A series of four experiments investigated Neill's (1977) claim that there are inhibitory mechanisms in selective attention. It was demonstrated that the evidence supporting the inhibitory theory, namely, the diminished availability of distractor responses during a discretetrials version of the Stroop task is complicated by a number of strategic adaptations to various contingencies within the trial sequence. These results do not support a simple interpretation of response inhibition during the Stroop task.

A series of three experiments employed a modified version of the Stroop task in order to document some conditions under which attention may be differentially deployed. It was shown that, within limits, subjects may choose either to selectively attend to a single stimulus or to distribute processing resources over several stimuli. Thus, attentional strategies may be actively chosen to suit prevailing conditions. Finally it was demonstrated that, once established, these different modes of attending are, themselves, differentially open to strategic modulation. RESUME Trois experiences, dans lesquelles est utilisee une version modifiee de la tache de Stroop, sont menees dans le but d'etudier les conditions dans lesquelles I'attention peut etre centree de facon differentielle. Les resultats montrent que, dans une certaine mesure, les sujets peuvent choisir, ou de porter leur attention de facon selective a un seul stimulus, ou de la distribuer sur plusieurs. Les strategies attentionnelles peuvent done etre choisies activement en fonction des conditions. Enfin les resultats montrent aussi qu'une fois etablis, ces differents modes d'attention sont euxmemes differentiellement sujets a des modulations strategiques.

Three experiments employed a discrete-trials version of the Stroop task to investigate further the claim that there are inhibitory mechanisms in selective attention (Neill, 1977). The results i~dicated that the evidence supporting the inhibitory theory (i.e., the diminished availability of distractor responses) cannot be explained by subjects' attempts to physically match successive stimuli (Lowe, 1979). In light of all the available evidence, it was suggested that the suppression effect does not index attentional inhibition, but may be attributed to a problem of code coordination (Keele & Neill, 1978).Deployment of selective attention permits an observer to respond optimally to chosen aspects of stimulation while excluding other, unwanted aspects. An important question concerns how this selective processing is achieved. Treisman (1964) hypothesized that selection of one message for analysis and response may sometimes involve the attenuation of others. According to this account, selection of some information may entail direct and specific inhibition of other information.Evidence for the existence of some form of inhibition in the context of selective attention comes from recent experiments (Neill, 1977) employing the Stroop (1935) color-word task. The Stroop task requires subjects to name the ink color of a letter display while overcoming the tendency to read a different color name spelled by the letters. Employing a series of discrete trials, Neill (1977, Experiment 1) found that when two successive Stroop stimuli were related such that the color name of the second item (~2) matched the distracting word of the first (Sl), observers took longer to respond than when the sequential inputs were unrelated (the suppression effect). Neill (1977) reasoned that when the color for Sl is named, the name of the conflicting color word is actively suppressed. Then, if that inhibited name is the correct response for S2, the inhibition slows responding, producing the suppression effect. However, subsequent results did not support a simple inhibitory account. For example, Neill (1977, Experiment 2) found that S1 distractor responses remained continuously available, and facilitated responses for S2, when manual responding was substituted for vocal color naming. Lowe (1979, Experiment 3) also demonstrated a facilitatory effect during related trials when color patches were substituted for color-word stimuli as S2. Lastly, results by Neill (1977Neill ( , 1979) also indicated that inhibition varied with the relative demands for speed versus accuracy. More pronounced inhibitory effects were obtained when instructions emphasized accuracy of responding. It is evident that the suppression effect is extremely sensitive to a number of variations in experimental context. This phenomenon can be influenced by the nature of the response requirements as well as by other sequences and stimuli that appear in the task.In light of these complications, Lowe (1979) suggested that subjects did not treat the successive stimuli as discrete, but employe...

A series of four experiments investigated previous findings which suggest that a patterned masking stimulus presented immediately following tachistoscopic presentation of letter rows produces large decrements in die recall of letters from the central positions in the rows but has little effect on recall from either end of the displays. The present experiments confirm the existence of a selective masking effect. The effect was obtained following exposure durations which varied from 30 to 200 msec and with both full-report and partial-report techniques. In addition, the selective masking effect was limited to multfletter displays in that it was shown that single letters were masked equally well across the positions used for an entire row. The results suggest that both ends of multfletter displays are processed and identified before the center positions of the displays are processed.IN SEVERAL PREVIOUS STUDIES Merikle, Lowe & Coltheart, 1970) involving tachistoscopic presentation of rows of letters, followed by either a blank field or a patterned masking stimulus, it was found that the masking stimulus had a selective effect on recall accuracy. Supplementary findings in these studies indicated that the mask produced large decrements in the recall of letters from the central positions in the display but had little or no effect upon the recall of letters from either end of the rows. Since Ss in these studies fixated at the center of the letter rows, the results indicate that masking is greatest at the point of fixation and decreases with increases in foveal eccentricity.This selective masking effect was somewhat unanticipated for several reasons. First, Stewart and Purcell (1970) have reported that single letters are more susceptible to a mask the greater distance they are from the point of fixation. Second, several previous studies involving probes for single letters from eight-letter displays have shown that when no mask is presented identification accuracy for the center positions is among the best of all the positions in the displays Haber & Standing, 1969). Such findings have suggested that the identification of letters may proceed on the basis of which retinal locations provide the best information for resolving the letters (Merikle, "This research was supported by Grants APA-231 and APA-350 from the National Research Council of Canada. Dawn Redmond deserves thanks for collection and tabulation of the data.

Seven-letter sequences, representing either first-order or second-order approximations to English, were presented for 150 msec and masked immediately after presentation. For different groups of 14 Ss, performance was evaluated by either a partial-report (PR) or a full-report (FB) procedure. In general, second-order approximations were better recalled regardless of the method of report, and the overall magnitude of the familiarity effect was approximately the same under both PR and FR procedures. The different methods of report, however, led to quite different accuracy functions across the letter rows. The results were taken to indicate that familiarity, as defined by differences in sequential redundancy, has its effect during processing but that left-to-right sequential processing may not necessarily be involved.