596 



SCIENTIFIC THOUGHT. 



phenomena reduces us, distinguishes between probable 

 and improbable events or arrangements of a crowd 

 of elements i.e., between such as are of an average 

 and such as are of an exceptional character. Any 

 highly improbable arrangement though possible will 

 be followed by a gradual settling down to more prob- 

 able or average arrangements. And as in nature you 

 are forced to introduce the conception of availability, so 

 in the calculus of chances you can introduce a certain 

 mathematical quantity which is the measure of the proba- 

 bility. The more improbable, i.e., exceptional, the begin- 



not a refutation, but a confirmation, 

 of our theory. But one must not 

 consider the matter thus : as if 

 two gases . . . which were initially 

 unmixed, then became mixed, after 

 a few days again unmixed, then 

 again mixed, &c. We find, rather, 

 that . . . only after a period 

 which, even compared with 10 1 10 

 years, is enormously great, a per- 

 ceptible unmixing would take place. 

 That this is practically equivalent 

 to never, we see, if we consider 

 that in this period there would 

 be, according to the laws of prob- 

 ability, many years in which, by 

 mere chance, all the inhabitants of 

 a large city would, on the same 

 day, commit suicide, or fire break 

 out in all its buildings ; whereas 

 the insurance companies are in so 

 good an agreement with facts that 

 they do not consider such cases at 

 all. If even a much smaller im- 

 probability were not practically 

 identical with impossibility, nobody 

 could rely upon the present day 

 being followed by night and the 

 latter again by day." And further 

 (p. 255) : " If we, therefore, repre- 

 sent the world under the figure of 

 an enormously large mechanical 

 system, composed of enormously 



numerous atoms, which started 

 from a very perfectly ordered 

 condition, and exist still mainly 

 in an orderly condition, we arrive 

 at consequences which actually 

 stand in perfect harmony with 

 observed facts " ; and (p. 258), 

 " That in nature the transition 

 from a probable to an improbable 

 condition does not happen as fre- 

 quently as the reverse, can be ex- 

 plained by the assumption of a 

 very improbable initial state of the 

 whole surrounding universe, in 

 consequence of which any arbitrary 

 system of interacting bodies is, in 

 general, in an improbable condition 

 to begin with. But one might 

 say, that here and there the 

 transition from probable to im- 

 probable conditions must, after 

 all, be observable. . . . From the 

 numbers regarding the inconceiv- 

 ably great rarity of a transition 

 from probable to improbable con- 

 ditions, happening in observable 

 dimensions and during an ob- 

 servable period, it is explained 

 how such a process within what 

 we, cosmologically, call a single 

 world, or, specially, our world, is 

 so extremely rare that any experi- 

 ence of it is excluded." 



