152 CREATIVE EVOLUTION [chap. 



have said that the function of intelligence is to establish 

 relations. Let us determine more precisely the nature 

 of these relations. On this point we are bound to be either 

 vague or arbitrary so long as we see in the intellect a faculty 

 intended for pure speculation. We are then reduced to 

 taking the general frames of the understanding for some- 

 thing absolute, irreducible and inexplicable. The under- 

 standing must have fallen from heaven with its form, 

 as each of us is born with his face. This form may be 

 defined, of course, but that is all; there is no asking why 

 it is what it is rather than anything else. Thus, it will be 

 said that the function of the intellect is essentially uni- 

 fication, that the common object of all its operations is to 

 introduce a certain unity into the diversity of phenomena, 

 and so forth. But, in the first place, "unification" is a 

 vague term, less clear than "relation" or even "thought," 

 and says nothing more. And, moreover, it might be asked 

 if the function of intelligence is not to divide even more 

 than to unite. Finally, if the intellect proceeds as it does 

 because it wishes to unite, and if it seeks unification simply 

 because it has need .of unifying, the whole of our knowledge 

 becomes relative to certain requirements of the mind 

 that probably might have been entirely different from 

 what they are: for an intellect differently shaped, know- 

 ledge would have been different. Intellect being no longer 

 dependent on anything, everything becomes dependent 

 on it; and so, having placed the understanding too high, 

 we end by putting too low the knowledge it gives us. 

 Knowledge becomes relative, as soon as the intellect is 

 made a kind of absolute. — We regard the human intellect, 

 on the contrary, as relative to the needs of action. Postu- 

 late action, and the very form of the intellect can be deduced 

 from it. This form is therefore neither irreducible nor 

 inexplicable. And, precisely because it is not independent, 



