A CRITIQUE OF MODERATE FORMALISM

Moderate formalism is the view that all artworks which have aesthetic properties have formal aesthetic properties, and some but not all of those works also have non-formal aesthetic properties. Nick Zangwill develops this view in his Metaphysics of Beauty after having argued against its alternatives – extreme formalism and anti-formalism. This article reviews his arguments against the rivals of moderate formalism, and argues that the rejection of anti-formalism is unjustified. Zangwill does not succeed in proving that the broadly determined (context-determined) properties of artworks are in some cases irrelevant to their aesthetic properties – and following that, interpretation and assessment. A historical argument presented here shows how aesthetic properties of every work must partly supervene on this work’s contextual properties. In particular, this disproves Zangwill’s claim that epistemological matters are unessential in determining the artwork’s properties, and exposes some problems his account has with explaining relations between nonaesthetic and aesthetic properties.

on the non-aesthetic as follows: the conjunction of narrow non-aesthetic properties entirely determines the formal aesthetic properties, and the conjunction of the broad non-aesthetic properties partly determines the non-formal aesthetic properties. Thus there can be two kinds of aesthetic property: formal and nonformal. Zangwill holds that the former is loosely comparable to Kant's free beauty, and the latter to dependent beauty. 2 An example might clarify the above distinction: El Greco's View of Toledo has the following physical properties: sharp colour contrasts, blurred contours, and elongated shapes, as well as the following relational properties: not accurately depicting the contemporary city of Toledo and being created in late sixteenth-century Spain. The former kind narrowly determines formal aesthetic property of being mysterious. Together with the latter, they broadly determine a non-formal aesthetic property, for example, being disturbing.
The question now is whether there really are these two kinds of aesthetic properties. It seems undeniable that all works of art have both narrow and broad non-aesthetic properties. But can they both give rise to aesthetic properties? This question is of vital importance -for if there were no non-formal aesthetic properties, then the history of production, the cultural context, representative qualities, and so forth, of a work would all be irrelevant to its aesthetic assessment; or if there were no formal ones, colour and internal structure alone would be of no aesthetic importance. Extreme formalism is the view that all aesthetic properties are formal; and anti-formalism is the view that all aesthetic properties are non-formal (MB,58). Extreme formalists would naturally agree that an object has both narrow and broad non-aesthetic properties, but would deny that there are any broadly determined aesthetic properties. Anti-formalists, on the other hand, would deny that any aesthetic properties are determined solely by narrow non-aesthetic ones. Moderate formalism attempts to stand in between. As Zwangill writes: 'Moderate formalism is the view that while some aesthetic properties of a work of art are formal, others are not […] but […] there are some works of art that only have formal aesthetic properties. ' (MB,59) On one hand, then, there are works like El Greco's The View of Toledo, and on the other -Kandinsky's Composition VII or Bach's Fugue in E BWV 854. The aesthetic properties of the former are based on its physical properties (that is, colour contrasts, vertical composition) and the fact that it is a representational painting, created A Critique of Moderate Formalism 2 Only loosely, however. Non-formal properties need not involve the concept of an end or purpose (MB, 60) -a representative painting is beautiful not because it represents something, but as a representation; similarly, contextual works do not need any extraaesthetic purpose; their non-formal properties are broadly determined simply by standing in some relation to other works. The definition of being determined by broad non-aesthetic properties does not, however, imply anything of this sort, so I will disregard Zangwill's invoking Kant. at a certain time and place, and so forth (that is, it is disturbing as a depiction of Toledo created in sixteenth-century Spain). 3 Abstract paintings or absolute music, on the other hand, have no broad non-aesthetic properties that could influence aesthetic judgement: Bach's fugues are elegant, and so forth, only in virtue of their perfect structure, regardless of where, when, or by whom they were created, save the fact that they are not representational at all.

II
Zangwill tries to give his account some plausibility by criticizing the alternatives -extreme formalism and anti-formalism. I reconstruct his arguments briefly, partly because similar arguments can be used to criticize moderate formalism, and partly to soften his conclusions, showing that they are not as inevitable as they might seem. Zangwill's strategy to disprove extreme formalism is simply to falsify it with examples of works of art that uncontroversially have non-formal aesthetic properties. First, it seems simply false to say that representational properties of paintings are irrelevant to their aesthetic value, since surely a painting can be better or worse depending on how well it depicts its subject.
