SCIENCE AND THE UNOBSERVABLE — DINGLE 223 



light that enters our eye, and not necessarily where it actually is. 

 We are, therefore, not only confined within the universe, but neces- 

 sarily without the possibility of the experience of being at the bound- 

 ary; however we move, and wherever we go, we must always see the 

 same kind of thing — stars more or less uniformly distributed all round. 



All that is of the nature of simple, traditional physics; now comes 

 the crucial point. Since the region outside the system is physically 

 inaccessible and unobservable, and the experience of being at its 

 frontier is physically unattainable so that it is never possible for us 

 to know whether we are there or not, we leave these things outside 

 our description of the universe. We give the name "space" to the 

 volume which contains what we can observe, and describe it simply 

 according to our experience as both finite and boundless. We can 

 still, of course, conceive that there is an infinite region outside (wher- 

 ever "outside" may be), but that is merely another way of saying 

 that the region is not logically unobservable. The rejection of in- 

 finite space is, in fact, precisely similar to the rejection of absolute 

 simultaneity: it represents an economy of ideas — we are to introduce 

 no conceptions which are not necessary for the description of the 

 physically observable. 



It may be worth while to point out that this account of finite bound- 

 lessness differs from the statement that space is curved in the fact 

 that it says nothing about any intrinsic property of space. There is 

 no need to try to think of emptiness with a curvature ; we have simply 

 to think of possible experience, and keep within its bounds. Of 

 course, for mathematical purposes the conception of curvature is 

 useful because it allows us to employ the technique of Riemannian 

 geometry to solve particular problems, and for the mathematician it 

 may have some esthetic value, as the poetry of Mr. T. S. Eliot is 

 said to have for those "in the know." But for the purpose of under- 

 standing it is worse than superfluous; it is definitely misleading. 



What, now, of the validity of this principle, which has taken 

 charge of physics and threatens to direct all future philosophical 

 thought? Let me repeat the principle; it says that nothing which is 

 logically or physically unobservable is significant. This statement 

 must be appraised on prescientific, general rational grounds, and we 

 can best approach the task by returning to the rival arguments set 

 out earlier and considering them in the light of our rigorous statement. 



The essence of the argument for the principle is that it is needed 

 to prevent arbitrary invention of existences or arbitrary dogmas about 

 them; and it meets this requirement by saying, in effect, that every- 

 thing that exists is observable by known means. The argument 

 against the principle is that we have no right to assume omniscience; 

 that although it may be granted that the logically unobservable can- 

 not exist, it is presumptuous to say the same of the physically unob- 



