Academic journal article Memory & Cognition

Response Suppression Contributes to Recency in Serial Recall

Academic journal article Memory & Cognition

Response Suppression Contributes to Recency in Serial Recall

Article excerpt

Published online: 4 May 2012

© Psychonomic Society, Inc. 2012

Abstract Serial recall is often assumed to involve response suppression: the removal or inhibition of items already recalled so that they are not recalled again. Evidence for response suppression includes repetition inhibition and the separation of erroneous repetitions. Some theorists have suggested that response suppression, by eliminating competing responses, also contributes to recency in forward serial recall. We present experiments in which performance on the final item was examined as a function of whether or not the preceding retrievals entailed suppression of potential response competitors. In line with the predictions of response suppression, recency was found to be reduced when the earlier recall errors consisted of intrusion errors (which leave list items unsuppressed) rather than transposition errors (which involve suppression).

Keywords Working memory . Short-term memory . Recall . Inhibition

Contemporary models of serial recall have assumed various mechanisms to represent order among list items, such as a primacy gradient (Brown, Preece, & Hulme, 2000; Farrell & Lewandowsky, 2002; Lewandowsky & Farrell, 2008b; Page & Norris, 1998) or positional coding (Brown, Neath, & Chater, 2007; Brown et al., 2000; Burgess & Hitch, 1999; Henson, 1998b). The models also differ with respect to their representational assumptions, relying on either localist (Burgess & Hitch, 1999; Henson, 1998b; Page & Norris, 1998) or distributed (Brown et al., 2000; Farrell & Lewandowsky, 2002; Murdock, 1995) representations. Notwithstanding this diversity, there is some notable convergence between different theories (see Lewandowsky & Farrell, 2008b). One nearly universal assumption is the notion of response suppression (Brown et al., 2000; Burgess & Hitch, 1999; Farrell & Lewandowsky, 2002; Henson, 1998b; Lewandowsky, 1999; Lewandowsky & Murdock, 1989; Nairne, 1990; Page & Norris, 1998). This is the assumption that once an item is recalled, it is somehow prevented from competing for report again during subsequent retrievals on a trial.

Although direct evidence for response suppression is difficult to obtain, much indirect empirical support for the notion has been found. For example, the observed patterns of erroneous repetitions (i.e., repeated reports of an item that occurred only once on a list) strongly suggest that an item is suppressed after its first report (Duncan & Lewandowsky, 2005; Henson, 1998a; Vousden & Brown, 1998). Likewise, people's reluctance to report an item twice when it was repeated on the list (i.e., the repetition inhibition, or Ranschburg, effect) is also commonly taken to reflect response suppression (Duncan & Lewandowsky, 2005; Henson, 1998a).

Some theorists have attributed an additional role to response suppression in the recency effect, the enhanced recall commonly observed for the final one or two list items (Brown et al., 2000; Farrell & Lewandowsky, 2002; Lewandowsky, 1999; Lewandowsky & Farrell, 2008b; Lewandowsky & Murdock, 1989). According to this view, as recall proceeds, more and more items will be suppressed, thus reducing competition in the pool of recall candidates during the remainder of recall. When the last few output positions are reached, only one or two unsuppressed items will be leftto compete for report, thus increasing the probability that the final items will be correctly recalled. This occurs despite the fact that, according to most models, the final list items have the weakest representations (Brown et al., 2000; Farrell & Lewandowsky, 2002; Lewandowsky & Farrell, 2008b; Page & Norris, 1998).

The theoretical linkage between response suppression and recency does not deny that other factors, such as temporal distinctiveness (Brown et al., 2007; Brown et al., 2000; Burgess& Hitch, 1999; Henson, 1998b; Lewandowsky & Farrell, 2008b), might contribute to recency. …

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


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.