Kant has once and for all fathomed the depth of aesthetic disagreements among humans: they are tantamount to a denial of the other party’s humanity […] The amazing thing is that he grasped that an issue of such magnitude—are we capable of living in peace?—was at stake in a sentence as anodyne as “this rose is beautiful.”

– Thierry de Duve1

Thierry de Duve is one of very few art theorists who was active in the period following the launch of the influential journal October in 1976 who rejected the then-standard artworld choice between an anti-aesthetic postmodernism or late modernist aestheticism by seeking to do justice to both Clement Greenberg and Marcel Duchamp. As anyone familiar with how such debates typically break down will be aware, this is a highly original undertaking and one of the main appeals of his oeuvre for anyone unpersuaded by standard artworld doxa on these divisions. Despite this, de Duve’s work has yet to receive the sustained critical attention it deserves. This is a highly distinctive body of work, but philosophers are either unaware or, with one or two exceptions, uninterested in it; and art historians typically lack the training or expertise to interrogate de Duve’s more philosophical claims.2 Alongside Greenberg and Duchamp, Immanuel Kant is the third figure in the triumvirate that animates de Duve’s thought and the one I focus on here.

The cover of Aesthetics at Large, de Duve’s most recent collection, features an only somewhat tongue-in-cheek button declaring that “Kant got it right.”3 But what exactly did Kant “get right?” Answering this question requires examining the proposals for updating Kant in light of Duchamp’s readymades that de Duve has now been fine-tuning for over a quarter century. I focus on two questions in what follows. My primary question is whether de Duve’s proposals amount to a plausible interpretation of Kant. Or, if not strictly a plausible interpretation of Kant’s own claims, then at least a plausible extrapolation of those claims in light of subsequent art history. This is to ask whether the “Kant after Duchamp” approach to aesthetic judgement is still recognizably Kantian—at least in spirit. My second question is more foundational though, for the purposes of this paper, subsidiary: it is whether de Duve’s theory is philosophically coherent. For even if the answer to the former question turned out to be no—that is, even if de Duve’s proposals are neither a plausible interpretation nor extension of Kant—they might nonetheless be philosophically coherent, independently of de Duve’s claim to a Kantian provenance for his theory.

I. De Duve’s Kant: Aesthetic judgement after Duchamp

The basic claim of Kant after Duchamp (1996) is that bringing Kant up to date, given the transformations in art since Kant’s day, involves substituting the “modern” aesthetic judgment “this is art” for its “classical” Kantian counterpart “this is beautiful.”4 Or, as de Duve now puts it: “Beauty in nature was for Kant the place where hope in the ethical destiny of the human species was put to the test. It no longer is […] Art […] is now this place.”5 This shift, to be marked by substituting the word “art” for every iteration of the word “beautiful” in the third Critique and observing the consequences for Kant’s argument, de Duve maintains, captures the transformation in art embodied by Duchamp’s readymades. Judging the readymade for its beauty, traditionally conceived, is clearly not to the point. Duchamp himself rejected such a reading, claiming his choice of readymades was grounded in “visual indifference […] in fact a complete anaesthesia.”6 And even had he not, doing so would involve a kind of category mistake: it would be to judge the beauty of an everyday urinal, snow shovel, or bottle rack when Duchamp’s works cannot be identified with these everyday artefacts.

The question posed by the readymade concerns its standing as art, not its beauty. Readymades not only bring this shift from beauty to artistic status into the open; they insist on it. As a kind of art, readymades reduce to the artist’s judgement “this is art,” which their viewers are asked to endorse by judging in turn. For this reason, de Duve claims, readymades narrow the gap between making and judging: if the viewer concurs by endorsing the artist’s identification of the object in question as art, she does no more or less than the artist himself. For the artist did not make it either, he merely chose it.7 By endorsing that choice, the viewer chooses it anew, thereby admitting it into in her personal artistic canon. De Duve is withering about the oft-expressed view that the readymades are somehow responsible for this shift from beauty to art. The readymades do not bring about this transformation; that they are possible at all is a consequence of it: “In front of a readymade, there is no longer any technical difference between making art and appreciating it […] this ‘fact’ is not a consequence of the readymade, rather, it is its condition. The readymade only reveals it” (KD, 290). In sum, Duchamp had the acuity and wit not only to recognize but to capitalize upon a change that was already underway.8

In Kant after Duchamp, de Duve presents this transformation in the nature of aesthetic judgement as reflecting a shift from the specific to the generic internal to the development of art itself.9 In the more recent Aesthetics at Large, consistent with his original goal of providing an “archaeology” of modernism, he presents the same shift in more foundational terms. It signals a transition between distinct artistic epochs or “epistemes,” from a Beaux Arts to a Generic System of the Arts that he now calls “Art-in-General.”10 That is, the move from a closed, or at best gradually evolving, set of conventions tied to individual art forms within which works of art are made and judged to an open-ended set of conventions about which judgement is now asked to pronounce. Modernism, which he conceives as a kind of hinge period between the two, just is this transition for de Duve.11 That something like the readymade was even capable of showing up as art, at a given moment of history, reflects these underlying transformations in the nature of art.12

De Duve’s reason for believing that the judgement “this is art,” understood in light of these developments, still counts as aesthetic is that we make it by consulting the complex array of feelings elicited by the particular judged.13 By conceiving the judgement in this way, de Duve is trying to preserve the basic Kantian thought that judgements taking the form “this x is beautiful” do not predicate a concept of x so much as refer the intuition of x to the feeling it gives rise to in the judging subject. To this end, de Duve proposes transferring the general outlines of Kant’s analysis of beauty in judgements of the form “this x is beautiful” to the idea of art in judgements of the form “this x is art.” Rather than subsuming the candidate in question under the concept “art,” the judgement “this is art” instead expresses the feeling that it warrants inclusion in one’s personal canon or “imaginary museum.” In making the judgment, one holds the feelings elicited by the candidate work up to the feelings elicited by objects or events to which one has previously accorded that status, in order to compare the feelings so elicited.

