Apkil 17, 1914] 



SCIENCE 



563 



IX. APPLICATIONS OP THE STATISTICAL VIEW 

 TO THEORIES OP NON-MECHANICAL 

 SYSTEMS. AGGREGATION AND 

 ASSIMILATION AS STATIS- 

 TICAL TENDENCIES 



It at first seems, I have said, as if the 

 statistical methods of the kinetic theory- 

 were applicable only to mechanisms whose 

 complications were too vast to make it pos- 

 sible to follow in individual detail their 

 necessary sequences of movements. 



But this seeming is unfounded. Let me 

 summarize in my own words a few consid- 

 erations which Peirce summarily states, 

 and which, to my mind, get a constantly 

 increasing importance as the statistical 

 view of nature comes to be applied to wider 

 and wider fields of research. 



Suppose an aggregate of natural objects 

 which contains a very great number of 

 members, each one of which is subject to 

 some more or less exhaustively definable 

 range of possible variations. These objects 

 may be things or events, at your pleasure. 

 They may be molecules or stars or cells or 

 multicellular organisms or members of a so- 

 ciety or observations of a physical quan- 

 tity or proposals of marriage or homicides 

 or literary compositions or moral agents 

 or whatever else you will. The essential 

 basis which is needed for a statistical view 

 of such an aggregate is this : 



First, the members of each aggregate 

 must actually form a collection which is, 

 for some physical or moral reason, a gen- 

 uine and therefore in some way a definable 

 whole. 



Next, some more or less systematic tend- 

 ency towards a mutual assimilation of the 

 fortunes, the characters or the mutual re- 

 lations of the members of this aggregate 

 must exist. This tendency toward mutual 

 assimilation may be of very various sorts. 



The policyholders of an insurance com- 

 pany tend to assimilate the fortunes of 



their various investments when they all of 

 them pay their premiums to the same com- 

 pany. The stars tend to a certain assimi- 

 lation of the mutual relations amongst those 

 photographs of their various spectra which 

 chance to get collected on the photographic 

 plates of the same astronomical observatory. 

 For, as a consequence of this aggregation 

 of photographs, the stellar spectra in ques- 

 tion may tend to be classified ; and the log- 

 ical, as well as the other socially important, 

 and the physical fortunes of objects which 

 are once viewed or arranged or tabulated 

 as objects belonging to the same class, tend, 

 in general, to a further mutual assimila- 

 tion. 



Birds of a feather not only flock together, 

 but tend to get statistically similar for- 

 tunes, when they come into chance contact 

 with other birds or with breeders, with 

 hunters or with biometrical statisticians. 



All objectively well-founded classifica- 

 tion is not only founded upon real similari- 

 ties amongst the objects which belong to an 

 aggregate, but tend to some increase of 

 these similarities, in so far as these objects 

 are not changeless mathematical entities, 

 but are natural objects, whose fortunes are 

 subject to change. 



One of the most widely applicable laws 

 of nature is, in fact, the law, wholly inde- 

 finable in mechanical terms, but always ex- 

 pressible in terms of statistical tendencies 

 - — the law that aggregation tends to result 

 in some further and increasing mutual as- 

 similation of the members of the aggregate. 

 This assimilation may express itself in the 

 fact that one classification or aggregation 

 leads both logically and physically to 

 another and deeper and also wider aggre- 

 gation. 



If the stars are already physically classi- 

 fied into two distinct drifts, which move 

 through each other in two different direc- 

 tions, and if the stars in question tend to 



