Ian Cameron (QP1), The Diachrony of Word-Final Obstruent-Sonorant Clusters in North Germanic
Abstract: Manuscript evidence suggests that word-final rising sonority clusters were permitted in early varieties of Old Norse. By the 15th century, all North Germanic languages had developed a vowel separating these clusters involving the apical trill /r/, but not the other sonorants /n/ and /l/. Traditionally, such cluster repairs have often been analyzed as a result of the Sonority Sequencing Principle (SSP) (Selkirk, 1982; Page, 1995). However, an analysis motivated by the concept of sonority falls short of explaining the North Germanic facts, as the apical trill /r/ is less sonorous than /n/ and /l/ (Parker, 2008). We argue that these patterns are best explained by an analysis utilizing Articulatory Phonology (Browman & Goldstein, 1986). In our analysis, gestural mistiming between heterorganic, voiced obstruent-sonorant clusters gave rise to an intrusive vowel (Hall, 2006). This is in line with the observation that heterorganic clusters are most prone to developing intrusive vowels, as they exhibit less gestural overlap than homorganic clusters. The order by which this affected other obstruent-sonorant clusters follows from morphological boundary effects; gestural mistiming is more variable between morpheme boundaries than within morphemes (Cho, 2001).
Byung-doh Oh (QP1), Incremental parsing for semantically-sensitive psycholinguistic predictors
Abstract: Expectation-based theories of sentence processing posit that processing difficulty is determined by predictability in context. While predictability quantified in the form of information-theoretical surprisal (Hale 2001; Levy 2008) has gained empirical support, this representation-agnostic measure leaves open the question of how to best approximate the the human comprehender's latent probability model. This work presents an incremental parser that incorporates both syntactic structure and propositional content into a single probability model. Regression analyses show that the surprisal measures derived from the semantically-sensitive parser make an independent contribution to predicting human reading times over those derived from a purely syntactic parser, suggesting the role of propositional content in incremental sentence processing.
