ABSTRACT
Connectionism is a style of explanation, a framework for thinking about cognition; it is not a particular topic in memory research in the same way as autobiographical memory, working memory or prospective memory are topics. Frameworks can be viewed as more abstract than a particular theory or model of a cognitive process (Slack 1987). What a framework does is to provide a set of ideas that can be used in constructing a wide range of more specific theories or models. Several other ideas commonly used in cognitive psychology can be seen as frameworks. One example of this is the schema and related variants such as scripts and frames (Bartlett 1932; Minsky 1975; Schank and Abelson 1977). In this chapter we try to show the qualities of the connectionism approach that might make it an appealing framework for constructing models of memory. Frameworks are not open to the same kind of direct empirical test as models, so whilst a particular finding might be of difficulty for connectionism, it is unlikely to be sufficient to discard the whole approach. However, the popularity of a particular framework is likely to be affected by the empirical successes and failures of the models to which it gives rise.