Like reflective judgment in Kant, this is based on an act of comparison, though what is compared on de Duve’s account are the feelings that different works have occasioned, rather than the feelings that the particular in question does in fact give rise to and the feelings that it might in principle have given rise to. That is, in Kantian terms, the disposition of the faculties elicited by the particular judged and an ideal but indeterminate norm for the ‘free formal interplay’ of those same faculties. So even here, in his apparently most Kantian moment, de Duve’s way of conceiving such judgement is subtly different from Kant’s. For de Duve it is a comparison between the feelings to which different cases have in fact given rise. For Kant it is a comparison between the free play of the faculties, as sensed in feeling, to which a particular does give rise, and some ideal such relation to which it might in principle have given rise.14

But even taken on its own terms, it is hard to see by what criteria past feelings, as non-cognitive and logically private, could be reliably reidentified over time for the purpose of such comparison.15 De Duve’s understanding of the word “art” aside, this reading shares Greenberg’s tendency to marginalize the reflective dimension of aesthetic judgment in Kant. That is, de Duve underplays the necessity to critically reflect on the grounds of our pleasure in aesthetic judgment, and hence the warrant for imputing such pleasure to others. But such reflection is a minimal requirement for laying claim to the agreement of others for Kant. By echoing Greenberg’s stress on the “immediate” and “involuntary” nature of such judgment, de Duve unnecessarily deprives himself of the most obvious criterion for distinguishing, at least in principle, between judgments of the beautiful and the agreeable. Granted, this will always be a moot point in practice, since one can never be sure that one has in fact abstracted from any idiosyncratic basis one might have for the pleasure one takes in a given object or event, rendering it merely agreeable. Nonetheless, if aesthetic judgments were really as immediate and involuntary as de Duve suggests, what basis would we have for contesting the skeptical rejoinder that all claims to universality are in reality exercises of power designed to mask the merely subjective preferences of their maker?

Introspection cannot help us here, as de Duve is clearly aware, because the feelings occasioned by the agreeable and the beautiful need not be discriminable in experience. But de Duve fails to draw the implications of this fact for his own contention that it is the claim to universality that serves as our best indication of a judgment’s disinterestedness and not vice versa; for this begs a question as to how anyone could know that claim to be warranted.16 I agree with de Duve that we do feel strongly about the universality of our judgments and that it is therefore not a matter of indifference to us whether those whose judgments matter to us share our feelings. In this respect the pleasure we take in the agreeable and the beautiful does appear to be distinct, and the phenomenology of their respective judgments correspondingly different. Nonetheless, the fact that I feel sufficiently passionate about some of my judgments to declare their universality could have any number of psychological grounds to which I am blind. So, the mere fact that I feel moved to demand assent from others concerning some of my feelings but not others does nothing to mitigate the fact that all claims to universality are in principle open to corruption and hence defeasible. My conviction, for all its passion, may be nothing more than bloody-mindedness.

But something more fundamental that passes curiously unremarked in both de Duve’s own account and the critical attention it has to date received is what we should do with all occurrences of the word “art” (specifically schöne Kunst, but also Kunst überhaupt) in the third Critique itself, once we have performed the proposed substitution of this word for all occurrences of the word “beauty.” For if we do this without also modifying every iteration of the word “art” in some way, we will erase a distinction that is fundamental not only to Kant’s account of aesthetic judgement but also his theory of art. That is, the distinction between “free” and “dependent” beauty (C, §16). This is what enables Kant’s account of aesthetic judgement, which would otherwise be exceptionally restrictive—applying solely to cases in which we either have no knowledge of the object judged or abstract entirely from such knowledge in so judging—to be applied to cases in which we do take the nature of what we are judging into account when judging it beautiful as or for an object of that kind. That is, beautiful as or for a church rather than a townhall, given their very different functions, or for a print rather than a watercolour, given their different general characteristics.17 Note that such distinctions may in principle be run at ever finer levels of grain: one can judge the same building beautiful as or for a Protestant rather than a Catholic church, given their different norms of decoration, the print beautiful as or for an aquatint rather than an etching, given their different technical constraints, and so on. This does not, moreover, hold only with respect to the fine and applied arts, and nor is it limited to artefacts more generally. One may judge something beautiful as or for an Ash rather than an Oak or beautiful as or for a Thoroughbred rather than a Dray, given their very different forms and functions.

A moment’s reflection makes it apparent that although Kant himself presents judgements of “adherent” or “dependent” beauty [anhängende Schönheit], beauty that “depends” or “hangs upon” a concept of what the object judged “is meant to be,” as almost a marginal addendum to the real business of the Analytic, it is arguably what saves his account of aesthetic judgement for any practical purpose. For the vast majority of aesthetic judgements made by humans do not—and arguably should not—abstract from all knowledge of the objects judged, making judgements of dependent beauty by far the more common kind of aesthetic judgement in practice. Imagine the impact of learning that what you took to be a naturally occurring sunset is only possible due to pollutants in the atmosphere, or that a charming, ostensibly natural play of light on the water’s surface is in fact a by-product of an oil slick, will have on your aesthetic judgements thereof. Although taking any of this into account is ruled out on Kant’s account of pure aesthetic judgement, some knowledge of what is being judged typically plays a major role in most aesthetic judgements—especially with respect to art. What happens to judgements of this kind, in which knowledge of the object judged does impact the judgement, as opposed to judgements in which it does not, once aesthetic judgement (in general) confers the status “art” on any object so judged? That is, once the word “art” has been substituted for every iteration of the word “beauty” but without modifying every occurrence of schöne Kunst in Kant’s text accordingly.

In Kant after Duchamp, de Duve limits himself to the parenthetical remark: “(Although [Kant] has interesting things to say about art, they will not come under review here)” (KD, 302). But given that de Duve’s proposals have implications, both for what Kant says about judgements of artistic beauty as one species of dependent beauty and for how such judgements are to be distinguished from judgements of free beauty that may prove untenable, it is not respectable simply to duck the question in this way.18 This question of philosophical probity aside, given that de Duve is concerned, unlike a more traditional Kant scholar, to make Kant’s aesthetics productive for more recent debates about art, it is genuinely odd that he neglects to consider judgements of dependent beauty. Not only is this one of the few places in the third Critique where Kant offers a clue as to how his account might function in practice, is also where he clarifies how his general theory of aesthetic judgement applies to works of art in particular.19

Setting this reticence about judgements of dependent beauty to one side, De Duve’s focus on updating Kant’s account of judgements of free beauty requires him to recast Kant’s “Antinomy of Aesthetic Judgement” accordingly. “Antinomies of Reason” arise when our attempt to fulfil Reason’s demand to locate the unconditioned ground for a given cognitive power pushes us beyond the domain in which finite rational beings, as sensibly conditioned cognizers, can make genuine knowledge claims. As such, the task of a “Dialectic” is always negative for Kant: it seeks to dispel the conflicts that appear to arise between the principles underlying our capacity for a given kind of judgement. Since this threatens to undermine the a priori basis of the capacity in question, such conflicts have to be defused. Kant initially presents the “Antinomy of Taste” informally, as follows:

• Thesis: everyone has his own taste.
  (Taste is subjective, so makes no claims on others.)

