PHILOSOPHY OF PHYSICS MILNE 227 



event occurs to me I can prick off on the clock disk the position of 

 the hand, and so attach some definite time nmnber to the event. 

 There is no question of my " resolving my experiences into time and 

 space " — there is no arguing about my own temporal experience. 

 We remark further that the only events the observer actually experi- 

 ences are events at himself. He may be informed of other events, 

 or infer their existence from his own experience, but he has direct 

 experience only of events at himself. 



The observer, myself, is quite unable to superimpose two different 

 intervals on the clock. Once gone, an interval on the clock is gone for- 

 ever. Thus from my own experience alone I cannot select one par- 

 ticular system of graduation of my clock subdivisions and call them 

 " uniform time." So far uniform time is not definable. 



I have certain experience of the relative proximity of objects 

 near me, namely, tactual experience. I can extend my arm, and 

 ascertain that I have to extend it more for one object than for an- 

 other. I can also walk to an object, and count my paces. But it 

 would be difficult to build up consistent measures of " distances " of 

 objects from these observations. If, on the other hand, I use a 

 meter scale and read off positions of objects against its gradtiation, 

 either I assume its graduations are equal or that the scale itself is 

 unaltered by displacement — in each case invoking the indefinable 

 concept of the rigid body. As we are endeavoring to proceed without 

 using indefinable concepts, we must exclude the use of the meter 

 scale. 



The assigning of a distance to a distant object can be accom- 

 plished in principle as follows: I can send a signal (e.g., a light 

 signal) from myself to the object, despatching it at time t^ by my 

 arbitrary clock. I can then receive back the echo, and note the time 

 of its arrival, say to. If necessary I can arrange another observer at 

 the object who returns my signal immediately he receives it. The 

 position now is that I have two data of observation, /'i and tn- What 

 can I do with them ? 



Out of two numerical magnitudes two other independent numbers 

 may be constructed in an infinite number of ways. I want to con- 

 struct two numbers, one of which I can call the epoch of the event 

 constituted by the arrival of my signal at the object, the other of 

 which I can call the distance of the same event, i.e., the distance 

 of the object at the instant of occurrence of the event. Though the 

 epoch of an event occurring not at myself is completely undefined, 

 so far, it is clear that it will be convenient to assign to this event 

 an epoch or time number which is greater than /i, the epoch of 

 despatch, and less than ^2. the epoch of return. For in any sense we 

 can reasonably attach to " earlier " and " later ", the event is later 



72774—35 16 



