Adam Liter in Phonology
August 24, 2020 Linguistics
What biases do phonological learners use in the face of data that is ambiguous between generalizations of differing complexity?
Now out in Phonology, "There is a simplicity bias when generalising from ambiguous data”, by Karthik Durvasula (Michigan State University) and our own Adam Liter. With results from three artificial language learning experiments, the paper argues that, while learners are able keep track of multiple generalizations related to the phonotactic sequence patterns, the generalizations they end up learning are only the simplest ones consistent with the data.