Crítica, Revista Hispanoamericana de Filosofía, Volume 11, number 33, December 1979
Nota sobre la similitud y los colores
Adolfo García Díaz
Universidad de Zulia

Abstract: The aim of the paper is to give an account of the similarity of colors as well as of their identity and individuation, even if the author’s comments will get their full share of meaning only within the horizon of a broader theory which so far is incomplete.
In what follows we will take colors as they are given to us when we are directly acquainted with them, i.e., we will leave aside any consideration of colors in terms of physical or scientific theories and our way of dealing with their shades of color will be to take any two different ones as being two different colors. On the other hand, the problems we are confronted with will spring if these colors are given to us as simples.
Let us start by assuming that we certainly know two different colors, A and B, which are such that their difference is the least one sufficing for them to be two colors given to us and not just one. Now, it can be asked whether these two colors are similar. What we do have is two different colors, even if they only differ minimally. To give an account of the two colors being similar one might be tempted to appeal to a traditional answer: they must have something common to both and something else, besides, which is why they are not the same but only similar. But we do not have this answer at hand since, in this case, to say that the two colors are minimally different is to say they are different altogether. But as least —it could be objected— they must have something in common for them both to be colors and not something else; e.g., they must have one and the same relation as intentional objects of vision, or be ideas intromitted by sight, as Locke understood this point, or something like this.
To the above it can be pointed out that in saying that A and B are colors, we have not referred to anything common to both but, as opposed to this, we have just considered their being different; A and B are different but in a more basic way than that in which red, say, would differ from a sound. A and B are mutually different and, because they do differ, they are given to us; but red has to be already given to differ from a sound, i.e., it has already to be different from a color for it to be different from something else.
Now, if we add to our list a third color, C, such that the three of them are again minimally different among themselves, this does not seem to help us to conclude that they are similar. Since, assuming A minimally to differ from C and C minimally to differ from A, we have again a case that can be reduced to the former one in which we had just two colors whit a minimal difference between them. But now, if we consider another possibility, one in which three colors, A, B, and C, are given to us such that A differs minimally from B, B differs minimally from C, but where C differs not minimally from A, we have an order, a line which can be lengthened by adding colors D, E, such that D differs minimally from C but not from B or A and E differs minimally from A but not from B, C or D etc. We claim that such an order is an example of similarity and, more precisely, of similarity among colors.
A difficulty one is faced with in taking the above line of though is that one has a problem in finding a limit, unless an arbitrary one is fixed, in the ordering of colors similar to a given one. Experimental research, by Berlin and Kay, has shown that the above is certainly the case: once a color was chosen as the reference of a term, the totality of color samples similar to the one chosen could only be arbitrarily limited.
But Berlin and Kay also found that there was nothing arbitrary in the selection of the best exemplars of color X. These so called color focuses seem to be constituted around a paradigm and to be integrated, firstly, by exemplars not similar to, but which differ minimally from, it and, secondly, although not always, by the paradigm itself.
Now, which one of those colors functions as a paradigm, is arbitrary. If, besides, it coincides with the referent of a color term, the referent can be whichever of those colors which, among themselves, differ minimally; but once the referent is chosen it will be the same in any further occasion.
A paradigm might be represented as a point, a focus, in a three dimensional space, around which would be located its possibly best copies at a minimal distance, each one of them differing minimally among themselves as well as from the paradigm. But, as has been said, there will also be copies which will not qualify as the best in virtue of their not being minimally different from this paradigm.
But now, by what has been said, the paradigm color would fulfill its function with regard to colors different from it, but the possibility of its fulfilling the same function with respect to colors which are the same as it, cannot be excluded. If this is so, what would be the sense of talking of identical colors? Would they be the same and not several or would they just be similar in an ultimate degree?
When we say, e.g., that at a certain time a surface has or shows a unique color, what we mean is that part of that surface has the same color as the other; it is the same red, say, as the other. In such a case, the color we are referring to by means of “red” is functioning as a paradigm with respect to the identity of the color of both parts of the surface. I.e. it is functioning, in this context, as an identity criterion and not as a criterion of exemplarity.
But now, what can be done whit Putnam’s dictum “‘yellow’ as a thing predicate is indexical”? Since, certainly, in the case we have considered “red” has been taken as a predicate about surfaces, it carries it an indexical commitment. But if “red”, as a predicate, is indexical, this precludes that it names, each time it is used, a unique and indivisible property, some sort of sensible universal of which this and that surfaces would partake. No, what in each case it names is the property of this and of nothing else; the property of that and of nothing else. But with respect to which this is not distinguished from that. And now, again, we are confronted with a problem already mentioned: Are we dealing with just one or with several properties? Can several properties be identical without being just one? If several, will their identity be some sort of similarity of maximum degree?
Husserl, as an example, gave a common answer to these questions: sameness is the limit approach by degrees of similarity. Since he thinks of similarity as a spatial coincidence by means of which a part of A covers a part of B —in that in which they are the same— while at the same time another part of A is different from another part of B —in that in which they differ— he can proceed with the image of a partial eclipse of B by A to obtain a total eclipse in which A completely covers B —when they are identical— but with both terms of this equality being transparent to each other.
Even if this image is to be rejected, Husserl’s intention is to be taken into account: the diversity and individuality of colors which present themselves as the same on this and that surfaces has to be preserved, maintaining their sameness and, at the same time, their similarity and, so, their difference.
The air of paradox within this project will settle once we consider this matter within the boundaries within which we have stated it.
What we have said is that a color is given to us only if it is not to be confused with another color, i.e., where the distinction is made within the bounds of a certain category, that of color in our example. And, again, a color is individuated if it is different from other members of the same category, i.e., from another colors. And this can give us a starting point to account for the fact two surfaces are identical in color, in red, say, but different between themselves.
Let us suppose, now, that we break into two halves a one colored surface. Surface A will be different from surface B. But the color of A and the color of B, with respect to their category, are still one and the same individual, since they do not differ.
What happens is that the individuation of surfaces is not the same as that of colors. At a certain level we do have two individuals —the two red surfaces A and B; at another level we just have one individual —a color: red.
And now, what we do have is that red, as a property of the two surfaces A and B, is two colors; but as being that color and no other, is just one individual color.
We have thus answered the questions about the conflictive unity and diversity of color identity. With respect to colors, it cannot be said that they are the same F, where F would always be a general term. As has been shown, this F might, sometimes, be an individual term as when we say: “This color is the same red as that one.”
As another result our former hypotheses, we might have the possible view that a color predicate be indexical or deictical since every time it will co-mention, so to say, a color which as a property always will be the same as this, the one pertaining to this individual, the one taken as a paradigm, so that there will be no possible world in which that color is not the same as this, know and chosen, in an actual world, as the referent of my color term.
[J.A. Robles]

| PDF en Spanish (228 Kb)