ENGINEERING AND PURE SCIENCE — SWANN 213 



sense in natural philosophy repatterns itself from age to age. At 

 each stage of its development it seeks to generalize the ideas born of 

 the experience of the immediate past and to weld them into bonds 

 which sometimes restrain the future. Thus, the breeders of error in 

 the epoch to come are sometimes the truths of the days which have 

 gone. So beware how you extol too much the virtues of good old 

 horse sense, for I fear that, in the last analysis, horse sense is the 

 kind of sense that a horse has. 



I think with some regret of many experiments which in my youth 

 occurred to me as worthy of being tried, but which on further medita- 

 tion I failed to perform because in terms of common sense, as exempli- 

 fied by the theories of the day, they seemed destined to give trivial 

 results. Sometimes I wish that I had not thought so much and had 

 done the experiments in any case, for in later years many of these ex- 

 periments have been done and have given results of value, results which 

 are harmonious in the frame of thought of today, however strange 

 they would have appeared in the light of our concepts of 40 years ago. 

 If a proposed experiment lies far within the boundaries of established 

 facts and theories and would, according to their diagnosis, give a 

 perfectly clear-cut result, I do not deem it worth while to do the 

 experiment. If someone comes with a scheme for perpetual motion 

 involving the utilization of ordinary machinery and the like, I give 

 him but little attention. If he claims that, since the experiment has 

 never been done, one cannot predict with surety what result it would 

 give, I have to agree with him, as I would also agree with him if he 

 claimed that the sun might not rise tomorrow morning. I then 

 answer to the effect that while no one can predict the result of an 

 untried experiment, the chance of his perpetual motion experiment 

 giving a valuable result is, in my judgment, so small that we should 

 not be justified in discontinuing the other experiments which we are 

 doing in its favor. If, however, I am confronted with an experiment 

 which lies near the horizon between the known and the unknown, 

 I prefer to give it the benefit of the doubt ; for it is such boundary 

 dwellers as these that establish connection with the knowledge of to- 

 day and that of tomorrow. 



EMPIRICISM 



And now perhaps it will be well if I get back to what is supposed 

 to be the main theme of this discourse, the relation of engineering 

 to physics. 



In speaking of the empiricism I have perhaps implied that today 

 we have something better than empiricism. In the last analysis, this 

 is not true. All our laws of nature are empirical. However, the 

 desire of the theoretical physicist is to limit empiricism in the sense 

 that from the fewest number of empirical starting points he can 



