THE RATIONAL ELEMENT IN GEOGRAPHY 469 



ination," as Tyndall put it. The second or structural class has 

 at least the negative merit of not being dangerous, but it fails to 

 satisfy the student who has left empiricism in search for rational- 

 ism in his geographical work. The third or explanatory, ra- 

 tional, and genetic class is stimulating to the investigator, but it 

 is objected to by conservatives as involving grave risk of error, 

 because the explanations on which its terms are based may be 

 incorrect. It is interesting to inquire which of these classes of 

 terms a geographer shall employ in his studies or which a teacher 

 shall use with his scholars. 



The practical worker will at first probably employ some terms 

 from all three classes, because he finds no one class complete in 

 itself. According to his temperament, he will feel a preference 

 for one class or another, and he will very likely venture now and 

 then to suggest new terms appropriate to his favorite class, if his 

 attention and interest are directed closely to a special field of 

 research where existing terms are insufficient for his needs ; but 

 as soon as he begins to use the genetic class he finds that the 

 success of his work is marked by the freedom and confidence with 

 which he can use explanatory terms, and that just as the greater 

 includes the less, so the genetic include the empirical and the 

 structural. The teacher, as well as the investigator, will then 

 feel an increasing discontent with the blind and dull empirical 

 terms, however safe they may be, and with the structural terms, 

 however essential they may be. He will press forward in the 

 hope that all the land forms with which he is concerned may in 

 due time be vouchsafed as full and certain explanatory descrip- 

 tion as many of them have already received. 



A double reward comes to the teacher who leads his scholars 

 beyond empirical description toward rational explanation: a 

 much greater interest is excited in geography as its meaning is 

 found to be richer, and soon afterward a greater power of obser- 

 vation is developed in response to the discovery of many corre- 

 lations among the elements of land forms that springs from their 

 explanation. Herein lies the practical value of a method that 

 may thus far seem chiefly theoretical. 



A concrete case may be illustrated by the diagrams on pages 

 470 and 471. Empirical description will see a narrower and a 

 1 1 i-i >ader canyon : the walls of the first consist of cliffs and slopes ; 

 the walls of the second consist of cliffs, slopes, and platforms. No 

 correlation is sought for between structure and form, for empirical 

 description does not concern itself with correlations. 



