ii MIND AND MECHANISM 15 



obstacles, adapts itself to an indefinite number of varying 

 circumstances in accordance with their relation to the 

 matter in hand. If a philosopher from another planet, 

 ignorant of all forms of life as they exist upon this world, 

 were to watch a stone rolling down hill and a man 

 running to catch his train, he would come to the con- 

 clusion that the stone and the man were actuated by very 

 different principles. He would, for example, see the man 

 go round the obstacle which caught up the stone, and if 

 he proceeded to compare their behaviour under many 

 circumstances and in different relations, he would arrive 

 at the result that the broad difference could be most 

 easily formulated by conceiving the stone's action as 

 determined always by the reaction of its inherent qualities 

 upon the forces directly impressed upon it without regard 

 to the ultimate issue, while the man's action would be, in 

 the majority of cases, determined by its relation from 

 moment to moment to some result more or less remote. 

 If, further, our observer was to extend his investigations 

 to the operation of man-made machines, he would 

 discover in their working a co-operation of parts, a 

 correlation of processes, in subservience to certain results. 

 But if he carefully analysed and compared different cases, 

 he would arrive at the conclusion that the working of 

 every part of the machine rested on causes fundamentally 

 identical with those which urge the rolling stone. He 

 would see that the bearing of its act upon the outcome 

 never affects the behaviour of any cog in the machinery. 

 If, however, he looked into the workshops where the 

 machines are made, and especially into the offices where 

 they are designed, he would discover that the behaviour 

 of the beings that made them is continually adjusted and 

 readjusted in accordance with the relation of the act 

 to the result. That is to say, proceeding purely by 

 inferences based on comparison of behaviour, he would 

 discover two fundamentally distinct types of correlation, 

 one in which each element of behaviour is conditioned by 

 its relation to its result, the other in which no such 

 relation is operative although the result is in fact pro- 

 duced. Now he might ultimately decide that these two 