• Antithesis: there is no disputing about taste.
  (Even if it is objective, it cannot be rationally disputed.)

As it stands this hardly creates an antinomy: prima facie the two claims look compatible: there is no disputing about taste precisely because everyone has their own taste. But Kant takes the antithesis to contain an ambiguity: we can and do argue about taste even if we cannot, strictly speaking, dispute taste. Dispute is based on concepts and so admits of proof; that we cannot dispute about taste means that disagreements about taste cannot be based on concepts. But the fact that we nonetheless argue about it suggests we cannot regard it as merely subjective either: we act as though there were something extra-subjective that our aesthetic judgements latch onto, even if we cannot establish as much decisively. Were this not so, there would be little point to our arguing; for what would be arguing about? This gives rise to the Antinomy proper:

• Thesis: the judgement of taste is not based on concepts, for otherwise it would be possible to dispute about it (decide by means of proofs).
• Antithesis: the judgement of taste is based on concepts, for otherwise, despite its variety, it would not even be possible to argue about it (to lay claim to the necessary assent of others to this judgement). (C, 56, Ak. 338–39)

Formulated in this revised way, the antinomy is clear. If the thesis was not true one could prove one’s aesthetic judgements and taste would be objective. If the antithesis was not true we could raise no claim on the agreement of others, and taste would be subjective. Kant’s goal is to show that thesis and antithesis each mistake one aspect of judgements of taste for the whole; the task of the Dialectic is to show that, since both are needed, they must be compatible. This requires disambiguating the notion of a “concept” at stake. Thesis and antithesis, it turns out, use the term in different ways. If so, there need be no conflict between them. The thesis, Kant claims, appeals to a “concept of the understanding”: that is, a determinate concept that admits of sensible presentation in intuition. The antithesis, by contrast, appeals to an “idea of reason”: that is, an indeterminate and indeterminable concept that cannot be presented to intuition in sensible form (God, freedom, the immortality of the soul, and so on). The antinomy arises from the mistaken assumption that the notion of a “concept” is being used univocally throughout. As such, it may be defused as follows:

• • Thesis: judgement is not based on a determinate.
  • Antithesis: judgement is based on an indeterminate. (C, §57, Ak. 340–41)

Note that this solution is purely formal: it removes any suggestion of logical contradiction. What remains controversial, at least for anyone unsympathetic to Kant’s “Transcendental Idealism,” is Kant’s identification of the indeterminate and indeterminable concept at stake with the “supersensible substratum of appearances.” For this renders his own solution of the antinomy dependent upon the metaphysics of Transcendental Idealism. The issues here are complex. I shall hold off discussing them until de Duve’s own proposals for updating Kant’s Antinomy are on the table. Consistent with his more general account, this requires replacing Kant’s classical formulation of the Antinomy, couched in terms of beauty, with its modern formulation, couched in terms of art:

• Thesis: The sentence “This is art” is not based on concepts, for otherwise we would be able to demonstrate probatively what is art and what is not.
• Antithesis: The sentence “This is art” is based on concepts, for otherwise we would not even bother to quarrel about what is art—and about what art is.20

Stated more pithily:

• Thesis: The sentence “this is art” is not based upon concepts.
• Antithesis: The sentence “this is art” is based upon concepts. (KD, 304)

Or, more pithily still:

• Thesis: Art is not a concept.
• Antithesis: Art is a concept. (KD, 304; CA, 143)

Set aside here the question as to whether all these formulations are compatible. De Duve associates all versions of the thesis with formalism (particularly as theorized by Greenberg) and all versions of the antithesis with Conceptual Art (particularly as theorized by Joseph Kosuth). In what de Duve calls its modern form, the aesthetic judgement “this is art” not only judges but also produces the work of art: a urinal, say, is initially baptized as art through this judgement, and that baptism is endorsed, or not, by the viewer’s judgement in turn. Reduced to a mere nominator of objects as art, the artist is—as least technically—no different to the work’s judge. Baptizing objects as art requires no apprenticeship, no acquisition of special skills, no knowledge of artistic rules or conventions: in theory, at least, anyone could do it.21 Before a readymade, genius and taste are one and the same: an artist, who did not make the object, judges it to be art, putting it forward as such for the judgement of others who did not make it either. In submitting that object to judgement, viewers assess whether, when judged as art, it elicits an “art feeling” sufficiently like those elicited by the works previously accorded that status in their personal canons. Because readymades reduce in this way to the implication “this is art,” de Duve takes them to be presented “in the name of art,” or what he today calls “the idea of ‘art itself.’” In being so presented, they are to be judged aesthetically as to their standing as such. His solution to the antinomy is formulated accordingly:

• Thesis: The sentence “this is art” is not based on the concept of art; it is based on the aesthetic/artistic feeling.
• Antithesis: The sentence “this is art” assumes the concept of art; it assumes the aesthetic/artistic idea. (KD, 321)

At least, this is how it appeared in Kant after Duchamp. It is recast in the slightly modified form that de Duve now prefers in Aesthetics at Large:

• Thesis: The sentence “this is art” is not based on the concept of art; it expresses the feeling of art.
• Antithesis: The sentence “this is art” is based on the concept of art; it postulates the idea of art. (CA, 149)

Because the concept denied in the thesis is determinate and the concept affirmed in the antithesis indeterminate, there is no formal contradiction. The feeling endorsed by the thesis is the aesthetic feeling that what one is judging is art; the idea endorsed by the antithesis is not merely compatible with such a feeling but required by it: it is “the idea of ‘art itself.’” That is, the idea that there is something indeterminate and indeterminable that everything called “art” must have in common, and that the art feeling implicated by the thesis is capable of picking out.22 So construed, the thesis denies that aesthetic judgements of the form “this is art” are based on a determinate concept of art, while the antithesis concedes that such judgements are indeed based on a concept, but only on the condition that the concept at stake is understood as the indeterminate and indeterminable idea of “art itself.” As de Duve now puts it: this is “the idea in the name of which the candidate to your art judgement is summoned to appear” (CA, 152).

