198 Science Religion and Reality 



would be more accurately described as a forecast that these bio- 

 logical studies may ultimately be reduced to an exact science ; for 

 it is not so much a question whether the physical entities recognised 

 as such to-day suffice to account for everything observed, as 

 whether in supplementing them we can keep exclusively to entities 

 of the same category — the category to which exact science applies. 

 The accomplishment of this vision now appears very unlikely, 

 because we have recently realised that the claim of physics to be an 

 exact science is only allowable because its subject-matter is much 

 more restricted than is commonly supposed. To show the kind 

 of knowledge which physics can handle in an exact manner, let us 

 examine critically a problem in physics such as might be set in an 

 examination paper. 



The examiner, exercising his ingenuity, begins (let us say) as 

 follows : " An elephant slides down a grassy hillside . . ." The 

 experienced examinee knows that he need not pay heed to this ; 

 it is only a picturesque adornment to give an air of verisimilitude 

 to the bald essentials of the problem. He reads on: " The weight 

 of the elephant is two tons." Now we are getting to business ; 

 henceforth the elephant can be dropped ; it is " two tons " that 

 the examinee will really have to grapple with. What exactly is 

 this two tons — the real subject-matter of the physical problem ? 

 It connotes according to some code a property, which we can only 

 vaguely describe zs ponderosity^ occurring in a certain region of the 

 external world. But never mind what it connotes ; what is it ? 

 Two tons is the reading which the pointer indicated when the 

 elephant was placed on a weighing-machine ; it is just a pointer- 

 reading. Similarly with the other data of our problem. The 

 mountain flank is replaced by an angle of 60" — the reading of a 

 plumb-line against the divisions of a protector ; and its verdant 

 covering is replaced by a coefficient of friction, which though 

 perhaps not directly a pointer-reading is of kindred nature. No 

 doubt there are more roundabout ways used in practice for deter- 

 mining the weights of elephants and the slopes of hills, but they 

 are justified because they are known to give the same results 

 as would be obtained by direct pointer-readings. If then only 

 pointer-readings (or their equivalents) are put into the machine of 

 scientific calculation, how can we grind out of it anything but 

 pointer-readings ? But that is just what we do grind out of it. 

 The question was, say, to find the time of descent of the elephant, 



