While text readability has been a focus in research on computational linguistics for English and other resource-rich languages, there is still very little work conducted with regards to understudied languages like Vietnamese. We propose a semi-automated method to build the corpus of text readability for the Vietnamese language. Not only does this method address the lack of corpora that are graded on various difficulty levels in Vietnamese, but it can also be extended to the analysis of other resource-poor languages without requiring major adjustments.
Coreference between a pronoun and a subject in its local clause is typically prohibited across languages. However, I have shown that this supposedly universal grammatical constraint is not strictly enforced in Vietnamese. This vein of research investigates the real-time comprehension and final interpretation of pronouns in Vietnamese. Results from 6 different experiments reveal that there are two distinct mechanisms involved in resolving pronoun reference in Vietnamese. While Vietnamese speakers allow pronouns to corefer with local referential subjects, they prohibit pronouns from referring to local quantified subjects. The coexistence of these two mechanisms aligns with the key predictions of pronominal competition theories. Consequently, I argue that pronoun resolution in Vietnamese operates in accordance with a competitive model.
Tense has been traditionally analyzed as mandatorily overt grammaticalized expressions that impose constraints on temporal relation. Contrary to this emphasis on overt tense markings, many languages, including Vietnamese, lack overt tense morphology. In these studies, I address the question of how temporal reference is determined in the language. I argue that Vietnamese is tenseless in the sense that a bare verb predicate can receive a past or a present, but not a future construal. I propose that while past and present interpretation can be expressed covertly through a non-future tense operator, future reference has to be marked overtly. These findings suggest that Vietnamese employs a mixed tense system, directly contrasting with the traditional view that a language only has true tenses when it expresses tense overtly.
In this work, I address the question of whether different linguistic components are processed in separating processing systems. To test word-level difficulty, I examine the processing of low frequency words, which are predicted to be reliably slower than that of high frequency words. To investigate sentence-level difficulty, I compare negative polarity items, which must be licensed in negated contexts, with regular articles, which does not require special licensing. Results from 3 experiments show that individually, the presence of either a low-frequency word or a negative polarity item significantly increases reading times and decreases the naturalness of a sentence. However, when they co-occur in the same sentence, one type of complexity obscures the other, suggesting that there are independent processing systems targeting two linguistic levels.
The literature on Algonquian agreement displacement often discusses the person hierarchy in which second person outranks first person for all of the agreement slots. However, many Algonquian languages have been reported to establish more than one hierarchies in their agreement system. This project concerns itself with the typological classification as well as the syntactic derivations for person hierarchy effects, with a focus on number agreement. In particular, I propose that in order to capture the plural paradigms in both one- and two-person hierarchy languages, two distinct probes for person and number are required. Crucially, what drives one-person hierarchy languages away from two-person hierarchy ones is the availability of the goal for subsequent matching with the number probe after their features have been checked by the person probe.
This project examines the syntax of yes-no questions in Vietnamese. I show that yes-no questions in the language are constructed from negative sentences, sharing similarities to negative particle questions in Hokkien. The proposed syntax argues that the formation of these types of questions involve the movement of the negation to the end of the sentence, advocating for an interconnection between negation, polarity, and yes-no questions.
I coded, transcribed, annotated, and analyzed data for a National Science Foundation-funded collaborative project. This study investigates the patterns in morphological marking on verbs in child African American English. The results inform linguistic developmental stages and identify syntactic patterns that signal a language disorder, advancing our understanding of dialectal variation and its connection to language acquisition.