Academic journal article Canadian Journal of Experimental Psychology

Repetition Priming and Repetition Blindness: Effects of an Intervening Distractor Word

Academic journal article Canadian Journal of Experimental Psychology

Repetition Priming and Repetition Blindness: Effects of an Intervening Distractor Word

Article excerpt

Repetition blindness (RB) is a tendency for observers to fail to report repeated occurrences of briefly presented visual stimuli. For example, in an early demonstration of the effect (Kanwisher, 1987) participants viewed sentences such as "When she spilled the ink there was ink all over," in a rapid serial visual presentation (RSVP) task, with the words being displayed one at a time for 117 ms each. When reporting sentences of this kind, participants omitted the second occurrence of the repeated word in over three quarters of the trials. In contrast, if repetition was eliminated by replacing the first word with another (e.g., ink with liquid), participants were much better at reporting the second of these words, omitting it in only one fifth of the trials. RB with words occurs not only when two targets (referred to as C1 and C2) are identical, but also when they overlap substantially in their orthography or phonology (Bavelier, Prasada, & Segui, 1994). The effect is most often studied using words, letters, and digits, but it has also been demonstrated with pictures of objects (Bavelier, 1994). The presence and extent of RB depend on the duration of the interval between C1 and C2 (Bavelier et al., 1994; Kanwisher, 1987; Luo & Caramazza, 1995).

Most hypothesised mechanisms of RB involve processes in identifying a word and making it available in working memory for report. These "perceptual" accounts are the focus of the present investigation. They can be distinguished from memory reconstructive accounts that attribute failures with repeated items to memory reconstructive processes occurring largely at the end of the RSVP stream. Such failures have been reported in the memory literature. An example is the Ranschburg effect, a deficit in accuracy for a repeated item when a short list is presented for immediate serial recall (Crowder, 1968; Greene, 1991). Although reconstructive accounts make no specific predictions about the manipulations used here, it is worth noting that there is evidence that such processes contribute to RB. For example, Armstrong and Mewhort (1995) found that RB could be eliminated by changing the way in which participants were asked to report the critical item. In an initial experiment, participants were asked to report all stimuli in order, and an RB effect emerged. But in a second experiment, participants were instead shown one of the items from the RSVP stream as a cue and were asked to report the item which had followed it. In this task, despite identical conditions of stimulus presentation, no RB effect was observed; participants were just as accurate in reporting repeated as unrepeated items when shown the preceding item as a cue (Masson, 2004).

In the present experiments C1 and C2 were words. The key manipulations of interest were the duration of C1 and the nature of a single distractor item that separated C1 and C2. There is already some evidence that RB is affected by the duration of C1, with RB increasing with C1 duration (Burt, Kipps, & Matthews, 2014; Huber, 2008; Luo & Caramazza, 1995). Additionally, there is evidence that RB is affected by the nature of the intervening, distractor items (Whittlesea & Masson, 2005). As we will discuss below, there is no clear consensus on the source of these effects.

Theories of RB began with Kanwisher (Kanwisher, 1987; Kanwisher & Potter, 1989), who proposed a token individuation (perceptual) account of her RB effect. Specifically, RB results from a failure, under the time-constraints of online perceptual processing, to encode repeated occurrences of a stimulus as distinct events. Under this view, an observer's ability to report the contents of an RSVP stream depends on the activation of type nodes, the formation of episodic tokens, and on type-token binding (Bavelier & Potter, 1992). In Kanwisher's explanation, RB is said to occur because of a limit on type-token binding, specifically on the rate at which a given type can be bound to multiple tokens. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.