558 



SCIENCE 



[N. S. Vol. XXXIX. No. 1007 



view of things the killing resembles an 

 eclipse. Unique as it is, it is supposed to 

 have been essentially predictable. Perhaps 

 if you had known the precise configuration 

 and the accelerations of all the physical 

 particles in the world at some appropriate 

 moment, then this killing could have been 

 calculated in advance. It is a mere case of 

 a law — an eclipse, so to speak, of some sun 

 ■ — a point on some curve. 



But for a statistical view the single kill- 

 ing of -B by A is an event against which an 

 insurance provision could have been made 

 in advance — not because any mortal could 

 have predicted whether or no A would kill 

 B, but because the death-rate of men of 

 5's age and occupation can be statistically 

 known with approximate and probable ac- 

 curacy, so as to make a policy insuring B's 

 life a contract whose value is calculable, 

 not on mechanical but upon statistical 

 grounds. 



Now you will easily recognize that the 

 actual knowledge of vital phenomena which 

 science possesses is, in the main, a statis- 

 tical knowledge. It is the sort of knowl- 

 edge which the mortality tables of the in- 

 surance companies exemplify. We know 

 little of the history of individual organ- 

 isms, and less of their mechanism, but we 

 can and do study the statistics of groups 

 of organisms. In such statistical terms 

 heredity and variation are now constantly 

 investigated. In such terms growth and 

 disease, as well as death, economic prosper- 

 ity and social transformations, financial 

 and political processes, the geographical 

 distribution of organisms and the gradual 

 accumulation and change of the material 

 as well as the mental products of civiliza- 

 tion — in such statistical terms, I say, all 

 such things come to be the objects of scien- 

 tific description and explanation. To give 

 an account of the special phenomena of life 

 in terms of mechanism remains in practise 



a remote ideal, despite all the proofs that 

 the vital processes, being subject to phys- 

 ical and chemical laws, must be, in some 

 sense, if not wholly, then very largely me- 

 chanical in their nature. Life may be a 

 case of mechanism; but its phenomena are 

 best known to science in terms of statistical 

 averages, of laws which hold approxi- 

 mately true regarding these averages, and 

 of probabilities which are definable in 

 such terms as are used when the insurance 

 value of a life-policy is computed. The 

 logic of the insurance actuary is essentially 

 the same as the logic which is consciously 

 or unconsciously used in dealing with all 

 forms and grades of vital processes. 



This general rule regarding the methods 

 of the sciences of life is well known to you. 

 For it is also known that, just as a me- 

 chanical theory of the details of the phe- 

 nomena of life still remains a remote ideal, 

 so too an historical knowledge of the indi- 

 vidual events of the life of an organism is 

 something which may possess upon occa- 

 sion great moral or social or perhaps clin- 

 ical interest, but can occupy but a part, 

 and usually a very small part, of the inter- 

 est of the sciences of life. 



Into the study of human history itself, 

 devoted as such a study naturally is to the 

 sequences of individual events, natural 

 science enters in so far as something of the 

 nature of statistical knowledge is acquired. 

 And therefore the use of deliberately sta- 

 tistical methods in historical study, the use 

 which Dr. Woods has recently proposed — 

 such a use, I say, is in principle nothing es- 

 sentially opposed to methods long since in- 

 exactly and unconsciously employed. For 

 the historiometry of Dr. Woods is in prin- 

 ciple a legitimate extension and a logically 

 legitimate refinement of the long since well- 

 known disposition to explain human his- 

 tory in terms of "historical tendencies" 

 and of "historical forces." 



