appear in. Adults and children as young as three years of age are guided by verb bias in language comprehension and production. In this dissertation, I explore the effect of direct experience with verbs and sentences on verb bias learning. Specifically, I focus on the adaptation of dative verbs that adults and children are familiar with (e.g., give, show, throw). In six production experiments and two comprehension experiments, I exposed adults and children to new experiences with familiar dative verbs in dative structures. Experiments 1 and 2 showed that adults and children can change the biases of familiar verbs after being constrained to produce some verbs in one dative structure (Dora gave the apple to Boots) and other verbs in the other dative structure (Mickey showed Minnie the block). They then use the updated biases in making production choices at test. Experiments 3, 4 and 5 investigated the mechanisms underlying this verb bias learning. Specifically, I asked whether verb bias learning, like abstract syntactic priming, could be driven by error-based implicit learning. All three experiments showed that the magnitude of the training effect varied with the likelihood of sentence structure and with pre-existing verb bias. Unexpected structure and verb-structure combinations resulted in larger training effects, suggesting the operation of error-based implicit learning. Experiment 6 examined multi-level syntax learning in children, asking if verb bias adaptation is influenced by the distribution of biased versus alternating verbs in the language. I found verb bias training effects when most verbs in children’s new experience were biased and not when most-verbs-alternated between the two dative structures. Finally, Experiments 7 and 8 found that adults and children can also change the biases of familiar verbs by listening to one verb in one dative structure and another verb in the other dative structure. When listening to new dative sentences at test, they used their modified biases to generate expectations in online comprehension. My data offer converging evidence that verb bias learning is continuous from childhood to adulthood, and is driven by syntax learning at multiple levels of analysis.

Graduation Semester

2020-12

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/109327

Copyright and License Information

Copyright 2020 Yi Lin

Owning Collections