DISPERSION 15/3 



lines of behaviour. Thus Pavlov conditioned a dog so that it 

 gave no salivary response when subjected to the compound 

 stimulus of : 



the experimental room, the harness, the feeding apparatus, 

 the sound of a metronome beating at 104 per minute, and 

 the sound of a No. 16 organ pipe, 

 but gave a positive salivary response when subjected to : 



the experimental room, the harness, the feeding apparatus, 

 the sound of a metronome beating at 104 per minute, and 

 the sound of a No. 15 organ pipe. 

 Such a ' discrimination ' has been considered by some to be beyond 

 the powers of mechanism, but this is not so : all that is neces- 

 sary is that the system should be complex and should contain 

 part-functions. 



15/3. The same point of view helps to make clear the physio- 

 logical concept of ' adding ' stimuli. In the simple case it is 

 easy enough to see what is meant by the ' addition ' of two 

 stimuli. If a dog has developed one response to a flashing light 

 and another to a ticking metronome, it is easy to apply simul- 

 taneously the flashes and the ticks and to regard this stimulation 

 as the c sum ' of the two stimuli. But the application of such 

 ' sums ' was found in many cases to lead to no simple addition 

 of responses : a dog could easily be conditioned to salivate to 

 flashes and to ticks and yet to give no salivation when both were 

 applied simultaneously. Some physiologists have been surprised 

 that this could happen. Let us view the events in phase-space. 

 Suppose our system has variables a, b, c, . . . and that the 

 basal, c control ', behaviour follows the initial state a Q9 b , c , . . . 

 Suppose the effect of stimulus A corresponds to the line of 

 behaviour from the initial state a l9 b , c , . . . , and that of 

 stimulus B to the line from a , b l9 c 09 . . . Then the behaviour 

 after the initial state a l9 b lt c , . . . would correspond to the 

 response to the simultaneous presentation of A and B. If we 

 know the behaviours after A and after B separately, what can 

 we predict of the behaviour after their presentation simultane- 

 ously ? The possibility is illustrated in Figure 15/3/1, which 

 shows at once that the lines of behaviour from II and J in no 

 way restrict that from K, which represents, in this scheme, the 

 4 sum ' response. 



167 



