Dr. Aleksei Nazarov
Dr. Aleksei Nazarov
Assistant Professor
I work on questions around how we learn the patterning of speech sounds in languages. I study this using machine learning techniques based on existing theories of sound structure. My particular focus is on so-called hidden structure: structure that is necessary but cannot be determined based on sound alone. This includes exceptions to sound structures in a language, structures underlying linguistic stress (emphasis) patterns, and exception-like interactions between overlapping patterns (also termed “opacity”). Such structure requires systematic analysis of the data, which has to take place subconsciously for human learners, and may be modelled using computer learners. The goal is to study how the psychological representation of language interacts with data, and ultimately to evaluate and argue for specific hypotheses on the mental representation of such structures and patterns.