One of the key concepts of Russian formalism is defamiliarization.1 Here is the most succinct exposition of what this concept means:

If we start to examine the general laws of perception, we see that as perception becomes habitual, it becomes automatic….Such habituation explains the principles by which, in ordinary speech, we leave phrases unfinished and words half expressed….The object, perceived in the manner of prose perception, fades and does not leave even a first impression; ultimately even the essence of what it was is forgotten….Habitualization devours work, clothes, furniture, one’s wife, and the fear of war….And Art exists that one may recover the sensation of life; it exists to make one feel things, to make the stone stony. The purpose of art is to impart the sensation of things as they are perceived, and not as they are known. The technique of art is to make objects “unfamiliar,” to make forms difficult, to increase the difficulty and length of perception because the process of perception is an aesthetic end in itself and must be prolonged.2

A distinctive feature of Russian formalism, something we do not see in Bell and Fry or in Wölfflin and Riegl (or see it more rarely, see Section IV below), is this emphasis on the analysis of everyday perception and the ways in which art encourages us to perceive differently. But it is difficult not to read the concept of defamiliarization as a naïve early statement of what art historians and aestheticians of the second half of the 20^th century criticized as the “Innocent Eye” tradition. As talking about the Innocent Eye is now universally frowned upon, defamiliarization has not been taken as seriously as it should have been. The aim of this paper is to revive the concept of defamiliarization by showing that it has nothing to do with the “Innocent Eye” tradition and to propose a new interpretation of it in terms of distributed attention. I conclude that this reinterpreted concept of defamiliarization can be very useful for contemporary post-formalist accounts of the history of vision (and imagination as well as visual attention).

Shklovsky, De Chirico, Proust, Musil 

What I take to be the central idea of defamiliarization is not primarily a claim about art or aesthetics. It is a claim about perception: everyday perception is different from the way we perceive (or should perceive) works of art. This idea (with its undeniable Kantian overtones) was quite widespread around the time when Shklovsky first wrote about defamiliarization (in 1914,3 although his most widely known description of the phenomenon I quoted above is from 1917). Here is a famous passage by Giorgio de Chirico from 1912, for example:

One clear autumnal afternoon I was sitting on a bench in the middle of the Piazza Santa Croce in Florence. It was of course not the first time I had seen this square….The whole world, down to the marble of the buildings and the fountains, seemed to me to be convalescent. In the middle of the square rises a statue of Dante draped in a long cloak, holding his works clasped against his body, his laurel-crowned head bent thoughtfully earthward. The statue is in white marble, but time has given it a gray cast, very agreeable to the eye. The autumn sun, warm and unloving, lit the statue and the church façade. Then I had the strange impression that I was looking at all these things for the first time.4

This experience of seeing something he had seen many times but as if he saw them for the first time is very close to what Shklovsky means by defamiliarization. What I take to be the big move of Russian formalists is that they connected this duality between everyday perception and these special, defamiliarized instances of perception with the idea of formalism. This connection, in the case of Shklovsky is that one way in which we can achieve defamiliarization is to use unusual or exaggerated formal properties. As he says, “the technique of art is to make objects “unfamiliar,” to make forms difficult, to increase the difficulty and length of perception.”5

The general idea of some kind of connection between defamiliarization and forms is not new.6 Here is Proust, in 1913:

But even the ugliness of faces, which of course were mostly familiar to him, seemed something new and uncanny, now that their features,—instead of being to him symbols of practical utility in the identification of this or that man, who until then had represented merely so many pleasures to be sought after, boredoms to be avoided, or courtesies to be acknowledged—were at rest, measurable by aesthetic coordinates alone, in the autonomy of their curves and angles.7

A lot is going on in this quote. We have the (broadly Kantian) insight that this experience is devoid of practical utility. We also have a formalist spin with the emphasis on curves and angles. And also the seeing of something familiar in a new light, with fresh eyes. And we see versions of all these three ideas in the writings of Shklovsky (even in the longish quote above). But the way Proust makes the connection between defamiliarization (the idea of seeing of something familiar in a new light, with fresh eyes) and forms (“the autonomy of…curves and angles”) is different from Shklovsky’s. Proust seems to suggest that defamiliarized seeing is the seeing of formal properties. When we see something as if we saw it for the first time, we see curves and angles. For Shklovsky, in contrast, the experience of seeing something as if we saw it for the first time is achieved by means of formal qualities that are difficult to process. For Shklovsky, form is the means to the end of defamiliarized vision. For Proust, forms and defamiliarized vision are two sides of the same coin.

