AESTHETIC SUPERVENIENCE VERSUS AESTHETIC GROUNDING

The claim that having aesthetic properties supervenes on having non-aesthetic properties has been widely discussed and, in various ways, defended. In this article, I aim to demonstrate that even if it is sometimes true that a supervenience relation holds between aesthetic properties and ‘subvenient’ non-aesthetic ones, it is not the interesting relation in the neighbourhood. As we shall see, a richer, asymmetric, and irreflexive relation is required, and I shall defend the claim that the increasingly popular relation of grounding does a much better job than supervenience.

against this core, intuitive claim; I think it is correct. There is indeed a relationship between the aesthetic and the non-aesthetic properties of an object, where the latter are somehow the basis for having the former. The incorrect claim that I do have something against is that this relation is supervenience.
Supervenience is a relation well known from many philosophical debates, including philosophy of mind, metaphysics, and ethics. It is typically defined as follows: 'A set of properties A supervenes upon another set B just in case no two things can differ with respect to A-properties without also differing with respect to their B-properties. In slogan form, "there cannot be an A-difference without a B-difference". ' 2 Given the serviceability of supervenience in many philosophical debates, the idea of appealing to the same kind of relation in aesthetics to explain how aesthetic properties supervene on non-aesthetic ones does indeed look appealing, and has been largely influential and widely defended. Zangwill, for instance, defends this claim, 3 and so does, famously, Levinson: 'aesthetic attributes of an object are supervenient on its non-aesthetic ones.
[…] I will take the thesis of aesthetic supervenience to be roughly this: […] Two objects (e.g. artworks) that differ aesthetically necessarily differ non-aesthetically. ' 4 To make these claims plausible and informative, they must be supplemented with more data about the nature of the supervenience basis -the question is what exactly do the aesthetic properties of an object supervene on? Indeed, Levinson, Zangwill, and many others all agree that in addition to colours, shapes, and the like (bearing in mind paintings and sunrises; other types of things, like symphonies, would of course possess other first-order non-aesthetic properties), some relational properties also have to be included in the 'supervenience basis' . These typically include the history and context of the production of an artwork. 5 In Walton's terms, aesthetic properties of an object depend not only on its narrow non-aesthetic properties, but also, importantly, on broad non-aesthetic relational properties, like the process and history of the production of an artwork as well as the context in which it was created. 6 Two indistinguishable paintings, exact duplicates, that is, exactly the same arrangements of paint on canvases of the same size, shape, texture, and so on, would (or, at least, could) still possess different aesthetic properties depending, for instance, on the period when they were created. 7 As Zangwill correctly points out, this is in no way an objection to the claim of aesthetic supervenience; it merely shows that the supervenience basis is not only constituted by non-aesthetic non-relational intrinsic properties of an object, but also by some relational properties -all one has to do is to allow for a broader and richer supervenience basis. 8 This 'supplement' to the claim of aesthetic supervenience also solves a problem raised by Scruton, who criticizes the supervenience thesis when he says that 'different emergent "properties" can depend on precisely the same set of "first order" properties' . 9 What he has in mind here is that one and the same artwork can be context-dependently characterized as sad or joyful, without contradiction. 10 Here again, the reply to such an objection is that once we include the context of production and the context of evaluation in the supervenience basis, it is not the case that 'different emergent properties could arise from the same supervenience basis' . The notion of 'context of evaluation' corresponds here to the Humean idea of taste: not only are the history and the context of the production of an artwork relevant to the attribution of aesthetic properties to it, but so is the individual taste of the evaluator. 'Taste' means not simply a 'liking' , but a more elaborate capacity of the evaluator (one that can be trained), as, for instance, Sibley (as well as Hume himself ) 11 makes clear: 'When I speak of taste […], I shall not be dealing with questions which center upon expressions like "a matter of taste" (meaning, roughly, a matter of personal preference or liking). It is with an ability to notice or discern things that I am concerned. ' 12 What this means is that aesthetic properties are response-dependent. Their instantiation depends not only on features of the objects themselves but also on the responses they bring about in (trained, competent, serious) evaluators.
