ASA 126th Meeting Denver 1993 October 4-8

5aSP27. Lexical effects in nonwords?

Rochelle Newman James R. Sawusch Paul Luce

Dept. of Psychol., SUNY, Buffalo, NY 14260

Previous research has reported the existence of a ``lexical effect'' in identification tasks. For instance, one series might range from ``dice'' to ``tice,'' while the other varied from ``dype'' to ``type.'' The ambiguous stimuli from each series were more likely to be classified as members of the category which makes a word. (Thus, they would be classified as ``d'' in the dice-tice series but as ``t'' in the dype--type series.) However, these results have sometimes been inconsistent, suggesting that factors other than lexical status are involved. In a series of experiments, pairs of nonword--nonword continua were presented to listeners, where one end of each series had a higher neighborhood density than the other end. That is, one end of each series was similar to a greater number of words, weighted according to frequency of occurrence, than the other. Ambiguous items were classified by listeners with the phonetic label corresponding to the higher density end of the series. The results suggest that neighborhood effects may contribute to the variability in previous ``lexical'' effects. [Supported by NIDCD Grant No. DC00219 to SUNY at Buffalo and an NSF Graduate Fellowship to the first author.]