SCIENCE AND NATURE 88 



But I mention this matter for another reason. There 

 are persons under what we may call permanent hallu- 

 cinations : colour-blind people are an instance. There 

 are people who say that, to them, two objects, which to 

 normal people appear one as pink and the other as greenish- 

 blue, appear exactly the same colour ; and no traps 

 set for them will show any inconsistency in their judg- 

 ment. They will maintain their position when all their 

 interests lie in an ability to distinguish the colours. In 

 such cases universal agreement cannot be obtained. Are 

 the judgments to be excluded from the subject-matter of 

 science ? The answer is, Yes ; they are excluded. 

 And the fact that they are excluded is a support for the 

 definition of science which has been offered ; for there is 

 no doubt that they would have to be included if science 

 studied simply the properties of the external world. 

 Strange as it may appear to the uninitiated, colour, 

 judged by simple inspection, is not a scientific conception 

 at all, and it is not a scientific conception because 

 universal agreement cannot be obtained about it. The 

 procedure adopted is this. We find that normal people 

 regard the objects A, B, C, . . . as all pink and the 

 objects X, Y, Z, . . . as all blue ; colour-blind people, 

 on the other hand regard A, B, C, . . . X, Y, Z, as indis- 

 tinguishable in colour. But we find also that there is 

 some other property in which both normal and abnormal 

 people find that A, B, C, . . . agree and that X, Y, Z, . . . 

 agree, while in respect of this property both normal and 

 abnormal find that A, B, C, . . . differ from X, Y, Z, . . . 

 When we find that, we regard this new property as 

 the true and scientific test of colour ; for about this 

 property we can obtain universal agreement. And we 

 call some people abnormal, not merely because they fail 

 to agree with the majority, but because the}^ fail to make 

 a distinction where it is universally agreed that there is 

 a distinction, 



