II

So much for preliminaries. Now for a little scaffolding. Consider two very popular claims drawn from recent theoretically minded views of texts (or, more narrowly, artworks). In one, Rosalind Krauss, pressing into service what (reflecting on the views of Roland Barthes and Jacques Derrida) she takes to be the postmodernist intention to "blur the distinction between literature and criticism," speaks of "a kind of paraliterature," that is, a literature that is now neither criticism nor noncriticism but a sort of analogue of what criticism would have been for the modernist (preeminently, for Clement Greenberg) now that the distinction between artworks or texts and criticism has been blurred:

The paraliterary space [she says] is the space of debate, quotation, partisanship, betrayal, reconciliation; but it is not the space of unity, coherence, or resolution that we think of as constituting the work of literature. For both Barthes and Derrida have a deep enmity toward that notion of the literary work. What is left is drama without the Play, voices without the Author, criticism without the Argument. It is no wonder that this country's critical establishment—outside the university, that is—remains unaffected by this work, simply cannot use it. Because the paraliterary cannot be a model for the systematic unpacking of the meanings of a work of art that criticism's task is thought to be . . . there is not behind the literal surface, a set of meanings to which [the paraliterary] points or models to which it refers, a set of originary terms onto which it opens and from which it derives its own authenticity.^[6]

Clearly, in opposing the views of theorists like Greenberg, Krauss means to dismantle altogether (not merely to reverse) the high modernist the sis in all of its forms—for instance, as it appears in T. S. Eliot's famous remark that a work of art "is autotelic" and that "criticism by definition is about something other than itself" (a notion Eliot considerably changed in due course);^[7] or to oppose the thesis of New Criticism—for instance, as it appears in Monroe Beardsley's so-called Principles of Independence and Autonomy—"that literary works exist as individuals and can be distinguished from other things" and "that literary works are self-sufficient entities whose properties are decisive in checking interpretations and judgments."^[8]

Krauss's thesis is at least a first specimen of what I have called the constructive or "second" sense of interpretation, although to admit that is neither to support her particular thesis nor to suggest particular weaknesses in the second sort of theory as a result of weaknesses in her version of the theory. What is easy to miss, what Krauss misses, is that: (1) the rejection of the fixed disjunction between criticism or in-

terpretation and text or (2) the rejection of the fixed nature or fixedly bounded nature of texts independent of particular judgments or interpretation is not tantamount or equivalent to (3) the rejection of any functional (or logical) distinction between criticism and text—or, of course, not tantamount to simply rejecting all reference and predication. This may not be obvious, but it bears on the difference between description and interpretation and the logic of reference. It also bears, by example at least, on the distinction between postmodernism and poststructuralism.

Roughly put: the "paraliterary" need not—indeed cannot, logically—disallow, at any moment at which it is pertinently pursued, distinction (3). The very nature of assertion forbids it. Krauss risks—the evidence of her extended discussion indicates that she more than risks, she actually loses—the point of the paraliterary insertion itself. When she says that what is "left" is "drama without the Play, voices without the Author, criticism without the Argument," we may understand that she thereby opposes the pertinence or adequacy of modernist theories of the would-be referents of criticism—the objector distinctions (1) and (2); but, in dismissing them, she must hold (so must we all) to some version of (3), the logical distinction of paraliterary comments (criticism or interpretation, if you like), if, as she obviously also means, she means to speak (and does speak) of the work of Duchamp and Pollock and Stella and Serra and LeWitt and others. She must preserve reference, in short. She apparently does not grasp this and wrongly attributes the same opposition to Barthes and Derrida. I do not wish to deny that there are other uses of language besides that of making statements; but discourse concerned with truth and falsity cannot, I contend, avoid the resources of reference and predication. The matter is logically but not philosophically trivial. (I acknowledge that many philosophers—in our time, Quine, notably—have thought to retire reference in favor of supposedly adequate predicative resources. But, quite apart from the success or failure of such maneuvers, the intended work of ordinary reference is certainly admitted—for instance, by Quine himself. Krauss risks, therefore, impoverishing our discursive resources altogether, in a peculiarly profound way.)

Thus, reviewing a variety of postmodernist work, Krauss speaks of the "index," the "shifter," "traces, imprints, and clues," and similarly attenuated referential devices.^[9] She shows by her discourse that she cannot—she is hardly disposed to—abandon the devices of reference (captions and titles included); but her dialectical maneuvers against mod-

ernists are intended to leave the impression (there is reason to think she herself is convinced by the argument) that she has actually abandoned the logical referent we call "the play," in abandoning the high, complex, modernist entity "the Play"—as well as authors and criticism and the rest. Simply put: the logical distinction and pairing between interpretive discourse and interpreted referent is both entirely different in purpose from and perfectly compatible with the so-called postmodernist (or, more generally, the poststructuralist) insistence on denying an unbridgeable disjunction between criticism and text or artwork. The denial of the inherent fixity of texts and artworks is not equivalent to the denial of the heterotelic work of criticism; and modernist and (so-called) postmodernist criticism both rely on the same devices of reference and predication.

