METHODS OF DISCOVERING CAUSES. 421 



remains, it is probably due to a new cause, and is, there- 

 fore, important. We can often detect a cause by measur- 

 ing the amount of energy which disappears in producing 

 given amounts of the effect. 



There are several general methods of discovering 

 causes set forth in different books on logic, and they 

 are usually termed the Method of Agreement, the Method 

 of Difference, the Combined Method of Agreement and 

 Difference, the Method of Coincident Variations, and the 

 Method of Kesidues. They may be described as follows, 

 but each method as described is only applicable provided 

 all separable conditions have been previously excluded. 



The Method of Agreement is If two instances of a 

 phenomenon have only one circumstance in which they 

 perfectly agree, that circumstance is the only invariable 

 concomitant, and therefore the cause of the phenomenon. 

 But as it is rarely the case that two instances perfectly 

 agree in only one particular, a sufficient number of instances 

 are taken, so that by means of one or another the whole of 

 the conditions are excluded except the one which is the 

 invariable concomitant; so that there only remains in all 

 the cases one circumstance to which the phenomenon can 

 possibly be due. Agreement in absence of the cause and 

 effect largely confirms and strengthens the conclusion 

 derived from their agreement in presence. This method 

 enables us to discover all invariable connection, whether of 

 causation, necessary condition, or mere coincidence ; every 

 additional instance also strengthens a conclusion obtained 

 by its means. If there remain, as is usually the case, 

 several inseparable circumstances, one or more must be the 

 cause, and the remainder are only conditions of the result. 

 The Method of Difference is If two instances are 

 alike in all respects, except that a circumstance is present 

 and a phenomenon occurs in one but not in the other, that 



