206 ETHICAL ASPECTS OF EVOLUTION 



ment. But subjective states are not measurable ; exact 

 similarity is not, therefore, demonstrable with regard to any 

 two of them, and it is impossible to bring them under the 

 law of uniform sequence. Moreover, as, in order to the 

 establishment of any proposition under that law, both 

 terms must be defined, they cannot be predicated either as 

 cause or effect in any proposition of which the other term 

 is an event in the external world ; or, rather, if so predicated, 

 we cannot conclude that the sequence is invariable, any 

 more than we can conclude from the observation that even 

 a large number of eggs are good that all eggs are good. 

 It should be added that the measurements of objective 

 states are as exact as our practical demands require them 

 to be. They enable us to predict the future with a fair 

 degree of certainty; and this is all that we ask of them. 



Impossibility has been defined as that, the truth of 

 which would conflict with a complete induction, that is, 

 with the most conclusive evidence we possess of universal 

 truth. 1 But this seems to overlook the fact, which is else 

 where fully recognized by the author, that even the best 

 established inductions are in a sense hypothetical ; that is 

 to say, they hold good only in the absence of causes which 

 may have hitherto escaped our notice, or which (such as 

 an unknown comet) have not yet entered the field of our 

 experience. When the possible existence of unknown 

 disturbing causes is recognized, it will be seen that there 

 is no known induction which may not be contradicted without 

 any interference with the general law that like follows like. 

 The introduction of fresh matter alters the terms of a pro 

 position but does not affect the validity of the law by which 

 it is tested. In order that that law may be contradicted, 

 it would be necessary to show that from a single perfectly 

 defined antecedent, or from two which were exactly alike, 

 one effect has followed at one time or in one place, and 

 another differing effect at another. We should then, 

 and only then, find a real contradiction to the law of uni 

 formity, and this, if compliance with that law is the sole 

 1 J. S. MiU, Logic, III, xxv. 3 (note). 



