PRAGMATISM AND THE FORM OF THOUGHT 207 



of the terms in question. But at any stage of scientific progress 

 all this remains an abstract possibility; and the degree in which 

 the statement of a relation is actually comprehensive of the other- 

 wise known content of its terms is capable of indefinite variation. 

 And with respect to thought and conduct it must be said that 

 the very indirectness and equivocality of the reference of the 

 former to the latter gives thought a character of its own, which 

 is as independent of aught beyond as can well be imagined. The 

 more meaning is read into this particular doctrine, the less truth 

 there is n it. Apart from the reference of thought to conduct, 

 that is to say, in the limitless interrelations of concepts with each 

 other, thought has as distinctive a form as any abstractly con- 

 sidered entity whatsoever. 



What, then, shall be said of logical validity? Is it true that 

 this does not attach to thought considered in abstraction from 

 the control of conduct that its only test is the practical one, 

 the cessation of thought itself when its task of readjustment is 

 done? For the reasons just given we cannot assent to this. The 

 very indirectness of the reference of concepts to modes of 

 of reaction implies that the interrelations of concepts which me- 

 diate the ultimate practical reference must have a character of 

 Tightness or wrongness in themselves. To say that without the 

 ulterior test of workability all other Tightness or wrongness would 

 be fictitious is to interpose an idle objection. For the point 

 precisely is that without a characteristic organization of the con- 

 tent of thought the practical significance of thought would itself 

 disappear. 



The fact is that according to the common pragmatist view a 

 chain of reasoning would be altogether impossible. For in such 

 a chain each link must be valid if the whole is to have any 

 strength. But the test of practice obviously cannot apply to 

 the separate links; it can only indicate in a general way the 

 profitableness of the whole procedure. If the test fails, that 

 alone does not determine where the difficulty lies. It is, indeed, 

 implied, that each valid link, if separately tested or if tested 



