The Phenomenological Meaning of Normal Illumination

The first example of divergence between phenomenological theories of per­ception regards the question of color constancy, which set the qualifying fea­tures of a phenomenological account since Hering. Katz (1930) quantifies the phenomenal magnitude of color constancy by designing a condition for the perceptual comparison of objects with varying illumination. A partition wall divides a room so that two Maxwell discs are presented in an environment illuminated by the light entering the room from the outside through a win­dow or in an environment shuttered against the light. The surface of a Maxwell disc is divided into sectors of various adjustable sizes of different colors. If the disc rotates at a speed beyond 60 rotations per second, the colors and the light­ness of sectors are no longer seen, rather they give rise to a “fusion colour.”

For example, by adjusting white and black sectors, it is possible to perceive various shades of grey (the relation between the proportion of sectors and per­ceived color and lightness was formulated by Talbot). The surfaces of the two discs have the same proportion of white and black sectors; hence, the disc ap­pears darker in the shadowed environment. If the black sectors of the darker grey disc are reduced to zero, the black sectors of the illuminated disc can be ad­justed according to the subjects’ reports until a perceived equality is obtained. Suppose that the quantity of black sectors needed amounts to 245° (a measure in angular degrees of the darkening of the disc in the shadowed environment). Katz discovered that this darkening is not dependent on the physical reduc­tion of the light intensity that is reflected by the disc in the shuttered environ­ment. If subjects observe the discs with a reduction screen through which they can only see their surfaces, the disc in the illuminated ambient looks darker than previously. In order to establish a perceptual equivalence with the disc in the shadowed ambient, the black sector has to be incremented up to 356°. This is a measure of the darkening that the disc in the shadowed ambient should have presented had its color and lightness depended solely on the decrement of the physical intensity of light. If subjects are asked to remove the reduction screen and look at the discs again, the disc in the shadowed environment now appears much lighter than the disc in the illuminated environment. The dif­ference between the perceptual equivalence with and without the reduction screen points to the phenomenal magnitude of brightness that is needed by the disc to preserve its color across the conditions of ordinary life in which it is seen in a shadowed or dark environment. It is the phenomenal magnitude at which things keep their color constant in spite of the reduction of light in­tensity due to the decrement of the illumination component. This magnitude amounts to the white sector remaining in the first equivalence task (115°). It is noteworthy that it differs from the value obtained for a disc that appears white on the grounds of its reflectance (a disc with a 360° white sector) but with the reduced light intensity in the shadow (a disc with a 4° white sector).

Katz accounts for this phenomenal magnitude by introducing the con­cept of “normal illumination.” He holds that even if stimulation results only from the light reflected by surfaces, subjects see things immersed in environ­ments that are illuminated by the sun or artificial lights with a particular hue, or that are darkened and shadowed.^[1] Like von Helmholtz, Katz claims that the “genuine” colors of things are experienced because of the connection with the illumination that turns out to be normal across various perceptual contexts. Normal illumination is “neither twilight nor direct sunlight,” rather the one with the “intensity of illumination of the open air when the sky is lightly cloud­ed,” which “is not characterized by a single value but by a certain range of values” (Katz, [1930] 1999: 83). Unlike von Helmholtz and like Hering, Katz suggests that normal illumination has a thoroughly perceptual nature, be­cause it provides the optimal conditions to distinguish most of the quali­ties of things with the least difficulty ([1930] 1999: 83; Hering, 1878: 73, for a similar argument on form). In normal illumination, things usually show their genuine colors. The disc presented in the illuminated environment is not subject to variation of color if seen with or without a reduction screen, be­cause it shows its genuine color in the light entering from outside through the window. The disc in the environment shuttered from light tends to hold the color constant, that is to say to preserve the color that it would show in normal illumination. For this reason, the peculiar amount of whiteness (115°) derives from a kind of compromise between the reduced light inten­sity at the proximal stimulation and the color that the disc would display in normal illumination due to its reflectance. Katz introduces the notion of “intermediate object” (Zwischengegenstande) to denote the color yielded by this compromise (Brunswik, 1933: 383, gives a principled justification of such a term).