Second, most representational paintings operate with an illusion of depth -they appear to be three-dimensional, although they are in fact flat. (Zangwill calls this the 'plastic property' of painting.) Surely there must be some cultural or historical conventions that determine our viewing a flat picture as having depth, and indeed the history of art knows several different types of perspective, each one being thought to be 'natural' at some point in time. 4 To argue that three-dimensional properties are narrowly determined by what is on the canvas alone would either beg the question against the variety of ways to represent perspective or require us to broaden the definition of formal properties to include plastic properties. Last, in architecture, the function of the building partly determines correct aesthetic judgements about it. A mosque may be judged beautiful, but if the same structure were to be a post office, it is quite likely that it would not be assessed along the same lines. It is not 'beautiful and functions well (non-aesthetically) as a mosque, but [it is beautiful partly] because the building aesthetically expresses or articulates the religious function of a mosque' (MB,68). In such cases the object's beauty depends on the category it belongs to -the artwork here is not beautiful simpliciter, but beautiful as a member of a category. 5 Although an aesthetic judgement can be 3 The notion of 'being beautiful as' is taken from the anti-formalists, specifically from Kendall Walton, 'Categories of Art ' , Philosophical review 79 (1970): 334-67. 4 See Erwin Panofsky, Perspective as Symbolic Form, trans. Christopher S. Wood (1927;New York: Zone Books, 1991). made on the basis of the narrow properties of the building without taking its function into account (for example, when the function is unknown to the observer, the building is being treated as an abstract sculpture), such a judgement would be at best incomplete.
To undermine anti-formalism, Zangwill criticizes the arguments supporting it.
Two of the seven arguments that he discusses and dismisses with the charge of irrelevance seem central to his case (MB, 104-6). First, judging an artwork seems impossible without basing it on one's knowledge of other works. This knowledge is always present and thus it seems that every aesthetic judgement is partly based on the relations between the object judged and other works. Zangwill's response is that one's judgement may indeed rely on one's broader knowledge, but this has nothing to do with an object actually having any determined kind of properties, only with one's knowledge of these properties. Zangwill builds the following analogy: one can use a Geiger counter to determine whether an object has the property of being radioactive; one has to know about radioactivity and Geiger counters to do that, but this does not mean that using a Geiger counter on an object is part of what it is to have the property of being radioactive. Thus the knowledge of other works seems to be an epistemological matter that tells nothing about the metaphysical nature of the properties in question, but only about how we come to know about them. I find this solution deeply unsatisfactory, and will discuss it in the following sections.
Second, to judge an artwork accurately one has to know whether it has representational or contextual content or not (even if only to be sure that it does not), and this means knowing about the history of its production. In fact, nonrepresentational and non-contextual works seem to have a negative non-formal property of lacking non-formal aesthetic properties. Zangwill's counterargument is to dismiss if not the very possibility of negative properties, then their relevance to the matter, for anti-formalists claim that aesthetic judgements are to be dependent on positive representational or contextual properties. And if this is the case, the argument can again be reduced to the irrelevant epistemological point that in order to be able to make an aesthetic judgement one has to know whether the object is (non)representational or (non)contextual, which is not the same thing as the object itself having a property of being (non)representational or (non)contextual.
This dismissal seems too hasty. I can see at least two accounts on which the supposedly negative properties become important. Numerous avant-garde programmes 6 provide evidence that being abstract can reasonably be thought For example, with the rise of representational expressivism the property 'is expressive' became definitional of some works, but since it was the only property that differentiated it from earlier art, 'is not expressive' became relevant for definitions of other works. There is, it seems, no reason to think that it should be different for representation -it can be aesthetically relevant for expressive works if they are representational (as in fauvism) or not (abstract expressionism). 7 Further on, Zangwill attempts to show that not all aesthetic judgements depend on non-formal properties, by discussing Walton's 'guernicas' example, in which a culture that knows no other works of art except various three-dimensional and extremely dynamic versions of Picasso's guernica judges the flat guernica, which we would describe as violent and dynamic, as dull and static. Zangwill gives his own real-life example of Minoan and Mycenaean seals (MB, 90-96) and argues that Walton's view has odd consequences -we cannot compare items from different categories, that is, we cannot say whether the Minoan seal is more dynamic than the Mycenaean seal; we can only say that it is more dynamic as a Minoan or as a Mycenaean seal. As Zangwill notes, however, there is nothing easier than broadening the categories and judging all Minoan and Mycenaean art as belonging to the same category of prehistoric Greek art, thus providing grounds for comparison. I believe that such an answer is in fact sufficient, and the arguments Zangwill uses against it are inconclusive. He first writes that such a broader category would be somewhat mysterious, since the subject would apply it subconsciously, and we have no independent reason to believe that such a subconscious application of categories is taking place. I do not see why using broader categories must be subconscious. It seems perfectly acceptable to apply such categories consciously, and surely the category 'prehistoric Greek art' is hardly mysterious. But even if this were true, I cannot see an independent reason to disbelieve the application of such subconscious categories either. In fact, there seem to be a number of subconscious mechanisms involved in aesthetic judgements and people are often unaware, for example, why exactly they find certain things dynamic.
