CHANCE. 89 



l^ut as regards the number tt the question presents 

 more difficulties, and for the moment I have no 

 satisfactory explanation to give. 



There are many other questions that might be 

 raised, if I wished to attack them before answering 

 the one I have more especially set myself When we 

 arrive at a simple result, when, for instance, we find 

 a round number, we say that such a result cannot be 

 due to chance, and we seek for a non-fortuitous cause 

 to explain it. And in fact there is only a very slight 

 likelihood that, out of 10,000 numbers, chance will 

 give us a round number, the number 10,000 for in- 

 stance ; there is only one chance in 10,000. But 

 neither is there more than one chance in 10,000 that 

 it will give us any other particular number, and yet 

 this result does not astonish us, and we feel no hesita- 

 tion about attributing it to chance, and that merely 

 because it is less striking. 



Is this a simple illusion on our part, or are there 

 cases in which this view is legitimate? We must 

 hope so, for otherwise all science would be impossible. 

 When we wish to check a hypothesis, what do we 

 do? We cannot verify all its consequences, since 

 they are infinite in number. We content ourselves 

 with verifying a few, and, if we succeed, we declare 

 that the hypothesis is confirmed, for so much succe.ss 

 could not be due to chance. It is always at bottom 

 the same reasoning. 



I cannot justify it here completely, it would take 

 me too long, but I can say at least this. We find 

 ourselves faced by two hypotheses, either a simple 

 cause or else that assemblage of complex causes we 

 call chance. We find it natural to admit that the 



