[You are reading Objectivism and the Corruption of Rationality. This e-text version is copyright © 2002 by Scott Ryan. All rights reserved. Permission to make printed copies solely for personal use -- not for distribution or sale -- is hereby granted. Permission to make copies for any other reason may be sought by contacting the author by e-mail at SandGRyan@att.net. This information is repeated at the end of this page.]
[Back to Introduction: Why Critique Ayn Rand's Epistemology?]
[On to Chapter 2: The Optical Illusion of Objectivism]
[Back to Table of Contents]
Most people believe that an issue of this kind is empty academic talk, of no practical significance to anyone -- which blinds them to its consequences in their own lives. [Ayn Rand, "The Metaphysical Versus the Man-Made," in Philosophy: Who Needs It, p. 29.]
A "universal" is any property, quality, relation, characteristic, attribute, or what-have-you -- generally, any "feature of reality" -- which may be identically present in diverse contexts. The "problem" of universals is -- to put it in any of several ways -- whether there are, or can be, strict identities between disparate contexts; whether two objects can literally have common attributes; whether universals (i.e., repeatable predicables, or qualities that can be "predicated" of more than one object) are really and genuinely present in their apparent "instances" or whether the mind merely behaves as though they are.
(We are not concerned in this volume with the Hegelian and neo-Hegelian understanding of the "concrete universal". Basically, a "concrete universal" in this sense is a coherent system, and although we shall be concerned with coherent systems we shall not be using this term for them.)
What is odd about universals is that they seem to be able to be in more than one place at the same time. (Somehow being present at more than one time doesn't seem so counterintuitive -- even though it probably should.) If this book and that one are the exact same shade of green, the shade itself appears to be identically present in each book; if I have five dollars and you have five fingers, the number "five" seems to be present in both my set of bills and your set of digits. For that matter, if my cat is exactly the same color this evening as she was this morning, that color seems to be identically present at two different times. If not, what is it about these apparent instances that makes them "instances" in the first place?
We shall have more to say later about why the problem is important; for now I simply want to have it clearly stated. I have tried discussing universals with Objectivists before, and I have found that even getting the issue straight is something of a chore.
This is a surprising difficulty to have in dealing with a philosophy expressly devoted to solving the "problem of universals". For Objectivism does take the solution of this problem as its central task.
In her essay "For the New Intellectual," by way of introducing her own allegedly groundbreaking insights, Rand characterizes "post-Renaissance philosophy" as "a concerted attack on man's conceptual faculty" [the italics are hers]. According to Rand, even those philosophers who did not "intend to invalidate conceptual knowledge" nevertheless "did more to destroy it than its enemies" -- precisely because these philosophers "were unable to offer a solution to the `problem of universals,' that is: to define the relationship of concepts to perceptual data -- and to prove the validity of scientific induction" [in For the New Intellectual, p. 30].
This Rand believes herself to have done. The preface to the same volume indicates that Rand is "working on . . . a treatise" which will present the full system of Objectivism; this treatise "will deal predominantly with . . . epistemology, and will present a new theory of the nature, source and validation of concepts" [ibid., p. vii]. The full treatise never appears -- but the theory in question is the one she later presents in the monograph (originally a series of essays) published as Introduction to Objectivist Epistemology, and we shall begin dealing with that work in our next chapter.
First we must be clear, at least, that Rand departs from standard philosophical usage in using "universal" as a synonym for "concept" and "abstraction" (which, we shall soon see, is just what she does, although she never quite gets around to offering us a definition of "universal"). Here, partly to make this point firmly and partly to introduce a few references for interested readers, are some excerpts from relevant sources.
Objects around us share features with other objects. It is in the nature of most such features that they can characterize indefinitely many objects. Because of this the features are called universals and the main problem is to describe their status. [A.R. Lacey, A Dictionary of Philosophy, entry for "Universals and particulars," p. 368 of the 1996 edition. Antony Flew's dictionary of philosophy gives a similar definition.]
The problem of universals has a rich tradition that dates back, at least, to Plato. It is a distinctively philosophical problem[, as is] demonstrated by the fact that people other than philosophers are generally unaware that the problem even exists. Nevertheless, it is a real problem because particulars are, and can only be, described by their characteristics. Such characteristics are qualities and qualities are what are generally understood to be universals. . . . [I]t is indubitable that relations exist, e.g., that San Francisco is north of Los Angeles. Once it is understood that qualities and relations are ontologically inescapable, it remains to determine the nature of such beasts. [Andrew B. Schoedinger, The Problem of Universals, p. ix (Introduction). This volume includes an extensive collection of topical readings from throughout philosophical history.]
