THE PERCEPTION OF SIGHT. 263 



sensation, it could not be disturbed by the mental operation 

 which refers the two impressions to the same object. But we 

 avoid the difficulty, if we suppose that the coincidence in localisa- 

 tion of the corresponding pictures received from the two eyes 

 depends upon the power of measuring distances at sight which 

 we gain by experience that is, on an acquired knowledge of the 

 meaning of the ' signs of localisation.' In this case it is simply 

 one kind of experience opposing another ; and we can then 

 understand how the conclusion that two images belong to the 

 same object should influence our estimation of their relative 

 position by the measuring power of the eye, and how in conse- 

 quence the distance of the two images from the fixed point in 

 the field of vision should be regarded as the same, although it 

 is not exactly so in reality. 



But if the practical coincidence of corresponding points as 

 to localisation in the two fields of vision does not depend upon 

 sensation, it follows that the original power of comparing 

 different distances in each separate field of vision cannot depend 

 upon direct sensation. For, if it were so, it would follow that 

 the coincidence of the two fields would be completely established 

 by direct sensation, as soon as the observer had got his two 

 fixed points to coincide and a single meridian of one eye to 

 coincide with the corresponding one of the other. 



The reader sees how this series of facts has driven us by 

 force to the Empirical Theory of Vision. It is right to mention 

 that lately fresh attempts have been made to explain the origin 

 of our perception of solidity and the phenomena of single and 

 double binocular vision by the assumption of some ready-made 

 anatorrr'cil mechanism. We cannot criticise these attempts 

 here : it would lead us too far into details. Although many of 

 these hypotheses are veiy ingenious (and at the same time very 

 indefinite and elastic), they have hitherto always proved insuffi- 

 cient ; because the actual world offers us far more numerous 

 relations than the authors of these attempts could provide for. 

 Hence, as soon as they have airanged one of their systems to 

 explain any particular phenomenon of vision, it is found not to 



