Immediate serial recall is one of the best-known tasks in cognitive psychology. Participants are typically given a sequence of familiar verbal items that they are then asked to recall in the correct order. Among many theorists, Alan Baddeley and colleagues have perhaps been the most successful in identifying the cognitive components underlying this simple task (Baddeley, 1986; Baddeley & Hitch, 1974). Within their more general working memory (WM) model, they have identiﬁed a subsystem that they call the phonological loop. This comprises a phonological store and an articulatory control process that is assumed to refresh the contents of the store via a process usually identiﬁed with active rehearsal. The WM model has been extremely useful in identifying those components of short-term memory that subserve immediate serial recall (ISR). Nonetheless, Baddeley’s account is a verbal one and thus suﬀers the problems of a purely verbal model. In particular, it does not describe any detailed mechanism by which serial recall is actually implemented and hence cannot explain some of the detailed patterns present in the ISR data. A number of researchers have sought to provide more quantitative models of immediate serial recall. In some cases, these researchers have tried to keep their models close to Baddeley’s verbal characterization, essentially supplying a mechanism by which the verbal model might be implemented. In others, the modellers have tried to use their simulations to call into question some of the assumptions of the verbal model.