Another author who connects the discussion of defamiliarization and forms is Robert Musil.8 His concept of “other condition” is a version of the concept of defamiliarization: the “other condition” is a way of perceiving the world that is very different from the way we normally perceive it.9 In “Toward a new aesthetic,”10 first he complains how obscure writers and philosophers tend to be when describing this “other condition” and then proceeds to give a pretty obscure description himself:

There is neither measure nor precision, neither purpose nor cause: good and evil simply fall away, without any pretense of superiority, and in place of all these relations enters a secret rising and ebbing of our being with that of things and other people. It is in this condition that the image of each object becomes not a practical goal, but a wordless experience.11

In order to understand Musil’s account of the “other condition,” we need to also have a closer look at two related topics that very much informed his way of thinking about the Other Condition. The first one is his formalism. Musil was a formalist—somewhat surprisingly for someone who is credited as the par excellence philosophical (that is, content-based) novelist. Being a formalist in Austria/Germany in his time was also much more unusual than, say, in England or Russia. Nonetheless, he endorsed a fairly sophisticated version of formalism that was more similar to Roger Fry’s than to Clive Bell’s.12

But his formalism is even closer to that of Russian formalists and the Shklovsky himself—a theme that I can’t fully explore here. While Shklovsky put the emphasis on the “unfamiliar,” Musil is interested in what makes our perception of the “unfamiliar” different from everyday perception. And his answer has to do with a concept that have since had a long career in analytic aesthetics—that of uniqueness.

While the uniqueness in the domain of aesthetics, at least within the analytic tradition, has either been conceived as a metaphysical claim or as a claim about aesthetic evaluation (and neither way of cashing out the importance of uniqueness has been particularly successful), Musil focuses on the uniqueness of our experience. Here is what he says: “Every work of art offers not merely an immediate experience but an experience that can never be completely repeated, that cannot be fixed but is individual, even anarchic. [That is where] it takes its uniqueness and momentary quality from.”13

Musil’s view here is an application of a distinction he has been making throughout his life, between the “ratioid” and the “nonratioid” domain. Here is the best (and earliest) characterization of this distinction from his 1918 essay “Sketch of what the writer knows”: “Roughly delineated, this ratioid territory embraces everything that science can systematize, everything that can be summarized in laws and rules; primarily, in other words, physical nature….The ratioid area is characterized by a certain monotony of facts, by the predominance of repetition…; but above all the chief characteristic of this area is that in it facts can be unambiguously described and communicated.”14 “If the ratioid area of the domination is the “rule with exceptions,” the nonratioid area is that of the dominance of the exceptions over the rule….In this region facts do not submit, laws are sieves, events do not repeat themselves but are infinitely variable and individual.”15 Morality and aesthetics are supposed to be part of the nonratioid domain.

After introducing the concept of the “other condition,” Musil warns against understanding the contrast between ordinary perception and the “other condition” as a difference between intellect and the senses. This, he argues, would amount to embracing some kind of anti-intellectualist pan-romanticism, and this is certainly not something he would want to do. Instead, the contrast is between the repeatable (ratioid) and the nonrepeatable (nonratioid) domain. While our ordinary perception groups things together in categories the instantiations of which are repeatable and interchangeable, the other condition takes everything as a unique individual—not as something repeatable and interchangeable. So here is where Musil’s ratioid/nonratioid distinction meets his brand of formalism: when we are looking at the world in a practical manner, say, when we are hungry, we see things as belonging to these two categories only: edible or non-edible (or, maybe, edible in case of emergency…). This is clearly a phenomenon in the ratioid domain. But when we are in this “other condition,” we see things for what they are and for what they uniquely are: not as an edible thing that could be interchangeable with some other edible things, but as a unique, not interchangeable entity.16

This concept of course bears various traces of Kant’s thinking—a connection Musil would probably not like. At one point in The Man without Qualities, he even describes the “other condition” as “a state entirely free from purpose.”17   But it is a very specific and very unusual way of cashing out the general idea of “disinterested pleasure”: what makes it disinterested is that we appreciate the object’s uniqueness, without trying to find a template of what to do with it or how to think about it. Note that Musil is also against taking the contrast between ordinary perception and the “other condition” to be a conceptual/non-conceptual distinction, as, for him, this distinction stands at every level of perceptual processing—just like the ratioid/nonratioid distinction.

