Preliminaries to a distinctive feature analysis of handshapes in American Sign Language
Temas
Detalles
In this article we propose a set of 11 distinctive features for hand configurations (Dez) in the American Sign Language of the deaf. The model is based on the results of applying clustering and scaling analyses to confusion matrices for Dez identifications in visual noise. These matrices were obtained in much the same way as in Miller and Nicely's classic study of consonant perception in noise. The descriptive adequacy of the model is evaluated in several ways: The level of the node in the distinctive feature tree that dominates each pair of Dez is correlated with the confusion frequency of that pair (r = 0.6); The percentage of correct proximity predictions when the Dez are clustered according to the feature tree (77%) is compared with that obtained with the optimal empirical tree (77%); Shared feature assignments for each Dez pair are weighted according to various strategies, and the 11 features are correlated jointly with the confusion frequencies of the Dez pairs (R = 68). Finally, the predictive validity of the model is examined, bringing to bear data on the rated similarity, recall, and variants of signs, and on “slips of the hand.”