Abstract
Potter, Nieuwenstein, & Strohminger (2008) found that there was an attentional blink for report of the second of two red (or uppercase) words at a short SOA, in an RSVP sentence presented for 93 ms per word. If participants simply reported the whole sentence without looking for targets, they had no difficulty including the same two words along with the rest of the sentence. In new experiments, the task required report of the sentence together with report of the two targets (red words or uppercase words). Although the target words were reported in recall of the sentence, participants again showed an attentional blink—they frequently failed to identify which word had been red (or uppercase). In further experiments, the targets were strings of Arabic digits (e.g., 5555) or written digits (e.g., five). The easily discriminated Arabic digits showed little evidence of a blink in whole report, whereas with written digits there was an attentional blink in both whole report and partial report—and recall of the sentence was poor. Thus, when the targets are not visually distinguished from the words of the sentence and must be selected by meaning, sentence report and digit report are mutually interfering. Notably, in all the experiments there was lag 1 sparing (facilitation of T2 at an SOA of 93 ms) in all the conditions in which there was an attentional blink at lag 2 or 3, in accordance with Wyble's theory of AB (Wyble, Bowman, & Nieuwenstein, in press). Lag 1 sparing results because the attentional episode initiated by T1 extends for about 150 ms, benefiting T1 if it appears in that interval, but leading to an attentional blink at a longer SOA as the first episode is consolidated. Overall, the results show that the attentional blink is a blink of selection, not of perception.