De Duve’s solution, as formulated here, thus fulfils the formal task of the Dialectic; it defuses the apparent contradiction between two competing principles and does so, like Kant, by distinguishing between the determinate and indeterminate concept employed by each. But de Duve has a tendency to multiply solutions, creating unnecessary hostages to fortune. Here, for example, he muddies the waters by identifying the word “art” understood—by appeal to Kripke—as a proper name that rigidly fixes a reference but possesses no semantic content of its own, with the idea of art, understood in the Kantian sense as something that possesses so rich a semantic content that no sensible presentation could be adequate to it:23

• Thesis: Art is not a concept; it is a proper name.
• Antithesis: Art is a concept; it is an idea.24

The two notions appealed to here, in order to clarify the ways in which aesthetic judgement does and does not rely on concepts—Kantian “ideas of reason” and Kripkean “proper names,” respectively—are clearly incompatible with one another, rendering this version of thesis and antithesis incompatible. Hence what de Duve presents as an alternative, “more compact and more general formulation” of the solution just cited, turns out to be incompatible with it, and so counts less as a solution of the Antinomy than, at best, a restatement of it.25

II. De Duve on “sensus communis,” Kant on the “supersensible”

So much for fulfilling the formal task of the Antinomy. Some of de Duve’s formulations succeed, while others do not. The coherence of his account would benefit from restricting himself to one rather than multiplying solutions and attendant hostages to fortune in this way. As regards the substantive task of an Antinomy, which is to clarify not merely what kind of concept (an indeterminate and indeterminable one) underwrites the legitimacy of judgements of taste, but which specific concept of that kind, Kant and de Duve identify the idea at stake quite differently. In Kant after Duchamp, de Duve identifies it with the idea of “sensus communis” that Kant introduces in the “Fourth Moment” to underwrite the legitimacy of demanding agreement from others for judgements based on nothing more than feeling. Kant defines this idea as follows:

the idea of a communal sense, i.e., a faculty for judging that in its reflection takes account (a priori) of everyone else’s way of representing in thought, in order as it were to hold its judgement up to human reason as a whole and thereby avoid the illusion which, from subjective private conditions that could easily be held to be objective, would have a detrimental influence on the judgement. (C, §40, Ak. 5:293–94)26

By identifying the indeterminate and indeterminable idea at stake in Kant’s Antinomy with sensus communis in this way, de Duve is interpreting the Dialectic (C, §55ff, especially §§56–57) through the optic of more or less distant passages from the Analytic and the Deduction (C, §§20–22, §40).27 In Aesthetics at Large, by contrast, de Duve is noticeably more circumspect about identifying the notion of sensus communis with the idea of “Art itself.” That is, the idea that whilst it may not be conceptually specifiable, there must nonetheless be something that all art has in common, and that it is the role of aesthetic feeling to pick out.28 Just as it is the role of such feeling to register the presence of beauty for Kant, despite the variety of things that may exhibit it. The greater caution is welcome, as it is very hard to see how Kant’s conception of a shared capacity for judging which feelings are capable of being shared could be identified with the idea that there must be something that all art has in common, even if such a shared capacity is what enables us to pick this common something out.

Be that as it may, neither solution can be easily squared with the indeterminate and indeterminable idea of a “supersensible substratum” that Kant himself takes to be implicated by the antithesis, and this raises a question as to the rationale for appealing to the Antinomy at all. Given that, for Kant, Antinomies arise when reason’s quest for an unconditioned ground with which to arrest the regress of incomplete explanations leads us to overstep the bounds of what we may legitimately claim to know, the resolution to a Kantian Antinomy necessarily implicates Transcendental Idealism. A question thus arises as to whether any solution that neglects the metaphysics of Transcendental Idealism entirely still merits being called Kantian and, if so, in what sense.

De Duve has remained largely silent on metaphysics of Transcendental Idealism to date. Rather than endorsing or rejecting Kant’s own solution, he redirects his readers’ attention to the idea of a “sensus communis” introduced earlier in the text. Whatever else may said about this solution, it is not Kant’s own. Kant’s solution unequivocally implicates Transcendental Idealism: that is, the ideality of space and time for imperfectly (or “finitely”) rational beings such as ourselves, whose experience of the world is sensibly conditioned by space and time as “a priori forms of intuition.” It does so by identifying the indeterminate and indeterminable idea at stake with a “supersensible substratum.” That is, a non-sensible (or merely intelligible) ground within and without. This implication presents a significant challenge for anyone who would like to harness Kant’s aesthetics for more recent debate about art. The question, stated baldly, is whether appealing to Kant commits one to Transcendental Idealism and, if so, what being so committed entails.29 This is a point on which anyone who appeals to Kant eventually has to come clean. For one cannot seriously engage with any philosopher’s work, least of all a philosopher as systematic as Kant, by simply helping oneself to their conclusions while either bracketing the metaphysical premises they depend upon or ignoring the substantive commitments these commitments may entrain. Until such time, the cogency of de Duve’s own theory cannot really be assessed.

Setting aside the vexed question of how he understands the relation between sensus communis and the Antinomy, it is clear that the former still goes to the heart of de Duve’s project as he now conceives it.30 For what is significant about aesthetic judgement for de Duve is not, so to speak, the sheer fact that human beings aspire to make such judgements, but what he takes doing so to reveal: a universally shared capacity for agreeing on the basis of nothing more than shared feeling. Postulating such a capacity, or sensus communis, is all that legitimates our practice of demanding agreement from others for judgements based on nothing more than feeling. Because the judgement “this is art” does precisely that, de Duve now wants to claim that such judgements implicate our shared humanity. They speak to the fact that humans ought to be capable of living in peace, despite the pervasiveness of aesthetic disagreement attesting to our inability to do so. This is why art is now “the place” where hope in the ethical destiny of mankind is put to the test.

