In this paper, we report on experiments that investigated form-based similarity effects in visual spoken word recognition. Specifically, we tested whether accuracy of speechreading a word was related to the number of words (neighbors) perceptually similar to that stimulus word and to its frequency of occurrence. In the first Experiment, the Neighborhood Activation Model (NAM) [1,2] was adapted to generate predictions about the accuracy of visual spoken word identification. In the second Experiment, we used the concept of the Lexical Equivalence Class Size [3] to generate predictions regarding the accuracy of visual spoken word recognition. Both experiments provided evidence that words are identified more accurately if they have few neighbors and occur frequently in the language. Correlational analyses provided evidence that a word's neighbors, or close competitors, are based on perceptually defined similarity. The results of the current experiments are interpreted as evidence of a common spoken word recognition system for both auditory and visual speech information, which retains sensitivity to form-based stimulus similarity among words.
Cite as: Auer Jr., E.T., Bernstein, L.E., Mattys, S. (2001) The INFLUENCE OF THE LEXICON ON VISUAL SPOKEN WORD RECognition. Proc. Auditory-Visual Speech Processing, 7-12
@inproceedings{auerjr01_avsp,
author={Edward T. {Auer Jr.} and Lynne E. Bernstein and Sven Mattys},
title={{The INFLUENCE OF THE LEXICON ON VISUAL SPOKEN WORD RECognition}},
year=2001,
booktitle={Proc. Auditory-Visual Speech Processing},
pages={7--12}
}