PERCEPTION 



1649 



wik ratio, except that logarithms of R, S and C are 

 employed : 



TR 



log R — log S 

 log C — log S 



The meaning of the ratio values of o and 1.00 are 

 the same here as for the Brunswik expression. Leibo- 

 witz (312) has indicated how these traditional ratios 

 can be determined from power functions based on 

 experimental results in which the independent 

 variable (e.g. slant, distance, etc.) has been intro- 

 duced for a wide range of values. 



interpretations. Both von Helmholtz (501) and 

 Hering (200, 202) knew that colors tend to look "the 

 same' under rather wide variations in illumination, 

 but they differed profoundly in their interpretations 

 of these phenomena. Von Helmholtz (501) believed 

 that we learn to expect how any particular color 

 should look under 'normal' while illumination. 

 Changes in the light reflected by such objects, under 

 different intensities of illumination, or under dif- 

 ferent colored lights, were then corrected by recourse 

 to experience. The interpretation involved distinction 

 of basic sensory from elaborative perceptual processes, 

 with the latter acquiring a quasi-intellectual char- 

 acter; sensations became perceptions on the basis of 

 •unconscious inference.' 



Hering (200, 202) agreed that phenomena of ap- 

 proximate color constancy represented "one of the 

 most striking and important facts of physiological 

 optics . . ." since without this "... a piece of chalk 

 on an overcast day would appear as dark as a piece 

 of coal on a sunny day and in the course of a day- 

 would take on all possible brightnesses between 

 white and black." In accounting for the effect, how- 

 ever, lie tried to avoid judgmental factors and at- 

 tempted to reduce brightness constancy (<> such more 

 tangible physiologic mechanisms as changing pupil- 

 lary aperture, adaptation and contrast. In dealing 

 with constancy under chromatic illumination, he 

 nevertheless felt compelled to invoke individual ex- 

 perience: "All the things which are known to us from 

 past experience ... in respect to color are seen through 

 the spectacle of memory colors'' (200, 202). 



constancy in animals and children. Controversies 

 about the interpretation of perceptual constancies 

 have continued to this day. The role of higher judg- 

 mental processes has been made less likely when it 

 turned out that animals below man showed similar or 

 stronger constancy effects under comparable experi- 



mental conditions. Thus, Kohler (265, 266) showed 

 brightness constancy in chimpanzees, while Locke 

 (327) proved color constancy in rhesus monkeys; the 

 effect in that species clearly surpassed corresponding 

 effects in human observers. There are equally striking 

 demonstrations of color constancy for birds [employ- 

 ing domestic chicks (252)] and cyprinid fish (78). 

 There are similar, though less complete, results for 

 tests of size and shape constancy in subhuman forms 

 [Leibowitz (unpublished observations) and Zeigler & 

 Leibowitz (556) for rhesus monkeys, Gunter (176) 

 for cats and Gotz (160) for chicks]. Studies of higher 

 invertebrates are still to be made. 



In apparent contrast to these univocal reports of 

 strong constancy effects in subhuman forms are a 

 series of studies purporting to show smaller con- 

 stancy effects in young children than in older children 

 or adults, e.g. the studies of Beyrl (42), Klimpfinger 

 (258), Brunswik (7;;, 7b), Piaget (3731 and Lamber- 

 < iii- 1 297 1 These accounts are contradicted b\ those 

 of Frank (131 1, Burzlaff (82), Koffka (284) and 

 Akishige (4). 



1111 mid FOR PARAMETRK STUDIES. There may be 

 two main reasons for ibis lack of clarity: experiments 

 in this area have rarel) been sufficiently analytic, 

 and too main investigators have assumed that dif- 

 ferent constancies should have a common base. 



a) The first objection has been raised by Graham 

 ([lib] .ind Leibowitz (313) who point out that few 

 of the available studies have been 'parametric' in 

 design; variables are nut often sampled over a suffi- 

 ciendy wide range of values. For instance, by sampling 

 size constancy functions over a wide range of dis- 

 tances, lor groups of children and adults, Zeigler & 

 Leibowitz (555) were able to show that the size- 

 matching functions of the two groups were alike at 

 close distances and differed as the distance of the 

 objects increased. It was as if size constancy in young 

 children did not -reach out' as far into space as in 

 older children and adults. Such a result reconciles 

 several seemingly discrepant reports based on experi- 

 ments restricted each to .1 single but different distance. 



/)) The second source of difficulties in this area is 

 the common view that the many different constancy 

 effects should reflect a single process. It is true that 

 the varied constancies have identical biologic use; 

 an animal should not misperceive a large predator 

 as a small prey merely because the predator appears 

 at a greater distance (504), and it should distinguish 

 sounds that are soft but close from those that are 

 loud but far (350). These accomplishments, however, 



