The problem of lightness and color constancy is a clear example of this fundamental divergence. Hering’s account gives a different solution from von Helmholtz’s. It also vindicates the tenet that the terms, the concepts and the primitives of the theory have to fit the meaning of appearances, in the sense that their empirical specification has to be derivable from the observation of appearances.
Von Helmholtz ( 1925, 11: 28if.) accounts for lightness and color constancy with the inferential scheme of unconscious inductive conclusions based on the knowledge of normal conditions. A surface appears to be of constant white in shadow and in sunlight. However, the intensity of the achromatic sensation is the product of the light reflected by the surface and the intensity of illumination, so that in contexts with varying illumination the sensation may be consistent with any grey. Von Helmholtz suggests that subjects must have the capacity to separate unawares the sensory components of light and illumination. Otherwise, a white surface in shadow should appear darker than a grey paper in sunlight. Because subjects are interested in external things rather than in sensations, they discount the sensory effects that are due to a deviation from known normal conditions. Since there is less illumination in shadow, the sensed intensity is presumed to be less than in sunlight, which is the normal condition of the standard use of the eyes. The darker sensation is divided by the illumination in a particular region of the visual field, and the conclusion is drawn that in reality it is correspondingly lighter, thus bringing about the appearance of a white surface in shadow rather than of a grey surface in sunlight. The same holds for the constancy of chromatic colors. Subjects cancel out the effects that are inferred to be due to a difference in illumination from the whitish sunlight, given the implicit knowledge that the latter causes the surface to have the familiar colors.
Hering contests the logic of this account, its misuse of terms regardless of their phenomenal meaning, and its lack of description of the relevant phenomena. First, the argument is inconsistent because it is circular (1905: 20). The subjects need to know the illumination intensity to separate the actual lightness or hue of a surface, but they also need to know the actual lightness or hue to discount the illumination. Secondly, von Helmholtz improperly uses the coupled terms “white-black” and “light-dark,” and does not pay due attention to their different perceptual referents. In his theory these terms denote equivalent sensations, however they are not synonymous, because the phenomenal changes in the black-white dimension are so different from those in the lightness-darkness dimension that they cannot form a series of sensations as a function of the variation of light intensity (1878: 66). In naive experience, the corresponding terms mean the properties of things or the mode of illumination of the surrounding environment. For instance, if a shadow is cast on a part of a white paper, this is not said to be “grey,” but rather “darker,” although the reflected light has the same intensity and composition of that reflected by a grey paper. If a light spot is cast on a part of a grey paper, this is not said to be “white” but “lighter,” although the reflected light is the same as that reflected by a white paper. Finally, if this difference is recognized, it is possible to describe the phenomenal interaction between visual things and their visual surroundings and to account for the different relations of colors and shadows with surfaces (1905: 2of.).
If a piece of cardboard is suspended so that it intercepts the light cast by a light bulb on a white paper, it casts a shadow that is seen as a dark spot accidentally lying on the paper. If a large black line is drawn around the deeply shaded inner region of the shadow so that it covers completely the penumbra, a grey region appears within the black outline, as if a piece of a grey paper with black borders were stuck onto the white paper, or the white paper had become grey as if colored by India ink. If the cardboard or the paper is then displaced a bit and the shadow shifts accordingly with respect to the black border, the region that looked like a grey patch reverts to appearing as a part of the shadowed white paper.
Hering concludes that light takes different phenomenal values because subjects see the changes in the perceptual surroundings at the boundary of things. Shadows have a different kind of appearance from colors that are seen to spread over things. The dark shade that appears as a shadow is perceived as lying on the surface, and subjects mean to see through it the underlying white color. Instead, the dark shade that appears as a grey color merges with the white of the surrounding surface, thus giving rise to a new color. Hering emphasizes that it is just because the shadows are segregated from the colors of things that the fine changes of shadow gradients are perceived as contingent variations, which serve also as clues to see the shape, distance and depth of things, while they continue to enable perceivers to see the actual colors of things through them (1905: 11).
The same holds for chromatic colors. If, through a hole in the shutter of a room, which is also brightly illuminated through other open windows, sunlight falls on a limited patch of a black coat, it is seen as a grey spot that looks like dust than can be wiped off. As soon as the subject looks at that patch, she no longer sees a dusty stain, but rather a light spot resting on the black coat surface, and she is scarcely able to revert to the first perception. The light has different phenomenal values on the grounds of the distinct relations it ostensibly holds with surfaces. If a light spot is generated by a mirror surface so that it lies on a grey surface, the light spot does not merge with the grey, thus making it appear lighter; rather, it is an appearance that is added to and segregated from it as a mere light that rests on it, through which the perceiver means actually to see the underlying grey color.
