In Defence of the Acquaintance Principle in Aesthetics

Abstract

Making an adequate aesthetic judgment about an object or an aesthetic property requires first-hand experience of that object or property. Many have suggested that this principle is a valid epistemic norm in the epistemology of the aesthetic. However, some recent philosophers have argued that certain works of conceptual art and other counterexamples disprove the principle in question, even suitably modified. In this paper, I argue that these philosophers are mistaken and that, when properly qualified, the acquaintance principle (in some of its versions) is not threatened by their examples and arguments.

1. Introduction

Recent versions of the acquaintance principle (also sometimes called ‘The Principle of Acquaintance’, here shortened as ‘AP’) place a necessary condition on aesthetic judgments: experiencing an object or a property is a requirement for making appropriate aesthetic judgments about it.¹ This principle has been variously interpreted as articulating an epistemic, a non-epistemic, or even a peculiarly aesthetic requirement or norm – that is, as stating (i) an essential requirement for making epistemically adequate aesthetic judgments, (ii) a non-epistemic norm (e.g., a norm of use or even a moral norm) for making aesthetic judgments, or (iii) a norm that characterises aesthetic judgments as such.² Besides, the AP has been interpreted not only as stating different types of normative requirements for making aesthetic judgments but also as applying to different types of things classified as aesthetic – for example, as a requirement for having an aesthetic experience, for having aesthetic knowledge or for appreciating the aesthetic value of an object (Sauchelli 2016: 5). In what follows, I will focus on versions of the AP intended as epistemic norms regulating aesthetic judgments, but will also refer to aesthetic knowledge and understanding.³

In the contemporary debate, the AP has seldom, if ever, been defended in an unqualified form. Even Richard Wollheim's original formulation (which concerned judgments of aesthetic value) admits of exceptions (1980: 233). Subsequent theorists – notably, Paisley Livingston – have worked out such exceptions in additional detail. In particular, Livingston has pointed out that experiencing some (proper) aesthetic surrogates of the object we intend to judge aesthetically may suffice for making legitimate aesthetic judgments about it. For example, experiencing a photograph may suffice for making legitimate judgments about at least some of the visual aesthetic properties of the photographed object.

Updated versions of the AP would then include a clause to the effect that experiencing aesthetic surrogates of an object O (broadly understood) can make it ‘epistemologically safe’ to make adequate aesthetic judgments about O – or, more generally, can contribute to having justified aesthetic beliefs about O.⁴ However, recent philosophers have advanced several objections even to these versions of the AP. Some of these objections are based on alleged counterexamples deriving from conceptual art (e.g., Hanson 2015) and from cases of aesthetic testimony using descriptions (e.g., Robson 2018). For instance, some have claimed that conceptual art such as Duchamp's Fountain or certain ‘dematerialised’ works would show that experience is not required to make an aesthetic judgment in every case.⁵

In this paper, I distinguish two versions of the AP and clarify some other theoretical preliminaries required to understand the contemporary debate on the AP (Section 2). I then argue that the above recent criticisms fail (Sections 3 and 4). More specifically, in Section 3, I attempt to show that Louise Hanson's recent criticism of the AP based on examples coming from conceptual art fails in showing that the AP is false. In Section 4, I discuss Jon Robson's criticisms based on forms of aesthetic testimony. My conclusion will be that at least one version of the AP is not proven false even by his challenging arguments.

2. Theoretical preliminaries and two versions of the AP

Little has been said by contemporary aestheticians debating the AP to clarify what they mean by ‘acquaintance’ or ‘first-hand experience’.⁶ The term ‘acquaintance’, popularised by Bertrand Russell, has been at the centre of new interest in contemporary epistemology and philosophy of mind. As Thomas Raleigh has pointed out, ‘acquaintance’ has been used to refer to an epistemologically special state or relation or a peculiar feature of the way we are related to certain aspects of our minds (Raleigh 2019). Unlike Russell, contemporary aestheticians do not seem to understand acquaintance as a special epistemological relation we can have only with sense data, mental states, and perhaps universals. Nor are defenders of the AP willing to be committed to the idea that, in general, knowledge by acquaintance is more fundamental than knowledge by description. Further, most contemporary aestheticians do not understand ‘acquaintance’ as equivalent to sensory perception, if only because they seem to allow that we can be acquainted with literary works of art (which some understand as created and repeatable abstract types) or other abstract structures (e.g., mathematical structures or theorems).⁷ However, more positive characterisations of what is meant by ‘acquaintance’ are lacking in the contemporary literature on the AP. In what follows, the AP will be understood as involving first-hand experience (I will drop ‘first-hand’ in the rest of the paper), where this experience is not inferential or derived from other people's testimony, although beliefs based on testimony (e.g., art-historical beliefs acquired through testimony) may properly influence the experience itself.⁸

In addition, I distinguish two versions of the AP:

• Object-acquaintance: experiencing an object, O, or its proper aesthetic surrogates, S₁, …, S_n, is required to make proper aesthetic judgments about O.

• Property-acquaintance: experiencing an aesthetic property, P, or its proper aesthetic surrogates, S₁, …, S_n, is required to make proper aesthetic judgments involving P.

