256 THE EVOLUTION OF SCIENTIFIC THEORIES 



ing logical asymmetry. Suppose that among other predictions of 

 theory (MF) one, say cn, fails of confirmation. Then, as Popper 

 notes, we have the possibility of making the onli' logically valid in- 

 ference that is truly inductive: the syllogism modus tollendo tollens 

 draws a general conclusion from particular instance ( s ) : 



If (MF), then Cn 



Not Cn 



Therefore not ( MF ) 



This possibility of falsifying ( MF ) is of vital importance to a science 

 that progresses through a method of rejection. But falsification, so 

 clear and simple in logical principle, becomes in scientific practice a 

 matter of immense complexity. 



THE CRUCIAL EXPERIMENT 



Very large claims are sometimes made for the crucial experiment. 

 Some have even argued that, indirectly, it furnishes us with a method 

 of proof in the positive sense. Given two or more theories exhausting 

 all conceiv^able possibilities, can we not demonstrate one simply by 

 way of crucial experiment(s) falsifying the other(s)? That question 

 answers itself: the force of the proof depends wholly on the hidden 

 major premise that the theories actually conceived represent all ever 

 conceivable— and this premise is neither proved nor provable. On 

 occasion just two possibilities may well seem to exhaust all alterna- 

 tives. Light, considered as a something that travels, would appear to 

 admit of only two mutually exclusive interpretations: either light is 

 a wave, or light is a particle. Disqualifying one, do we not unequivo- 

 cally establish the other? In the 19th century Arago answered this 

 question with a categorical affirmative, and proposed an experimen- 

 tum crucis (measurement of the relative velocities of light in air and 

 in water) from which he thought one might draw conclusions as 

 follows : 



Does the image of the upper point appear at the left of the other 



image? 

 Light is a body. 



Is the contrary the case? Does the image of the upper point appear at 



the right? 

 Light is an undulation. 