At this point I think we can get a deeper view of the Problem of Universals. There are those philosophers who hold that when we say truly that two tokens [apparent instances] are of the same type, then sameness here should be understood in terms of strict identity. . . . Historically, these philosophers are called Realists and are said to believe in the reality of universals.
On the other side there are philosophers who . . . hold, with John Locke, that `all things that exist are only particulars.' There are no (strict) identities reaching across different tokens; there are no universals. Philosophers who take such a view are traditionally called Nominalists. [D.M. Armstrong, Universals: An Opinionated Introduction, pp. 5-6. A highly recommended volume in general.]
The phenomenon of similarity or attribute agreement gives rise to the debate between realists and nominalists. Realists claim that where objects are similar or agree in attribute, there is some one thing that they share or have in common; nominalists deny this. Realists call these shared entities universals; they say that universals are entities that can be simultaneously exemplified by several different objects; and they claim that universals encompass the properties things possess, the relations into which they enter, and the kinds to which they belong. [Michael J. Loux, Metaphysics, p. 20, from the overview of the first of two chapters on the topic "The problem of universals". One chapter each is devoted to realism and to nominalism, which are correctly presented as exhausting the possibilities. And note that the book's title is Metaphysics, not Epistemology. Robert Audi's work of the latter title in the same series (Routledge Contemporary Introductions to Philosophy) contains not a single word on universals -- for the very good reason that, as we shall see, the "problem of universals" is an ontological problem, not an epistemological one.]
The entry under "Universals" in Blackwell's A Companion to Metaphysics is also helpful, especially as regards the various subheadings of realism and nominalism. (And as above, note that this is the companion to "metaphysics," not to "epistemology". The latter volume has no entry for "universals," for the reason I gave just a moment ago.)
And since, in our discussion, I shall be relying on philosopher Brand Blanshard at numerous points, I may as well quote him too:
[W]hat we mean by a universal is a quality or relation or complex of these that may be identical in diverse contexts. [Brand Blanshard, Reason and Analysis, p. 392. In The Nature of Thought, v. I, p. 649, Blanshard argues that the existence of such intercontextual identities entails that space and time cannot be real just as they appear to us. No wonder Rand is worried.]
Anticipating our argument a bit: before one announces to the world that with half an hour of introspection one has solved a philosophical problem of some two thousand years' standing, it is advisable to make sure one has correctly understood the problem one is supposedly trying to solve. Since Rand does not do so, her sole contribution to the problem of universals is to confuse her readers, perhaps irreparably, about what it is.
In fact she does not solve, or even raise, the genuine problem of universals at all, as I have demonstrated to my own satisfaction not only by reading her works thoroughly but by asking Objectivists, "According to Rand, is it possible for two objects literally to have properties in common?" Usually, when I do so, there is a minor flurry of responses that say, in effect, "That's a trivial question to which the answer is obviously ______" -- and then half say "yes" and half say "no". My conclusion, then and now, is that Rand does not address this question.
Whatever Objectivists may say to the contrary, the question I quoted in the preceding paragraph is the problem of universals, and Rand not only fails to "solve" it, she does not appear even to know what it is. Here is a brief excerpt from Michael Huemer's "Why I Am Not An Objectivist" [online at http://home.sprynet.com/~owl1/rand.htm; page will be displayed in this window/frame]:
I have here two white pieces of paper. They are not the same piece of paper, but they have something in common: they are both white. What there are two of are called "particulars”" -- the pieces of paper are particulars. What is or can be common to multiple particulars are called "universals" -- whiteness is a universal. A universal is capable of being present in multiple instances, as whiteness is present in many different pieces of paper. A particular doesn't have "instances" and can only be present in one place at a time (distinct parts of it can be in different locations though), and particulars are not "present in" things. . . .
Also understand that I don't by a "universal" mean a certain kind of word, idea or concept. I mean the sort of thing that you attribute to the objects of your knowledge: Whiteness itself is the universal, not the word "white" and not the concept "white". . . . Whiteness is not a concept; it is a color. . . . I say this because the confusion between concepts and their referents is all too common, both inside and outside Objectivist circles.
Exactly. And we shall soon see how Rand deals -- or fails to deal -- with this topic.
First, though, by way of introducing several themes that shall concern us throughout this volume, we shall look briefly at a short passage that illustrates many of the points we shall be raising later.
What follows, interspersed with my own comments, is from an edited transcript of a "workshop" Rand conducted on some unspecified date with unnamed people, as edited by Harry Binswanger [Introduction to Objectivist Epistemology, pp. 137-141].