Although related, these two versions of the AP might come apart on some understandings of the notions involved (e.g., depending on how demanding we regard being acquainted with an object is). These principles could be made more precise; for example, by including answers to these questions: How long and detailed should a relevant experience be? Are these conditions meant to apply to all aesthetic judgments, or do they involve only judgments concerning, say, determinable or determinate aesthetic properties?⁹ In what follows, I will assume that the kind of relevant experience at issue is at least an experience in the correct representational format for the kind of aesthetic belief it is supposed to justify – for example, if we are to make aesthetic judgments about some visual aesthetic properties, the experience at issue must be a visual experience, and so on. Intuitively, tapping on a painting in a museum while blindfolded may count, in a broad sense, as being acquainted with the work, but not as having the right kind of experience for making aesthetic judgments about the painting's visual aesthetic properties. From the fact that someone is, in a broad sense, object-acquainted with O, it may not follow that, for all of O's aesthetic properties, they are also thereby property-acquainted with O – and thus be in an adequate epistemic state to make judgments about all of O's aesthetic properties.

What is, more precisely, an aesthetic surrogate for objects and properties? I will follow Livingston's definition of an aesthetic surrogate for objects:

some object, O₂, is an adequate aesthetic surrogate of another item, O₁, just in case O₂ directly presents enough aesthetic features qualitatively identical to the relevant features of O₁; or, where O₂ is a representation or depiction of O₁, enough of the relevant features of O₁ are observable in O₂ (Livingston 2003: 264).

In the case of properties, an aesthetic property P is an adequate aesthetic surrogate of another property Q just in case P and Q share a sufficient number of qualitative features. For example, where P is ‘being structurally balanced in the Titian's Diana and Actaeon way’, and Q is ‘being structurally balanced in the Titian's Diana and Callisto way’ P can be an adequate aesthetic surrogate of Q: the way in which Titian achieves balance in the representations of the mythical figures in the painting (e.g., through a horizontal disposition of twisting and twirling bodies showing dramatic tension caused by an event on the left side of the canvas) in Diana and Actaeon may give us a reasonably good sense of the way in which Titian achieves balance in Diana and Callisto.

Following James Shelley, we may even regard the relation of acquaintance as gradual: on this view, experiences can be more or less acquainting with properties or objects (Shelley 2023: 4). For example, we may argue that experiencing Diana and Actaeon can acquaint us to a reasonable degree with the way in which Titian achieves balance in Diana and Callisto, although an experience of the former may not always be sufficiently acquainting for more specific judgments or beliefs concerning the latter. In general, even if a brief encounter with an O under poor visibility conditions may count as an experience of O (albeit a not very ‘acquainting’ one), it can hardly count as the kind of experience that can justify proper judgments about O's visual aesthetic properties. Similar considerations apply to other sense modalities and related experiences.¹⁰

These two versions of the AP are meant to capture necessary, not sufficient, epistemic conditions for making legitimate aesthetic judgments. For the purposes of this paper, it is not necessary to specify what provides sufficient justification for an aesthetic judgment. In addition, both versions of the AP are understood primarily as regimenting aesthetic beliefs and judgments about objects or properties. I shall remain neutral regarding the plausibility of analogous higher-order versions of the AP (e.g., versions of the AP that regiment the normative significance of one's evidence).¹¹

As discussed in more detail in Section 4, distinguishing these two versions is helpful also to bring out one peculiarity of aesthetic concepts and beliefs. In addition, one of the conclusions of this paper, is that Robson's arguments may threaten at most object-acquaintance (depending on how we understand the notion of being acquainted with an object), but not property-acquaintance.

3. Hanson's criticisms

In the next two subsections, I examine Hanson's criticisms of the AP. Although connected, they are divided into two categories: arguments related to the ontology of conceptual art (Section 3.1), and criticisms focused on the notion of an aesthetic surrogate (Section 3.2).

3.1. Hanson on the AP and the ontology of conceptual art

The appreciation of some works of conceptual art has been proposed as a counterexample to the AP (Sauchelli 2016: 8–9). In reply, a distinction between the aesthetic and the artistic has been deployed to defend the AP against these cases. Hanson considers and seeks to disarm this defence (Hanson 2015: 248–49).¹² More specifically, the defence runs as follows: the AP is a principle governing the aesthetic, not necessarily the artistic; it applies to the artistic at most insofar as it applies to the aesthetic. For those artistic judgments about conceptual works of art that do not depend upon their aesthetic properties the AP may simply be irrelevant. After all, the principle does not state or presuppose that all art must have aesthetically relevant properties. Rather, the AP claims only that insofar as we want to make a proper aesthetic judgment about any object, we should at least experience it, its properties, or its aesthetic surrogates. So, the AP is compatible with the idea that some works of art may have artistic properties without thereby having artistically relevant aesthetic properties. It follows that, even if it is true that we can ascribe certain artistic properties to some works of conceptual art on the basis of a description or testimony without experiencing them for ourselves, this does not necessarily show that the AP is false.