But what is more interesting from the point of view of the present paper is the similarity with Russian formalism. But rather than describing the Other Condition on the analogy of our experience of unfamiliar objects, Musil takes a slightly different approach. We can, after all, experience unfamiliar objects in a variety of ways: if we have some strong practical inclination, we will experience this object as one that can or cannot be used for these practical purposes. But this has little to do with “what makes the stone stony.”

Nonetheless, it could be argued that the Russian formalists and Musil were really after describing and understanding the same experience. When Shklovsky says that “Habitualization devours work, clothes, furniture, one’s wife, and the fear of war….And Art exists that one may recover the sensation of life; it exists to make one feel things, to make the stone stony,”18 it is easy to see how habitualization is a part of the ratioid domain and defamiliarization, therefore, must be a nonratioid phenomenon. This is where Musil’s and Shklovsky’s way of thinking about the relation between defamiliarization and formalism seem to converge.

Defamiliarization and distributed attention

The reason why I spent so much time on the similarities between Russian formalism and Musil’s views is because I believe that Musil can help us interpret the concept of defamiliarization in a way that points beyond mere historical interest in this slippery notion.

Russian formalists are interested in the experience of the unfamiliar. But as encountering something that is unfamiliar seems to exclude the possibility of using well-practiced attentional resources for approaching them, it seems that encountering an unfamiliar object is likely to prompt one to attend to it in a distributed manner.

If we encounter an object that is not familiar, we don’t really know how to attend to it: what properties of it we should attend to and which ones we should ignore. We have no precedent of how to do this as we have not encountered this object before. So we have no blueprint to follow: we try out attending to all kinds of properties of the object—our attention is distributed. In other words, if we encounter an unfamiliar object, we tend to approach it with distributed attention.

Thus, the metaphor of defamiliarization could be interpreted easily in this framework. If an object is unfamiliar, we do not know how to approach it, therefore, we tend to attend to a number of its properties to figure out what to do with it or what can be done with it. If an object is familiar, we just attend to those of its properties that we need to attend to. When the Russian formalists describe the experience of the unfamiliar, they really describe a way of attending to this object that is less focussed than it normally would be. They describe distributed attention.

Let’s go back to the De Chirico quote in the previous section. The experience he describes is an experience that is very much akin to encountering something for the very first time. As encountering something for the very first time seems to imply some version of distributed attention (as we have no precedent to go by for approaching the object visually), these views seem to be consistent with the reading of defamiliarization as distributed attention.

And this is the point where we can return to the comparison with Musil. Here is what I take to be Robert Musil’s clearest characterization of the Other Condition, which appears to make the same connection with distributed attention:

Everything was shifted out of the focus of attention and has lost its sharp outlines. Seen in this way, it was all a little scattered and blurred, and yet manifestly there were still other centres filling it again with delicate certainty and clarity. For all life’s problems and events took on an incomparable mildness, softness and serenity, and at the same time an utterly transformed meaning.19

And Musil is not the only one who makes the connection between defamiliarization and distributed attention. Take Roger Fry’s famous description of his experience of watching a film (which also clearly indicates how close Fry was to Russian formalism before Bell’s influence):

If, in a cinematograph, we see a runaway horse and cart, we do not have to think either of getting out of the way or heroically interposing ourselves. The result is that in the first place we see the event much more clearly; see a number of quite interesting but irrelevant things, which in real life could not struggle into our consciousness, bent, as it would be, entirely upon the problem of our appropriate reaction. I remember seeing in a cinematograph the arrival of a train at a foreign station and the people descending from the carriages; there was no platform, and to my intense surprise I saw several people turn right round after reaching the ground, as though to orientate themselves; an almost ridiculous performance, which I had never noticed in all the many hundred occasions on which such a scene had passed before my eyes in real life. The fact being that at a station one is never really a spectator of events, but an actor engaged in the drama of luggage or prospective seats, and one actually sees only so much as may help to the appropriate action.20

Fry talks about irrelevant aspects of the perceived scene that, if the scene were observed in an everyday, habitualized manner, would have gone unnoticed. In other words, we are attending to aspects of the perceived scene we would not be attending to otherwise. Our attention is distributed attention. D. H. Lawrence gives a surprisingly similar characterization of the way attention is exercised when we engage with art: “The essential quality of poetry is that it makes a new effort of attention, and “discovers a new world within the known world.”21

The concept of defamiliarization is easy to confuse with an appeal to the “innocent eye,” which has a very bad reputation among the art historians of the last half-century (and rightly so). The difference between accounts of defamiliarized experience that emphasize the “as if encountered for the very first time” aspect of this experience and appeals to the “innocent eye” is the topic I now turn to.

