THE LAWS OF SCIENCE 47 



study in which propositions are deduced one from the 

 other in a direct line of descent from simple ultimate 

 assumptions to complex final conclusions, this double 

 role of laws, partly assumptions and partly conclusions, 

 would cause grave difficulty. All scientific arguments 

 would appear " circular," that is to say, they would 

 assume what they pretend to prove. But the result 

 that follows from our discussion is not that science is 

 fallacious*, because it does not adhere to the strict rules 

 of classical logic, but that those rules are not the only 

 means of arriving at important truths. And it is essen- 

 tial to notice this result ; for, since logic was the first 

 branch of pure learning to be reduced, to order and to be 

 brought to something like its present position, there has 

 been a tendency in discussions of other branches and 

 especially in discussions of science to assume that, if 

 they have any value and if they do really arrive at 

 truth, it can only be because they conform to logical 

 order and can be expressed by logical formulas. The 

 assumption is quite unjustifiable. ^Science is true, what- 

 p.ygr^pyrmfi may say : it has, for certain mimis t j|japt 

 for all, the intellectual value which is the ultimate test 

 of truth. If a study can have this value and yet violate 

 the rules of logic, the conclusion to be drawn is that those 

 rules, and not science, are deficient. Nevertheless, while 

 it is important to insist that science is not necessarily 

 bound by logical formulas, it may be well to point out 

 that the difficulty which we have been noticing can be 

 overcome to some extent. The difficulty arises because 

 we have regarded all the different laws of science as differ- 

 ent propositions, some of which give rise to others. J^ 

 would probably be more accurate to regard all the 

 so-called kwsjo^science as one single law which Isalways 

 being extended and refined ; and if we take "that view 

 there can be no question of deducing one law from 

 another ; the difficulty does not arise. Much might be 



