44 THE METHOD OF SCIENCE 



thinking on the result of the experiment, in other words, if we 

 reason, not towards the experiment, but from it, then our thinking 

 is not tested by it. We get information which may be very 

 accurate and very valuable ; but we get no more than this informa- 

 tion. The subsequent thinking, if any, may, or may not, be 

 accurate. To discover its accuracy we have to use tests which that 

 particular experiment does not furnish. Now it is a very common 

 delusion that, because experiment when used as a test tends to render 

 thinking accurate, therefore all thinking connected with experiment 

 tends to be accurate. Consequently a neglect to test thinking 

 founded on experiment, combined with an assumption that such 

 thinking is necessarily accurate, is very frequent. A hypothesis, 

 whether resulting from induction or deduction, which has stood a 

 valid experimental test, has stood a single test so nearly conclu- 

 sive that it can be made more conclusive only by continued testing 

 by other appeals to reality. But a hypothesis which is merely 

 founded on obscured facts revealed by experiment has no special 

 claims to accuracy. For example, if I, having a notion that Brown 

 wishes to enter my house surreptitiously, give him the opportunity 

 and entrap him, my hypothesis is tested by experiment. It is then 

 very nearly conclusively proved ; the only doubt remaining being 

 due to the fact that at the time that I supposed Brown wished to 

 enter my house he might not yet have thought of it. But if, on 

 the fact revealed by the experiment, I build the notion that he 

 came to steal the spoons, I must test my new hypothesis by a fresh 

 appeal to reality. It matters not, however, whether I discover the 

 spoons on him by a formal experiment (i.e. by searching him), or 

 whether I simply observe them sticking out of his pocket. In 

 either case the appeal to reality, if successful, is equally decisive. 

 In everyday life it is often not worth while to test our hypotheses, 

 but, however much science occupies our thoughts, our thinking is 

 never scientific unless we do so. 



74. The two functions of experiment, testing and discovery, are 

 sharply distinct. Nevertheless they are often confused by people 

 who claim for hypotheses jfo&TZdfo/ on experiment the high scientific 

 status of theories tested by experiment. We shall see that while many 

 important biological hypotheses have been founded on experiment, 

 there is hardly an example of one which has been tested byit,or which 

 it is practicable to test by it. Moreover, we must ever keep in mind 

 the manifest truth that the field of biological experiment is limited. 

 A suitable experiment, whether to test thinking or reveal facts, 

 cannot always be devised. Thus, in the case we have just considered 