(According to one source [online at http://www.bomis.com/rings/obj/17; page will be displayed in this window/frame], "Prof. A" was Harry Binswanger himself, "Prof. B" was Allan Gotthelf, and "Prof. C" was someone named Nicholas Bykovitz -- as recollected by Lawrence Gould, who was himself "Prof. M". Incidentally, it does not appear that all of them were "professors" at the time, if indeed any of them were, so it is unclear what "Prof." is supposed to stand for. Leonard Peikoff also seems to have been one of the participants: Gould identifies him as "Prof. E".)
I hesitate to rely too heavily on this passage, both because of its brevity and because of the practical impossibility of determining how much it has been edited. I shall therefore not base any arguments directly upon it at this point; I present it here only because it does represent, in microcosmic form, many of the errors and other difficulties we shall be discussing further on. But we shall be returning to certain of these "workshops" when we are further along in our arguments, and readers skeptical that I have understood this or that point of Rand's epistemology are advised to re-read this passage later as well.
Prof. A: Now, as I understand it, . . . measurement-omission is accomplished by means of differentiation. Take the concept of "blue." You begin as a child with two blue objects of different shades perhaps (so their specific color measurements differ), and, say, one red object. And then you are able to see that the two blues belong together as opposed to the red; whereas if you just consider the two blues by themselves, you would only be aware of the differences between them; you wouldn't see them as similar until you contrasted them to the red.
AR: That's right.
Note that the possibility that the two blues are identical is mentioned briefly and thereafter ignored. Note also Rand's claim that in order to form a concept, we need a third object as a sort of "foil" in comparison to which we can see two objects as similar. This point will come back to plague her account of "axiomatic concepts," which by the present standard are not concepts at all since -- as she explicitly tell us -- they are not formed by contrasting anything with anything else. [See IOE, p. 58: the two paragraphs beginning, "Since axiomatic concepts are not formed by differentiating one group of existents from others . . ."]
Prof. A: Now . . . [i]s it that by means of this differentiation you see blueness as a range or category of measurements within the Conceptual Common Denominator: color? . . . You see the blue of this object and the somewhat different blue of that other object; both have specific measurements, but those measurements fall into one category, as opposed to the measurements of some red object, which fall outside that category. So that the omission of measurements is seeing the measurements as falling within a given range or category of measurements . . . within the Conceptual Common Denominator.
AR: Yes, that's right. Now, the essential thing there is that you cannot form a concept by integration alone or by differentiation alone. You need both, always. You need to observe similarities in a certain group of objects and differences from some other group of objects within the common standard or kind of measurement. . . .
Here again, this point will come back to haunt her. She claims, with great definiteness, that we always require "foil" objects against which to perceive similarities among a group of nonidentical existents (even though the "always" admits of an exception in the case of "axiomatic concepts"). And note once again that she never raises the question whether the two objects might be an identical shade of blue.
Note also: we have been told that we cannot perceive "similarities" unless there is a third, dissimilar object present. But we are also told that we really do perceive the similarities in question (and we shall be told later that they are really there quite independently of our perceiving them). So the business about the third object is something of a red herring, having to do only with the psychological conditions under which we can recognize a similarity and not with the real existence of similarity relations themselves. Even if Rand is right about our requiring a third "foil" object, this fact is no more interesting epistemologically than the fact that we can't see colors in the dark.
Nor can I fathom why we "always" need a third object. If we have discriminated two objects, we have already "differentiated" them from their background, have we not? (Does the background count as an "object"?) Even if we require a "foil" of some sort, why may it not be simply another attribute (or complex of attributes)?
And speaking of attributes, note also that we are apparently granted the real existence of "ranges," "categories," "standards," and "kinds" -- at least as far as measurements of perceptually-given attributes are concerned. These brief remarks appear to rely on a perfectly ordinary "realist" understanding, not only of attributes, but of kinds of attributes. We shall see later that Rand has swept the problem of "natural kinds" entirely under the carpet and goes about blithely assuming the real existence of such kinds.
Prof. B: In forming the concept "blue," a child would perceive that two blue things, with respect to color, are similar and are different from some red thing. And he places the blues in a range of measurements within the broader category, red being somewhere else on the scale.
AR: Right.
Prof. B: Now, in fact, he doesn't have a category of measurements explicitly, so what actually goes on, as you indicate, is that he perceives similarities and differences directly.
AR: That's right. . . .
So according to Rand we directly perceive similarity relationships. And, moreover, we do it without knowing anything about the underlying measurements themselves:
Prof. B: To describe the process of concept-formation on a conscious level, one wouldn't have to refer to omitting measurements [because one does not ordinary possess, or need to possess, knowledge of such measurements]. Is the purpose then of discussing it in terms of omitting measurements to stress the metaphysical basis of the process?