Aesthetic Supervenience versus Aesthetic Grounding 7 It is less clear that a similar claim can be reasonably articulated in the case of the beauty of natural objects, like sunrises, since they are not human-created artefacts. Difference in context will there rather be a difference in the perceiver: think of a sunrise observed comfortably from your living-room and the same sunrise observed from the summit of the Matterhorn after long hours of hard climbing. (On this point, see the discussion on 'taste' later in this essay.) 8 Zangwill, Metaphysics of Beauty.  Miller, rev. ed. (Indianapolis: Liberty Fund, 1987), 226-49. Hume defends the claim that not everyone is a good critic of art, and adds (pp. 240-41) that the 'qualified judges' must also obey some additional constraints such as having some practice in the attribution of aesthetic properties to the type of objects they evaluate, having 'good sense' , and being intellectually honest (for instance, by trying not to be jealous of, or sympathic to, the maker of the object they are evaluating). 12 Sibley, 'Aesthetic Concepts' , 423.
If one thus accepts enriching the supervenience basis with the context and history of production, as well as the evaluator's disposition to certain responses, the concerns about the claim that aesthetic properties supervene on non-aesthetic ones dissipate. Now that we have seen how the claim of aesthetic supervenience can be accurately, completely, articulated, let us see where the trouble really lies.

III
In short, the trouble is that supervenience is not a relation suitable to play the theoretical role we want it to play here. Some think this is unimportant, that it is enough to have a loose, imprecise, intuitive understanding of the relationship between aesthetic and non-aesthetic properties in order to defend a claim of 'aesthetic supervenience' . This is the case, for instance, for Zangwill when he says: A fundamental principle is that aesthetic properties are determined by or are dependent on nonaesthetic properties. Things come to have aesthetic properties because of or in virtue of their nonaesthetic properties. For example, a performance of a piece of music is delicate because of a certain arrangement of sounds, and an abstract painting is brash or beautiful because of a certain spatial arrangement of colors. In the philosophical jargon, aesthetic properties supervene on nonaesthetic properties. This means that if something has an aesthetic property then it has some nonaesthetic property that is sufficient for the aesthetic property. (The relation of dependence or supervenience is a general one. I shall not probe the exact nature of the relation, although it can be formulated in different ways. The notion is important outside of aesthetics, in areas like moral philosophy and the philosophy of mind.) 13 Zangwill seems here to be well aware that there are many different types of relations in the neighbourhood of supervenience, and he only appears to think that it is not important for the aesthetician to decide which one best suits the relationship between aesthetic and non-aesthetic properties. 14 He thus speaks about this relationship as being one of supervenience, 15 or one of 'metaphysical determination' , 16 dependence, or explanation.
My intent here is not to criticize Zangwill's view (which, again, I find on the whole philosophically appealing and intuitively adequate), but rather to state my concern that not enough attention has been paid in the literature to the exact nature of the relation, often called 'supervenience' , between aesthetic and nonaesthetic properties of an object. Levinson, for instance, also says interchangeably that aesthetic properties supervene on non-aesthetic ones, or that aesthetic 13 Zangwill, 'Beauty' , § 2, italics in the original. differences are grounded in non-aesthetic differences, or that non-aesthetic properties are responsible for aesthetic ones. 17 But all of these relations are different, they exhibit different formal features, and they are suitable to playing different, incompatible, theoretical roles. Thus, in order simply to understand what the claim of 'aesthetic supervenience' amounts to, we have to understand not only the nature of the supervenience basis (which has been discussed at length and in detail in the relevant literature) but also the nature of the relation between such a basis and the properties that are said to arise from it. Let us start by seeing why supervenience cannot do the job.

