ABSTRACT
The shape of the ACT theory has been strongly influenced by preconceived notions that I have about the nature of cognitive functioning. The purpose of this initial section is to identify some of these predisposing biases and to try to motivate them. My reliance on preconceived biases in constructing ACT contrasts with a more conservative methodology for theory construction in which one attempts to construct a theory by means of simple generalization from data. t There are two justifications for my approach over simple generalization. The first is that there is reason to suspect that a procedure of simple generalization is not the optimal way to develop a theory of broad generality. Simple generalization is just one procedure of many for inducing a scientific theory: it involves starting with no theory and then developing the assumptions of the theory as the data come in. My approach is to start with my best guess and modify that as data come in. It can be shown (see Chapter 12) that no induction procedure is uniformly better than any other; it is just that different procedures are better for inducing different theories. Simple generalization would provide a better induction procedure only if my preconceived biases were largely incorrect assumptions about the human. As I do not think they are. it is obviously rational for me to start with a theory that
conforms to these biases. In a Bayesian inductive scheme, this would essentially correspond to giving one's prior probabilities their appropriate weighting.