AR: No, not only to stress the metaphysical basis, but to explain both the metaphysical and the epistemological aspects. Because, in modern philosophy, they dismiss similarity practically as if it were ineffable; the whole nominalist school rests on that in various ways. The nominalists claim that we form concepts on the ground of vague similarities, and then they go into infinite wasted discussions about what we mean by similarity, and they arrive at the conclusion that nobody can define similarity. So that one of the important issues here, and the reason for going into the process in detail, is to indicate the metaphysical base of similarity and the fact that it is grasped perceptually, that it is not a vague, arbitrary abstraction, that similarity is perceptually given, but the understanding of what similarity means has to be arrived at philosophically or scientifically. And similarity, when analyzed, amounts to: measurements omitted. . . .
We aren't told which "nominalists" make such claims and then "go into" such "infinite wasted discussions," of course. But note the advance confirmation of a point we shall make later: what Rand is doing is trying to provide a non-vague, non-arbitrary formulation of "similarity" or "resemblance". She has not granted that any two objects may literally have properties in common; her examples assume the contrary (and so, as we shall soon see, never get around to raising the problem of universals). She is dealing with one issue only: how we form concepts based on, and of, ranges of properties which are assumed not to be identical.
And her view of measurements is firmly realistic; the attributes in question, their measurements, and the relations between them that allow them to be ordered along a spectrum are acknowledged to be "really out there". The position for which she is actually arguing is simply that all cases of "perceived" similarity are ultimately based on real relations of commensurability whether we are aware of it or not.
Prof. C: I understand how one grasps similarity on the perceptual level. Aristotle, presumably, was unable to identify how we grasp similarity beyond that point . . .
AR: He didn't say you grasp similarities intuitively. He said you grasp the essence of things intuitively. . . . He assumed that there are such things as essences -- and that's the Platonism in him. But he didn't agree with Plato's theory that essences are in a separate world. He held that essences do exist, but only in concretes. And the process of concept-formation, in his view, is the process of grasping that essence, and therefore grouping concretes in certain categories because they have that essence in common. . . . He isn't concerned with perceived similarities and differences. And since he can't explain how it is that we grasp these essences, which are not perceived by our senses, he would have to treat that grasp as a direct intuition, a form of direct awareness like percepts, but of a different order and therefore apprehending different objects.
For Aristotle, she says, "essences" are "not perceived by our senses" and so must be grasped by "direct intuition". Rand seems to be bothered by the possibility that reason may provide us with direct intuition of the structure of reality; this point too shall concern us later. But for the time being, note that Rand is confusing two issues here: the possible existence of essences, and the process by which we allegedly apprehend such essences. Again she conflates the question what constitutes knowledge with the question under what conditions knowledge becomes psychologically possible. Even if she is successful in her arguments, she will have shown, not that there are no "real essences," but that we apprehend them through sensory perception rather than by rational intuition.
Here we see a hint of a deep problem in Rand's epistemology: note that little phrase, "perceived by our senses". We shall repeatedly have occasion to wonder whether the "senses" can do all the work Rand eventually heaps upon them. (We shall have a similar wonder about perception. Note Rand's earlier remark that similarity is "grasped perceptually". "Grasped" is an interesting term to use in the context of perception; it seems to imply that perception itself involves some sort of rational apprehension. Is Rand building reason into the "perceptual level"?)
And here also we see the beginning of a confusion that will concern us in our next two sections: the difficulty of deciding whether, in the final analysis, Rand is a nominalist or a realist as regards the existence of universals. Disagreement is surely possible here, and I have changed my own mind on this point as I have read and studied Rand's epistemological writings.
At one time I took her, as regards universals, to be a realist who was trying to show that the nominalistic understanding of "resemblance" actually rests on a foundation of realism. But my best opinion at this point is that Rand sets out to be a nominalist and falls into realism only by accident, neither knowing nor caring that her analysis of similarity presumes the existence of real universals. (Of course I mean these terms to have their standard philosophical meanings, not the tendentious misdefinitions Rand gave them.)
To show this, we must look very critically at her account of "concept-formation". To that task we now turn.
[Back to Introduction: Why Critique Ayn Rand's Epistemology?]
[On to Chapter 2: The Optical Illusion of Objectivism]
[Back to Table of Contents]
[You are reading Objectivism and the Corruption of Rationality. This e-text version is copyright © 2002 by Scott Ryan. All rights reserved. Permission to make printed copies solely for personal use -- not for distribution or sale -- is hereby granted. Permission to make copies for any other reason may be sought by contacting the author by e-mail at SandGRyan@att.net