44 SCIENCE (and common sense) 



^^'here new concepts are used, new constructs are made. Often 

 these scientific constructs may appear superior approximations to 

 "pure observation reports." Sufficiently taxed (e.g., by optical illu- 

 sions), even common sense will seek to reform its constructs— to 

 separate what is "given" from what, plausibly but perhaps ground- 

 lessly, has been inferred. But history has taught the scientist that a 

 great deal of effort may be invested profitably in this separation— 

 in trying to free constructs from their customary freight of uncon- 

 scious inference. This procedure the scientist calls "paying attention 

 to the observables," and as a procedure it has been of crucial impor- 

 tance to such 20th-century developments as quantum mechanics and 

 relativistic mechanics. 



The scientist's resolve to "pay attention to observables" is too often 

 mistaken for a profound new insight of modern science. That resolve 

 is already detectable in the remote beginnings of science with such 

 as Aristarchus and Copernicus. Galileo writes: 



... I cannot find any bounds for my admiration how reason was 

 able in Aristarchus and Copernicus to commit such a rape upon their 

 senses, as in despite thereof, to make herself mistress of their 

 credulity. 



Of course there was no rape upon the senses: what reason did was 

 simply to undo or redo what reason had formerly done. By reason, 

 using new concepts, Aristarchus and Copernicus managed to recast 

 or reconstruct what had long passed as the naked report of the senses 

 though it contained a large component of "credulity": plausible in- 

 ference with which it had been laden by common-sense reason. 



"The naked report of the senses," shown in Chapter I to fall be- 

 yond the reach of common sense, is for science, too, an utterly un- 

 attainable ideal. Also, I think, in some degree a sterile ideal. If I re- 

 port pointer readings and nothing more I report nothing of interest 

 to science. Interest develops precisely as the pointer readings are 

 associated with inferences. Consider Poincare's example: 



I obsei-\'e the deviation of a galvanometer by the aid of a movable 

 minor which projects a luminous image or spot on a divided scale. 

 The crude fact is this: I see the spot displace itself on the scale, and 

 the scientific fact is this: a current passes in the circuit. 



... if I ask an ignorant visitor: Is the current passing? he looks at 

 the wire to try to see something pass; but if I put the same question 



