QUANTITATIVE ESTIMATES OF SENSORY EVENTS 317 



of measurements on it to obtain the absolute extensions of the various 

 extensive magnitudes which are involved in its relation structure. A 

 relational construct built out of all this material is what we mean by ' the 

 house.' From the moment when we first notice that some of these re- 

 lations are repeated on different occasions of observation, we get the idea 

 of a permanent physical entity firmly rooted in our minds ; and thereafter 

 every sense impression which presents relations that may, without evident 

 contradiction, be regarded as part of a self-consistent relation structure is 

 linked in our minds, by the psychological process known as association, 

 not merely with the other relations comprising the structure but also with 

 this co-ordinating idea of a permanent physical entity — an object. Then, 

 when at any time we perceive any reasonably important group of the 

 relations for which such associative bonds have been established, it at once 

 calls up the idea of the whole object, and we recognise the object as a house, 

 or a cuckoo, or tea, or whatever it may happen to be. 



Now the strength of an associative bond is reinforced by every repetition 

 of an experience of the relations involved. It is only after many repetitions 

 of similar experiences that the association becomes instantaneous and 

 automatic as is the case with those associations by which we recognise the 

 presence of a well-known object or well-known type of object. It is clear 

 therefore that of all the relations pertaining to a relation structure, those 

 will be most important for recognition which are most often observed, and 

 will be those phenomenal relations of which our sense-impressions are least 

 affected by the variable conditions in which our miscellaneous observations 

 are made. Now if we consider the visual sense, it is at once obvious that 

 those phenomenal relations involving absolute size will apparently vary 

 every time we change our distance from the object. It appears bigger in 

 the perceptual field from one distance than from another and those apparent 

 sizes vary continuously as the distance changes and do not cluster round 

 any norm. Any relation involving absolute size will be but rarely repeated 

 in our miscellaneous observations of the object, and the associative bond 

 will be too feeble for us to recognise the relation as part of the relation 

 structure of the object. On the other hand, those phenomenal relations 

 involving only the relative lengths of different parts of the object are 

 apparently unchanged by distance, and the same relations will be observed 

 every time we view the object from any given direction. Strong associative 

 bonds between these relations and the idea of the permanent objects are 

 established, and we regard the experiencing of any important group of 

 them as ' seeing the object.' 



In so far, therefore, as its distribution in the dimensions of space is con- 

 cerned, an object is recognised by its apparent shape, not by its apparent 

 size. Apart from possible differences in the relation structure of other 

 magnitudes associated with the object, and which we are not considering 

 at the moment, objects of the same apparent shape observed separately 

 are absolutely indistinguishable from one another by direct perception 

 whatever may be their sizes. A series of objects of the same shape, but 

 in ascending order of size, seen in succession present merely a selection of 

 the appearances which any one of them would present as we approach it 

 from a distance. It is this which makes sameness of shape a uniquely 

 important phenomenal relation for perception. If we are shown a large 

 number of triangles of miscellaneous sizes and shapes, each is a relation 

 structure of the same general character, involving the same number of 

 samples of the same magnitude, length, arranged with respect to each other 

 in the same manner, i.e. end to end and completely enclosing a space. In 

 addition to perceiving the relation structure constituting each triangle we 



