FALLACIES OF GENERALIZATION. 429 



word into buoyancy, to turn the semblance of argu- 

 ment involved in Bacon's illustration directly against 

 himself. 



A metaphor, then, is not to be considered as an 

 argument, but as an assertion that an argument exists ; 

 that a parity subsists between the case from which the 

 metaphor is drawn and that to which it is applied. 

 This parity may exist though the two cases be ap- 

 parently very remote from one another : the only 

 resemblance existing between them may be a resem- 

 blance of relations, an analogy in Ferguson's and 

 Archbishop Whately's sense. As in the instance 

 quoted from Mr. Carlyle : there is no resemblance 

 between convulsions of the body and fits of passion 

 in the mind, considered in themselves ; the resem- 

 blance is between the relation which convulsions of the 

 body bear to its ordinary motions, and that which fits 

 of passion in the mind bear to its steadier feelings. 

 Thus, where the real difference between the two cases 

 is the widest ; where the metaphor seems the most 

 far-fetched, the analogy the most remote; and where, 

 consequently, a limited and literal understanding 

 would be most apt to shut itself up within its in- 

 trenchment of prose, and refuse admittance to the 

 metaphor, under an idea that cases so very unlike can 

 throw no light upon each other ; it is often in those 

 very cases that the argument which the metaphor 

 involves and suggests is the most conclusive. 



8. To terminate the subject of Fallacies of 

 Generalization, it remains to be said, that the most 

 fertile source of them is bad classification; bringing 

 together in one group, and under one name, things 

 which have no common properties, or none but such 

 as are too unimportant to allow general propositions 



