I am delighted to see Tom Rockmore's Art and Truth After Plato in print.1 It is a compendium of enormous value and heroic brevity, written by a man who has a genuine interest in the fate of truth and knowledge and the social role of art in our time and across the entire span of civilizational time (ATP 6). If I understand him rightly, he opposes the implications of Plato's strong disjunction between art and truth (cast in terms of art's incapacity to represent reality and the divine madness of those poets who may be thought to have succeeded in revealing more than they themselves could possibly comprehend); he thinks there is no reason to deny the social role and contribution of art to the Bildung of mankind: and he ends his argument with the important but very modestly tendered verdict that the social contribution of art, read in accord with G. F. W. Hegel's notion of mankind's reflexive "concretizing" of itself through its growing knowledge of its historied world and itself, draws, must draw, on art's capacity to convey the evolving truth of that world (ATP 273). I would say Rockmore makes the case in spades; and I confess, if the compliment will not be misunderstood, that I shall now view his book under the implicitly augmented title, Art and Truth after Plato, after Rockmore! Nevertheless, I must add that I find myself inclined to begin where Rockmore ends: that is, with the state of play regarding the theory of truth and knowledge and what it means to affirm art's serving as vehicle or medium of truth and knowledge. I am aware that, even as I venture to broach a corollary inquiry, which I am certain Rockmore would wish to engage at once, I may not have made my intention entirely clear.

Let me begin, then, with an outrageously distant association of ideas, which may try your patience a little, until you see what it is meant to bring into view. Donald Davidson is rather famous—perhaps that's a poorly chosen compliment—for having one-upped W. v. O. Quine's well-known theory of "radical translation" by replacing it with his own notion of radical interpretation (in, of course, his paper by the same name).2 He chides Quine, who theorizes that we understand an alien speaker's meaning entirely in terms of the linguistic resources of our own language, as having (it seems) argued circularly or in a question begging way or by way of a vicious regress, since Quine does not indicate how to confirm that we rightly suppose that we understand what we ourselves say. (Davidson does not, I must add, say anything about the ubiquity of bilingualism. Nor does Quine.) But he pursues the scruple of his complaint by insisting that "radical interpretation should rest on evidence that does not assume knowledge of meanings or detailed knowledge of beliefs" (RI 135). The matter is not lacking in technical complications, which I shall spare you: except to say that Davidson warns us that, even if we suppose that "the totality of [Tarskian] T-sentences should… optimally fit evidence about sentences held true by native speakers" (on Quine's thesis), "we cannot assume that a T-sentence satisfies the translation criterion" (RI 139). Just so.

The real trouble goes deeper, and neither Quine nor Davidson addresses it; though Quine is more reasonable in his intuition that he (and we) cannot escape the circularity of assuming that mature speakers of a language must, in the large, understand the meaning of what they say. There is, I should add, an impressively disciplined, very large and detailed, volume produced by two devoted readers of Davidson, Ernie Lepore and Kirk Ludwig, in large part committed to determining whether Davidson's emendation is viable at all, or valid; they come to the conclusion, in the final paragraph of a 400-plus-page book that Davidson's thesis "is impossible."3 Full stop.

Of course it's impossible. Because natural language is an emergent phenomenon of the kind that is irreducible in nonlinguistic terms—that is, in terms that admit (or might have admitted) only nonlinguistic (or nonlinguistically qualified) things as obtaining in the actual world. (Davidson tends to treat intentionality as confined within the space of alternative descriptions that can be finally replaced by the physical description of mere bodily movements, which are basic to the human world and which he also calls "primitive actions.") Now, I think this raises the general problem affecting an entire run of puzzles that we associate with "first philosophy," particularly in the modernized form that draws on Immanuel Kant's Copernican revolution. If the point I am making is valid with regard to understanding meanings, then I expect you will find it holds true for truth, knowledge, judgment, understanding, evidence, confirmation, proof, reality and the like as well—the entire family of questions that has exercised Kant and the German Idealists so sorely. Once we abandon cognitive privilege of every kind, we cannot completely stalemate or defeat the skeptical rejoinder: the threat of skepticism is simply the verso of circular reasoning that, I suggest, we cannot supersede. So it is not so much that we wonder whether the arts can convey truths of any kind: it is rather that we know that we cannot fashion a determinate theory of truth or knowledge that can be confirmed or demonstrated without presupposing the underlying competences we wish to establish.

I take this to be tantamount to saying that to be a person or self—not merely a member of Homo sapiens—is to have been transformed by some suitable form of societal Bildung that issues in the mastery of one or another language; and that, accordingly, to challenge our being persons or persons reasonably apt in speaking a home language is ordinarily preposterous, though not for that reason a synthetic a priori truth. It signifies only that the questions of first philosophy trail off, informally, into guesses about different ways of speaking or altering or making or doing meaningful things that we know no way of testing with assurance that is not question begging in the extraordinary sense Davidson finds unacceptable. The upshot is, roughly, this: we cannot demonstrate that we understand the meaning of what we standardly claim to understand, in understanding language; and we cannot distinguish, disjunctively, between our understanding language and our understanding the meaning, significance, signification, import, semiotic function or the like of any of the arts, which presuppose linguistic fluency even where language is not the obvious medium of this or that particular art (architecture, music, painting, or dance). The same is true of history and action—and for the same reasons.

