PHYSICAL METHOD. 581 



observation, can only be relied upon as applicable to 

 unobserved cases, so long as there is reason to tbink 

 that no change has taken place in any of the remote 

 causes on which the immediate . causes depend. In 

 making use, therefore, of even the best statistical 

 generalizations for the purpose of inferring (though it 

 be only conjecturally) that the same empirical laws 

 will hold in any new case, it is necessary that we be 

 perfectly well acquainted with the remoter causes, in 

 order that we may scrupulously avoid applying the 

 empirical law to cases which differ in any of the cir- 

 cumstances on which the truth of the law ultimately 

 depends. And thus, even where conclusions derived 

 from specific observation are available for practical in- 

 ferences in new cases, it is necessary that the deduc- 

 tive science should stand sentinel over the whole 

 process ; that it should be constantly referred to, and 

 its sanction obtained to every inference. 



The same thing holds true of all generalizations 

 which can be grounded on history. Not only there 

 are such generalizations, but it will presently be 

 shown that the general science of society, which 

 inquires into the laws of succession and coexistence of 

 the great facts constituting the state of society and 

 civilization at any time, can proceed in no other 

 manner than by making such generalizations after- 

 wards to be confirmed by connecting them with the 

 psychological and ethological laws on which they 

 must really depend. 



6. But (reserving this question for its proper 

 place), in those more special sociological inquiries 

 which form the subject of the separate branches of the 

 social science, this two-fold logical process and reci- 

 procal verification is not possible ; specific experience 