Zangwill's second claim is that aesthetic properties are gradated, that is, an artwork can be more or less elegant. Thus when we judge the Minoan seal to be dynamic, our judgements do not have to involve categories -that is, the Minoan seal may be dynamic as a Minoan seal, but not very dynamic when considered by itself, and thus we can make category-neutral aesthetic judgements. But what are the universal criteria for judging that a certain work is dynamic per se? There are no independent reasons to believe that there are universal (that is, categoryneutral) criteria for judging the degree to which artworks are dynamic any more than there are for category-specific ones. In fact, I believe that a very simple answer to this puzzle can be given on Walton's account: what seems to be a category-neutral judgement is in fact embedded in an even broader category of all the works that the judging subject knows. One does not have to be aware that one judges a work relative to all other works one knows, but surely guernica is judged differently in both our culture and other cultures precisely because we place it in the context of all the art we know and they place it in all the guernicas they know. This stands in opposition to the first of the reconstructed arguments against anti-formalism by Zangwill, and is one of the reasons to deny it.

III
Zangwill claims that the perceiver's knowledge of the historical background of the work or his or her own background is irrelevant to whether an object actually possesses certain properties. In other words, an object does not change its properties relative to the perceiver's knowledge. 8 Walton, on the other hand, claimed that the epistemological position of the perceiver does influence his or her aesthetic judgements -one has to learn what properties a given artwork was supposed to have and recognize them in order properly to appreciate the work. 9 But, since the formal properties supervene on the object's physical properties, this implies that if the judgement is not history-dependent, these formal properties have to supervene on the same physical properties at all times. I think, however, there is a lot of evidence against this claim -historically people do change their judgements on the aesthetic qualities of some works, and not just locally. For example, Vasari and Petrarch despised all medieval art and the art of the Northern Renaissance, by this expressing the general change in the way certain physical properties of the works were perceived -for example, preference for idealistic representation, classical composition, geometrical perspective, and bright colours, over naturalistic representation, allegorical composition and perspective, and allegorical or toned down colours. This in turn gave rise to differences in the supervening aesthetic properties. But since on Zangwill's account the same artwork cannot change its aesthetic properties depending on how people view it, the aesthetic property of being, say, elegant supervenes either on all of these physical properties or on just some of them. Thus either the Italian Renaissance critics somehow failed to see that relation holding for medieval and Northern artworks or the works were not elegant at all. One could naturally argue that Vasari may have simply been wrong or blind to the merits of medieval art, but such an explanation would require one to hold that virtually all his contemporaries were wrong as well and that for hundreds of years most people misjudged a style with regards to its properties. Since people changed their views on past art rather often, a moderate formalist would need to hold that for most of history most people were wrong about most works. But that would be a somewhat revisionary thesis, considering that there is a simple and quite intuitive alternative -namely, to admit that some kind of change, perhaps only a change in the general perception of art, not art itself, has actually taken place.
So how exactly are the historical and cultural properties of art aesthetically relevant? First, the same aesthetic property can supervene on one set of nonaesthetic properties at one time and another set at another time (or, to avoid unnecessary metaphysics, the criteria for applying aesthetic predicates vary with time). For example, the aesthetic property of being consonant supervenes on certain acoustic features of the sounds produced -the relations between pitches.
