135 



this case it is a fact ; and I unhesitatingly say that any inference 

 which is not a fact, is utterly worthless as a foundation on which 

 to build an explanation of natural phenomena, or indeed of 

 anything else. What can the superstructure be worth if the 

 foundation is not sound ? and every inference that is not a logical 

 inference must be unsound. But logical inferences drawn from 

 facts we know are facts — facts about which we are as sure as we 

 are of those with which we become acquainted through the 

 instrumentality of the senses. Mr. Kitchin's definition, therefore, 

 when reduced to simple terms, reads thus : — A theory is an explan- 

 ation of natural phenomena founded on facts drawn from facts — 

 which I think you will agree is only a roundabout way of putting 

 the definition given by me. I should like you to bear this in 

 mind, because it is the source of nearly all the differences that 

 have arisen between Mr. Kitchin and myself in this discussion. 



The next point of difference between us is where Mr. Kitchin 

 says that it can hardly be assumed that the explanation of the rise 

 of a column of water in a tube from which the air has been 

 exhausted is an "inference drawn from facts," but that it is rather 

 a statement of facts themselves. Now, whilst I agree with Mr 

 Kitchin that it is a statement of facts, it is at the same time quite 

 clear to me that it is also an inference drawn from facts. What is an 

 inference ? A deduction from premises. It is something there- 

 fore which is outside of observation. No man ever arrived at an 

 inference through observation alone. Deduction is the business 

 of the reason. We may observe the pressure of the air at the 

 earth's surface. Or we may observe, under suitable conditions, 

 the equality of action and reaction. Observation, too, may show 

 us that water presses equally in all directions, and that it rises to a 

 certain height in an exhausted tube, but it can never tell us why it 

 so rises. The senses may say what is, but it is the intellect 

 alone that can tell us 7ej/iy it is, when it is possible to satisfy the 

 why. The rise of a column of water in an exhausted tube is not 

 by any means self-evident. It requires explanation, and the way 

 it is explained I have already stated. In that explanation we start 

 with certain isolated facts as premises. From these, by a process 



