If the reader is still with me, I would like now to point out that because the addition and subtraction of colors are such different processes, they involve different conditions for complementarity. When this difference is overlooked, one easily makes faulty assumptions or is confounded by apparent contradic tions between statements that actually are referring to different facts. In a noteworthy article on the technique of the impressionists,
J.
Carson Webster has scored the widespread but mistaken belief that these painters obtained the effect of bright green by placing dabs of blue and yellow side by side andletting them fuse in the eye of the viewer. Webster finds that the impressionists
did no such thing, for the good reason that the juxtaposition of blue and yellow would produce the additive effect of white or gray. Only by mixing blue and yellow pigments will one obtain green.
We distinguished between generative and fundamental primaries. The
same distinction must now be applied to complementary colors.
Generative
complementaries
are colors that in combination produce a monochromatic white or gray.Fundamental complementaries
are colors that, in the judgment of the eye, require and complete each other. To confuse these two notions is to invite unnecessary trouble. Thus a diagrammatic color circle derived from the results of the optical superposition of lights will designate yellow and blue as a complementary pair by presenting them in diametrical opposition. This will arouse protest from the painters, who will assert that in their color system the pairing of yellow and blue produces a partial and incomplete effect; for thepainter, yellow is complementary to a violet or purple, and blue to orange.
There is no contradiction here. The two parties are talking about different things.
Generative complementaries can be verified by various methods. Offhand
C O L O R 343
one would not necessarily expect that the colors that add up to white or gray in the combination of lights are the same that do so when colored surfaces are
rotated on a wheel. However, as far as one can tell from the published results, the different additive methods all yield the same results. Woodworth and Schlosberg ·give the following complementary pairs:
red and blue green orange and green blue yellow and blue
yellow green and violet green and purple.
The results seem to agree also with the complementary pairs obtained by physiological mechanisms operating in the nervous system. This is true for simultaneous contrast, by which, for example, a small piece of gray paper placed on a green ground looks purple, and for afterimages, which according to Helmholtz yield the following complementary pairs:
red and blue green yellow and blue green and pink red.
Minor differences may be obscured by the fact that color names point only approximately to the exact hues observed in experiments.
It is remarkable that the results for generative complementaries should be in such consistent agreement since in at least one obvious respect they fail to correspond to the system of fundamental complementaries on which artists have insisted for good reasons. As I mentioned before, in that system the
colors blue and yellow are by no means acceptable as complementaries, be
cause red, the third fundamental primary, is missing from the combination. Apparently we are dealing here with a principle of visual relations which does not simply reflect the basic physiological opposites, manifest in the coatrast phenomena, and which is not even disturbed by them.
For simplicity's sake I have spoken mostly of complementary pairs. But of course any number of colors, if suitably chosen, can combine to produce a monochromatic effect. The triplets operating in color vision, color printing, color television, are complementaries: any two of the three colors are comple
mentary to the third. And Newton's principal discovery amounts to saying that every hue of the spectrum is complementary to all the rest of them to
gether. Finally, it must be noted that complementarity holds not only for hue but also for brightness. A black square will produce a white one as its after image; and a light green will be contrasted by a dark red.
344 C O L O R
A Capriciou1 Medium
Remarkably little has been written about color as a means of pictorial or ganization. There are descriptions of the palette used by particular painters; there are critical judgments praising or condemning an artist's use of color. But on the whole, one can only agree that, in the words of the art historian Allen Pattillo, "a large part of what has been written about painting, it is fair to say, has been written almost as if paintings were works in black and white." In some university art departments, black-and-white slides are preferred, either because colors "distract attention" from the shapes or, more sensibly, because the reproductions cannot be trusted.
Anyone working with color slides knows that no two slides of the same object look alike and that the differences are often far from subtle. Even under optimal conditions, the projection of transparencies on the screen transforms the subdued surface colors of paintings into rhapsodies of luminous jewelry, and the change in size also influences appearance as well as composition. The color reproductions in art books and magazines vary from excellent to miser able. Most of the time the viewer cannot judge how much of a truth or a lie
he is being told.
Apart from false testimony, the originals themselves let us down. Most masterpieces of painting can be seen only through layers of darkened varnish, which has absorbed the dirt of the ages. We can have a more reliable view of fishes swimming in the muddy green water of an aquarium than we can of the Mona Lisa. No one has seen the Titians and Rembrandts for centuries, and the cleaning and restoring of paintings lead to notoriously unreliable results. Moreover, pigments arc known to change chemically. By the time one has seen aggressive blues play havoc with the compositions of a Bellini or Raphael, or has seen a Harunobu print or Cezanne watercolor bleached beyond recog nition by sunlight, one realizes that our knowledge of the pictures we possess is based to a considerable extent on hearsay and imagination.
I mentioned how thoroughly color is modified by illumination. Such mod ifications are not mere transpositions: light of a given color will affect different colors in a picture differently. Even more fundamental is the constant percep tual interaction among colors by contrast or assimilation. Place a triangle next to a rectangle, and you will find that they remain what they are, although the shapes influence each other somewhat. But a blue color placed next to a strong red veers toward the green, and two paintings hanging side by side on a wall may profoundly modify each other's colors.
A green hue that looked conservatively restrained in the sample booklet
C O L O R 345 at the paint shop will overwhelm you when it covers the walls. The color trees and cones de.signed by Munsell and Ostwald as systematic presentations of colors according to hue, brightness, and saturation serve admirably to make us understand the complex interaction of the three dimensions· but a color see
�
in the context of its neighbors will change when placed i�
a different environment.In .no relia
�
le sense can we speak of a color "as it really is"; it is always determined by its context. A white background is by no means a zero back ground b�
t has strong idiosyncrasies of its own. Wolfgang Schone has pointed out that m European paintings of the sixteenth toeighteenth centuries the light
'.
s mo�e important than the color and that therefore they are badly s:
rvcd by being displayed on white or very bright walls. Such mistreatment occurs, he says, at museums such as the Louvre, the Uffizi, the National Gallery in
�
ondon: and the Hamburg Kunsthalle under the influence of modern paintmg, which stresses color rather than light-an effect enhanced by light-colored walls.
T� all t.hese .uncertainties we must add the problems of perceptual and verbal 1dentificat1on. When observers are presented with a continuum of the ra
'.
nb?
w colors, e.g. with a light spectrum, they do not agree on where the pnnc1pal colors appear at their purest. This is true even for the fundamental primaries, especially for pure red, which may be located by observers any where between 66o and 76o millimicrons. Accordingly, any color name refers to a range of possible hues, so that verbal communication in the absence of direct perception is quite imprecise. Newton, for example used "violet" and" purp e interc angea y-not a negligible matter since according to modern l " . h bl '
�
sage, v.io�
ct is contained in the spectrum of light but purple is not. In our own time Hilaire Hiler has compiled a chart of color names which indicates, for example, that the color corresponding to the wavelength of 6oo millimicrons is described. by various authors as orange chrome, golden poppy, spectrumorange, bittersweet orange, oriental red, Saturn red, cadmium red orange, or
red orange.
It will be evident why the discussion of color problems is fraught with obstacles and why so little useful discussion therefore takes place. However,
these facts should not be taken to mean that what we see when we look at a painti
�
g is elusive, accidental, or arbitrary. On the contrary, in any successfully �rga�
1zed composition, the hue, place, and size of every color area as well asits bnghtncss and saturation are established in such a way that all the colors
toge