Abstract

That this is one of the most distinguished books in the excellent Bradford Books cognitive science/philosophy series is suggested by the March 1983 issue of Behavioral and Brain Sciences, in which we find a precis of the book, some twenty commentaries, and Dretske's replies. Physicalists and anti-physicalists in psychology have both stressed the importance of "top-down" strategies and have debated, prospectively, about the likelihood that we eventually will have suitable reductions, or explanatory instantiations, of psychological generations in neurophysiological terms. Dretske takes the more straightforward approach, starting with an avowedly physical notion of information and then trying to show how one can make, building on this foundation, a reasonable construal of knowledge, perception, and belief. Dretske starts with the following information-theoretic specification of a signal's information content: a signal r carries the information that s is F = the conditional probability of s's being F, given r, is l. Thus, to use Dretske's humble example, "My gas gauge carries the information that I still have some gas left, if and only if the conditional probability of my having some gas left, given the reading of the gauge, is l." The parenthetical clause makes the causal character of this analysis clear: the state of the source is related by physical laws to the occurrence of the signal--the gas in the tank causes the gauge reading. Insisting on the conditional probability of l has the advantage of preserving the properties of certainty, transitivity, and addition that we associate with knowledge. If the gas gauge does carry the information that my tank has gas, then my tank must have gas. If my statement that my gauge indicates gas carries the information that it does so indicate, then information flows intact from signal to signal. And finally, if the signal carries the information that s is F, and also carries the information that s is G, then it carries, at the same level of probability assigned each, the information that s is F and G. Now we can, eventually, say that K knows that s is F = K's belief that s is F is caused by the information that s is F. The "eventually" part is important, for what is most instructive and provoking about this book is the way in which Dretske lays out the steps in constructing viable construals of knowledge, perception, and belief out of the root notion of information.--Justin Leiber, The University of Houston.