IV
Recall that supervenience is typically defined as follows: 'A set of properties A supervenes upon another set B just in case no two things can differ with respect to A-properties without also differing with respect to their B-properties. In slogan form, "there cannot be an A-difference without a B-difference".' 18 One problem that arises when we attempt to apply this notion of supervenience to the relationship between, say, the beauty of Dalí's Temptation of St Anthony and its non-aesthetic (intrinsic and relational) properties is that many other non-aesthetic properties of Dalí's painting also supervene on the very same properties (the very same 'supervenience basis'), with the same sort of necessity, and consequently the claim of aesthetic supervenience fails to single out what's special about the way the aesthetic properties of the painting arise from their supervenience basis. This is simply because, given the nature of the relation of supervenience, necessary properties of any object supervene trivially on any properties it has. 19 Stock examples include properties such as being self-identical, which any object necessarily has, and since it has them necessarily it cannot differ with respect to them, and so cannot differ with respect to them without differing with respect to any other property the object has. It is impossible for Dalí's Temptation of St Anthony not to be self-identical, and so it is impossible for this painting not to be self-identical without, say, featuring five elephants. Thus the property of being self-identical supervenes on the property of featuring five elephants, or any other of the painting's properties and, relevantly to our present concerns, it supervenes on the very same set of properties (including paint distribution, canvas size, history and context of production) on which the beauty of The Temptation of St Anthony supervenes. Since the very same relation of supervenience thus holds between aesthetic properties and their supervenience basis, and between many other (necessary) properties of the same object and the very same supervenience basis, the claim of aesthetic supervenience becomes trivial and uninformative when one tries to understand the nature of the relationship between aesthetic and non-aesthetic properties. To make it perhaps clearer: the trouble here is that we want to say that the painting has its aesthetic properties in virtue of having its non-aesthetic properties, but saying that this is so because its aesthetic properties supervene on its non-aesthetic ones is of no help at all, since supervenience holds in a trivial way even between totally unrelated properties, such as the property of being self-identical and the property of featuring five elephants. Note, as an aside, that this problem with necessary properties extends to having 'impossible properties' as well, since no object can differ with respect to properties that no object can possibly have, like the property of being a talking donkey and not being a talking donkey, and so no object can differ with respect to (not) having such properties without differing with respect to (any) other properties -thus not having the property of being a talking donkey and not being a talking donkey, which Dalí's painting (as any object) trivially does not have, supervenes on the very same properties (paint distribution, canvas size, history, and context of production) on which the painting's beauty supervenes.
V A second type of concern arises with respect to some formal features of the relation of supervenience. When saying that aesthetic properties supervene on non-aesthetic properties, what we mean is that aesthetic properties arise from non-aesthetic ones (the supervenience basis) in the sense that the latter are somehow prior to the former. is prior to another, then the latter cannot be prior to the former (nothing can be prior to anything that is prior to it). Supervenience, however, is neither symmetric nor asymmetric. It can be symmetric, since any instance of reflexive supervenience is trivially symmetric, and, more controversially, it can also be said to be asymmetric, for example, if one thinks that mental properties supervene on physical properties, and that mental properties are multiply realizable, which means that physical properties do not supervene on mental properties.

VI
Far from being a logician's or metaphysician's quibble, which would be unimportant for the aesthetician, the technical concerns discussed in the last two sections show us that something more than 'mere supervenience' is required to help us understand the nature of the relation between aesthetic and non-aesthetic properties. Indeed, supervenience is simply not the right kind of relation to play the role we need it to play. As McLaughlin rightly notes, a supervenience claim does not automatically entail an 'in virtue of' claim. 22 As we have seen in our discussion so far, supervenience is not a relation of priority, and it simply does not say anything genuinely informative about the relationship between aesthetic and non-aesthetic properties -it does not say that an object has its aesthetic properties in virtue of or because of its non-aesthetic ones, since it is generally fair to say that a set of properties of x supervenes on another set of properties of x while x does not have the former in virtue of having the latter. (Trivially, for any object, and for any property F, x's having F supervenes on its having non-F since an object cannot differ with respect to F without also differing with respect to non-F, but clearly an object is not said to have F in virtue of having non-F.) We need more than just supervenience. We need a relation that links aesthetic and non-aesthetic properties of an object in a more intimate manner -in a way that makes sense of Dalí's Temptation of St Anthony's having such and such aesthetic properties because it has such and such non-aesthetic (intrinsic and relational) properties. Its having these non-aesthetic properties must somehow explain that it has the aesthetic ones. Supervenience is merely a form of covariation of (sets of ) properties, 23 but lacks any element of explanation about why such covariations occur. As Kim aptly puts it: As Correia puts it: 'The concept of supervenience involves the notion of "covariation": that which supervenes (the set of supervenient properties) "covaries" with that on which it supervenes (the set of subvenient properties), i.e. there can be no "variation in" the supervenient properties without some "variation in" the subvenient properties. ' Correia, 'Ontological Dependence' , 1028. when we look at the relationship as specified in the definition between a strongly supervenient property and its base property, all that we have is that the base property entails the supervenient property. This alone does not warrant us to say that the supervening property is dependent on, or determined by, the base, or that an object has the supervening property in virtue of having the base property. 24 Thus, the thesis of aesthetic supervenience, when understood as a thesis of aesthetic supervenience, is rather frustrating and uninteresting. Supervenience is not a relation of priority; it is merely covariation; it does not have the right formal features, and it cannot do the job that we need it to do. But grounding can.

