June 17, 1910] 



SCIENCE 



937 



As a fiirther contrast, all the many mem- 

 bers of an extended empirical series of 

 ideal types must be learned arbitrarily 

 and separately, for no mnemonic aid from 

 explanation attaches to any of them. All 

 the members of an extended explanatory 

 series may be divided into groups, so that 

 the groups themselves shall have certain 

 highly suggestive general relationships, 

 and so that the members of each group 

 shall be treated as systematically inter- 

 dependent and easily remembered. . The 

 development of the explanatory series is 

 immensely aided by the mental process of 

 deduction, which may be carried on- by a 

 trained student anywhere and at any time 

 at his convenience; but deduction has no 

 significant place in the preparation of the 

 empirical series, each member of which 

 must originally be learned by some ob- 

 server, traveling about in the actual 

 world. 



Having now pointed out the strong con- 

 trasts between these two kinds of type 

 forms, in terms of which the descriptions 

 of natural landscapes and regions must be 

 made, let me hasten to state that no one 

 to-day uses either kind in its purity. The 

 most conservative empiricist will introduce 

 some explanatory types and terms in con- 

 nection with forms of which the origin 

 is manifest, such as sand dunes, deltas, 

 volcanoes and .sea cliffs ; while the most de- 

 termined rationalist will not infrequently 

 find certain actual features which he can 

 not explain, and for which he can there- 

 fore establish no corresponding explana- 

 tory types. The difference between the 

 empiricist and the rationalist is therefore 

 not so much in their practise as in their 

 intention. The empiricist introduces ex- 

 planatory terms as it were by accident ; he 

 makes no conscious effort to substitute ex- 

 planatory tjTDes for empirical types, and 

 he has no definite intention of introducing 



explanation as the most effective means of 

 description. The rationalist, on the other 

 hand, consciously and intentionally strives 

 to find out the origin of every form that 

 he observes, and then tries to describe 

 every observed form systematically in 

 terms of deductively developed type forms. 

 The conservative empiricist condemns the 

 rash rationalist as using a dangerous 

 method, in that it must often be unsafe to 

 describe what one sees in terms of what 

 one does not and can not see ; and in that 

 it is unwisely venturesome to introduce- 

 theoretical considerations, which are in. 

 many cases necessarily more or less doubt- 

 ful, instead of holding to direct observa- 

 tion which is essentially safe. The san- 

 guine rationalist criticizes the hesita- 

 ting empiricist as using a blind method, in 

 that it is short-sighted to describe only 

 those things which can be seen with the 

 outer eyes, and unreasonable to omit all 

 those illuminating explanatory considera- 

 tions, theoretical though they be, by which 

 so much light is thrown on empirical facts, 

 and by which the way is indicated to many 

 facts which the empiricist overlooks. 



My own preference for the explanatory 

 method is so strong that the preceding 

 paragraphs have probably done some in- 

 justice to the empirical method. Be this 

 as it may, it seems to me a plain duty to 

 use to the utmost every explanatory rela- 

 tion that we can discover, in so far as it 

 aids us in describing existing landscapes. 

 If the explanation seems assured, it may 

 be used without qualification ; if it appears 

 somewhat ventiiresome, explicit notice may 

 be given of its insecurity by introducing 

 warning words; for example, "as if." 

 The extraordinary advances made in the 

 understanding of the evolution of land 

 forms in the last half century, particularly 

 those advances made by the government 

 geological surveyors in the arid southwest- 