In reply, Hanson claims that (i) there are people who support an understanding of ‘artistic’ as equivalent to ‘aesthetic’, and (ii) there is an interesting version of the AP which is taken to regulate ‘art-critical judgments’. In addition, she claims that:

Even if this usage of the term ‘aesthetic’ turns out not to be correct, the fact still remains that many who subscribe to AP in some form are subscribing to it as a principle governing the formation of art-critical judgments. Even if it is wrong to use the word ‘aesthetic’ to articulate such a principle, the fact remains that the principle is widely held, and so if conceptual art really does present an example of works whose artistic properties we can judge without being acquainted with them, this is an interesting finding. (Hanson 2015: 248)

Roughly for these reasons, Hanson focuses on a version of the AP governing ‘how judgments of artistic value and artistic properties may legitimately be formed’ (Hanson 2015: 248). Shortly after this passage, she further clarifies that the claim she intends to defend is that ‘one can be in a position to form judgments about the artistic value and artistic properties of works of conceptual art on the basis of a description’ (Hanson 2015: 249).

However, this attempt to set aside disputes about whether the artistic and the aesthetic are equivalent or whether they only partly overlap has decisive consequences for the plausibility of the AP. First, if the artistic and the aesthetic only partly overlap and are not equivalent (which is the view of most contemporary philosophers of art), the version of the AP Hanson intends to focus on is not particularly interesting. Consider this statement: ‘Roman Opalka's and On Kawara's treatments of the phenomenon of time seem to be closely linked with their own biographies. For the conception of his series of Infinity Paintings, Opalka used the simple principle of progression. In 1965 he completed the first picture, which bore the title 1965/1 – ∞ (Detail 1–35327), which begins in the top left with the white figure 1 on a black background […]’ (Marzona 2005: 20). Or consider the statement: ‘Duchamp's Fountain is a seminal work in the history of contemporary art.’ These are examples of descriptions, art-historical, or art-critical statements in which artistic but not aesthetic properties are ascribed to some works of conceptual art. Now, supporters of the AP would agree that we can make proper artistic judgments of the previous kind solely on the basis of, say, descriptions contained in respectable books of art criticism without being acquainted with the works themselves. To reiterate, few, if any would dispute that I may learn that Duchamp's Fountain or Leonardo's Mona Lisa are seminal works in the history of art without being acquainted with the works themselves. A version of the AP suggesting otherwise would be so implausible that it would not even be worth defending.

Fortunately, Hanson's criticisms can also be applied to a more plausible version of the principle (i.e., a version that focuses only on the aesthetic). In particular, she suggests that one way of countering the claim that works of conceptual art constitute counterexamples to the AP is to hold that at least the troublesome ones (e.g., works that seem to be fully ‘dematerialised’, i.e. works that allegedly have no material parts) are identical with ideas. The reasoning supporting this defence is that if ideas, which is what these troublesome works would be, take a linguistic form, there can be accurate paraphrases of these ideas. Such paraphrases may count not only as descriptions of the relevant ideas but also, plausibly, as their aesthetic surrogates – provided that these surrogates satisfy certain requirements (such as being of roughly the same length, reflecting the tone and style of the original, and so on). If a work is an idea, and if a paraphrase of such an idea counts as both a description of it and an (aesthetic/artistic) surrogate for it, then, in those cases in which we are acquainted with a proper paraphrase and that seem to constitute a counterexample to the AP, we do after all seem to be acquainted with the work or, at least, with a proper surrogate of it. And, crucially, this satisfies the necessary condition imposed by the AP. As Hanson puts it: ‘[d]escriptions of conceptual works [understood as ideas], then, are not necessarily an alternative to acquaintance with these works after all, but may rather be a means of achieving it’ (Hanson 2015: 249). In reply to this argument, Hanson claims that (i) there are good reasons to deny that works of conceptual art are ideas and (ii) descriptions cannot be proper aesthetic surrogates. In the rest of this subsection, I will focus on the first, whereas the second criticism will be addressed in Section 3.2.

Now, it is not my purpose to defend a specific view on the ontology of conceptual art, although I find a pluralistic approach plausible (e.g., Dodd 2016; Irvin 2009; Irvin and Dodd 2017; Thomasson 2010). On this approach, works of conceptual art need not all belong to the same ontological category. Instead of arguing that all works of conceptual art belong to the same ontological category (e.g., ideas), the pluralist suggests that works of conceptual art belong to different ontological categories. For example, some works may be compounds of a particular object and an idea, and other merely ideas. In short, on this approach, ‘conceptual art’ classifies an ontologically heterogenous set of entities (Dodd 2016). Still, discussing Hanson's arguments against the work = idea thesis (WIT) is valuable also because it sheds light on a series of tangled issues at the intersection between the appreciation of works of conceptual art and the AP.¹³ So, although I reject Hanson's arguments against the WIT, this does not mean that I also claim that this thesis is true for all works of conceptual art.¹⁴

Hanson's first argument is based on metaphysical considerations: ‘the persistence conditions of the artworks look to be exactly those of the object. But if the work is the idea, and the object is just a documentation of it, then the destruction of the object would leave the work intact’. However, Hanson claims, this is counterintuitive (Hanson 2015: 249). On one understanding of the quoted passage, the argument does not seem to be valid. In fact, Hanson seems to be claiming that, where ‘object’ means a physical object or something that can be perceived with the senses, the following holds:

