ABSTRACT
This chapter reviews a good deal of empirical evidence, which has accumulated for such a distinction in the human mind. Declarative information is represented in terms of data structures which are basically the chunks that G. A. Miller described in 1956 and which were the focus of a great deal of research in psychology until the mid-1970s. It describes each of these properties of chunks; this is followed by some discussion of what representational notation might be appropriate for them. The experiment involved crossing three conditions of the initial experience with two transfer tasks. Subjects were given no prior experience, prior experience in generating such code, or prior experience in evaluating such code. A chapter on representation has to come early on cognitive architecture because one needs to establish a notation for describing the elements of cognition. The actual theoretical claims of this notation are relatively few, but they are important.
