THE EVOLUTION OF SCIENTIFIC THEORIES 261 



vacuum. Ascending the peak, Perier had found a progressive diminu- 

 tion in the height of a barometric cokimn of mercury. This diminu- 

 tion Pascal held to signalize a decline of pressure, with increasing 

 altitude, wholly inexplicable by the older theory. For, said he, one 

 can hardly maintain that Nature abhors a vacuum any less at a 

 mountain's top than at its base. But any who might have wished to 

 maintain the theory of horror vacui were not driven to any such 

 "absurd" contention. They could perfectly well have held the central 

 theory intact by challenging the ancillary assumptions involved in 

 the test. For example, Pascal simply assumed that the density of 

 mercury is much the same at mountain top and base. However plau- 

 sible, the assumption was not checked ( and might have failed if, for 

 example, mercury contracted rather more than it does with decreas- 

 ing temperature). Pascal assumed that the ratio of length of a 

 column of mercury and a wooden measuring stick remain unaltered 

 with changes in their position, which is amply plausible but uncon- 

 firmed. He assumed further that bodies are drawn to earth no more 

 strongly on mountain top than at mountain base, and so on. Had an 

 attempt been made to check such assumptions as these, still other 

 assumptions like in kind would have been involved in the checks, and 

 so on ad infinitum. 



When experimental finding conflicts with theoretic prediction, we 

 learn that something is amiss in the complex conceptual structure 

 used to produce that prediction. But, as Duhem long ago empha- 

 sized, the exact locus of error is just what the experiment does not 

 teach us. The logician Quine makes the same point with an excellent 

 metaphor. 



. . . total science is like a field of force whose boundary conditions 

 are experience. A conflict with experience at the periphery occasions 

 readjustments in the interior of the field. Truth values have to be re- 

 distributed over some of our statements. Reevaluation of some state- 

 ments entails reevaluation of others, because of their logical intercon- 

 nections— . . . But the total field is so underdetermined by its 

 boundary conditions, experience, that there is much latitude of choice 

 as to what statements to reevaluate in the light of any single contrary 

 experience. . . . 



. . . Any statement can be held tme come what may, if we make 

 drastic enough adjustments elsewhere in the system. 



Clearly this extreme statement fails entirely to render due account 



