In this dissertation, I establish a research program that uses computational modeling as a testbed for theories of phonological learning. This dissertation focuses on a fundamental question: how do children acquire sound patterns from noisy, real-world data, especially in the presence of lexical exceptions that defy regular patterns? For instance,…

A strong predictor of children's language is performance on non-word repetition (NWR) tasks. However, the basis of this relationship remains unknown. Some suggest that NWR tasks measure phonological working memory, which then affects language growth. Others argue that children's knowledge of language/language experience affects NWR performance. A…

Humans can attribute beliefs to others. However, it is unknown to what extent this ability results from an innate biological endowment or from experience accrued through child development, particularly exposure to language describing others' mental states. We test the viability of the language exposure hypothesis by assessing whether models…

Decades of psycholinguistics research have shown that human sentence processing is highly incremental and predictive. This has provided evidence for expectation-based theories of sentence processing, which posit that the processing difficulty of linguistic material is modulated by its probability in context. However, these theories do not make…

Linguistic meaning is generated by the mind and can be expressed in multiple languages. One may assume that equivalent texts/utterances in two languages by means of translation generate equivalent meanings in their readers/hearers. This follows if we assume that meaning calculated from the linguistic input is solely objective in nature. However,…

During language acquisition children generalise at multiple layers of granularity. Ambridge argues that abstraction-based accounts suffer from lumping (over-general abstractions) or splitting (over-precise abstractions). Ambridge argues that the only way to overcome this conundrum is in a purely exemplar/analogy-based system in which…

Data-driven learning uncontroversially plays a role in human language acquisition--how large a role is a matter of much debate. The success of artificial neural networks in NLP in recent years calls for a re-evaluation of our understanding of the possibilities for learning grammar from data alone. This dissertation argues the case for using…

The syntactic behaviour of human beings cannot be explained by analogical generalization on the basis of concrete exemplars: analogies in surface form are insufficient to account for human grammatical knowledge, because they fail to hold in situations where they should, and fail to extend in situations where they need to. [For Ben Ambridge's…

Ambridge argues that the existence of exemplar models for individual phenomena (words, inflection rules, etc.) suggests the feasibility of a unified, exemplars-everywhere model that eschews abstraction. The argument would be strengthened by a description of such a model. However, none is provided. I show that any attempt to do so would immediately…

In this commentary, I suggest that it may be helpful to think about the formidable problem space that Westergaard's (2021) Linguistic Proximity Model seeks to address at the three levels of analysis that Marr (1982) famously proposed are needed to understand any complex cognitive system. I argue that at the computational level of analysis, where…

This dissertation investigates how bilinguals use their two grammars to comprehend written intra-sentential codeswitches. I focus on adjective/noun constructions in Spanish and English where I manipulate the congruence of grammatical word order in the two languages across the codeswitch boundary. This allows me to test three codeswitching…

Ambridge's proposal cannot account for the most basic observations about phonological patterns in human languages. Outside of the earliest stages of phonological production by toddlers, the phonological systems of speakers/learners exhibit internal behaviours that point to the representation and processing of inter-related units ranging in size…

Ambridge calls for exemplar-based accounts of language acquisition. Do modern neural networks such as transformers or word2vec -- which have been extremely successful in modern natural language processing (NLP) applications -- count? Although these models often have ample parametric complexity to store exemplars from their training data, they also…

Ambridge reviews and augments an impressive body of research demonstrating both the advantages and the necessity of an exemplar-based model of knowledge of one's language. He cites three computational models that have been applied successfully to issues of phonology and morphology. Focusing on Ambridge's discussion of sentence-level constructions,…

Language is one of the hallmarks of intelligence, demanding explanation in a theory of human cognition. However, language presents unique practical challenges for quantitative empirical research, making many linguistic theories difficult to test at naturalistic scales. Artificial neural network language models (LMs) provide a new tool for studying…