62 W J MCGEE — UNIFORMITARTANISM AND DEFORMATION. 



the consideration of terrestrial deformation. Excluding those montanic 

 regions in which the rocks are crumpled, then, it may be affirmed that 

 the prevailing movements of the earthcrust are of the kind observed 

 along a score of coasts and inferred from raised beaches and from forma- 

 tions and unconformities — that the normal corporeal movements of the 

 earth are essentially radial, only subordinately tangential. 



Methods of Acquiring Knowledge. 

 the empiric method. 



There are three interrelated modes of acquiring empiric knowledge con- 

 cerning the processes of nature : Stated in the inverse order of develop- 

 ment, the first and most trustworthy mode is direct observation; the 

 second mode is inference through homology, or reasoning from the like '■> 

 the third and least trustworthy is inference through analogy, or reason- 

 ing from the unlike. 



The value of direct observation depends on training. In order to be 

 fruitful, observation must be fertilized by inference, while inference can 

 be made safe only through frequent checking by comparison with the 

 facts, and trained observation is a composite process in which inference 

 plays an essential part. 



The value of inference through homology, or reasoning from like to 

 like, depends on several conditions : Tlie first condition resides in the 

 closeness of similarity among the things compared in their five primal 

 attributes of number, extension, motion, duration, and serial succession • 

 other conditions reside in habits of observation or in methods of inter- 

 preting the things compared, and still other conditions reside in the degree 

 of accuracy of the pictures, or records, or concepts, of the things compared . 

 Thus, while inference through homology is the safest of all reasoning, 

 there are nevertheless, in this mode of acquiring knowledge, degrees of 

 trustworthiness, grading down from a certainty hardly below that of 

 trained observation to an uncertainty hardly above that of inference 

 through analogy. 



The value of inference through analogy, or reasoning'from the like to 

 the like only in part (and therefore in strict sense to the unlil<:e), simi- 

 larly depends on a variety of conditions. The primary condition resides 

 in the number and variety of things compared or contrasted, and other 

 conditions reside in habits of observation and in methods of interpreting, 

 as well as in the accuracy of records and concepts. Thus reasoning 

 through analogy does not directly tend toward the identification of 

 things, but rather toward the multiplication of ideas ; and while analogic 

 inference sharpens and thereby improves observation, it adds little 