VII
Aesthetic grounding is what we need instead of aesthetic supervenience. The relation of grounding, increasingly popular in metaphysics, is also one that we can appeal to in order to understand the relationship between aesthetic and nonaesthetic properties.
Grounding cannot be analyzed successfully in terms of supervenience or in other terms and it is best taken as being a primitive relation. 25  metaphysical claims such as 'the temporal extension of a material object is grounded in its temporal parts' or 'a trope is grounded in its bearer' , 27 but also claims like 'moral features are grounded in natural features of a situation' , 28 and, importantly for us, claims like 'aesthetic facts are grounded in non-aesthetic facts' , which Audi cites as being 'among the most compelling examples' of grounding. 29 One can directly see that the relation of grounding does not have the frustrating features of the relation of supervenience. If a set of properties A is grounded in a set of properties B, and if B is more fundamental than A, then we not only secure the claim that there cannot be an A-difference without a B-difference, but we also have an answer to why an object has A when it has B. That is, grounding is a relation of priority and ordering, it is irreflexive and asymmetrical, it does not yield the unwelcome result that necessary properties are grounded in anything, and it does not force us to say that being F is grounded in being non-F. This is how we want the relationship between aesthetic and non-aesthetic properties of Dalí's Temptation of St Anthony to be. Indeed, while the grounding relation works with the same 'base properties' of the painting as the supervenience relation (including the broad non-aesthetic relational properties, like the process and history of the production of an artwork as well as the context in which it was created), 30 it does not amount to a mere covariation between the higher-order properties and the 'base' ones. The concern here is not only to use proper terminology -and to stop using 'supervenience' as a term for talking about the relation between aesthetic and non-aesthetic properties -but to make it clear, In one sense, these concerns cannot be dispelled. Indeed, if grounding really is a primitively postulated tool, introduced by the philosopher who realizes that supervenience cannot do the job and that something else must do it, then it is in some sense impossible to ask for more: one cannot dig deeper than a 'primitive' .
Primitives are the pillars that sustain the architecture of our philosophical theories, and it is in some sense unfair to ask one's opponent to go beyond what she takes to be a non-analyzable central feature of her view. But then, the concern becomes different and perhaps even more pressing. Indeed, the theoretical job of the grounding view, as proposed above, is achieved through a primitive. (This is actually also true of the supervenience view, as well as for many other Why does lightning occur just when there is an electric discharge between clouds or between clouds and the ground? Because lightning simply is an electric discharge involving clouds and the ground. There is here only one phenomenon, not two that are correlated with each other; and what we thought were distinct correlated phenomena turn out to be one and the same. Here the apparent correlation is understood as identity. 31 The relation between the explanandum and the explanans is thus here simply identity. But not all identities are explanatory, as Ruben has rightly remarked. 32 For instance, the identity 'lightning = lightning' is not explanatory, whereas the identity 'lightning = atmospheric electric discharge' is, because, as Ruben points out, even if there is only one phenomenon involved in the latter identity, it is conceptualized in two different ways. What we learn here is that explanation, unlike identity, is like grounding, an irreflexive relation. But, it is also reasonable to argue that even the identity 'lightning = atmospheric electric discharge' is not explanatory. The concern is that if the relation between the explanandum and the explanans is identity, how does this explain anything? Not surprisingly, this concern is 31 Jaegwon Kim, Philosophy of Mind, 2nd ed. (Boulder, CO: Westview, 2006), 85. 32 David-Hillel Ruben, Explaining Explanation (London: Routledge, 1990), 219. familiar also from the discussion of the mind-body problem; for instance, when discussing the psychoneural identity theory, Kim says: Our conclusion, therefore, has to be that both forms of the explanatory argument are open to serious difficulties. Their fundamental weakness lies in a problematic understanding of the role of identities in explanation, an important topic that has not received much attention in the literature. The only clear (and also simple) view is that identities function simply as rewrite rules in explanatory derivations -or any derivation, for that matter.
