106 W. M. DAVIS RELATION OF GEOGRAPHY TO GEOLOGY 



Empirical concepts — I still speak especialh' regarding land forms — 

 are known only as far as direct observation can penetrate, and they are 

 therefore necessarily superficial in space, short-sighted in time, and rigid 

 in definition. Explanatory concepts are known through and through, 

 fore and aft: the farther side of the concept of a ridge is seen just as 

 well as the near side, by the eye of the imagination, which takes any point 

 of view that it desires ; the inside of the ridge is seen as well as the out- 

 side, the past and the future forms of the ridge as well as the present 

 form, for all these concepts are avowedly mental concepts only and not 

 matters of fact. Explanatory concepts are, moreover, most elastic and 

 adaptable, so that they may be easily made to match the facts of nature. 

 Such concepts may be fanciful in the sense of not being necessarily coun- 

 terparts of any natural forms; they may be erroneous in the sense of 

 being incorrectly deduced from unsafe generalizations, and such chances 

 of error must be recognized and guarded against. But the prime fact 

 remains that explanatory concepts, deduced from general principles, are 

 much more intimately and reasonably knowable than empirical concepts 

 or even than facts of observation usually are, and in this quality of being 

 intimately and reasonably knowable lies their highest value. It is as if 

 one located them by sighting from many different points along the path 

 of time, and thus fixed their position by the intersection of many con- 

 verging lines of sight, while empirical concepts are located only by a 

 single line of sight running in one direction from the viewpoint of the 

 momentary present. 



GROUNDS OF CHOICE BETWEEX EMPIRICAL AXD EXPLANATORY METHODS 



Let it, however, be clearly understood that the object of explanatory 

 geographical descriptions is not the presentation of inferred facts con- 

 cerning the past history of the earth, but the presentation of the most 

 carefully defined concepts concerning the present eartli. This is essential. 

 The land forms of today may, of course, be used, just as stratified rocks 

 or fossil prints are used, to constitute an observational basis for inferences 

 regarding the past history of the earth, and the study of land forms with 

 this object is a part of geology just as much so as the similar study of 

 strata or of fossils, for it is, as already stated, the object of a study and 

 not the thing studied or the method of studying it that determines the 

 place of the study in a classification of the sciences. Inferences based on 

 the form of the upland of southern ISTew England as to the relatively 

 modern date of its uplift from a long residence at a lower altitude belong 

 under geology ; but descriptions of the uplands of southern New England 

 as the product of prolonged erosion during a lower stand of the land, fol- 



