Introduction

Research on memory has provided considerable evidence for a verbal short-term memory (STM) system that is involved in memory tasks, such as span, free recall, probe recognition, and the Brown–Peterson paradigm, in which subjects must retain small amounts of linguistic information over brief periods of time. There is evidence that the representations in STM include phonological forms (Conrad, 1964) and that these representations are maintained in STM in part through a process that involves articulatory rehearsal (Baddeley, Thomson, & Buchanan, 1975). The specific character of the STM system is the subject of investigation and debate. One view (see Shallice & Vallar, chapter 1, and Baddeley, chapter 2, this volume) maintains that auditorily presented items are entered into a phonological store (PS) directly (Salame & Baddeley, 1982; Baddeley, Lewis, & Vallar, 1984; Greene & Crowder, 1984), while printed items are entered, at least in part, through a controlled process of subvocal rehearsal (Murray, 1968; Levy, 1971; Peterson & Johnson, 1971; Estes, 1973; Baddeley et al., 1984). There appears to be a progressive diminishing of the strength of the phonological representations in the phonological store over a period that has been variously estimated to last between 2 and 20 sec (Crowder & Morton, 1969; Wickelgren, 1969; Darwin, Turvey, & Crowder, 1972). However, the strength of these representations can be increased through the controlled use of articulatory rehearsal processes (the articulatory loop [AL]: Baddeley et al., 1984; Baddeley and Lewis, 1984).