(1.) work = object

(2.) work = idea

(3.) object = documentation

So, (4.) it is possible to have a situation in which we have the work but not the documentation, and (5.) this conclusion is counterintuitive and/or implausible. However, the alleged problematic conclusion does not follow from the premises of her argument. What follows is that it is not possible to have the work without the documentation since from (1.), (2.), and (3.) it follows (by the transitivity of identity) that the work = the documentation. On a different understanding of her argument, Hanson would be claiming that if a work of conceptual art is an idea and the object is merely a documentation and not the work itself, it is possible to have a situation in which we can have the work without the object. However, she may be arguing that this is not possible since the identity conditions of the work coincide with those of the object. Thus, a work is not an idea. This argument does not seem very persuasive. In fact, a crucial premise (that the identity conditions of the work coincide with those of the object) can be one direct way of denying the main tenet of the WIT – on the assumption that the WIT is incompatible with the claim that an idea is an object. Leaving aside this last point, also the conditional (‘if a work is an idea and the object is only a documentation and not the work itself, it is possible to have a situation in which we can have the work without the object’) may turn out to be false in its present form. For example, the WIT supporter can adopt ontology of ideas according to which ideas, although ontologically dependent on documentations or other forms of material embodiment (e.g., speech acts or neuronal activities) for their existence, are not identical with them.¹⁵ On this view, similar to an Aristotelian universal, an idea would require for its existence at least an instantiation. So, on this alternative view, it would not be metaphysically possible to have an idea without its proper material realisation(s) or manifestation. If this is correct, Hanson's argument seems to be at best underspecified with respect to the motivation for accepting ontology of ideas as entities that can exist independently of their instances. To the extent that the supporter of the WIT finds problematic the claim that a work can continue to exist without a corresponding object or documentation, they may adopt a different and more amenable ontology of ideas to account for this intuition.

The second objection against the WIT we will consider is: ‘if the object is just a documentation and is not the work, we have no resources to explain why it has any different status than photographic reproductions of it’ (Hanson 2015: 249). If the object is not the work, it may turn out that our practices of buying, selling, and going to (dematerialised) conceptual art exhibitions are irrational or in need of an explanation. This argument is enthymematic, but presumably what is meant here is that if (1) the work is an idea and the object is only (at most) its documentation and (2) the documentation of the work is not an idea, then (3) the object is not the idea and therefore not the work. Why, then, would it be rational to engage in the current practices that seem to be focused on objects? According to Hanson, the best (or only) explanation of why we rationally care is that the object is the documentation and, crucially, the work. This argument does not seem persuasive either.

First, the vehicular media of many conceptual works of art have been destroyed and artists, for example, simply replaced them or even made multiple instances of some of their works. In the case of Duchamp's Bottle Rack, this would not have been possible if we metaphysically identify Bottle Rack (the work) with the first object that ‘embodied’ this work, namely, the specific bottle rack that Duchamp allegedly bought at a department store in Paris. Similarly, the original urinal used for Fountain has been lost or destroyed. I find it intuitive to claim that, on the supposition that no new significative creative artistic acts were required to instantiate again the relevant work-idea, Duchamp did not create an entirely new work of conceptual art each time he replaced the lost or destroyed objects with similar material objects (as he would have had the work been identical with the object).¹⁶ A plausible account of this case is that Duchamp made or commissioned different versions of the same work or that he instantiated different tokens of the same (created or ontologically dependent) abstract kind by indicating (e.g., through the act of signing) different vehicular media (i.e., numerically different urinals).¹⁷

Alternatively, a defender of the AP may slightly modify the WIT and claim that works such as Bottle Rack or Fountain should be understood as compounds of indexed ideas-cum-object-types. On this view, it is not the original object (the particular object that was indicated by the artist) that is necessary for the identity and persistence of a work of this kind, but rather its originating idea and the type of object(s) indicated by the artist. After all, not any object-type would successfully refer to a fountain in the same ironic way that distinguishes Fountain. A defender of the AP would then claim that a proper aesthetic judgement about this kind of work should involve an experience of a proper instance of the relevant object-type. Of course, this modified version of the WIT would also easily explain why caring about the work may make it rational for us to care about the object as well: the object-type is part of the work and the object is one of its instances. So, attending to the object is one way of attending to the work.

The main point of the third argument we will consider here is that if the artist had chosen to document the same idea in a different way, this would have affected the artistic value of the work, and the WIT would seem to preclude us from appreciating the import of such differences. Hanson seems to suggest that the WIT would misrepresent or not allow the appreciation of the ‘more intimate link between the artistic value of the work and the nature of the object’ (Hanson 2015: 250). I agree with the rationale behind this argument; still, the WIT seems compatible with its main motivating intuition. For example, the supporter of the WIT may claim that ideas, understood as ‘concepts’ or ‘conceptions’, essentially have a ‘mode of presentation’, or a way of appearing (whether intellectually or sensorily).¹⁸ They could then argue that it is an idea as presented in a certain way – under a certain mode of presentation – that is identical with the work. The documentation of a work of conceptual art may then be or stand for a way in which the idea is presented and thus be relevant to the value of the work, provided that they are in the same representational format. So, to the extent that the documentation is important for appreciating a work, such an importance can be explained by the supporter of the WIT by claiming that appreciating it is the most appropriate means through which the relevant mode of presentation of the idea can be apprehended and appreciated. In conclusion, the supporter of the WIT can successfully reply also to this criticism. If the supporter of the AP can adopt ontology of conceptual art compatible with the claim that some paraphrases can be proper aesthetic surrogates, they can claim that experiencing appropriate definitions or paraphrases of works of conceptual art – including the ‘dematerialised’ ones – would suffice for satisfying the AP, since these works are, at least partly, ideas.