In the sixteenth century, the set of intervals and harmonies considered consonant (call it A) was relatively narrow, whereas by the nineteenth century the set (call it B) had been enlarged by a number of intervals and harmonies previously considered dissonant. Thus the intersection of A and B (and since A is a subset of B -all of A and the part of B which intersects with A) contains intervals that are consonant at either time, but the part of B which does not intersect with A contains intervals that are consonant only in the nineteenth century. It therefore seems that music featuring intervals included in the remaining part of B is consonant later, but not earlier, for example, Wagner's Tristan und Isolde played at the court of Elizabeth I would not meet with much applause. But for a moderate formalist, 'being consonant' must supervene on the same set of non-aesthetic properties at all times, that is -either Wagner's audience or the Tudors were wrong. It could be argued that the set was always broad, but it took some time before people learned to appreciate it (which would stress the distinction between the aesthetic and epistemological judgements). But in other cases similar sets seem to shrink -an aesthetic property of being colourful supervened on a much narrower set of colours and relations between them at Rubens's time than it does now, for example, a Rubens is not considered particularly colourful nowadays, but a Matisse would have been outrageously colourful in the seventeenth century. Or it may happen that two works will have the same aesthetic property supervening on completely different sets of physical properties, for example, Romanesque figurative sculpture's being mysterious or sublime supervened on slightly unnatural symbolic body proportions, while in Renaissance figurative sculpture it did so on idealized natural proportions. 10 All this suggests that the base of supervenience of one and the same aesthetic property can change over time, in which case knowledge of the history of production of every single work is relevant to its assessment.
Importantly, it is not only the case that a single aesthetic property can supervene on two or more sets of non-aesthetic properties -this would hardly be problematic, as surely a non-history-relative property such as 'having the sum of 12' can supervene on infinitely many sets of numbers. But aesthetic properties are more than that -their bases of supervenience change, that is, 'being elegant' (α) may supervene on a set of formal properties A at one time but not at another.
Both '6 + 6' and '5 + 7' have 12 as their sum, but it is never the case that once 12 starts supervening on '5 + 7' it stops supervening on '6 + 6'; but once α started supervening on bright colours and geometric perspective in the Renaissance, it stopped supervening on dark colours and allegorical perspective. This suggests that either the relation between aesthetic and non-aesthetic properties is not that of supervenience or it is history relative: we should admit that α supervenes on more than just the formal properties, and we should talk not about 'being elegant' simpliciter, but 'being elegant as a Renaissance painting' . Second, the same set of non-formal properties can give rise to different, possibly even mutually exclusive, aesthetic properties at different times. For example, Penderecki's St Luke Passion uses cluster and microtonal techniques to determine narrowly the properties of being, again, mysterious and sublime. But if the same music were by the sixteenth-century composer Clemens non Papa, his contemporaries would, arguably, hold it to be boisterous and terrifyingwhich would hardly be compatible. A variation of this case is even more striking -most of the avant-garde works, even those now as 'classical' as abstract painting, have physical properties that we consider to determine aesthetic properties, A Critique of Moderate Formalism 10 A great art history description of this question was presented by Panofsky in his analyses of different kinds of perspective, body proportions, iconography, and so forth, which were used at different times. Panofsky himself was very careful not to imply that there was an aesthetic (and not just artistic) change involved here, but in the light of the current discussion his examples may serve to back my view. but would not have been seen as such if created and seen two hundred years ago, that is, the same set of physical properties would give rise to different sets of aesthetic properties in the year 2000 and in the year 1800, and the latter set would be empty. 11 It seems now that the aesthetic properties that arise from a given set of physical properties are contingent on when the work was produced and when it was perceived. But for moderate formalists this is unacceptable, since some aesthetic properties supervene on narrow non-aesthetic ones independently of historical properties -and if so, the same set of narrow non-aesthetic properties should give rise to the same set of aesthetic properties at all times. Once again, it seems that either the relation is not that of supervenience (since the same supervenience base should always give rise to the same property) or the supervenience bases actually differ -and since they include only the same formal properties, they must differ in the non-formal, historical properties. Lastly, art is full of clearly history-dependent processes of changing relations between aesthetic and non-aesthetic properties. 12 Consider the following simplified example: until the nineteenth century an aesthetic property, elegance (α), supervened on a set of non-aesthetic properties: toned down colours (A), complexity (B), and being representational (C). With Impressionism and the emancipation of bright colours, the base of supervenience for α changes slightly, it is now the set {¬A, B, C}. Furthermore, fauvism simplifies the composition to attain maximum expression, thus the set changes to {¬A, ¬B, C}, and abstract art finally brings α to supervene on {¬A, ¬B, ¬C}. 13 In fact, all combinations of these three non-aesthetic properties can give rise to α. 14 Clearly, in this case the relations between the physical and the aesthetic properties of the work change quite drastically, and the change is allied with the paradigm in which the work is created. In other words, the aesthetic property is relative to the time of the work's creation -it is a non-formal property. To hold otherwise would be to agree that the aesthetic property supervenes on an inconsistent set of physical properties, which threatens Zangwill with the relativism he wants to avoid.

