204 DOGMATISM AND EVOLUTION 



istic of actual human thought, at least to the observation of the 

 trained student of human nature, is the more or less limited fixity 

 and stability of its terms. They are products of an evolution 

 which still proceeds. And though we cannot in many instances 

 distinguish, or even imagine, the particular changes that may 

 have taken place within the period of human history, and must 

 even grant that certain concepts have, in all probability, remained 

 substantially unchanged for ages, we cannot avoid recognizing 

 at least the possibility of their future modification. In no case 

 have we sufficient warrant to guarantee the permanent fixity of 

 the existing forms; and, in fact, it is only within the domain of 

 the mathematical sciences that such fixity could be claimed with 

 any show of reasonableness. Of the great mass of our concepts 

 we can scarcely doubt that they are changing now more rapidly 

 than ever before. 



But where concepts are undergoing an evolution, a precise 

 clearness cannot be expected. Where distinctions are hardening 

 and melting away again and shifting generally, it is impossible 

 that dividing lines should be shadowless and unbroken. Bacon's 

 aphorism, that ultimately satisfactory definitions belong, not to 

 the initial stages, but to the consummation of the sciences, is 

 significant to us as the description of a never to be attained ideal. 

 The conviction of clearness is common enough. But we have 

 well learned that there is no more suspicious indication of shallow- 

 ness of mind. The nearer any concrete reasoning approaches 

 the mathematical type, the readier we are to condemn it as 

 doctrinaire. 



The weakness of the syllogism, that supposed universal form 

 of thought, is now evident. The possibility of drawing a con- 

 clusion depends upon the exact identity of the middle term in 

 the two premises. But who shall vouch for this? Not to the 

 satisfaction of common sense alone, but in accordance with the 

 canons of the syllogism itself? For by these canons the least 

 variation constitutes a quaternio, and no valid inference is then 

 possible. In fact, so far from being an absolutely certain mode 