3.2. Hanson on aesthetic surrogates

Can descriptions (qua linguistic representations) serve as aesthetic surrogates and, if so, would this be a problem for the AP? Below I argue that (i) Hanson fails to rule out descriptions as aesthetic surrogates and (ii) the fact that some aesthetic judgements about some works of art (e.g., works of conceptual art mostly in a linguistic representational format) can be justified on the basis of an appropriate description of a work poses no threat to the AP. The discussion of these arguments will start from the important concept of an aesthetic surrogate for objects.

Hanson's account of what counts as an adequate surrogate elaborates ideas originally discussed in Livingston (2003) but contains a misunderstanding of what Livingston means when he discusses forgeries (Livingston 2003: 263). To illustrate what this misunderstanding is and how it affects Hanson's discussion of aesthetic surrogates, I will briefly outline some of Livingston's main points in his seminal paper on the topic. First, Livingston claims that (i) perceptual equivalence is not necessary for being an aesthetic surrogate, and (ii) perceptual indiscernibility may not be a sufficient condition for X to be a proper surrogate for Y because X may be visually indistinguishable from Y and yet bear no relevant causal/historical connections with Y (as two qualitatively identical pebbles may be indistinguishable but located on different beaches and produced by largely independent geological processes). Presumably, what Livingston means here is that, for any numerically non-identical but indistinguishable X and Y, given that X and Y could have differed in all their aesthetic properties without affecting each other in any way, to deploy either one in an aesthetic evaluation of the other (or make any aesthetic inference from the one to the other) without further qualifications or conditions would fail to meet certain basic norms of epistemic safety. This point does not exclude that one could become an aesthetic surrogate for the other.¹⁹ Also, in cases in which certain works are not fortuitous copies of others – cases in which some sort of epistemic safety condition is met – Livingston seems to have to claim that a (good, causally connected, etc.) forgery could be an adequate aesthetic surrogate for the original since the copy would presumably present ‘enough aesthetic features qualitatively identical to the relevant features of’ the original (Livingston 2003: 264). Hanson considers the idea that these conditions for adequacy suffer from an underlying worry to the effect that a ‘forgery could never be an adequate surrogate’ because ‘there will be some judgments of the original for which the forgery would be an inadequate or misleading guide’. The reason is that even if the two are superficially alike, there will presumably be differences, and ‘so it would be irresponsible to take a referential forgery to be a good guide to the original work’ (Hanson 2015: 253). One problem with this reasoning is that no contemporary participant in the debate has argued or would (reasonably) argue that a faithful, properly causally connected, etc. forgery could never be an adequate surrogate on the ground that the forgery may contain minute differences – after all, if a photograph can be an aesthetic surrogate for a painting, why wouldn't also an (almost) exactly similar and causally connected forgery in the same vehicular or artistic medium? Besides, even though some artistic judgments based on a forgery may differ, many (though not all) aesthetic judgments about the original based on an experience of its forgery would be justified. Perhaps looking at a reproduction or forgery of a work would not give us a full appreciation of the artist's specific aesthetic/artistic achievement, but if, say, a translation can be a good aesthetic surrogate (as participants to the debate seem to accept), why not a copy that has enough of the relevant aesthetic qualities of the original?

What Hanson concludes from the foregoing considerations is that being a surrogate is a three-place relation between two objects and the kind of judgment to be made by using the surrogate (Hanson 2015: 254). After all, she argues, even just being perceptually similar might sometimes suffice for being an adequate surrogate, but this depends on the kind of judgment we want to make by deploying the surrogate. Although this previous point seems correct, in what follows, I object to one further consequence Hanson draws from her discussion. More specifically, according to Hanson, one consequence of the reasoning in the previous paragraph is that we now have no reason to see room ‘for descriptions of works to count as adequate surrogates’ (Hanson 2015: 254). For, Hanson claims, if being an adequate surrogate consists only in perceptual similarity, since descriptions are not perceptually similar to anything and do not have perceptual properties, they can never be adequate surrogates (Hanson 2015: 254).

Now, this argument is not persuasive. First, token-descriptions do have sensorial properties and can be similar to other things (how else could we read them?). Besides, assuming that Hanson understands ‘perceptual properties’ as referring to properties that can only be perceived by our external senses, Hanson has nowhere shown at this point that only perceptual (sensorial) similarity should be the only criterion for being an adequate aesthetic surrogate. We have already mentioned that good translations can be adequate aesthetic surrogates, partly in virtue of having similar aesthetic properties despite the differences in non-aesthetic perceptual properties (Livingston 2003: 263). On a charitable interpretation, perhaps Hanson believes that the problem with descriptions is that they are in a linguistic representational format that purports to represent objects in other and sometimes radically different representational formats. However, insofar as X and Y are relevantly causally-historically connected and in representational formats having a sufficient quantity of aesthetic qualities in common, being in a linguistic format as such has not been shown to make something an inadequate aesthetic surrogate. To reiterate the point, consider translations of certain novels – for example, Thomas Mann's Death in Venice. It is extremely likely that the German version of the novel contains some determinate aesthetic qualities irreproducible in its English translation. However, certain sufficiently general artistic and aesthetic properties can be shared by the work in German and by its translation. Now, if linguistic representations can be adequate aesthetic surrogates, there do not seem to be conceptual problems with regarding a description of an idea aesthetically similar to the idea itself as an aesthetic surrogate. So, at least for those works of conceptual art properly classifiable as (at least partially) entities in a linguistic representational format, some descriptions can be proper aesthetic surrogates.

