ABSTRACT
This chapter suggests that stimulus-response theory, or at least a mathematically precise version, can indeed gives an account of the learning of many phrase-structure grammars. It shows how any finite automaton, no matter how complicated, may be constructed purely within stimulus-response theory. Classical conditioning may be represented by an automaton having a one-letter alphabet and a single internal state. The next simplest case corresponds to the structure of classical discrimination experiments. The chapter helps to use the external alphabet of the automaton a two-letter alphabet consisting of a black or a white card. It describes both the reinforcement schedule and the transition table of the automaton. The correspondence to be made between letters of the alphabet that the automaton will accept and the appropriate objects within stimulus-response theory is fairly obvious. The letters of the alphabet correspond in a natural way to sets of stimulus elements presented on a given trial to an organism.
