Evidence from studies of intentional learning suggests that the accuracy of recall is not assisted by appropriate enactment at retrieval, as opposed to encoding. In the present study, long-term recall of spatial arrays following incidental learning (text messaging or calculator use) was tested under three different motor conditions at retrieval. For both letter and number arrays, the accuracy of recall was found to be improved by relevant enactment at the time of retrieval, relative to retrieval with no movement. In contrast, irrelevant movement was found to produce an impairment in accuracy. The overall accuracy of recalling a letter array was found to be a power-law function of the frequency of exposure to the array. The findings are discussed in terms of the hypothesis that appropriate movement during memory retrieval recruits egocentric representations that supplement allocentric representations subserving longer term spatial recall.
Early studies of human memory (e.g., Atkinson & Shiffrin, 1968) were in the forefront of psychology's paradigm shift from behaviorist to cognitivist explanation. One legacy of this may have been to discourage exploration of the possibility that overt behavior-that is, physical movement-can influence the operation of human memory. Nevertheless, evidence has been provided within the working memory framework that irrelevant physical movement may disrupt the initial encoding of spatial information (see Logie, 2003), and it has also been shown conversely that recognition of spatial arrays can be facilitated when they are appropriately pointed at during encoding (Chum, Bekkering, Dodd, & Pratt, 2007). The further question of whether the recall of previously stored information can be influenced by movement at the time of retrieval has been relatively little investigated, although a notable exception is provided by the subject-performed task (SPT) paradigm.
In the SPT method, participants receive simple sentences (e.g., fold the paper), which they either do or do not also enact. Research has demonstrated that subsequent recall of such sentences is enhanced when the sentences have been enacted at encoding (for a review, see Nilsson, 2000). Noice and Noice (2001) have shown that a similar result occurs if actors learn their lines while moving according to stage directions rather than remaining static, even though their movements bear a less close relation to the text. However, in a particularly relevant study, Kormi-Nouri, Nyberg, and Nilsson (1994) found that there was no further advantage from participants' enacting also at retrieval. In the absence of such an advantage, Kormi-Nouri and Nilsson (2001) proposed that the basic SPT enhancement effect instead derives from increased self-involvement when enactment occurs at encoding. An alternative interpretation of encoding enhancement has been proposed by Noice and Noice (2007) in terms of the concept of embodied cognition, whereby language use is facilitated by concurrent activation of appropriate affordances of the physical environment (Glenberg & Kaschak, 2002).
Were Kormi-Nouri and Nilsson (2001) correct to reject the hypothesis that enactment at retrieval can improve the accuracy of recall? Our consideration of this hypothesis was prompted by informal observation of some of the participants during testing for a recent study of spatial memory (Martin & Jones, 2007). Prior to formally recording their responses, these participants spontaneously ranged their hands over the blank response templates. They appeared to believe, at least implicitly, that activating these motor patterns might lead to more accurate retrieval performance. Of course, it is alternatively possible that in reality these participants' motor activity had no effect on retrieval, or even interfered with it. However, consideration of the distinction between egocentric and allocentric spatial representations (Byrne, Becker, & Burgess, 2007) does provide a theoretical rationale for the possible observation of an enactment effect at retrieval. …