The notion of “normal illumination” was soon disputed. Kaila (1923), who makes reference to Katz (1911), recognizes the epistemological meaning of a phenomenological account of the relationship between color and illumina­tion, which are correlated concepts. If such multifarious appearances as ach­romatic and chromatic colors, light spots, shadows and mirror images had the same value of object colors in normal conditions, the objects could not have the phenomenal outstanding character that they show in the ordinary expe­rience of the world. He remarks that von Helmholtz and Hering, albeit with different accounts, tried to understand whether the separation of colors and illumination was a special case of the scission or the splitting (Aufspaltung) by which two colors may be seen one behind the other along the same visual di­rection. He then investigated the conditions in which in a mirror image, which reflects colored things while it is illuminated with a chromatic artificial light, these two components appear to merge or split. Since the phenomenological explanation of color constancy takes only what is given in appearances into account, color constancy can be reduced to the conditions at which an object color is seen through the mirror color. The color of the mirror spreads con­tinuously over its surface. It appears as a volume color and makes the space in the mirror look like a colored fog. The reflected object colors are more or less shifted towards the color of the fog according to its intensity. Furthermore, the more noticeable the color of this “intermediate medium” (Zwischenmedium), the more conspicuously the colors split and the more intense the “abnormal illumination” of the mirror is. In this case, however, Kaila reports that object colors lose somewhat the quality that Katz called “insistence” (Eindringlich- keit), which depends on the relation between color brightness and apparent distance. Katz suggested that surfaces with more insistent colors appear to stand out with more relief than surfaces with less insistent colors. Therefore, in the mirror images object colors have less relief. In order to restore the insis­tence that colors would display in ordinary conditions, Kaila reports that it is sufficient to cast a shadow on the mirror and illuminate the reflected objects so that their contribution to the light stimulation is respectively diminished and strengthened. The splitting of mirror surface and object colors tends to disap­pear, while the insistence of the object color increases proportionally.

Kaila claims that the scission of what is reflected by the mirror surface in two distinct phenomenal components underlies the perception of the col­ors of things that tend to remain constant across the changes of ambient illumination. He puts Katz’s account into question, for it implies that normal illumination should be as much a phenomenal datum as the abnormal chro­matic illumination in his experiment. If normal illumination is the optimal condition for the color of things to appear as such, it should be seen as an autonomous, fairly achromatic appearance detached from the color of objects. For this reason, Katz’s theory requires that a colored thing, or something that counts as an object (Gegenstandlichkeit), is noticeably given to perception so that the phenomenal compromise between the color it has or should have in normal illumination and the actual illumination can be automatically real­ized. Kaila objects that the perceptual evidence does not satisfy this demand. If the wall of a house is first seen in the optimal condition of normal illumi­nation and then in the reddish light of sunset, nothing allows one to say that the bright diffuse illumination of daylight is perceptually detachable from the white color of the wall as the reddish one that instead appears to lie somehow over the white wall. Indeed, normal illumination lets the colors of surfaces ap­pear purely, namely not veiled by shadows, haze or fog and not disturbed by reflected light or reflections; hence, it cannot be a phenomenal datum distinct from things’ colors, like abnormal illumination. Therefore, Kaila infers that it cannot accomplish its function. Normal illumination turns out to be only a construct derived from the physical light condition.

Gelb (1929) and Kardos (1934) provide the classical support to the criticism of the concept of normal illumination by replicating Katz’s experiment. Gelb adds another partition wall casting shadow on the entire room, so that the whole room is darkened except for one of the discs that is still illuminated from the outside. Gelb proceeds like Katz. He also has subjects observe the discs by means of a reduction screen in order to adjust the black sector of the darkened disc, which was set to zero, for both discs to show a perceptually equal grey. However, when the reduction screen is removed, the two discs dis­play an inverted appearance with respect to Katz’s findings. The shadowed disc surrounded by the dark preserves an almost white appearance, while the disc in the external normal illumination becomes darker. This inversion should not have taken place were normal illumination the cause of color constancy. Moreover, if the light of a projector illuminates a black disc hanging from the ceiling at the middle of a dim room, it will appear white or light grey for a high intensity of light. The disc color is not segregated from illumination, but it depends on the local intensity of the reflected light. If a white strip is brought in front of the disc, it suddenly appears black and in an intense illumination. When the strip is removed, the disc suddenly appears white again. Kardos (1934) repeated the experiment with inverted conditions. A white disc hanging from the ceiling in a brightly illuminated room appears black and in a suitably good illumination if a screen casts a dark shadow on it by intercepting the projector light. If the screen is partially removed so that the shadow does not cover the whole disc, the disc appears white and lying partially in shadow. As soon as the screen is brought back to the original position, the disc suddenly becomes black again. Just as Gelb had shown that seeing what normal illumi­nation was like, when the projected light was revealed by the white strip, was not sufficient to segregate the real color appearance from ambient illumina­tion, so Kardos showed that normal illumination may not even be necessary.

