Gwendolyn Hildebrandt will be defending their dissertation, "Distributional cues to syntactic structure: case studies in Korean" on Monday June 24th at 9:00am EDT.

The defense will take place in the Linguistics department library, and on Zoom.


The dissertation abstract is provided below.


Title: Distributional cues to syntactic structure: case studies in Korean

Supervisor: Charles Yang
Committee: Julie Anne Legate, Martin Salzmann


Abstract: How can syntactic and learnability analyses inform each other, and thus deepen our understanding of syntax and its acquisition? This dissertation illuminates this question through three case studies in Korean syntax. I examine cases in which two structures that display distinct syntactic properties share extremely similar surface forms, thus restricting the space of possible explanations for how learners may arrive at distinct underlying structures. As differences in the form of the structures are not surface-level apparent, I posit that the learner must be led to the conclusion of their distinct syntactic identity on the basis of their linguistic input--distributional cues.

These case studies draw on both grammaticality data from consultations, as the basis for syntactic analyses, and corpus data, to model the linguistic input. I utilize the Tolerance/Sufficiency Principle (Yang 2016) to predict what generalizations a learner makes about syntactic properties of the structures at hand.

I begin with the case study of Korean verb-doubling in Chapter 2. Two types of verb-doubling in Korean have previously been given unified analyses, which claim that both structures arise from a single underlying syntax. I show that the two verb-doubling structures have substantially different syntactic properties and propose a new analysis that accounts for these empirical differences. Learnability analysis then recoups the major differences between the two types of verb-doubling, showing that we independently predict a young learner to conclude that these structures are underlain by distinct syntax.

In Chapter 3, I examine Korean periphrastic causatives and purposive clauses. I find that previously observed variable case-marking on the embedded subject of purposives is deceptive, while the embedded subject in causatives displays genuine structure-specific variable case-marking. This is reconfirmed by learnability analysis, which shows that only for causatives, and not for purposives, do multiple case-marking options reach the Tolerance/Sufficiency Threshold for generalization.

Chapter 4 then introduces Korean intentives and intent clauses. I find that the class of intentives comprises two distinct syntactic structures: volitional intentives and imminent future intentives. Identifiable through their respective interpretations, these two types of intentive are formally indistinguishable but display distinct syntactic properties. In contrast, the formally distinguishable class of intent clauses appear to share the same structure as that of volitional intentives. Without reliable formal cues the learnability analysis is limited, but provides evidence that in this case the learner will utilize interpretation to effectively distinguish between volitional intentives and imminent future intentives.

In Chapter 5 I conclude, discussing similarities and differences between the case studies, and how taken together they indicate the varied and flexible potential for the parallel application of syntactic and learnability analysis.