Thus Hering disputed the assumption underlying von Helmholtz’s explanation of constancy. There is no need for a sensation to be corrected to make the appearance correspond to the reflected light, because it is the meaning of perceiving colors and illumination that differs according to the relations that colors, surfaces and their visible surroundings hold in the perceptual scene. On the basis of these relations, the subjects may perceive whether the changes of appearances depend upon either the properties of things or the environment. For instance, qualities such as warm or cold are accidental properties of things because they are perceived to be dependent on the external cause of the bodies’ becoming hot or cold. Colors, on the other hand, are perceived as stable properties of things. The changes of color appearances do not have the continuity of change characteristic of an accidental property. Indeed, the subjects perceive a change of the mode of appearance of the same chromatic quality, hence an accidental variation due to changes in natural and artificial conditions of illumination. Hering explains constancy on the grounds of this capacity and the underlying structure of the perceptual scene (1905: 6ff., 16). For instance, he reports the following experience. A subject stands at a window and holds two flat, matte white and grey cardboards next to each other in their hands on the horizontal plane and at a little distance. If the grey cardboard is inclined towards the window and the white one away from it, the light intensity that the grey one yields on the retina is in reality higher than the intensity of the white one. Nonetheless, constant grey and white surfaces appear. Next the subject looks at the two cardboards through a tubular reduction screen. If the screen is held so that two sections of either cardboard appear to the subject like two bordering coplanar surfaces that do not cast shadows on each other, the grey cardboard looks lighter than the white according to the difference of light intensity on the retina. If the grey and the white cardboards are alternately inclined towards and away from the window, the increase in lightness or darkness of their surfaces is again seen as an accidental variation of color. In contrast to the reduction screen condition, the subjects see a contingent change in the mode of appearance, because the cardboards appear segregated from each other, each with its independent properties and relation to the window light. The change due to the increase or the decrease of light intensity is seen to depend on the concomitant conditions and is separated from the stable appearance of colors. It is noteworthy that the relation between accidental changes and stable colors is different also from the relation of overlapping between shadows and colored surfaces.
Hering suggests that the visual spatial relations play a meaningful role in the perception of dependence underlying the phenomenal distinction between accidental change and stable color properties. Consider objects located in the background of a room, which is dimly illuminated, in comparison with things located near a window through which the light enters the room. Since the illumination diminishes as the distance from the window increases, the farther back the objects are located, the more blackish they should appear. Let two identical white cardboards be placed one behind the other at a suitable distance from the window and parallel to it. If they are observed monocularly through the circular aperture of a tubular reduction screen, so that the halves of two distinct cardboards appear to be adjacent on the same visual plane, one half will appear white, the other grey. If the reduction screen is removed, and the binocularity restored, both cardboards will instead again appear colored of the same white.
Hering reports the following experiment with the Bouguer’s photometer to show that space also works as a reference system for color perception. This device was designed to equalize the brightness of two light sources by looking through a hole and modifying the conditions, like distance, that are supposed to influence it. Placing the lights at the two bottom ends of the device, which are divided by a partition, prevents interference between them. Hering put a brown paper at one end and an ultramarine paper at the other, and illuminated them with two light sources through the device apertures at opposite sides. An Edison bulb lighted the brown paper, while skylight reflected through a mirror lighted the ultramarine paper. At a suitable light intensity of the bulb, the ultramarine paper looked like the brown one to a subject looking at them through a hole in the tube at the top of the device. Hering ascribes this appearance to the mixture of blue and overwhelming yellow components of the resulting radiation. After removing the pieces of paper and exposing them in a room with all the windows closed and illuminated by an Edison bulb, the ultramarine paper again appears blue, although a bit darker than it did in full daylight, while the brown paper continues to appear brown. This experiment shows that the same radiation does not affect the same color any longer if it is perceived as an independent source of variation, as a quality of the surroundings.
Hering emphasizes that whether observers know the actual color of the two pieces of paper does not have any influence on how their appearances change in the artificial and natural conditions. He generalizes this evidence, contending that the nature of the factors determining the color constancy is phenomenological, against von Helmholtz’s claim that subjects make an inferential estimate of the seeming of the colors on the grounds of knowledge (1905: 19-20). Nonetheless, Hering does not deny the contribution of past experience, when the influence of the illumination is so strong that it alters the quality of colors rather than the mode of appearance. This happens when mountain peaks look red in the alpenglow, faces look pale in light from a sodium lamp, spots on the floor look variously colored due to sunlight passing through colored windowpanes. To account for the constancy in such cases, Hering (1905: 8) does introduce the construct of “memory colours,” that is, the colors which things are known to have shown in past usual conditions of illumination. The subjects know that snow appears white, soot black and gold yellow. These are memory colors, in the sense that they are representatives of the colors that things have typically shown in perception. Since they are evoked any time something is repeatedly seen or expected to appear, they enable subjects to see the mountain peaks as white even through the perceptual spectacles of a strongly deviant illumination. Unlike for von Helmholtz, the white does not result from an inferential correction, nor is the red glow a sensory effect that has to be canceled out. The white is seen as the stable independent property of the peaks but as it was altered in the deviant reference system of the glow. The different nature and role of this kind of perceptual knowledge is confirmed by the fact that Hering reverses the direction of causality between knowledge formation and perception. If the world were bereft of colors as independent properties of things, memory colors would not have emerged, because there would have been no stable colors across various circumstances from which to pick up memory colors as representatives (1905: 16).
Source: Calì Carmelo (2017), Phenomenology of Perception: Theories and Experimental Evidence, Brill.