118 HABIT AND INTELLIGENCE. [chap. 



caused by my having eateu salt meat an hour ago. We 



must consequently modify the definition of perception 



the sources offered above, and call it that act hy ivJiich the mind refers 



presfut sensations to sources present in time. This is the most 



ill time, accurate definition I can offer, but it is not perfectly 



accurate ; it includes as perceptions some acts which are 



really inferences. Thus, I may cognise a local pain arising 



from disease, and I may know with perfect certainty the 



state of the diseased part which causes the pain ; yet the 



knowledge by which I assign the painful sensation to its 



source may be a case, not of perception, but of very 



circuitous inference. 



1 believe, however, that it is in the nature of things 

 A per- impossible to frame a definition which shall be such as 

 aecurate ^^ include all that are really acts of perception, and to 

 defiuitiou exclude all that are really acts of inference. And the 

 sible. reason of this, in my opinion, is that perception and 

 inference graduate into each other. Perception indeed 

 is an inference ; it is only the first and simplest case 

 of inference. When we refer a sensation to its source 

 ' — as, for instance, when we refer light to a star, and 

 in so doing perceive the star — I do not regard the per- 

 Percep- ceptiou as a simple inexplicable act of the mind ; I 

 inference, regard it as an inference, instantaneously and spon- 

 taneously made. This view is supported by the fact that 

 The same the same act may appear to be a perception or an in- 

 bethc^oue feience, according as it is performed at once and spon- 

 or the taneously, or with hesitation and with some effort of 

 cording to thought. Thus, uogs perccive by the smell. A dog will 

 smell a dead animal and perceive at once where it is ; 

 wlien a man may ascertain its presence only by con- 

 sidering for some little time what the source of the 

 ■unpleasant smell can be, and even then may remain in 

 some degree of doubt. It is, I think, quite impossible to 

 point out any difference between the dog's unhesitating 

 perception and the man's hesitating inference, except the 

 mere circumstance that the one is instantaneous, while the 

 other occupies an appreciable time. It may be said that 

 perception is an unconscious process, while inference is a 



circum- 

 stances. 



