Note to Rev. W. V. Harcourt's Letter. 37 



disentangle, as far as we can, the knot into which three distinct 

 questions — namely the investigation of the laws of nature, the 

 prospect of their continuance, and the degrees of certainty and 

 presumption on which we judge, and on which we act — are 

 bound and mixed up together in the passages now referred 

 to: let us then take the principles of evidence here assumed 

 by Butler, and subject them to a strict analysis. 



For this purpose we will vary the subject of his illustration. 

 We will suppose a man to have observed, instead of the ebb 

 and flow of the tide, a flash of lightning, or the fall of a stone 

 from the air. Shall we say that the having observed one of 

 these phenomena today affords some sort of presumption, 

 though the lowest imaginable, that it may happen again to- 

 morrow ? On the supposition that we are in the dark con- 

 cerning the causes and circumstances which determine these 

 events, does not the doctrine of chances teach us, on the con- 

 trary, that there are an infinite number of presumptions to 

 one, that the same event will not happen again tomorrow, or 

 on any other day that can be named ? The solitary event does 

 indeed suggest to us, not " the lowest imaginable presump- 

 tion," but the highest possible certainty, that there are causes 

 in operation which may at any time produce the same result; 

 but apart from the suspicion of some permanent cause, it does 

 not afford a shadow of a presumption that the same event will 

 happen again at all. 



Dismissing then the notion that a certainty is a bundle of 

 low presumptions from single facts, let us consider what the 

 nature of the evidence really is which is furnished by the re- 

 petition of an event. 



If a die, presumed to be on all sides of uniform weight, be 

 thrown, there is no presumption that the same number will 

 come up on the next throw, the presumptions being on the 

 contrary as many against that as against any one of the numbers 

 marked on the die. If however in several successive throws 

 the same number does come up, a presumption does presently 

 arise, and increases rapidly with the repetition of the throws, 

 that the same number will continue to come up. The origin 

 of this presumption is obvious. The violation of the indiffer- 

 ence of the chances has indicated a definite cause, which is 

 conceived to determine the result of the throw. 



If I have seen the crater of a volcano constantly smoking 

 for fifty years, I shall entertain an expectation of its smoking 

 tomorrow : but the presumption that it will smoke tomorrow, 

 or that it will continue to smoke for fifty years to come, is 

 very far from being as violent as the presumption that a die 



