190 THE PRINCIPLES OF SCIENCE 



materialize if the parts considered are insufficiently "natural." And 

 du Noiiy underlines this difficulty with his vigorous assertion that 

 completely natural parts are completely unattainable, just as Aris- 

 totle contended ( see p. 152 ) : 



When we speak of a phenomenon, we speak only of an event, or of 

 a succession of events, arbitrarily isolated from the universe whose 

 evolution they share. By isolating a fact in order to study it, we 

 give it a beginning and an end, which are artificial and relative. In 

 relation to the evolution of the universe, birth is not a beginning, and 

 death is not an end. There are no more isolated phenomena in nature 

 than there are isolated notes in a melody. 



Undeniably, our first eflForts at theoretical comprehension ordinarily 

 center not on the complex general case actually encountered, but on 

 some highly restricted special case— an ideal "phenomenon" con- 

 spicuously artificial in its simplicity. To begin almost any theoretical 

 problem may then demand something of that ruthlessness, reckless- 

 ness, noted in one case by Pirie. 



Nowadays most thoughtful biologists realize that the evolutionary 

 origin of species is bound to give intermediate types and there may 

 be a transient continuity between adjacent species. But anyone com- 

 batively aware of this in the eighteenth century was a nuisance. 

 Then the most useful approach was a certainty that this is this and 

 that that, and that the trouble with the small proportion of doubtful 

 specimens was simply due to lack of knowledge. 



So slashing away with conceptual scalpels, we can count as success 

 only an operation in which we do recover our patient alive, entire. 

 We know we have succeeded when, after thinking about parts, we 

 can in thought reconstitute wholes. After abstracting from "non- 

 essential complications," we must in the end render just account even 

 of them— precisely as in the case just noted, where we do at last re- 

 integrate the intermediate types, and so grasp both the distinctness 

 and the continuity of "species." 



Superposition, not summation. The whole we seek to comprehend 

 may well differ from any sum of parts. Even common sense recog- 

 nizes that a mob is no mere sum of constituent individuals: it has 

 certain new "qualities," and lacks others found in individuals. Thus a 

 correct superposition may sometimes deiucind— beyond addition of 

 parts— due attention to the mutual interactions of those parts. En- 



