In all domains in linguistics, categorical variation exists, e.g., English be vs. get passives. Compared to other linguistic variation, phonological variation is much more gradient, because it encompasses subcategorical phonetic adjustments, with many categories occupying a continuous range of values, e.g., English / fronting. Moreover, many phonological patterns are grounded in physiological properties, e.g., /ki/ → [tʃi] alternation, which is motivated by articulatory, acoustic, and perceptual factors that enhance its learnability. This talk presents a series of phonological variation learning studies, from which I show how phonological patterns’ inherent gradience and physical grounding shape learning. Based on the findings, I propose [...]
This talk is concerned with how components in memory structures and online structure building processes interact by investigating the online processing of Wh-Filler-Gap Dependencies (WhFGD) and ellipsis constructions. Resolving long-distance dependencies involves linking the dependent element to the controlling element. In the case of Wh-gap dependency formation, the wh-element is linked to the gap. In the case of ellipsis resolution, the ellipsis site is linked to the antecedent. In the processing of long-distance dependency resolution, I point out that two component processes are involved: the storage/maintenance component and the retrieval component. A series of studies on WhFGD formation reveals that the [...]
People process language at an incredibly fast speed. A speaker can easily produce several words per second, while a listener must decode the meaning of each word and integrate the meaning into the context at roughly the same speed. Native language (L1) speakers usually do this efficiently, and one way they achieve this efficiency is by predicting upcoming language. In contrast, non-native language (L2) comprehension is less efficient compared to L1 comprehension. A possible reason for less efficient L2 comprehension is that L2 speakers tend to predict to a lesser extent than L1 speakers. In this talk, I will present studies [...]