NATURE 



[March 7, 19 iS 



If our speed through the aether happens not to be 

 .so great as we have supposed, the contraction is 

 smaller; but it escapes notice in our practical life, not 

 because it is small, but because from its very nature 

 it is undetectable. And because this real world is 

 undetectable we do not as a rule attempt to describe 

 it. Not merely in everyday life, but in scientific 

 measurements also, we describe the world of appear- 

 ance. We do this by imagining natural objects to be 

 placed, not in the absolute space, but in a quite 

 different framework of. our own contriving — a Sipace 

 which corresponds with appearance. In the space of 

 appearance a rod does not seem to change length when 

 its direction is altered ; and we use that property to 

 block out our conventional space, counting the length 

 occupied by the standard yard-measure as always a 

 \ard however its true length may vary. It is found 

 also that in like manner our time is a special time 

 ■of our own, different from the time we should adopt 

 if our motion through the aether were nil. This is a 

 perfectly right procedure; it introduces no scientific 

 inexactness, and it ig more m accordance with the 

 •ordinary meaning attached to space and time ; the only 

 thing to remember is that this space and time frame- 

 work is something peculiar to us, defined by our 

 motion, and it has not the metaphysical property of 

 absoluteness, which we have often unconsciously attri- 

 buted to it. 



. Now let us visit for a moment the star Arcturus, 

 which is moving relatively to us with a velocity of 

 more than 200 miles per second. Consequently its 

 motion through the aether is different from ours, and 

 the contraction of objects on it will be different. It 

 follows that our conventional space would not be suit- 

 able for Arcturus, because it w-as specially chosen to 

 eliminate our own contraction effects. There is a 

 different space and a different time proper to Arcturus. 

 We must then imagine each star carrying its own 

 appropriate space and time according to its motion 

 through the £ether. The space and time of one star 

 will not fit the experience of individuals on another 

 star. 



The exact relation between the appropriate space and 

 time of one star and the space and time of another 

 was first brought out clearly by Minkowski ; it is a very 

 remarkable one. We recognise three dimensions of 

 space, which we may take as up-and-down, right-and- 

 left, backwards-and-forwards. If we go over to Ire- 

 land we still have the same space, but Ireland's up- 

 and-down no longer corresponds with ours. The direc- 

 tions are inclined ; and what is vertical to them is partly 

 vertical and partly horizontal to us. Now let us add 

 a fourth dimension, imaginary ^ time, at right angles 

 to the other three. There is no room for it in the 

 model, but we must do our best to imagine it in four 

 dimensions. In Ireland the three space-dimensions 

 will have rotated, as I have said; but the time will 

 be just the same. But if we go to Arcturus, or to any 

 body moving with a velocity different from our own, 

 the time-dimension also has rotated. What is time to 

 them is partly time and partly imaginary space to us. 

 It is a change in the space-time world of four dimen- 

 sions just analogous to the change in the space-world 

 between here and Ireland. That is Minkowski's great 

 result; space-time is the same universally, but the 

 orientation — ^the resolution into space and time separ- 

 atelv — depends on the motion of the individual experi- 

 encing it, just as the resolution of space into horizontal 

 and vertical depends on his situation. In Minkowski's 

 own famous words — " Henceforth Space and Time in 



- Imaginary in the mathematical sense, i.e. involving \/ - x. It is much 

 simpler to consider imaginary time ; and throughout the lecture I have ven- 

 tured to omit reference to the complications which arise when our results 

 are restated in terms of real time. 



themselves vanish to shadows, and only a kin.i 

 union of the two preserves an independent existen 

 From our original point of view it seems ver\ 

 markable that in the Michelson-Morley experimeni 

 contraction should have been of just the right ami n,;, 

 to annul the expected effect of our motion throu-h 

 the aether. Many other experiments, which seem . I 

 likely to show such an effect, have been tried >i 

 then, but in all of them the same kind of compens.i 

 takes place. It looks as though all the forces 

 Nature had entered on a conspiracy together with the 

 one design of preventing us from measuring or even 

 detecting our motion through the aether. It is still an 

 open question whether one force, the force of gravita- 

 tion, has joined the conspiracy. Hitherto gravitation 

 has stood aloof from ail* the other interrelated pheno- 

 mena in majestic isolation. We have become almost 

 reconciled to leaving it outside every physical theory. 

 A new model of the atom is put forward which accounts 

 for a whole host of abstruse and recently discovered 

 properties ; but it ^vould be considered unfair to suggest 

 that it ought to account for the simple and universal 

 property of gravitation. Dare we think that gravita- 

 tion has so far forgotten its dignity as to join this con- 

 spiracy ? There is certainly not enough evidence for a 

 jury to convict; but yet I think we shall have to intern 

 it on suspicion. Recently Sir Oliver Lodge, believing 

 that gravitation was innocent of the conspiracy, showed 

 that a very famous astronomical discordance in the 

 motion of Mercury might be an effect due to the sun's 

 motion through the aether, and might afford a means 

 of estimating its speed. It is difficult in a brief refer- 

 ence to deal quite fairly with an intricate question, but 

 it seems now that we should rather lay stress, not on 

 this single discordance, which can perhaps be other- 

 wise explained, but on the exact agreement of Venus 

 and the, earth with theory; for they also should show 

 evidence of the sun's motion through the aether if 

 gravitation had not joined in the consipiracy to conceal 

 all such effects. It may be that the effects on Venus 

 and the earth are not found because the sun's motion 

 through the aether happens to be very small; but on 

 the whole it appears more likely that the effect of the 

 motion is null, just as in the Michelson-Morley experi- 

 ment, because there is a complete compensation in the 

 law of gravitation itself. 



The great advantage of Alinkowski's point of view 

 is that it gets rid of all idea of a conspiracy. You 

 cannot have a conspiracy of concealment when there is 

 nothing to conceal. W'e cut Minkowski's space-time 

 world in a certain direction, so as to give us separately 

 space and time as they apipear to us. We have been 

 imagining that there exists some direction which would 

 separate it into a real and absolute space and time. 

 But why should there be? Whv should one direction 

 in this space-time world be more fVindamental than any 

 other? We do not attempt to cut the space-world in a 

 particular direction so as to give us the redi horizontal 

 and vertical. The words "horizontal" and "vertical" 

 have no meaning except in reference to a particular 

 spot on the earth. So for a particular observer the 

 space-time world falls apart into its four components, 

 up-and-down, right-and-left, backwards-and-forwards, 

 sooner-and-later ; but no observer can say that this 

 division is the one and only real one. 

 j Our idea of a real space more fundamental than our 

 own was, however, not entirely metaphysical ; we had 

 materialised it by filling it with an aether supposed to 

 be at rest in it. We now- deny the existence of any 

 unique framework of that kind. We have failed to 

 obtain experimental knowledge of such a framework 

 since we cannot detect our motion relative to it. What- 

 ever may be the nature of the aether, it is devoid of 

 those material properties which could constitute it a 



NO. 



2523, VOL. lOl] 



