| A recorded word repeated over and over is typically heard to undergo a series of illusory changes (verbal transformations) to other syllables and words. When a second image of the same repeating word is added dichotically, with an interaural delay sufficient to prevent fusion, the resulting distinct lateralized images undergo independent illusory changes at the same rate observed for a single image. However, when the words differ contralaterally, the transition rate is dramatically suppressed (Lenz et al., 2000). The present study investigated the factors responsible for this suppression, utilizing a single target word and a variety of contralateral stimuli. Experiment 1 examined the transition rate of the target in the presence of four contralateral competitors: silence, non-speech (a door slam), the same word, and a different word. Suppression of the target only took place in the presence of the different-word competitor, indicating that suppression is due to speech-specific factors. Experiment 2 employed planned comparisons of conditions with competitors differing from the target in semantic relatedness, phonetic similarity, and lexicality. An omnibus analysis indicated that all competitors produced suppression relative to the same-word baseline condition. Planned comparison 2A examined semantic effects utilizing semantic opposite, semantic collocate, and semantically unrelated competitors. All competitors produced statistically equivalent suppression upon the target's transition rate, indicating that semantic relatedness is not a source of the suppression effect. Planned comparison 2B examined the effect of featural disparity (one linguistic feature vs. three) between phonemes of the target and its competitor. Competitors with three featural differences in a single phoneme suppressed the target's transition rate more than those with a single featural difference. Planned comparison 2C examined the effect of the lexicality (i.e., non-word vs. word) of the competitor upon suppression of the target. It was found that nonlexical competitors had a greater suppressive effect upon the target's transition rate than lexical competitors. These findings suggest that dichotic verbal transformations provide access to sublexical, phonetic levels of language processing that may be otherwise obscured by higher, lexical levels of processing. The use of the verbal transformation to study and evaluate various models of language processing is discussed. |