The previous reasoning supports the claim that Hanson's criticism (understood as a criticism against the idea of linguistic aesthetic surrogates) has not shown that descriptions, qua linguistic representations, cannot be adequate aesthetic surrogates. Some linguistic representations can have aesthetic properties and can be proper aesthetic surrogates (e.g., accurate translations of works of literature). Now, this point is important also because some works of conceptual art can be regarded at least partially as being in a linguistic format. Since we can make proper aesthetic judgments based on linguistic representations and there can be proper aesthetic surrogates in this format, the AP would still be true in cases in which we make aesthetic judgments based on some proper aesthetic surrogates in a linguistic representational format. In cases in which a linguistic representational format is adequate to capture the relevant aesthetic properties of a work, some descriptions can be appropriate aesthetic surrogates – and this would not show that the AP is false.²⁰

In conclusion, barring further elaboration or other examples of allegedly problematic descriptions, the AP is not threatened by examples from conceptual art as presented by Hanson.

4. Robson and the canonical route

Another recent forceful attack on the AP is contained in Robson (2018). To begin with, however, Robson's (2018) presupposes that ‘acquaintance’ is just another name for a form of perception. As mentioned in Section 2, only a few contemporary defenders of the acquaintance principle have specified what they mean by ‘acquaintance’ or ‘experience’. Still, most of the relevant works on this topic do not use the term ‘perception’ in discussing the AP or, at least, distinguish between acquaintance and some forms of sensory perception – if only implicitly. For example, Malcolm Budd does not use ‘perception’ in his (Budd 2003: 388), as Robson (2018: 660) incorrectly reports. Rather, Budd consistently uses ‘first-hand experience’ or ‘acquaintance’, presumably to highlight a contrast between this and a second-hand or testimonial form of epistemic state. Budd uses the term ‘perception’ only to discuss Pettit (1983) (one work where ‘perception’ is used). Similarly, Livingston's two possible successors of the AP do not include the term ‘perception’. For example, he spells out the weak AP in the following way: direct experience is necessary to the adequacy of some, but not all, aesthetic judgments because of the limitations of our descriptive capacities and theoretical resources (Livingston 2003: 276, emphasis mine). Without mentioning the AP directly, Keren Gorodeisky (2010) discusses Kant's rejection of certain forms of testimony – which seems to be motivated by a form of the AP – in terms of ‘engagement’, again without mentioning ‘perception’. Lastly, Robert Hopkins (2011) uses, not ‘perception’, but ‘experience’ in formulating the AP.

In any case, Robson provides a definition of ‘perception’ that is intended to be ‘very broad’: it includes all perceptual modalities, perceptions of surrogates (including photographs), and sensory imagination (Robson 2018: 658). To judge from the subsequent discussion, it seems that what Robson mostly has in mind is a form of sensory perception, with the inclusion of inner perception of mental images. Connected to this, the claim Robson endeavours to criticise is that there is something special or privileged about perception as a route to aesthetic judgment and that such a special connection is peculiar only to the aesthetic. Now, not many contemporary aestheticians and even fewer supporters of the AP have endorsed such a strong claim, at least if we understand ‘perception’ in terms of sensory or imagistic perception. Even Bence Nanay, one of the leading contemporary supporters of the idea that perception and aesthetics are strictly connected, recognises that not all aesthetic experiences are necessarily perceptual. For instance, he writes that not ‘only perceivable entities can be experienced aesthetically’ (Nanay 2016: 9). Still, Robson's arguments are relevant to our discussion of the AP. In fact, by arguing that aesthetic testimony based solely on descriptions – not understood here as possibly being a kind of aesthetic surrogates – is sufficient to justify an individual in making an aesthetic judgment, Robson's reasoning would show, as a further consequence, that the AP is false.

