Morpho-phonological regularities influence the dynamics of real-time word recognition: Evidence from artificial language learning

Phonological alternations are attested in many of the world’s languages. In production, these robustly generalize to new words and contexts, suggesting that talkers and listeners of a language have internalized them in some form. However, it is unclear whether listeners’ knowledge of phonological al...

Full description

Bibliographic Details
Main Authors: Ashley Farris-Trimble, Bob McMurray
Format: Article
Language:English
Published: Open Library of Humanities 2018-01-01
Series:Laboratory Phonology
Subjects:
Online Access:https://www.journal-labphon.org/articles/41
Description
Summary:Phonological alternations are attested in many of the world’s languages. In production, these robustly generalize to new words and contexts, suggesting that talkers and listeners of a language have internalized them in some form. However, it is unclear whether listeners’ knowledge of phonological alternations is used during real-time spoken word recognition. The present study asks whether listeners use knowledge of phonological alternations to modulate activation of competitor forms during real-time word recognition. In two experiments, listeners learned an artificial language with phonological alternations. We then used eye-tracking in the visual world paradigm to assess real-time spoken word recognition. We examined fixations to competitors that would be a match to the input because of the learned phonological alternation. Results showed that listeners do use phonological alternations in real time. Given a [t] ~ [d] alternation and an auditory stimulus with a surface [d], listeners fixated the [t]-competitor more than one that could not alternate with [d]. They were even able to generalize this to words that had not been learned in their alternated form. However, not all alternations showed the same pattern; listeners did not use a [d] ~ [z] alternation in the same way. Implications for various models of word recognition are discussed.
ISSN:1868-6354
1868-6354