Innocent eye versus unprompted eye

Statements of defamiliarization are often difficult to detach from other seemingly similar claims in the vicinity. Here is a famous quote by Stan Brakhage, avant-garde filmmaker par excellence, which I take to be a clear statement of the defamiliarization view:

Imagine an eye unruled by man-made laws of perspective, an eye unprejudiced by compositional logic, an eye which does not respond to the name of everything but which must know each object encountered in life through an adventure of perception. How many colors are there in a field of grass to the crawling baby unaware of “Green”? How many rainbows can light create for the untutored eye?22

Brakhage talks about the untutored eye and this sounds dangerously similar to the concept of “innocent eye.” In general, Brakhage’s emphasis on seeing something without employing concepts (like the concept of “green”)—an emphasis he shares with Aldous Huxley and many others23 —makes this general approach to defamiliarized experience be jeopardized by being subsumed under the heading of “innocent eye” tradition. I argue that the kind of picture Shklovsky, de Chirico, Proust, Musil and Brakhage are after is very different from the concept of “innocent eye” (in spite of some unfortunate statements that muddy the waters—including Shkovsky’s own, see his talk of “the sensation of things as they are perceived, and not as they are known”). But in order to do so, I need to say something brief about the concept of innocent eye.

A significant proportion of the art history and aesthetics of the second half of the 20^th Century can be described as a long fight against the concept of the “innocent eye.”24 The general idea of the innocent eye originates from John Ruskin, who famously said that:

The whole technical power of painting depends on our recovery of what may be called the innocence of the eye; that is to say, of a sort of childish perception of these flat stains of colour, merely as such, without consciousness of what they signify, as a blind man would see them if suddenly gifted with sight.25

A lot is going on in this famous quote. It is easy to spot the classic formalist assumptions: what we should pay attention to is the formal properties of paintings (for example, the “flat stains of colour”), not what they signify. But there is also an allusion to seeing things for the first time—an idea I want to salvage. To make things even more complicated, Ruskin’s subsequent discussion of how our non-innocent eye sees green grass as green even though it would look yellow if we were to focus merely on its look brings in various issues about color constancy. Nonetheless, what I take to be the core of Ruskin’s view is that what he labels the “innocent eye” is vision that is untouched by our knowledge of the world: of what we know about the color of grass in general, for example.

This is the view Gombrich and Goodman (among others) reacted against. Here is Ernst Gombrich: “The innocent eye is a myth. […] seeing is never just registering.”26 And “we must doubt all the more whether such an achievement of innocent passivity is at all possible to the human mind.”27 And again: “Reading an image, like the reception of any other message, is dependent on prior knowledge of possibilities; we can only recognize what we know.”28 Nelson Goodman is equally dismissive when he says that “the innocent eye is blind.”29 This stance is not exactly new, arguably it goes back at least to the early 18^th Century, to Jonathan Richardson’s Theory of Painting, where he says: “For ‘tis a certain maxim: no man sees what things are that knows not what they ought to be.”30 As Picasso summarizes, “our knowledge influences our vision.”31

From a psychological perspective, it should be clear that this “innocent eye” debate is about the cognitive penetrability of vision—about whether vision is an encapsulated process, sealed off from what goes on in the rest of the mind. And it should also be clear that the contemporary consensus is that things are not looking good for the proponents of “innocent eye” from that point of view. The view according to which perception is cognitively impenetrable—that is, according to which our eye is innocent seems to be plainly false.32 There are top-down processes that influence perceptual processing as early as the primary visual cortex33 or the thalamus.34 And these top-down just become more numerous and more diverse as the perceptual processing continues.

Here is a famous (and old) experiment):35 if we have to match the color of a picture of an orange heart to color samples, we match it differently (closer to the red end of the spectrum) from the way we match the color of a picture of some other, orange shapes. This shows that our recognition of the object in question (the heart) influences the color we experience it as having. A more recent experiment:36 Two pictures of identical (mixed race) faces were shown to subjects—the only difference between them was that under one the subjects read the word “white” and under the other they read “black.” When they had to match the color of the face, subjects chose a significantly darker color for the face with the label “black.” In general, one’s experience is not determined in a bottom-up manner by the perceptual stimulus: it depends on language, attention, the contrast classes and one’s expectations.37

It is important to highlight that in order to deny the innocence of the eye, it is enough to endorse a very weak sense of cognitive penetrability—so much so that it wouldn’t even count as cognitive penetrability under many formulation of cognitive penetrability because all it implies is that our perceptual experience is subject to top-down attentional influences —something even those who deny the cognitive penetrability of perception would accept.38 Even those who maintain that perception is not cognitive penetrable allow for top-down influences before and after “early vision.” But from the point of view of the “innocent eye” debate, it is irrelevant whether there is one brief part of perceptual processing that is sheltered from top-down influences. Proponents and opponents of “cognitive penetration” agree that our eye is not innocent.

