SCIENCE (and common SENSE ) 55 



can now teach us to be wary in applying the simple law to systems 

 highly "non-ideal": e.g., with "beams" made of a roll of paper or a 

 stick of chewing gum. We learn that if we have a reasonably light, 

 reasonably rigid, reasonably straight beam with reasonably good 

 bearing surfaces and reasonably heavy weights— then we can make 

 reasonably accurate predictions of its behavior. Most important of 

 all, we must learn to judge what is meant by "reasonably." Just as in 

 common sense, then, denotation remains in science something of an 

 art, the proper practice of which we learn only by experience. 



The ensemble of colligative relations. Compare the colligative re- 

 lations of science with those of common sense. The "laws" of science 

 are far more subtle, far more consistent with one another, far more 

 sophisticated. We may suppose them also far superior in reliability 

 and generality but— in view of the tension between these two factors 

 —the superiority cannot be absolute. Given the reliability we expect 

 of it, a "law of nature" must be so formulated that it is almost com- 

 pletely inapplicable in the tangled circumstances of everyday life. 

 Here the rough-and-ready common-sense maxim— however less re- 

 liable—is far more generally applicable. Outside his laboratory even 

 the most doctrinaire scientist bases the vast majority of his decisions 

 on common-sense maxim ("horse sense") rather than on scientific 

 law. On the other hand, given the generality we expect of a "law of. 

 nature" we must, as we have seen, accept a looseness in the denota- 

 tion of its conceptual terms resolved only by the exercise of an art. 

 That art being ever necessary— and necessarily ever uncertain— even 

 the "law of nature" must then reduce to the general form of the 

 colligative relation: // A, then probably approximately B. If a "lever"^ 

 is so loaded that ividi ^ t02d2, then probably it will be in balance; or, 

 if we wish a "lever" to balance, then the second weight, W2, must be 

 attached at a distance from the fulcrum approximately equal to 

 Widi/toz. Considering what we demand of a scientific law, consider- 

 ing too the range within which we check its predictions and the pre- 

 cision with which these checks are made, the qualifications ". . . 

 probably . . . approximately . . ." are needed here as urgently as 

 when we write a maxim of common sense. 



I conceive that the crucial diflFerence between the colligative rela- 

 tions of science and those of common sense is irreducible to dif- 

 ferences in generality or reliability— or subtlety or consistency or 



