Simple: Further applications of a local distinctiveness model of memory

It has long been common practice to divide memory into one system that operates from 0 to 500 ms (sensory memory), a second system that operates for up to a few seconds (short-term or working memory), and a third system that operates on all temporal durations from minutes to years (long-term memory). A scale invariant view of memory, in contrast, suggests that the fundamental principles of memory apply regardless of the time scale, whether remembering occurs immediately after an event or several years later. Brown, Neath, and Chater (2002) proposed a model, Scale Invariant Memory and Perceptual Learning (SIMPLE), which explicitly states that key principles of memory retrieval hold regardless of the time scale and thus regardless of the presumed underlying memory system. Brown et al. (2002) applied the model to a range of data from serial and free recall, and here we extend the range of application of the model. Our aim is to illustrate the basic properties of SIMPLE by examining its ability to account for data and paradigms not studied by Brown et al. (2002). We first provide an overview of the model and then apply it to new data on serial position effects in absolute identification tasks, arguing for a link between discrimination in absolute identification and discrimination of memories in traditional serial recall paradigms. Next we fit key data that support the idea that people represent items based on temporal rather than ordinal position in many paradigms and then offer a unified account of serial position effects that traditionally would have been described as being due to iconic memory, long-term memory, and semantic memory. We briefly show how SIMPLE is beginning to be applied to data used to support the idea of working memory and then return to the important issue of whether items are best thought of as being encoded using ordinal or temporal position cues. We end by addressing some arguments against the idea that unitary memory principles may apply over different time scales. © 2006 Elsevier Inc. All rights reserved.