Southern Appalachian Mountain Field Ruined by Erosion 



them ; the reasoning is from general 

 principles to particulars. Omitting 

 metaphysical subtleties, it may be said 

 that the .general principles which form 

 the starting points may be obtained in 

 either of two ways : they may be truths 

 which the mind perceives or arrives at 

 without experiment or observation, 

 such as the fundamental truths of math- 

 ematics and mechanics ; or they may be 

 truths which themselves have been ar- 

 rived at by induction and which are then 

 used as the basis of deductions. 



Some minds instinctively prefer one 

 of these methods, some prefer the other. 

 Some people are only satisfied by statis- 

 tical proof, or what appears to be such. 

 I presume some people would require a 

 statistical proof of a future life; they 

 would desire to have records of inter- 

 views with departed spirits, to know 

 how many of such departed spirits said 

 they were alive, and how many said they 

 were dead, and whether the proportion 

 of those who said they were alive had 

 increased in recent years. Some minds, 

 on the other hand, seem almost incapa- 



ble of inductive reasoning. They must 

 start with fundamental principles and 

 reason from them, and if they have no 

 fundamental principles to start with, 

 they will create them from their own 

 imaginations. 



Clearly, a proper combination of both 

 methods, using each where it is applica- 

 ble, is the basis of reasonable scientific 

 investigation. 



DIFFICULTIES WITH THE INDUCTIVE 

 METHOD 



There are several serious difficulties 

 which are likely to be encountered in 

 the application of the inductive method, 

 and many precautions which must be 

 taken if error is to be avoided. 



In the first place, a very essential 

 point to be observed is, that in order to 

 determine the effect upon Phenomenon 

 B which is exerted by Phenomenon A, 

 the observations or experiments must be 

 so conducted that, of all the elements 

 which may effect P>, all remain constant 

 excepting A. In other words, only one 



225 



