ABSTRACT
In this chapter and later ones the question with which we are essentially concerned is this: under what conditions can an observer increase the probability of receiving one message at the expense of others which may. be present at the same time? In terms of the notation developed in Chapter 2, given that we are conducting an experiment which may be described by
I t:p(raRr)JI{a]+J{b] > p(raR1)ji{a]+J{b] what sorts of signals may constitute a and what regions of signal space may be the location of I?