Now, Robson argues against the claim that only perception allows us to make adequate aesthetic judgments with respect to certain determinate aesthetic properties (Robson 2018: 661). Let us explore the case for that claim before turning to Robson's counterargument. The idea is that some objects have a certain aesthetic value in virtue of having determinate aesthetic properties: a landscape, a painting, a poem, etc., is not just beautiful but is beautiful in a specific way and has aesthetic value in virtue of that way of being beautiful (e.g., in virtue of being serene in x-way or moving in y-way) (Livingston 2003; Sibley 1974; Tanner 2003). The point is that aesthetic qualities (at least those that confer merit) can be determined in different ways: two landscapes can both be scenic and beautiful, but they will inevitably be scenic in different ways and beautiful in virtue of these specific ways. Given that, for various contingent reasons (e.g., we cannot entirely capture the x-way of being visually scenic of a landscape by means of a linguistic representation), the specificity of the way in which a landscape is scenic does not seem to be accessible to anyone who has not experienced the landscape or one of its adequate aesthetic surrogates (the accessibility premise), experience of the landscape or of a proper surrogate is required to make a legitimate aesthetic judgment of the landscape. The accessibility premise can be variously justified; for instance, we can argue that certain determinate aesthetic properties can be perceived or conceptualised only by means of or as consequence of experiencing sufficiently fine-grained representations. We might contend, for example, that a determinate visual aesthetic property can be properly perceived or conceptualised only by means of a fine-grained visual representation. Most, if not all, linguistic representations, such as those deployed in descriptions used in some forms of testimony, are not sufficiently fine-grained to capture the specificity of, say, determinate visual aesthetic properties. Insofar as the only way in which we can acquire and deploy representations sufficiently fine-grained to reason and make judgments about some determinate aesthetic properties is by experiencing them, experience is required for making aesthetic judgments involving determinate aesthetic properties. Now, Robson argues that this reasoning is mistaken. More specifically, his main criticism is that we can ‘certainly’ get some idea of the particular ways in which an object is beautiful without perceiving the object first-hand. For example, Robson argues, learning that an object belongs to a certain artistic category would already provide us with some idea about the way in which it is beautiful:

It is certainly the case that we can get some idea of the particular ways in which an object is beautiful, without perceiving the object for ourselves. Merely learning that the object we are discussing is a musical work rather than a painting, or a requiem rather than a lullaby (knowledge that can easily be acquired in the absence of perceptual experience), would already give us some idea as to the specific ways in which the object is beautiful. (Robson 2018: 662)

This point is puzzling – and seems to trade on a possible ambiguity in the claim that objects can have aesthetic properties in a specific way. To better clarify what is at issue, I will expand on this point. To begin with, we can understand the idea that an object can have aesthetic properties in a specific way at least as follows:

With respect to an object O,

1. (a) O may instantiate in S-way the determinable aesthetic property P, or

2. (b) O may instantiate in S-way the determinate aesthetic property P-in-Q-way, or

3. (c) O may instantiate the determinate aesthetic property P-in-S-way.²¹

Robson may be suggesting that we can have an idea of the ways in which some aesthetic properties are instantiated, or of the kind of determinable properties Q in virtue of which O is beautiful or of the kind of determinate ways in which O is beautiful. For example, if I know that O is a film and R is a poem and that both are moving, I may have an idea that the way O instantiates the property of being moving will be different from the way R does. In particular, O may be moving partly because of its sad soundtrack, whereas R may be moving because of its powerful metaphors or themes. The quoted text, however, seems to be suggesting that we may have an idea also of the determinate aesthetic properties in virtue of which O is beautiful, the reason being that knowing that O belongs to an artistic category would come with the knowledge that O would be likely be beautiful because of a kind of determinate properties. With some qualifications, we can grant this to Robson. One of these qualifications is that, unless we are acquainted with the work or with an appropriate surrogate of the relevant kind, we still won't have a clue about their respective determinate ways of being moving – for example, we will not know the property of being-moving-in-O-way or being-moving-in-R-way, nor probably the determinate aesthetic/non-aesthetic properties in virtue of which O (or R) is moving. Merely knowing the artistic category of O is not usually helpful in these respects. Still, this concession does not show that the AP is wrong – and I will focus on property-acquaintance to show why this is the case.²²

First, recall that, according to property-acquaintance, experiencing an aesthetic property or its proper aesthetic surrogates is required in order to make aesthetic judgments involving it. To begin with, notice that Robson's specification of what we can obtain from knowing the artistic category of an object (i.e., ‘some idea of the particular way in which something is beautiful’) is far from constituting the basis for ‘being justified in making a legitimate aesthetic judgment’, whether applied to determinate or determinable properties.

For instance, merely knowing that M is a horror movie would hardly give me any reliable sense of the determinate aesthetic properties of the scenes composing the movie upon which I could make an adequate aesthetic judgment (here understood as involving the formation of a belief that can rationalise or motivate certain attitudes in a fitting, right, or proper way). The prospects of knowing the kind of determinate aesthetic properties merely by knowing an artistic category appear even more desperate once we consider finer-grained aesthetic determinate properties.

In addition to this, an adequate aesthetic judgment may plausibly require that the subject should display at least some understanding of the aesthetic concepts involved. For any artistic category C the belonging to which essentially involves possessing some aesthetic properties (e.g., being soothing and calming in the case of lullabies or being sombre and solemn in the case of requiems), S could hardly make proper sense of the description ‘O belongs to artistic category C’ without any previous acquaintance with at least some of the relevant determinable aesthetic properties that characterise category C. For example, someone who has never experienced a requiem or a lullaby is not in a position to make an adequate inference and related aesthetic judgment regarding the requiem-kind or lullaby-kind of aesthetic properties. One reply is that it may be possible that for S to make an aesthetic judgment about O with respect to at least some already experienced determinable Ps, it may suffice knowing that O belongs to C. However, at least property-acquaintance is not affected by this reply. However, to reiterate, without being acquainted with the relevant aesthetic properties or surrogates, we would not be able to properly understand and make adequate judgments starting only from a merely descriptive categorisation of an object under a category.

