PERCEPTION 



1657 



tactile-kinesthetic) situation, but this interpretation is 

 called into question by the fact that congenitally 

 blind children obtain strong haptic illusion effects, 

 e.g. on presentation of the Miiller-Lyer pattern in 

 relief. These observations create various difficulties 

 for the ingenious approach to the visual illusions 

 proposed by Tausch (462) and von Hoist (504). 



figural aftereffects. The findings on the con- 

 genitally blind favor the argument (276) that some 

 illusions at least, and their decrement on prolonged 

 inspection, can be treated as instances of figural 

 aftereffects (271, 281 ). We have already dealt with 

 the related earlier observations on slow diminution of 

 apparent curvature for straight lines that had been 

 made to appear curved by a prism (151, 552). Gibson 

 (1 51-153, 158) has studied these changes in detail, 

 showing that they occurred, without the use of 

 prisms, on mere inspection of curved or tilted lines, 

 and that there were measurable aftereffects, in which 

 objectively straight lines appeared curved or tilted in 

 an opposite direction. He also showed that these 

 effects had a tactile-kinesthetic counterpart; on rub- 



Far 



Near 



A 



r 



L 



fig. 36. The Ponzo illusion (above) ;is an inappropriate 

 constancy effect. The distant log lying across the path in the 

 scene below is actually of the same length as the one parallel to 

 it which lies in front. If perspective is taken into account, the 

 more distant log seems much larger. However, the illusion is 

 effective in ohser\ ers who arc shown the drawing in 180 

 rotation and are thus unaware of its meaning; the elongation 

 of the line crossing the denser gradient of converging lines 

 seems therefore essentially like the effects illustrated in tig. 35. 

 [Modified from Ponzo (385 1 and von Hoist 1 504).] 



tions) ; once established, the decrement is transferred 

 (at least in part) to the visual form. Strangely, how- 

 ever, there is less transfer in the reverse direction, i.e. 

 after massed visual trials to subsequent tactile presen- 

 tations. This curious asymmetry of intermodal trans- 

 fer is still unexplained. 



Haptic analogues exist for most of the optic-geo- 

 metric illusions (398). They are sometimes dismissed 

 as results of transfer from the visual to the haptic (i.e. 



Q 



r 



L. 



L 



fig. 37. Illusions as misapplied constancies. .-1: The appear- 

 ance of a square surface leg. a table top) seen from in front 

 and slightly from above is distorted by the laws of geometrical 

 optics as shown in B. The perceptual process is thought to 

 operate upon this pattern by counterdistortion, resulting in the 

 changes schematized in C. Such a process would invoke appar- 

 ent lengths, sizes and angles. [Modified from Tausch (462 