In fact, this divergence is strictly empirical and the different results do not depart from a common phenomenological approach. Katz gives a phenome­nological justification of the concept of normal illumination. He indeed claims that if one has to choose the values range of normal illumination, it should be that which prevails “most of the time when we are perceiving objects” because it allows one to perceive the genuine colors “most frequently” ([1930] 1999: 90). Under this respect, identifying it with the essentially achromatic daylight il­lumination is sound. The changes in daylight illumination regard intensity values much more than qualitative values. Chromatic illuminations in natural conditions with high intensity are rare. Besides, Kaila’s claim implies that illu­mination has more phenomenal value as a perceptual datum the more it devi­ates from the most frequent one, hence a non-illuminated dark room should be considered “as the prototype of the experience” of real illumination ([1930] 1999: 86).

Katz remarks on the phenomenological rationale of the concept of normal illumination by pointing out the connection with the microstructure of things’ surfaces. In order to make correct judgements on things, subjects need to get as much perceptual information as possible on the material they consist of. Katz claims that the microstructure of surfaces plays this role. Even surfaces with the same hue show small color differences that become the bearer of surface structural features. Katz calls “normal” the quality and intensities range of day­light illumination because it allows optimal appearances of the microstruc­ture that provides information on the material of things with opaque surfaces. Therefore the characterization of daylight illumination as normal is not based on non-perceptual grounds. As normal illumination is the optimal condition for the microstructure to appear, so genuine colors are “distinguished espe­cially” from other color appearances because they occur in such optimal con­dition that they become the “representative” of the surfaces’ structural features ([1930] 1999: 84, 85). It is the “perceptibility of a microstructure in its details” that marks out a particular kind of illumination as normal. Katz claims that the relation between normal illumination and maximal distinction of the mi­crostructure in genuine colors is free from the ambiguity that may derive from illumination changes.

It is reasonable to think that this concept of optimality is derived from the analyses that Husserl was carrying out on spatial appearances in the period in which Katz attended his lectures. Husserl ([1907] 1997: i04f.) calls “optimal giveness” the mode of appearance that presents subjects with the “maximum points” or “regions” of perceivable properties. Since by definition a thing cannot be fully determined in any single appearance, the “optimal giveness” is achieved solely through a proper series of consistent and qualitatively co­herent appearances, whose connection is founded on the unification of their phenomenal features.

In accord with the tenet that the optimal value of appearances is specified through perceptual experience itself, Katz claims that normal illumination and genuine colors “form an indissoluble experiential unity” ([1930] 1999: 86). This is the reason for Katz’s reply to Gelb. He refers to an experiment in which a hanging rotating black disc changes from appearing white in weak light to appearing black in strong light as soon as rotation stops ([1930] 1999: 90). This observation has the same meaning as the other experiments. Katz quotes Gelb reporting that as the disc is still, the dust particles in the air and the micro­structure of the disc become visible, which may account for the sudden color change in relation to the prevailing ambient illumination. He then suggests that in the above-mentioned experiments, when the strip becomes visible, it reflects more light than the background disc and at the same time the air dust and the microstructure of two surfaces become visible. This explana­tion could account for Gelb’s findings, because the inversion would preserve the meaning of perceptual optimality for the appearance of the microstruc­ture, even though it would remain challenged by the coupling of Gelb’s and Kardos’s results. It is reasonable to conclude that this is a divergence of the observational meaning of theoretical constructs that can be settled only by empirical evidence. However, once the phenomenological core of this notion is specified, it could be expressed in terms of articulation of the visual field (see Gilchrist and Annan, 2002: 142) and its meaning as perceptual invariant could be emphasized. In such a case, the interpretations of Husserl and Koffka (1955: 224f.) would be reunited in a consistent and unitary account.

Source: Calì Carmelo (2017), Phenomenology of Perception: Theories and Experimental Evidence, Brill.