When Kant first introduces the idea of a sensus communis in the “Fourth Moment,” he presents it as a presupposition of making pure aesthetic judgements but withholds judgement as to whether it should be understood constitutively or regulatively. That is, whether it should be regarded as a necessary condition of making such judgements, hence something the mere fact that we make such judgements shows must obtain. Or whether it should be understood, more weakly, as an action guiding principle we are rationally constrained to act in accordance with, in so far as we aspire to make such judgements and, thereby, hope to bring into being.31 De Duve takes it to be regulative: we implicitly attribute such a capacity for shared feeling to others simply by virtue of demanding assent for our own aesthetic judgements. Indeed, that we do is precisely what makes such judgement significant: for it shows that we regard all others as similarly endowed with a capacity for shared feeling, such that they could, at least in principle, concur. As de Duve puts it: “Identified with taste itself, sensus communis thus names that very special faculty of agreeing that is based on the common feeling of belonging to the community defined by precisely this feeling.”32 This is what leads de Duve to claim, per my epigram, that Kant has fathomed the ethical significance of aesthetic disagreement; that it amounts to a denial of the other’s humanity.

As actually formulated, however, this claim seems ill-expressed or overblown. If we do indeed share a capacity qua human for judging on the basis of nothing more than feeling, and I deny someone else this very capacity, I might be thought to deny that person’s humanity.33 To the extent that I argue with her judgement, however, I acknowledge her humanity. Doing so requires that I recognize the force of her competing claim, that it has implications for my own judgement to which I am now called upon to respond. As such, the mere fact of aesthetic disagreement would seem to constitute an acknowledgement of the other’s humanity, rather than its converse.34

Even if this caveat about formulation is accepted, however, de Duve’s proposal still strikes me as implausibly overblown. To claim that competing judgements as to some object or event’s standing as art are now “the place” where the ethical destiny of mankind is put to the test hugely overburdens the significance of aesthetic disagreements about art. Not our treatment of migrants, not our tolerance of global poverty, not our collective failures over the climate catastrophe, but our judgements about urinals and hat racks presented as art—that is where the ethical destiny of mankind is put to the test? However deeply one may care about art, it is very hard to believe that it could shoulder the burden that de Duve’s theory asks it to carry. Indeed the account appears to so overburden the significance of our responses to art that it is hard to be sure whether de Duve even wants to be taken literally on this score.

But if he does, the claim seems curiously undermotivated by anything in de Duve’s own theory. For once Kant’s claims are shorn of the systematic significance their relation to his Transcendental Idealism confers on them, what would henceforth give such grandiose claims for aesthetic judgement their meaning or force? You must already care deeply not only about art but about modern Western art—a niche concern to put it weakly—before you have reason to care about anybody else’s conflicting judgement about Duchamp’s Fountain or one of Carl Andre’s Equivalents. But one would think that everyone has reason to care qua human about whether it is possible to exercise their will freely within the causal order or make systematic sense of their experience. Simply getting out of bed in the morning would seem to require it.

The significance of aesthetic judgement within Kant’s broader critical project falls beyond my concerns here, but very briefly, this can be understood in formal, systematic, or substantive terms. Formally, it is bound up with various “analogies of reflection” that Kant identifies between the respective structures of aesthetic and moral judgements: beauty functions as “symbol” (or indirect, analogical presentation) of morality because we reflect on beauty in ways that are structurally isomorphic to the ways in which we reflect on morality (C, §59, Ak. 353–54). Systematically, taste affords a “gentle transition” to the more demanding requirements of morality. Whereas taste requires merely that I judge apart from my own interests, morality often requires that I act contrary to those interests. For this reason, Kant claims that taking an immediate interest in natural beauty is “always the mark of a good soul” and, if habitual, indicates a mental disposition conducive to moral cultivation (C, §42, Ak. 299). Substantively, despite not being made by or for humans, natural beauty seems so well suited to bringing about a free formal interplay of the capacities required for cognition in general as to have been produced for that very purpose. This is an unexpected gift: it offers us a “trace or […] hint” that our most basic epistemic and moral goals of making sense of nature’s diversity and acting freely within its causally determined order may not be in vain (C, §42, Ak. 300).35 We cannot fail to take an interest in any such hint.

All of this, if de Duve’s proposals go through, should now be seen as hanging on competing judgements as to whether or not some object or event is art. But given that Kant’s own claims for the significance of aesthetic judgement accrue whatever force they possess from these implications of his broader project, it is difficult not to see the attribution of such significance to competing judgements about artistic standing, once shorn of these systematic considerations, as unearned.

Kant’s own solution to the Antinomy must be seen in light of these broader concerns. Given the context, one would expect the concept at issue in the Antinomy to be beauty: for what other concept could judgements taking the form “this x is beautiful” implicate? Then the Antinomy would concern whether or not beauty is a concept and, if so, what kind of concept: the thesis would maintain that beauty cannot be a concept since if it were judgements of taste would admit of conclusive proof, and the antithesis would maintain that it must be a concept for otherwise we could not so much as argue about such judgements. The solution would maintain, accordingly, that while beauty cannot be a determinate concept it can (and indeed must) be an indeterminate and indeterminable idea. As an idea, beauty could be put forward as the unconditioned ground of judgements of taste, but it cannot be further determined: it is neither a concept of the understanding under which instances may be subsumed in virtue of exhibiting the relevant traits, nor does it provide a decision mechanism for adjudicating between conflicting judgements in practice.

All this would be straightforward, but it is not how Kant identifies the idea at stake. Unlike de Duve, who identifies it with sensus communis, Kant identifies the indeterminate and indeterminable idea implicated by judgements of taste with the “supersensible within and without.”36 To anyone unfamiliar with the metaphysics of Transcendental Idealism this will sound highly obscure. What does it mean? Anything Kant characterizes as “super-sensible” must, as the term implies, pick out something extra sensible. So understood, the supersensible would include anything that could not be an object of possible knowledge or experience for humans as finite, or imperfectly, rational beings. As finite—sensibly conditioned, rather than purely rational—beings, our knowledge is constrained by what can be sensibly given in space and time. For this reason, we cannot claim to know the supersensible objects of which we form rational ideas: in this case the “supersensible substratum” (or “thing-in-itself”) that Kant takes to underlie appearances within and without.37

Despite this, we may postulate the objects of such ideas, as the unconditioned ground of everything we can know and do, on pain of a regress of conditions that would fail to provide a complete well-grounded account of the possibility a given a domain of human thought or action.38 It is important to grasp, however, that what is so postulated, despite going beyond anything we may claim to know, is nonetheless not empty. Were we not able to act under the idea of freedom, for example, we could not even aspire to act morally. Since aspiring to act in accordance with the demands of duty requires that we think of ourselves as responsible for our actions rather than determined. Thus, despite the fact that we cannot ultimately know that we are free—our noumenal nature not being, by definition, a possible object of phenomenal experience—the idea of freedom nonetheless has a regulative, action-guiding role for human beings.