Plato's charge does not seem vulnerable to this countermove. But I suggest it's only because we read his charge in terms of the theory of Forms. Certainly, Rockmore offers a very reasonable (and familiar) response to Plato, to the effect that, first, we do not rightly understand what the Forms are, from a cognitive point of view, or how to access them cognitively; and yet, secondly, it seems undeniable that the very attack on representationalism (which utterly subverts art's claim to knowledge and truth) depends on the theory of Forms, both as far as knowledge is concerned and as providing an explanation of what we should regard as knowledge among all of mankind's inquiries and activities that we suppose implicate and engender knowledge (ATP 38-45).

Now I think that if we accept this line of reasoning, which I certainly do, then Rockmore will have answered Plato more than satisfactorily in the opening chapter of his book. But then, we begin to see that the genuinely interesting question raised there is how to explain the sense in which the arts may be said to convey truth and knowledge, despite the fact that they appear not to do so discursively—even in the literary arts that exploit discursivity but do not do so primarily in the ways in which standard discursive treatments of truth and knowledge afford the paradigm. Try your hand, for instance, at reading Paul Celan or Stéphane Mallarmé or Gerard Manley Hopkins; or, for that matter, in theorizing, in any way reasonably informed by the aesthetic theories or theories of art provided or suggested (well or poorly) by Kant, F. W. J. Schelling, G. F. W. Hegel, Friedrich Nietzsche, György Lukács, Hans-Georg Gadamer, Roland Barthes, or Arthur Danto.

The key questions have, to a considerable degree, been overlooked. For one thing, if our primary interest is in an adequate theory of the arts, then the principal question concerns the sense in which meanings or meaningful structures may be rightly ascribed to artworks in the sense in which they may be validly and confirmably thus ascribed. Rockmore does not address this matter at all, though he notes that Plato speaks of the representational function of art, without attempting an analysis of how the function actually works (ATP 235, 241-2). He is occupied more with legitimate replacements of Plato's cognitive target, which, if conceded, would affect all of our cognitively qualified endeavours—politics, education, science and so on, in addition to the arts. But if Plato's or Lukács' objective is our concern, then, in a perfectly obvious sense, the essential nerve of the theory of art is not our concern, except incidentally. The theory of the licit forms of knowledge is an entirely separate issue.

If one protests that art is still our concern, then, of course, he would have to address the dual question: what are the valid kinds of knowledge and how can we confirm that those proposed (by Plato, for instance) actually afford proper and viable targets? Hence, if we went that route, we would also have to be able to show that the sorts of knowledge featured (possibly, knowledge of the Forms) were accessible and were able to be effectively conveyed by one or another form of art—accordingly, were discernible and confirmable as such. But that, of course, would entail our being able to answer the sort of question featured in my first suggestion: namely, what forms of knowledge can the various arts convey and how can they do so, particularly where the arts are not usually thought to be propositional at all?

You see at once that the two sorts of questions intersect. Cognition per art need not be representational; and representationality need not be cognitive (in the requisite sense); and "cognitive representationalism" (Rockmore's term) is not addressed to the same questions as is representation (in the context of artistic and aesthetic considerations local to the theory of art). Nevertheless, it is true that the proper use of these two concepts intersects just where (as with Plato) art is thought to claim (or to manifest) a cognitive function. The two notions are distinctly independent of one another, though, plainly, there are important contexts in which they must be conjoined.

I should say that representationality, whether cognitive or artistic is at the very least a syntactic or syntactic-like distinction—whether cast in terms of resemblance, imitation, intention, substitution, cognition, reference, semiotic function, or in other related ways—whereas the issue of what to count as true (in the arts), that satisfies the conditions of cognition in the realist sense, is entirely distinct from the issue of representationality in any of the variant forms conceded to function in the arts (even if we subscribe to a version of cognitive representationalism—as with Locke, say). My essential finding here is this: (i) we cannot really address Plato's charge satisfactorily, within the bounds of the theory of art, if we have no theory of art, in the sense I am proposing; (ii) the argument respecting the realism of Plato's Forms (or of any replacement) is not (except incidentally) about art at all; (iii) the theory of art includes at the very least an account of what, and in what way, art conveys information, beliefs, truths, claims, assertions, propositions, interpretations and other ideas about the world, imagined worlds, and our selves; and (iv) the question about truth in, or conveyed by, art cannot rightly be answered without an account of precisely how truth and knowledge (or belief, ideology, conviction and the like) can actually be discerned in, conveyed by, or otherwise retrieved from, artworks.