Memory behavior requires knowledge structures, not memory stores.

Abstract

Since the inception of cognitive psychology dominant theories of memory behavior have used the storage metaphor. In the multi-store models (e.g., Broadbent, 1958; Atkinson and Shiffrin, 1968; Baddeley and Hitch, 1974) the memory system comprises one or more short-term memory (STM) stores and a long-term memory (LTM) store. These stores are places where information is located for varying periods of time (i.e., seconds in the STM stores, and minutes to lifetime in the LTM store) and they have varying capacity limits: large for the LTM store, very limited for the STM store—4 to 7 items, see Miller (1956), Broadbent (1958), and Cowan (2001)1. Expertise research has shown that experts are able to remember a large amount of information presented immediately before testing their memory (e.g., more than 80 items in Chase and Ericsson, 1982 and in Gobet and Simon, 1996), suggesting that they are superseding the normal capacity limits of the STM store. However, given that this effect only occurs with domain-specific material expertise theoreticians (e.g., Ericsson and Kintsch, 1995; Gobet and Simon, 1996) explained these results in terms of the use of retrieval structures (see explanation below), but they retained the partition between STM and LTM stores. In this article I adumbrate an alternative explanation that builds upon three sources: (i) the behaviorist conception of memory as behavior (Delaney and Austin, 1998); (ii) models of memory that exclude the STM store (e.g., Nairne, 1992; Fuster, 1997; Neath, 1998; Cowan, 1999; Oberauer, 2002; Conway et al., 2005; McClelland et al., 2010); (iii) Gobet and Simon's (1996) and Ericsson and Kintsch's (1995) emphasis on the role of expertise in memory, and their pioneer theoretical conceptualization of retrieval structures. In the remaining of the article I briefly discuss these three sources, and then I present the alternative explanation and draw some conclusions.