But why identify the idea of beauty in judgements of the form “this rose is beautiful” with the supersensible so understood? Paul Guyer, no friend of Transcendental Idealism, dismisses this as an egregious product of architectonic considerations generated by Kant’s metaphysics, from which his theory of aesthetic judgement is therefore best detached. Why, Guyer asks, does the idea of beauty itself not suffice?39 If some further grounding is thought necessary, why not the free play of the faculties, which enables us to pick out its instances? As ever, Henry Allison is more charitable: the resolution of a Kantian Antinomy requires an idea capable of satisfying reason’s demand for an unconditioned ground for a given domain of human thought or action, but the object of which—for just this reason—cannot be given in experience. For, if it were so given, that would immediately raise a question as to its grounds. So understood, it is not immediately obvious that beauty, which may indeed be predicated of objects given in experience, could provide what is required.

But Allison takes Kant’s characterization of beauty to provide a clue to his thinking in the Antinomy, specifically his reasons for identifying aesthetic judgement (judgements of the form “this rose [etc.] is beautiful”) with the idea of a supersensible substratum. In §17, Kant defines beauty as “an object’s form of purposiveness insofar as it is perceived in the object without the representation of a purpose” (C, §17, Ak. 236).40 Perceiving such “purposiveness without purpose” [Zweckmäßigkeit ohne Zweck], the appearance of an indeterminate well-orderedness, absent the identification of any particular purpose, in a perceptual manifold is Kant’s explanation of the pleasure we take in the beautiful.

In §57 and its appended “Remarks,” Kant glosses the idea of a supersensible substrate he takes to be implicated by apprehending such indeterminate purposiveness in a bewildering variety of ways.41 These are some of the most obscure passages in the whole third Critique, no doubt because the implications of Kant’s broader critical project for his aesthetics are in play.42 I shall restrict myself to what Kant has to say about the supersensible “without” here, so as to remain with those formulations that are most clearly implicated by the critique of taste, rather than get waylaid by questions internal to Kant’s moral philosophy. Take Kant’s identification of “the principle of the subjective purposiveness of nature for our faculty of cognition” (C, §57, Ak. 346) as the supersensible idea implicated by pure aesthetic judgements.43 Something is beautiful or “subjectively purposive” for our powers of cognition if it occasions an experience of purposiveness without purpose. This happens when nature strikes us as so well suited to the optimal attunement of our powers of cognition as to appear designed for this very purpose. This is not to claim that nature is so designed, however, since that would go beyond the bounds of anything we can legitimately claim to know on Kant’s theory of knowledge. So, it should not be taken as a Kantian version of the argument from design. It is rather a claim about how some natural object or vista strikes us—that is, as if were so designed.

This is why Kant insists on the ideality of nature’s purposiveness in §58: “in such judging what is at issue is not what nature is or even what it is for us as a purpose, but how we take it in […] it is a favour with which we take nature in and not a favour that it shows to us” [wo es Gunst ist, womit wir die Natur aufnehmen, nicht Gunst, die sie uns erzeight].44 We “receive nature with favour,” then, to the extent that we approach it as if it were designed with our cognitive capacities in mind, and this is not to claim that it is so designed. To view nature as subjectively, rather than objectively, purposive for our powers of cognition in this sense is to take pleasure in those sensible manifolds that strike us as so well suited to bringing about a free formal interplay of the capacities required for cognition in general as to appear designed for this very purpose. Such a thought thus implicates the idea of a supersensible substrate of appearances, and in so doing provides a “hint or clue” that we fit into nature that we could not fail to take an interest in, even if it goes beyond anything we may legitimately claim to know. This is the final significance of beauty in Kant’s system, but it remains “indemonstrable.” As a mere idea, it is neither a basis for knowledge, not susceptible to proof.

Why I am trawling over such arcane material, the relevance of which may be far from obvious here? There is very simple reason for this: aesthetic judgement could only have this status on Kant’s account on the assumption, precisely, that it is not art that is at stake. That we are not judging something made by or for humans is crucial: the reason such judgements have the significance that they do, at least for Kant, is that they take as their object something we have no prior reason to expect will not only conform to the requirements of judgement but seem so well suited to doing so as to have been made for that very purpose. The beauty of nature is thus wholly unanticipated gift, and something that we receive with joy for precisely this reason. By contrast, because art is made for just this purpose, it rules out in advance the possibility of being surprised by its suitability or fit with our cognitive goals in the same way. But if the proposed substitution of “this x is art” for “this x is beautiful” goes through, all these implications of natural beauty must fall away, leaving us with a question: on what basis could art, absent the deep systematic significance conferred on natural beauty by Kant’s Transcendental Idealism, now be “the place” where the ethical destiny of mankind is put to test? So far as I can see, de Duve has yet to say.

Given that “updating” Kant’s Antinomy has been central to de Duve’s project from Kant after Duchamp through Aesthetics at Large, and the solution to a Kantian Antinomy always implicates Transcendental Idealism, in the third Critique through Kant’s identification of the indeterminate and indeterminable idea at stake with a supersensible substratum, this raises a question as to whether the proposed use of Kant still merits being called Kantian, de Duve’s contention that “Kant got it right” notwithstanding. This will remain true until such time as de Duve clarifies how he understands the relation between the “Kant after Duchamp approach” to aesthetic judgement and Kant’s own Transcendental Idealism.

III. Reconceiving aesthetic judgement: Could “art” be a proper name?

Set aside the question of whether de Duve can get what he needs from Kant. However one comes out on that, and it will be clear that I have my doubts, de Duve’s theory could still be coherent in its own right, independently of its claims on Kant. Before assessing this, one worry needs to be laid to rest at the outset: de Duve’s proposals for bringing aesthetics up to date by replacing the judgement “this is beautiful” with the judgement “this is art” might be seen as some kind of category mistake, if one understands the latter judgment to subsume a particular under a universal, namely, the concept “art.” But de Duve rejects this interpretation of what he is proposing, maintaining that the judgment “this is art” remains aesthetic by virtue of being singular (pertaining solely to the particular judged, and not to objects of its more general kind) and reached by dint of feeling alone. Rather than subsuming a particular under a concept, thereby predicating some property of that particular, de Duve maintains that the judgment “this is art” confers the name “art” on any object judged accordingly.

