xi KNOWLEDGE OF CONCRETE OBJECTS 265 



Upon this interpretation Mr. Lloyd Morgan remarks: 



" Now, here a course of action, which is quite readily inter- 

 preted on the hypothesis of sense-experience through association, 

 is explained in rational terms, and the dog is supposed to think the 

 therefore. But on similar principles every case of association may 

 be thus interpreted." l 



He proceeds to show how the refusal of a cinnabar 

 caterpillar by a chick may be interpreted as a rational act, 

 and thrown into syllogistic form. 



a Nothing is easier than to interpret this intelligent course of 

 action in rational and logical terms : Black and gold caterpillars 

 are nasty ; that is a black and gold caterpillar ; therefore it is nasty. 

 This does not seem to me to be at all more far-fetched than the 

 suggestion of Mr. Stone's that his dog reasoned : That dog is 

 barking out of doors, therefore he is not in this room." : 



I quite agree with Mr. Lloyd Morgan that there is no 

 evidence that a dog can " think the therefore." That is 

 to say, it has no explicit consciousness of the thought- 

 connection uniting one stage of the inference to the next. 

 I see nothing whatever to suggest that an animal can 

 analyse out this element in its own mental processes as a 

 man does when he uses an illative particle. But this does 

 not seem a sufficient ground for denying to the operation 

 the title of inference. 3 To the animal much less is explicit 

 than to the man, but to the man much less is explicit than 

 to the philosopher, and to the philosopher much less is 

 explicit than he could wish. Inference is essentially one 

 function, from the simplest case quoted by Mr. Morgan 

 of the chick, up to the highest elaboration of experience 

 by the human intellect. The differences are differences in 

 articulateness on the one side, and comprehensiveness on 

 the other ; and the question is, precisely what point in the 



1 Comparative Psychology, p. 300. 2 Id. p. 301. 



3 It is no doubt in part a question of terms. I have used the term 

 inference in a very wide sense to express a certain generic unity of 

 function exercised with very different degrees of articulateness. It is, of 

 course, possible for any writer, while recognising this generic unity, to 

 prefer to keep the term inference for the higher grades. This difference 

 of usage is, as I gather, responsible for making the divergence between 

 Mr. Lloyd Morgan's views and mine appear much greater than it 

 really is. 



