54 THE GRAMMAR OF SCIENCE 



he will draw from his perceptions this conception and not 

 its negation. It is in this sense, therefore, that we are 

 to understand the assertion that a conception to have 

 scientific validity must be deducible from the perceptions 

 of the normal human being. 



The preceding paragraph shows us how important it is 

 that the observations and experiments of science should 

 be repeated as often and by as many observers as possible, 

 in order to ensure that we are dealing with what has 

 validity for all normal human beings, and not with the 

 results of an abnormal perceptive faculty. It is not only, 

 however, in experiments or observations which can be 

 repeated easily, but still more in those which it is very 

 difficult or impossible to repeat, that a great weight of 

 responsibility lies upon the recorder and the public which 

 is called upon to accept his results. An event may have 

 occurred in the presence of a limited number of observers. 

 That the event itself cannot recur, and that it is totally 

 out of accord with our customary experience, are not in 

 themselves sufficient grounds for disregarding it scientific- 

 ally. Yet what an onus is laid on the individual 

 observers to test whether their perceptive faculties were 

 normal on the occasion, and whether their conceptions of 

 what took place were justified by their perceptions ! Still 

 greater onus is laid on men at large to criticise and probe 

 the evidence given by such observers, to question whether 

 they were men trained to observe, and calm and collected 

 at the time of the reported event. Were they not, 

 perhaps, in an exalted state of mind, biassed by pre- 

 conceptions or hindered by the physical surroundings 

 from clear perception ? In short, were or were not their 

 perceptive faculties in a normal condition, and were or 

 were not the circumstances such that normal perception 

 was possible ? It can scarcely be questioned that when 

 the truth or falsehood of an event or observation may 

 have important bearings on conduct, over-doubt is more 

 socially valuable than over-credulity,^ In an age like our 



1 A good example of another class of experiment, that which it is difficult 

 or unadvisable to repeat frequently, may be drawn from Brown -Sequard's 



