xiv.] FORBES' SCIENTIFIC WORK. 481 



legitimate inquiry to try to assign numerical values to 

 the force of belief in any given cases. All this folly, for 

 one cannot give it any other name, grows out of such 

 statements as "certainty is equal to unity/' and the like. 

 It belongs to the school of Condillac and the sensa- 

 tionalists they were in the ascendant when the theory 

 <>f probabilities received its present form, and there has 

 not yet been philosophy enough to expel it. 



* I hope you will notice the exquisite non sequitur in 

 Mitchell's argument. All that it can tend to prove is, 

 that there is some reason why so many stars appear very 

 r one aD other. Let this be granted, and that you 

 have a numerical expression of the cogency of the con- 

 clusion. What connection is there between this and the 

 conclusion that they are near or that there is a phy- 

 1 connection, to use Herschel's phrase, among them ? 

 How if the observer is drunk \ There will then be reason 

 enough why he should see double stars, and a reason 

 which applies to all the stars he looks at. Mitchell's 

 conclusion, therefore, would be justified in the case of 

 such an observer. Only the common reason here apply- 

 ing to all the stars is merely subjective and phenomenal. 

 And to set aside this illustration, how are we in any case 

 numerically to estimate the chance that the common 

 n really belongs to the essential nature of the 

 question ? ' . . . 



'BELGRAVE PLACE, BRIGHTON, 

 October lOtfi [1850]. 



' MY DEAR FORBES, 



' . . . I send you the second sheet, on which I 



less to remark than on the first, little as that waa 



I should rather have liked you to insist more on the 



t that such phrases as mare clwnce, at random, and 



lik- have n ally no meaning at all except with 



rence to the knowledge of the beervei and with his 



ng phenomena. I-'nr everything which 



lejinit. why it is what it i 



: the only question l'< in^ what analog ainmi- 



the causes of analogous \-.\u numena ; in other words, what 



r i 



