ABSTRACT
Of central importance to the ACT system is the process by which activation spreads through the memory network. Productions are evoked by having their conditions match the active partition of long-term memory. I have done considerable research in a paradigm that studies the spread of activation through memory as directly as possible. This paradigm is referred to as/act retrieval. In this chapter I present a fact-retrieval experiment and an ACT model for that experiment and contrast this with an older model for that paradigm (Anderson, 1974; Anderson & Bower, 1973; Thorndyke & Bower, 1974). I will also review some of the data favoring the new model, present a set of detailed tests of ACT predictions, and then discuss the generality of the model. Each of these tasks will occupy one section of this chapter.
