Emergent Grammar:
Reducing the Role of Universal Grammar in Phonology

Instructor(s): Diana Archangeli and Douglas Pulleyblank

Do phonological grammars emerge from the input without benefit from language-specific innate predilections (aka Universal Grammar)? Does Emergent Grammar mean anything goes or is it testably restrictive? A developing body of work challenges universal distinctive feature definitions: From language to language, the fine-grained phonetics of a feature vary, the feature values assigned to particular segments vary, the features required for segment classification may themselves vary.

A ripple effect with profound consequences is triggered: If features are not universal then how do language learners identify the needed distinctive feature categories? What happens with grammatical references to distinctive features (e.g. "universal" constraints posited in Optimality Theory)? Our empirical focus is "harmony systems": vowel harmony, consonant harmony, local assimilation, disharmony (OCP effects, morpheme structure constraints, etc.). Working from detailed datasets, we develop phonological analyses from an emergentist perspective and ask whether universalist and emergentist models make different predictions.

None. While the course is highly appropriate for advanced graduate students whose research focuses on sound systems, the material should also appeal, and be accessible to, students of other aspects of the cognitive sciences (especially psycholinguistics and acquisition), as well as those interested in computational modeling. We welcome a very diverse population of students!