290 THE EVOLUTION OF SCIENTIFIC THEORIES 



tially changes the balance of species in coastal waters, and leads to 

 a wholly new colonization of the land. But the shark— a very ancient 

 form of life— still continues his dominance of the oceanic deeps. A 

 major new scientific theory, with a much enlarged scope of correla- 

 tion, is well adapted to function throughout a vastly enlarged domain 

 of experience. We may be shocked to discover that the older theory, 

 to whose authority we formerly recognized no limits, is in certain 

 regions completely insufficient. Within its wonted domain, however, 

 it may conserve all its old effectiveness. Far from destroying classical 

 physics, says Weizsacker, 



The thesis of quantum mechanics is, rather, the persistence of classi- 

 cal laws; it asserts: when any classically defined quantity is known 

 through measurement, then all the consequences which can be drawn 

 from this knowledge according to classical physics, are exactly correct. 



In such cases the older theory is not supplanted, but only siibordi- 

 natedl Thus the overweening pretensions of a feudal lord are checked 

 by a strong sovereign newly risen. Having made his submission, the 

 feudal lord continues in full enjoyment of dominion over his tradi- 

 tional estate. Heisenberg remarks the parallel case in science: 



It is not the validity but only the applicability of the classical laws 

 which is restricted by modern physics. 



Subordination rather than falsification may perhaps obtain even 

 when there is no correspondence of new and old in the sense that 

 links modern and classical physics. Given the everyday experience in 

 the midst of which we are brought up, an older theory will often 

 seem a "more natural" way of handling the phenomena it was origi- 

 nally designed to construe— which is no doubt one reason it is the 

 older theory. Sometimes it will also offer a genuinely simpler way of 

 dealing with those phenomena, and so quite understandably con- 

 tinue in use long after its subordination. Conceptual tools will then 

 survive just as empirical tools survive: though today we measure 

 "distance" by a multitude of methods unknown to Archimedes, the 

 measuring stick remains useful to us as it was to him. 



I am very far from arguing that all older scientific theories survive 

 in one form or another: the eclipse of the phlogiston theory, for ex- 

 ample, seems total. On the other hand, I have some suspicion that. 