As de Duve presents it, the judgment “this is art” is akin to the original baptism through which a person acquires a proper name. Just as all people called “Paul” need have nothing in common in virtue of which they are so called, so works of art need have nothing in common in virtue of which they are called “art.” “Paul” is not a concept under which persons are subsumed in virtue of possessing the relevant properties or traits, and “art” is not a concept under which entities or events are subsumed in virtue of fulfilling the relevant conditions. According to de Duve, art “is of the same order as that which assembles all the Peters, Pauls, or Harrys: they have their name in common, but their name is not a common noun; it is their proper name” (KD, 52). Further: “Peter, Paul, or Harry, even though common to all the Peter, Pauls or Harrys, are labels that designate this Peter, this Paul, or this Harry, one by one, without presupposing shared qualities for them, without lending them a common sense” (KD, 54).

De Duve wants to support this claim by recourse to Saul Kripke’s account of proper names. Kripke understands proper names as “rigid designators,” terms that pick out the same person or natural kind across all possible (which is to say, counterfactual) worlds one might imagine, irrespective of whether their bearer in those worlds exhibits the same traits, or does the same things, as she does in this world. “Nixon,” for example, always picks out the actual Nixon, irrespective of whether in some other, counterfactual world the actual Nixon does not go into politics, and someone who is not Nixon does all the things in that world that Nixon did in ours. Because it designates rigidly, across all possible worlds, in this counterfactual world “Nixon” would not pick out a person forced from the Presidential office as a result of the Watergate scandal, but someone who lived a quiet and blameless life. There might well be a person fitting the former description in this alternative world, it just would not be Nixon, and this would remain true irrespective of whether he happened to be called “Nixon” in that world. Only those properties that are essential to Nixon being Nixon hold across all possible worlds, which is just to say that only that person born on the right date to Nixon’s parents in this world remains Nixon across all possible worlds.45 This holds irrespective of whether he happens to be called “Nixon” in those worlds, and regardless of whether in those worlds he does any of the things standardly attributed to him in this one.

To get at whether de Duve is right that the judgements of the form “this is art” do not subsume a particular under the concept “art” but rather confer the name “art” on any object judged accordingly, and hence whether art could be a proper name, consider the basis on which we assign this status according to de Duve himself. Rather than doing so in virtue of the objects to which it is applied fulfilling some minimal set of conditions, or exhibiting a sufficient number of properties or traits, as would standardly be required for the application of a concept, whatever merits conferral of the name “art” instead needs to sustain comparison to those entities upon which one has conferred that name in the past. One determines whether it does by judging whether the feelings to which the new candidate gives rise are similar, in enough respects, to those elicited by other bearers of the name in one’s personal canon:

The phrase “this is art,” as it is uttered by art lovers who use it to judge by dint of their feelings and of the conflict among their feelings, is a baptism. […] In baptizing a given thing with the name of art, they express the feeling that it deserves to be so called. […] What matters is that the word “art” expresses a feeling, or a set of feelings, but that it does not mean what it expresses. (KD, 57–59)

The last claim, which might otherwise sound paradoxical, requires some clarification. This is de Duve’s way of expressing the thought that names have reference (or denotation) but not sense (or connotation). That is, they do not pick out whatever they pick out in virtue of the meaning of the terms employed (proper names are not disguised or abbreviated “definite descriptions”) but solely in virtue of the original act of baptism that conferred the name, and its subsequent communication via a historical chain of speakers who intend to preserve its original reference when applying it.46 But what about the former claim, specifically, that in calling something “art” the speaker “express[es] the feeling that it deserves to be so called?” De Duve elaborates on this point as follows:

[W]ith the word ‘art’ you are pointing a finger at all the things that make up your critical collection, your personal, imaginary museum. In calling this thing art, you are not giving out its meaning; you are relating it to everything else you call art. You don’t subsume it under a concept; you don’t justify it by means of a definition; you refer it to all the other things you have judged through a like procedure […]

This is why aesthetic judgements are always comparative […] the feeling that comes to you in the presence of this thing, here […] is measured by the memory of past feelings. […] you know an “art feeling” when you feel one, and what you are attaching to this, precisely by calling it art, is certainly an “art feeling.” […] [W]hen you choose to express your aesthetic judgment with “this is art,” you claim such a comparison. (KD, 59–61)

Pause to consider the emphasis on comparison that runs like a thread through this account: “aesthetic judgements,” de Duve claims, “are always comparative.” We confer the name art because something deserves to be so-called. As to what generates such desert, de Duve’s response is that the feelings elicited by the present candidate must sustain comparison to the feelings elicited by those things on which one has conferred that status in the past.47 Is this a credible application of Kripke’s account of proper names? It seems unlikely. Proper names, according to de Duve himself, are precisely not comparative: people get called “Peter,” “Paul,” or “Harry” without regard to the qualities of any other bearer of these names. This, after all, is what makes such terms names rather than concepts. Names fix a reference rather than pick out some entity in virtue of it exhibiting the appropriate set of traits to satisfy the marks of a given concept. Yet when it comes to art, according to de Duve himself, each of us must not only judge for ourselves whether the term is warranted before either applying it ourselves or endorsing its application by others, we can only do so by comparing the feelings elicited by the case at hand to the feelings elicited by previous recipients of that status in our personal canons.48

But when it comes to fixing a reference, questions of warrant have no purchase. Acts of naming are neither warranted nor unwarranted. Questions of warrant find no application here because names are not conferred for reasons. There is no reason—that is, no reason internal to Paul’s nature or the meaning of the word “Paul”—that Paul gets named “Paul” rather than “Peter.” There may of course be any number of reasons why calling their son “Paul” appealed to Paul’s parents more than calling him “Peter,” but these are not reasons of the required sort. Consulting one’s own feelings about the appropriateness of their choice is, moreover, clearly otiose: Paul will remain “Paul” and Peter will remain “Peter” however anyone else feels about the matter. In sum, it looks as though questions of warrant have no purchase, and consulting one’s feelings no relevance, when it comes to the conferral of names, as opposed to judging art. But if that is correct, the two cases cannot be analogous.

