ABSTRACT
One of the great achievements of the Working Memory (WM) model is that it allows a large amount of data to be fitted into a compact theoretical framework. This is particularly true in the case of the phonological subsystem of working memory as applied to the task with which it has been very directly associated, the immediate serial recall (ISR) of verbal material. In the twentysix years since the first publication of the WM model, the conception of the phonological component of working memory, the so-called phonological loop (PL), has proved capable of being adapted to account for a variety of data from ISR tasks. These data bear on the effects of factors such as modality, phonological similarity, word length, concurrent articulatory suppression and irrelevant speech and, further, on the interactions between these factors. That the general patterns embodied in such a large body of data can be summarised within a reasonably concise theoretical framework, is a tribute to a style of verbal modelling that allows models to be adapted in the face of new and constraining data, but that preserves a clear sense of the core, defining features of the theory.