To further sustain this point, the previous reasoning regarding a subject who has never experienced the aesthetically relevant property relative to a given category can be supplemented as follows. In the case in which S only allegedly knows that O belongs to C without having experienced the relevant aesthetic properties Ps or relevant aesthetic surrogates characterising C, S's epistemic situation is intuitively analogous to Mary's in the notorious case in philosophy of mind.²³ Before exiting the black and white room and experiencing (phenomenal) red, the general intuition is that, with respect to the colour red, even in its determinable variety, Mary lacks some sort of knowledge crucial for the understanding of an essential or characterising phenomenal aspect of red – whether this may involve some relevant phenomenal concepts or some mental capacities to know what red is like (such as the capacity to imagine certain counterfactual situations involving red).²⁴ Leaving aside the complex debate in the metaphysics of mind, what matters here is that knowing or understanding phenomenal red seems to require at least knowing or understanding what it is like to experience it. Similar to the case of Mary's thought experiment, consider an individual who has never experienced, say, elegance or a solemn atmosphere and related aesthetic properties and is told that an object is elegant or a piece of music is solemn. Even after a linguistic description of someone's appearance, mannerisms, or of the fact that the piece of music has a slow tempo, restrained dynamics, and serious or mournful mood and so on, S would still lack the resources for a proper or full understanding of the aesthetic judgment that the object is (visually) elegant or that the music is solemn. Upon hearing from someone else that O is elegant, we would say that S may certainly utter the sentence ‘O is elegant’, however, we would reasonably doubt that S made a judgment that they would sufficiently understand to, say, rationalise their actions. In fact, it seems that there is something missing in S's conceptual repertoire had she not experienced the relevant aesthetic property in the appropriate format and, we may add, at the right level of determinacy.²⁵

I think that a plausible explanation of the analogy between the two cases is that, in the absence of a relevant experience, S would still be lacking the sufficiently fine-grained phenomenal concepts related to elegance or solemnity.²⁶ So, suppose that S has never experienced gracefulness, vividness, or any of the properties P generally associated with a category C. Can we say that, on hearing that O belongs to C, S can make a legitimate aesthetic judgment to the effect that O is P? Although part of the answer depends on what we mean by making an adequate aesthetic judgment and on the nature of belief, it seems that S does not possess those concepts required to form proper aesthetic beliefs about O. Since S has not had the relevant experience(s) and given that such experiences are required (at least given our mental capacities) for acquiring the kind of relevant concepts, S would be able to have at most a limited understanding of the thought that O is C or that someone claimed that O belongs to C. S would not be able, for example, to fully or even partially grasp why this attribution was made. Recall that the AP is here understood as an epistemic norm. Now, in coming to know that S has never experienced some or all of the characterising features of C (or their surrogates), it seems rational for us to decrease our epistemic trust in S's aesthetic judgements about C. For example, wouldn't you reduce the trust in a critic's aesthetic judgements about O possessing features characteristic of horror movies in coming to know that the critic has never experienced, say, the emotional characterising features of a horror movie or the way in which they are instantiated in them? My point here is that, without an experience of a relevant aesthetic property P, S would not grasp or possess part of the concept of P (i.e., its phenomenal part, which can normally be acquired only through experience) and would thereby at most have an inadequate (for epistemic purposes) grasp of the concept of P.²⁷

If the above reasoning is correct, then, after all, at least property-acquaintance is true: where P is an aesthetically relevant property for a category C, for S to make adequate aesthetic judgments about O involving P on the basis of a description that O belongs to C, S must have been acquainted with P or with an appropriate aesthetic surrogate in the relevant format and at the right level of determinacy. Robson mentions requiems and lullabies as possible categories which may suggest the kind of determinate aesthetic properties that works of these kinds may have. My reply is that he may be right, but knowing and understanding these categories requires having experienced the aesthetic properties that play an essential role in the understanding of such categories. Besides, in some case, we may also need to know the interplay of these qualitative aspects and how the representational means achieve such an interplay. Since this knowledge and understanding requires the possession of some aesthetic concepts having a phenomenal character, and given that we can obtain such concepts only through experience (due to our contingent way of forming the relevant concepts and capacities), we would need to have experienced the relevant aesthetic properties before deploying concepts referring to them in making adequate aesthetic judgements involving them.

This reasoning may have implications for judgments concerning the adequacy of aesthetic surrogates. When such judgments involve the possession of aesthetic concepts, and if we consider judgments about the adequacy of aesthetic surrogates to be aesthetic in nature, then property-acquaintance implies that such judgments require (a certain degree of) acquaintance with the relevant properties of both surrogates and what they stand for in order to be justified. It may be argued that this outcome implies that we have fewer justified aesthetic beliefs than we may expect as, in many cases, we lack knowledge about the adequacy of the surrogates themselves (e.g., when judging translated works that are originally in languages we do not know). However, if this outcome is deemed problematic, on an externalist view of epistemic justification, property-acquaintance may not imply that for S to have a justified aesthetic belief about O or P, S must also possess explicit knowledge of the appropriateness of the aesthetic surrogate upon which their belief depends. Therefore, property-acquaintance may not imply that we do not have justified aesthetic beliefs about O or P if we do not also know the adequacy of the aesthetic surrogates of O or P upon which our aesthetic beliefs are based.

In conclusion, I hope to have shown that the AP is not threatened by alleged counterexamples from conceptual art as discussed by Hanson and that at least property-acquaintance is not threatened by Robson's arguments.²⁸