What Krauss fails to notice (or to acknowledge) is that the constraints of discourse, whether paraliterary or high critical, must retain an effective reference to what is unitary (is individuatable in a logical sense) sufficient for making such discourse pertinent and operative; also, that it makes no difference at all what we suppose is the internal unity of artworks, or what internal order remains when art departs from the high unity modernist usage presumes, so long as our theory (and practice) permits reference and predication to succeed. What post-modernists of Krauss's conviction confuse—which is not equivalent to an accusation against either Barthes or Derrida—is the difference between merely judging or interpreting artworks and trashing modernism, or the difference between formulating the difference between modernist and postmodernist art and judging or interpreting works of either sort, or the difference between favoring or opposing, for cause, particular theories about art of either sort.

The denial of a principled disjunction between artwork and criticism has to do with the possibility of changing the "nature" and properties of particular artworks as a result of interpreting them (or reconstituting them by interpretation). That, the so-called postmodernist theme (really more "poststructuralist"), is in complete accord with what I have termed the "second" view of interpretation. But interpretive work could never be pertinent or effective if we could not fix the referent to which it applies—even under a change of "nature." I admit that speaking of reidentification under such a change is unorthodox. That is, what is unorthodox is the idea that the "nature" of a text may be altered by being interpreted, not that a thing may change over time (without yet changing its nature). That may be what has misled Krauss.

Nevertheless, to abandon reference and reidentification is to court logical disaster.

Fussy though it may be, the quarrel is a strategic one. (I am, I confess, more interested in the ineliminability of reference than in doing full justice to Krauss's entire picture of the postmodern. What I wish to make clear is the neutrality of the referential matter to the quarrel between the modernist insistence on the fixity and autonomy of artworks and the "postmodernist" denial of such fixity.)

Krauss's intent is to reject the fixed demarcation between criticism or interpretation and text or artwork. Fine. The idea is that what, in the "paraliterary" manner, is said about a would-be artwork at time t may need to differ from what may be said about that "same" artwork at t' later than t, as a result of already baying defensibly interpreted (or commented on ) the work at t. That hardly precludes reference; it actually presupposes it. The first notion is indeed very close to Barthes's well-known contribution about "writerly" texts, but it is also entirely distinct from the matter of referential resources.

I agree with Krauss, then, to this extent: texts may be altered by being interpreted. That is a heterodox thesis, both because of the strong reading of intentionality (or, better, Intentionality) and because of the strong discounting of fixed "natures" in the cultural world. There are, nevertheless, two mistakes that Krauss commits. First of all, she wrongly supposes, in rejecting what a modernist critic offers in "a reading [of a particular painting] by [as she says] proper names," that she is also somehow committed to rejecting the need for proper names and other referential devices in critically discussing that painting or its details. For example, she shows, regarding Picasso's La Vie, a 1904 Blue Period portrait of Picasso's friend Casagemas, who committed suicide but whose portrait was modeled on an earlier self-portrait of Picasso himself, that a standard, somewhat psychoanalytic interpretation of the "meaning" of the work pretty well trades on what she herself wishes to avoid and rightly condemns as "the art history of the proper name." In context, she actually mentions and briefly discusses the principal philosophical theories of proper names and links them to what she terms disapprovingly "an aesthetics of extension."^[10] But she confuses the requirements of reference with the presumptions of (modernist) privilege. We do not automatically settle the nature, essence, or boundaries of artworks merely by ensuring that we identify and refer to them. Reference, we may suppose, is a grammatical distinction, al-

though many have taken it to have metaphysical import (or to convey metaphysical intent).^[11] But that is hardly necessary.

The truth is, the extensional function of proper names and referential devices readily obtains within complex intentional contexts (for instance, as in the fragmenting of BEAUJOLAIS in a Juan Gris collage— which counts against modernist simplification); and where, as with titles or captions, it serves to individuate an artwork, we need not suppose that the very nature, structure, Intentional detail or unity of the work is fixed or bounded by, or somehow determinately specified or specifiable in accord with, or unalterable with regard to, or unalterably linked to, that extensional function. The extensional function of proper names (naming La Vie, for instance) is not the same thing as fixing the extension of what the name names (whatever we may suppose that to be—the "painting," say); and the formal extension of a name (whatever that is) is not the same as, and does not determine, the Intentional complexities of what the name names (for example, what one or many nonconverging interpretations of La Vie may reasonably impute, synchronically or diachronically, to La Vie ).^[12] Of course, in saying this, we see that the force of these distinctions depends on what we concede to be interpretable—what I have labeled "Intentional." But for the moment, I am more interested in drawing the connection than in debating the scope of the Intentional. Nothing will be risked by that economy.