I am not the first to query the compatibility between the stress on comparison at the core of de Duve’s conception of aesthetic judgement and the appeal to a theory of proper names with which he tries to underwrite it. In perceptive early reviews of Kant after Duchamp both Jason Gaiger and Barry Schwabsky point out the irrelevance of comparison to the conferral of names.49 Although the point is developed in greater detail by Gaiger, with reference to the relevant philosophical literature, de Duve has to date only responded to Schwabsky. De Duve concedes the point about the non-comparative nature of everyday names before dismissing it as a (conceptually redundant) product of poor presentation: “Schwabsky’s objection, it seems to me, addresses the clumsily didactic analogies with first names I admit having made but not the issue of comparison, which is what I would emphasize in Kripke’s theory today” (CA, 151). It is hard to know what to make of this reply. In response to an objection about what he actually says about proper names, de Duve maintains that he really had in mind something else, something that is better understood in light of what Kripke has to say about natural kind terms (which he also conceives as names). In the case of natural kinds, unlike everyday proper names, the case in question can be referred back to a paradigm sample: “[W]hen Kripke claims that names of colors, measures, natural kinds, and substances, such as gold, are rigid designators, it is precisely on the basis of samples being compared with a paradigm sample acting as standard of comparison.”50 In Kant after Duchamp itself, however, it looks as though de Duve wished to maintain just the opposite:

[I]t is preferable to speak of art as a proper name in the narrow sense rather than as a rigid designator (the Kripkean category that encompasses both “real” proper names and names of substances, natural species, colors, etc.) […] Whereas it is true that a putative piece of gold (like a putative work of art) is identified with reference to a paradigm sample transmitted over time through the linguistic community, it remains possible to call on an intensional definition of gold (atomic number 79) to verify its true nature. Not so with art. (KD, 57–58n22)51

In other words, unlike “art,” “gold” (allegedly) admits of intensional definition; it has a meaning that can be specified, whereas art does not. For this reason, it is preferable not to understand art on the model of terms like gold and restrict the account to everyday names like “Peter” and “Paul.” In fact this is a misreading of Kripke, whose account of natural kind terms is meant to parallel in all essential respects (and even more controversially) his account of proper names.52 Though descriptions can be used to secure reference for natural kind terms when ostension is unavailable, this should not be mistaken for providing a definition. One might fix the reference of the natural kind “water,” say, by stipulating that “water” shall refer to the natural kind comprising most of the liquid in the oceans; but “most of the liquid in the oceans” thereby fixes the reference of the term “water”; it does not give its meaning.

Setting his revisionism to one side, does what de Duve now says he meant settle the question of coherence? Here it may help to recall what Kripke himself has to say about how his account applies to mass names, names that unlike everyday proper names do not pick out unique particulars. According to de Duve, the two cases are different: everyday proper names are not conferred on the basis of comparison, whereas names like “gold” are. This is not Kripke’s view. On Kripke’s own view, terms like “gold” function in much the same way as everyday proper names. They are neither disguised nor abbreviated definite descriptions, but rigidly designate the kind in question across all possible worlds, irrespective of how things happen to stand in those other, counterfactual situations. An initial baptism fixes the term’s reference, typically by ostension. A sample of the substance or kind in question is picked out, and it is stipulated that “gold” or “tiger” will henceforth refer to the kind under which the items (or at least most of the items) in the sample fall. Note that, for Kripke, this holds true irrespective of whether we identify the kind in question on the basis of properties that subsequently turn out to be contingent. In that case, the real bearer of the name will be whatever possesses those properties that are in fact essential to those objects to which it was originally applied, irrespective of whether the name was originally applied in virtue of those properties.

Nor is anything else we might mistake for a tiger (even should it possesses all those properties that provided the basis of our original, mistaken identification) in fact one if it does not possess those properties that turn out on further investigation to be essential to the kind of entities initially baptized as such: “Since we have found out that tigers do indeed […] form a single kind, then something not of this kind is not a tiger […] The original concept of cat is: that kind of thing, where the kind can be identified by paradigmatic instances. It is not something picked out by any qualitative dictionary definition.”53 Just as “gold” rigidly designates any material that in its pure form, it turns out, solely comprises atomic element number 79, so “tiger” rigidly identifies the kind of animal originally picked out with whatever anatomical structure, hunting, mating, and territorial patterns it is subsequently discovered to possess by zoology. In neither case are the terms in question applied by consulting anyone’s feelings as to whether they are warranted. And, while applying the term to a particular case may involve reflection, if and when it does, such reflection is taxonomic, not aesthetic.54

Kripke’s account remains controversial, especially with respect to natural kind terms. Indeed, it is controversial in this latter respect precisely because it is supposed to mirror his account of everyday proper names. I am not concerned here with whether Kripke’s claims hold up: leave that to the logicians, metaphysicians, and philosophers of language to decide. My concern here is part exegetical, part conceptual: does Kripke’s account serve de Duve’s purposes? Now that the parallel structures of Kripke’s arguments for everyday proper names and natural kind terms have been set out, it should be apparent that natural kind terms such as “gold” and “tiger” do not serve de Duve’s purposes any better than everyday proper names such as “Peter” and “Paul.” Qua names, both natural kind terms and proper names are initially used to fix a reference to a particular entity or kind, not to pick out the properties of that entity or kind, both then rigidly designate that same entity or kind across all possible worlds, both are communicated by an historical-causal chain that not only preserves but intends to preserve reference, and neither is a disguised or abbreviated definite description. That being so, if Kripke’s account of proper names does not serve de Duve’s purposes, then neither will his account of natural kind terms. As rigid designators, the same considerations apply to both.55 Unlike art, moreover, no one need consult their feelings before applying, or endorsing the application of, a given name: how anyone may feel about the item or sample in question is entirely redundant.

So far as I can make out then, appealing to Kripke’s account of proper names as rigid designators appears to be in tension with much of what de Duve himself wants to say about the affective and comparative nature of our aesthetic practices. As such, it is not obviously helping de Duve to frame a coherent theory of aesthetic judgement or art and may even be actively getting in the way of his doing so. Perhaps he should give it up? “Art” is clearly a concept, and it is implausible to maintain otherwise. The more interesting question, as de Duve himself appears to have realized, is understanding precisely what kind of concept it is.56 Absent some response to these challenges, the jury must remain out as to whether de Duve has succeeded in retrofitting Kant’s aesthetics in such a way as to demonstrate its ongoing applicability to forms of art that Kant could not have envisaged—to my mind a wholly noble endeavour—his protestations that “Kant got it right” notwithstanding.57