ABSTRACT
Two-alternative forced-choice reaction-time experiments were conducted to determine whether preliminary information about a stimulus can begin to activate responses before recognition of the stimulus has completely finished, as assumed by continuous and denied by discrete models of human information processing. Targets were presented either visually or auditorally, with distractors on the other modality. When the auditory stimulus was a target, a visual stimulus similar to a target served as the distractor. If preliminary information about the visual distractor does activate responses before recognition is complete, it should cause a bias toward the response appropriate for the target to which the distractor is similar. This bias should facilitate responses when the distractor is similar to the visual target consistent with the auditory target (i.e., assigned to the same response) and inhibit responses when the distractor is similar to the inconsistent visual target (cf. Eriksen & Hoffman, 1973). The results suggest that recognition must finish before information about a stimulus begins to influence decisions in a two-choice task, contrary to continuous models. Biasing was not obtained with distractors that were similar to targets either physically or physically and categorically. It was only obtained with distractors that had the same names as targets, but this biasing was consistent with discrete as well as continuous models. Nor was there biasing when distractors shared one defining characteristic with each target, even though one characteristic was perceptually much more discriminable than the other.
