This project investigates the knowledge that native speakers have about the sound structures in their language. Languages differ in what sounds they use, and one of the structural properties of a language is the set of sounds that that language uses. Within a language, some combinations of the sounds of that language may not exist. For example, English has a sound 't' and a sound 'l', but the combination 'tl' does not exist at the beginning of a word; there are words like 'tad' and words like 'lad', but no words like 'tlad'. Speakers are generally aware of absent combinations of this type and find that they sound odd or foreign. This project investigates two main questions about speaker knowledge of non-existing structures: (1) Do speakers of a language find certain unattested structures more strange than others, and do these asymmetries correlate with which structures are more commonly found in the world's languages? (2) Do speakers notice only very specific combinations of sounds that are unattested, or do they make broad generalizations about the types of sounds that may not combine?
These questions are addressed by a series of experiments conducted by Dr. Gillian Gallagher in Cochabamba, Bolivia, with speakers of South Bolivian Quechua, a language with many systematic combinatorial restrictions on its sounds. The project contributes knowledge and documentation of the sound structures of this indigenous language, which will be made publicly available. The project also trains graduate students to carry out experimental work with indigenous languages.
This research contributes to understanding how humans make generalizations in the course of learning, a question central not only to linguistics but to the study of cognition more generally. The available learning data (e.g., the words of a language) are often compatible with multiple generalizations; this project investigates what criteria help humans choose between these possible generalizations in a manner that is consistent across learners.
