THE PEECEPTION OF SIGHT. 305 



the laboratory. So long as science is not experimental 

 it does not teach us the knowledge of any new force.* 



It is plain that, by the experience which we collect in 

 the way I have been describing, we are able to learn 

 as much of the meaning of sensible ' signs ' as can 

 afterwards be verified by further experience ; that is to 

 say, all that is real and positive in our conceptions. 



It has been hitherto supposed that the sense of touch 

 confers the notion of space and movement. At first 

 of course the only direct knowledge we acquire is that 

 we can produce, by an act of volition, changes of 

 which we are cognisant by means of touch and sight. 

 Most of these voluntary changes are movements, or 

 changes in the relations of space ; but we can also pro- 

 duce changes in an object itself. Now, can we recognise 

 the movements of our hands and eyes as changes in the 

 relations of space, without knowing it beforehand ? and 

 can we distinguish them from other changes which affect 

 the properties of external objects ? I believe we can. It 

 is an essentially distinct character of the Eolations of 

 Space that they are changeable relations bettueen objects 

 which do not depend on their quality or quantity, while all 

 other material relations between objects depend upon their 

 properties. The perceptions of sight prove this directly 

 and easily. A movement of the eye which causes the 

 retiral image to shift its place upon the retina always 

 produces the same series of changes as often as it is 

 repeated, whatever objects the field of vision may con- 

 tain. The effect is that the impressions which had 

 before the local signs a,,, a^, a^, a^, receive the new local 

 signs 60, 61, 62? ^3 5 ^^d this may always occur in the same 



* An interesting paper, applying this view of the 'experimental' cha- 

 racter of progressive science to Zoology, has been published by M. Lacaze 

 Duthiers, in the first number of his Archives de Zoologie. — Tb. 



