A random discrete-trials version of the Stroop color-word task was used to test the hypothesis that the conflicting response is directly suppressed to allow the appropriate color response. Two conditions were compared: (a) trials in which the current color matches the distractor word presented in the immediately preceding trial and (b) trials in which the current stimulus is unrelated to the previous stimulus. Vocal color naming (Experiment 1) was slower in the related than the unrelated condition, indicating that the preceding distractor response was selectively inhibited, making it relatively less available as an appropriate response during the next trial. However, manual key-press responses (Experiment 2) were faster in the related than the unrelated condition, indicating that the preceding distractor response remained more highly available than less recently activated responses. The results confirm the existence of inhibitory mechanisms in selective attention but suggest that their manifestation in an overt suppression effect is heavily dependent on experimental context.

Responses to recently ignored stimuli may be slower or less accurate than to new stimuli. This negative priming effect decays over time when delay is randomized within subjects, but not when delay varies between subjects. In Experiment 1, response-stimulus intervals (RSI) of 500 and 4,000 ms were randomized within subjects in a target localization task. Negative priming of ignored locations diminished with longer delay. However, no significant decay was obtained when RSI and the preceding RSI were equal. Similar results were obtained when RSI and preceding RSI were deliberately confounded by blocking (Experiment 2). Negative priming appears to depend on temporal discriminability of the priming episode.

Subjects identified target letters flanked by incompatible distractor letters (e.g., ABA). Distractor onset was randomly simultaneous with target onset or was delayed by 400 ms. In Experiment 1, one third of probe-trial targets were identical to the preceding prime-trial distractor. Responses were slower to repeated letters than to unrepeated letters {negative priming) only when prime and probe trials shared the same distractor-onset conditions. In Experiment 2, one third of probe-trial targets were identical to the preceding prime-trial target. Significant facilitation {repetition priming) occurred for repeated targets in all conditions but was again greater when prime and probe trials shared the same distractor-onset conditions. The results strongly support episodic retrieval theories of both negative priming and repetition priming and suggest that a common mechanism underlies both phenomena.A recurrent issue in cognitive psychology has been the degree to which cognition depends on the activation of abstract representations versus the retrieval of specific episodes or instances in memory. (See, e.g., Hintzman, 1986;Jacoby & Brooks, 1984;Logan, 1988;Medin & Shaffer, 1978.) One class of phenomena to which this debate is relevant is priming, the effect of a preexposed object (e.g., word, letter, or picture) on the response to a similar or related object. The simplest form of priming is repetition priming, referring to the facilitatory effect of a preexposed object on reaction time (RT) to an identical object. According to abstractionist views, such as Morton's (1969Morton's ( ,1979 logogen theory, the initial exposure to an object causes a preexisting mental representation of the object to become more highly activated or more easily retrievable. In contrast, the episodic view suggests that a subject may simply recognize the object as similar to one recently experienced, which may in turn speed its classification or the selection of a relevant response (e.g., Jacoby, 1983;Logan, 1990).Empirically, abstractionist and episodic views are contrasted by whether priming depends on specific trial details that should be irrelevant to an abstract representation-for example, the effect of typeface in word recognition or the context in which a target stimulus is imbedded. A recent These experiments were conducted while W.

The distractor-suppression effect is the relative slowing of Stroop (1935) color naming when the current appropriate response is identical to the inappropriate response activated by the distractor word appearing in the immediately preceding trial. Two experiments investigated aspects of the time course of distractor suppression. Experiment 1 found the suppression effect when subjects were instructed to maintain strict accuracy but not when subjects were encouraged to sacrifice some accuracy for greater speed. Experiment 2 traced the recovery from suppression by varying the interval between successive trials (20, 520,1020, or 2020 ms). The suppression effect was found to persist for at least a second; by 2 s the effect was completely dissipated. The results support the view of selective inhibition as an active, time-dependent control process that develops over time following the activation of distracting information and that is released after response to the task-appropriate information has been made. The results are interpreted in the context of a model in which widespread automatic activation in memory is followed by a process of "narrowing down" the range of activated representations to those specifically appropriate to current task demands (Keele & Neill, 1978).

Responses to recently ignored stimuli may be slower and less accurate than responses to new stimuli. Neill and Westberry (1987) found that such negative priming effects dissipated within a 2-s interval between response and the next stimulus, the response-stimulus interval (RSI). However, experiments by Tipper, Weaver, Cameron, Brehaut, and Bastedo (1991) found negative priming persisted unchanged over RSIs from 1,350 to 6,600 ms. Our experiments used a lettermatching procedure in which target letters were flanked by irrelevant letters. Negative priming was manifested by longer reaction times and more errors to letters that had appeared as flankers in the preceding trial. RSI was varied from 500 to 8,000 ms. Negative priming diminished over the RSI, particularly within the 1st second. This effect did not depend on the presence or absence of temporal uncertainty or on Ss' awareness of intertrial relations. We propose an episodic retrieval theory to account for negative priming phenomena.