11
Physical properties that were once thought not to give rise to aesthetic properties but are now thought to do so, for example, the free brushwork of the background of Titian's or Velázquez's paintings, are discussed by Richard Wollheim, Art and Its Objects, 2nd ed. (Cambridge: Cambridge University Press, 1980), 89. This lends even more power to my claim that aesthetic properties are not like 'having 12 as a sum' , simply multiply realizable, satisfied by a long disjunction of supervenience bases. If α can supervene both on {A, B, C} and {¬A, ¬B, ¬C}, such characterization would become trivial: one could always come up with some mad disjunctive account that gets the answer right without referring to the context, but the moral of the story is that the possibility of such disjunctive accounts can be discarded.

IV
In the light of what has been discussed here so far, I can return to the problem of the supposed aesthetic irrelevance of epistemological matters. Zangwill's claim simply seems wrong -the aesthetic properties of an object are partially constructed by the way they are perceived and depend on their historical or cultural background. 15 It is Zangwill's analogy with the Geiger counter that is irrelevantaesthetic properties are not like the property of radioactivity and are not assessed by the same means, that is, by objective, human-independent measuring devices.
An aesthetics Geiger counter would have to be recalibrated with every changing style and period, and the only possible justification for such recalibrations would be found in the cultural and historical background that the artists and critics stem from. Effectively, any accurate readings of such a counter would only be possible if the 'tested' artwork's non-formal aesthetic properties were taken into account.
Although this approach does not disprove moderate formalism altogether, it forces it to assume a very implausible position -if any given work's broad non-aesthetic properties were irrelevant to its aesthetic properties, then, to prevent the inconsistency arising from different interpretations of the nonaesthetic-aesthetic relations, one would have to claim that most art perceivers throughout history have been wrong. Moreover, one would have no guarantee that today's interpretations were the correct ones. Although, largely thanks to the work of such art historians as Alois Riegl, people are now more inclined to acknowledge the value of physical properties that were not thought to determine aesthetic properties when the artworks were produced (for example, early Christian art), or even truly appreciate all past works once thought not to have any aesthetic properties (unlike Vasari's interpretations of medieval art), one cannot be sure that this approach is correct either, because there may still be some relevant properties simply unknown to the present critics. In fact, even if the set of physical properties thought to determine aesthetic properties (call it A) A Critique of Moderate Formalism 15 As Wollheim put it, 'a clear separation cannot be made of fact and interpretation. For of many of the facts of art, it is required that they are interpreted in a certain way' . Wollheim,Art and Its Objects,90. contained the set of all those that actually do (B), it would still not be enoughthey would have to be identical, for otherwise one could not distinguish between an objectively true aesthetic judgement and an objectively false one, that is, a judgement that was made on the basis of considering properties included in the intersection of A and B (in which case it would be true) would be indistinguishable from a false one based on the properties included in the part of A that is not B.
The moderate formalist is therefore compelled to say that the objective relations between physical and formal properties are inaccessible to humans. This seems hugely revisionary, especially since a more intelligible explanation would be that these relations are historically (and culturally) relative, and thus a full interpretation of an artwork has to include an account of its broadly determined non-formal properties. 16 Moderate formalism is wrong in claiming that there are works that have only formal aesthetic properties. An analysis of the relations between the aesthetic and the non-aesthetic properties suggests that there is no fixed, objective bond assigning a given set of physical properties that determines only one aesthetic property -the relation is relative to the historical and cultural background in which the work is created and perceived. I hope to have shown that the epistemological questions related to an artwork's cognition should influence aesthetic judgements -aesthetic properties are not objective properties of artworks themselves, but are partially constructed by the perceivers. Consequently, we should reject moderate formalism as a valid theory of aesthetic properties of artworks. The additional virtue of this approach is that it explains how the original or anachronistic character of a work can be understood -a work that claims to have some aesthetic property claims to have it in virtue of possessing a set of physical properties, while, at the time that it is created, a different set of physical properties is realizing this aesthetic property. If the set of properties has determined the aesthetic property already in the past, it is anachronistic; if it determines it in the future, it is innovative.