This, the complexity of interpretation, is precisely what is to be accommodated by distinguishing between the unicity and unity of an artwork, where what "unity" designates may be contested by modernist and postmodernist theorists of art, all the while some referential fixity regarding the bare logical "unicity" (or "career") of a work enables that contest to be actually and first joined. It is entirely possible that the purely referential function may be secured by paying attention to reliable markings that are not even part, in any pertinent sense, of the painting in question (a peculiar mark on the reverse side of a canvas, say). In a word, criticism and interpretation require referentially successful discourse; but providing for that says absolutely nothing about, and sets no significant constraints on (though it does require constraints on), the intrinsic "nature" of artworks and other cultural entities.

By "unicity," I mean to flag the fact that things having complex "careers" through time and change are individuatable, have "number," as such, and that they are reidentifiable in virtue of that, and have

"natures" (are "unified" in some measure) in accord with their admitted careers. It is an extraordinary fact that standard theories of interpretation have been focused on the "natures" of interpretable things, whereas I am recommending that they focus instead on the "careers" of interpretable things. (I take this to be central to Hume's deliberately primitive "official" objections to discourse about numerical identity under conditions of change.)

A theory of how to interpret the Picasso, eschewing a literal-minded "art history of the proper name" applied to the representational content of the painting goes no distance toward demonstrating that the use of referential devices for fixing the painting's identity, or even for fixing certain of its details, commits us to the doctrine that paintings have or must have fixed natures.

No, that is an utter non sequitur that draws us on to Krauss's second mistake, namely, her supposing that the play of paraliterary criticism in what she believes conforms with Barthes's practice is, in its own turn, incompatible with the mere referential fixity of the artwork itself or is capable of proceeding without genuine referential resources. The truth is that many have been wrongly persuaded (it is the error common to Krauss's two mistakes) that the extensional function of reference somehow fixes once and for all the substantive or Intentional complexities (the nature) of whatever (referents) are thus individuated—if, indeed, they actually are the sort of entity that possesses such (Intentional) properties. Take a moment more to identify Krauss's mistakes as clearly as possible: the one maintains that postmodernist criticism (but not modernist criticism) is actually able to forego referential practices altogether; the other maintains that postmodernist criticism actually entails the abandonment of such practices in abandoning modernist fixities. It is easy to suppose that the two are the same, but they are not.

The referential (grammatical) fixity of a text or artwork is a matter quite distinct from the substantive fixity of what may be thus fixed. The two are doubtless closely linked in the sense that nothing could be referentially fixed that did not exhibit a certain stability of nature; but how alterable (or by what means altered) the life of a person or the restored Last Supper or the oft-interpreted Hamlet or Duchamp's theoretically intriguing Fountain or the marvelously elastic Sarrasine may be is not a matter that can be decided, or that is actually determined, merely by marking such texts or artworks as the reidentifiable referents they are. Modernism does indeed appear to have been too naive

or too conservative about the conceptual link between the two notions, and postmodernism may have Liberated us in that respect; but, for its own part, postmodernism has failed (in Krauss at least) to acknowledge an ontic conservatism (a need for adequation) implied in referential success insofar as the possibility of such success constrains the very nature of texts apt for reference—which of course (sadly perhaps) the modernists were never even tempted to disown.

Unicity and unity are yoked concepts all right; but they need not, running in tandem, be taken for the same horse. By the same argument, to say that interpretation (in the sense of the first theory) presupposes description is not to say that description must, to be valid or true, be timelessly fixed or unchangeable or designate the fixed or unchangeable properties of whatever we go on to interpret. That would depend on the particular nature of what we mean to describe or interpret—for whatever we describe or interpret must have a "nature" of some sort, must submit to predication. Admitting description—or, better, describability—is, first, a purely logical concession to the minima of discourse; it is only secondarily, beyond that concession, disputatiously, a further—a hardly entailed—-concession to modernism or any other privileged metaphysics of art.^[13] If so, then the requirement of the first sense of "interpret" cannot be denied; the modernist thesis is at least not entailed by the concession. The formal fixities of discourse, of reference and predication, have nothing to do with deciding what the intrinsic nature of texts or particular texts may or must be—except for the fact (the hardly negligible fact) that whatever we say is the nature of a text must be compatible with so saying and with the interpretive discourse it is meant to support. Interpreted texts must have somewhat stable properties but they need not have altogether fixed natures,^[14] So we must go beyond Krauss.