[…] We do not have to say that identities have no role to play in explanations. For they can help justify explanatory claims -the claim that we have explained something.
[…] It is only that identities do not generate explanations on their own. 33 The answer to this concern is, I think, that the relation of explanation is a lot like identity but is not identity. We have already seen that the relation of explanation is irreflexive, but it is also asymmetrical. Yes, lightning is atmospheric electric discharge, but the explanation does not consist just in pointing to this fact; it also points to the fact that the explanans is more fundamental than the explanandum, and this is what we gain, this is what we learn from a good explanation.
Explanations of this type are such that one of the two sides of the explanation relation is more fundamental than the other (and thus, explanation is not only irreflexive but also asymmetrical). And this is exactly how our relation of grounding works, and how it plays its useful theoretical role -even if it is a primitive relation, it has an explanatory role. The relation of grounding, or at least the way it fulfils its theoretical role, then becomes clearer and better understood, since grounding provides us with a kind of explanation that often goes under the name 'in virtue of' . Grounding is thus a relation of fundamentality and it is an explanatory relation: when a is grounded in b, then being a b explains being an a. As Fine, for example, puts it: 'We take ground to be an explanatory relation: if the truth that P is grounded in other truths, then they account for its truth; P's being the case holds in virtue of the other truths' being the case.' 34 Grounding then, unlike supervenience, is not just covariation. Thus if we have in mind a thesis of aesthetic grounding, rather than aesthetic supervenience, we can say that an object has the aesthetic properties it has in virtue of having its non-aesthetic properties, and this is what we wanted to say all along.
We encounter similar situations in very many cases of the workings of our philosophical theories. For example, consider exemplification: the explanandum is the sharing by two objects of the same property, the explanans is, say, the instantiating of the same immanent universal. It does, then, seem like the right thing to say, if you are an advocate of the universals view, that sharing the same property just is instantiating the same universal (comparably to the way lightning just is atmospheric electric discharge) where the latter is taken to be more fundamental than the former. Or, consider the debate about persistence through time: a's persisting through time is a's having temporal parts at different times (if you are an advocate of temporal parts) where again the latter is taken to be a more fundamental phenomenon than the former -one usually says: 'a persists through time in virtue of having temporal parts at different times' .
Granted, both supervenience and grounding (as well as many other primitively postulated relations, like exemplification) are primitives, and in a sense we then, frustratingly, cannot go deeper than them. But this does not mean that we cannot understand the ways in which they play their theoretical role. Both candidates for being the best relation that obtains between non-aesthetic and aesthetic properties of an object are primitive 'we-know-not-what' but they are 'we-know-what-it-does' , and the purpose of this article has been to establish that grounding is a more serviceable hypothesis than supervenience. It has the right formal features, it is a relation of explanation, or priority, of fundamentality -it is an 'in virtue of' relation known to be a useful hypothesis in many other areas of philosophy, and consequently it does a better job than supervenience in the case of aesthetic properties. It does not tell us much about itself, but it does tell us a bit about its relata: as Bricker, deRosset, and Schaffer claim, if a is grounded in b, a is nothing over and above b. 35 a, in other words, is an 'ontological free lunch' in Armstrong's sense; 36 the 'ontological price' , to use Schaffer's term, 37 you pay for a and b is just whatever you would pay for b alone. Only in this sense can one talk about identity between a and b. This does not mean that aesthetic properties are not real, no more and no less than lightning or two objects' sharing the same property or an object's persisting through time. Aesthetic properties, on this view, are as real as the non-aesthetic ones they are grounded in; it is just that the latter are the more fundamental ones.