It is important to point out that the distinction between “cognitive penetration” and the innocent eye claim is not the same as the somewhat technical distinction in analytic philosophy between conceptual and non-conceptual content. According to some philosophers of perception, perceptual content is conceptual: it has concepts as some of its ingredients. According to others, perceptual content has nothing to do with concepts: it is pre-conceptual or non-conceptual. Importantly, both of these views are consistent with cognitive penetration. Perception—whether or not it is conceptual—is influenced by our cognitive processes.

Again, there is no such thing as “innocent eye.” But the “innocent eye” claim is very different from the one I would like to salvage: the view that some experiences in the aesthetic context are akin to seeing things for the very first time. To make things simple, I will call this conception the “unprompted eye” and contrast it with the concept of “innocent eye.” Even if we assume, as we should, that the eye is never innocent—that our vision is always deeply influenced by our beliefs, previous experiences and expectations—it is still possible to see something as if we saw it for the first time.

But this distinction between the innocent eye and the unprompted eye will be mysterious unless we specify what makes the unprompted eye unprompted. And I argued that it is the exercise of distributed attention: if our eyes are “prompted,” we come with fixed preconceived ideas about how we should look at something. Suppose that you are hungry and you open your fridge. You will be likely to see the objects there as belonging to two different categories: edible and not edible. The way you’re attending to these objects is very much determined: you are zeroing in on very specific properties of the objects (roughly, their edibility or nutritional value) and you will actively ignore all their other properties. In other words, your perception is prompted: you will experience the objects in a way that is determined by pre-existing factors (like your quest for food).

Contrast this with experiences of a certain kind, where there is no such prompting of the eye: you have no fixed and preconceived way of visually approaching the object you’re looking at. In this case, your visual curiosity is not zeroing in on any fixed and predetermined set of properties: there is no pre-established prompt. You are trying to make sense of the object by trying out attending to a wide variety of its properties—just as you would do with an object you have never encountered before.

To be absolutely clear, we have very strong (empirical and philosophical) reasons to hold that both of these two kinds of experiences are non-innocent in the Gombrichian sense: both the experience of scavenging the contents of my fridge and the experience of artworks are influenced by various top-down factors such as my previous encounter with food and art, my beliefs about nutritional values and the exhibition I’m attending and my expectation of culinary or artistic value. But just because both experiences are cognitively penetrated doesn’t mean that both are prompted. The prompted versus unprompted distinction is made within the category of non-innocent eye (because that is the only kind of eye there is).

Post-formalist conclusion: The history of (defamiliarized) vision

What we got so far is an account of defamiliarized perception that avoids the pitfalls of the innocent eye tradition. Defamiliarized perception is seeing with unprompted eyes (not with innocent eyes, as no eyes are innocent). And seeing with unprompted eyes means attending in a distributed manner.

We have seen how these ideas of defamiliarization are related to formalism in Shklovsky’s writings: one way of achieving defamiliarization is by confronting the observer with challenging formal properties. But the idea of defamiliarization interpreted as unprompted vision and distributed attention makes it possible to make connections to another streak of formalist thinking—and one that seems more relevant these days, especially in the context of post-formalism:39 the historicity of vision.

According to an influential view within art history, the way the ancient Greeks saw the world was importantly different from the way we now see the world and part of what art history should study is exactly how human vision has changed in the course of history. If the ancients did see the world differently from the way we do now, then in order to understand and evaluate their art, we need to understand how they perceived it (and how this is related to the way they perceived the world). Thus, so the argument goes, the history of vision is a necessary precursor to art history. This general line of argument goes back at least as far as Tacitus,40 but it has become one of the most important premises of art history and aesthetics since the early 20^th Century.

The general idea behind the history of vision claim is that visual experience changes in various ways in the course of history. We should, therefore, not assume that people in ancient or medieval times perceived in the same way as we do now. Further, one important aspect of understanding the art of earlier times is to understand the way people perceived artworks then.

The most explicit statement of this claim comes from Heinrich Wölfflin, in one of the best-known passages in the history of art history: “Vision itself has its history, and the revelation of these visual strata must be regarded as the primary task of art history.”41 While Wölfflin’s provocative statement has become an important slogan for generations of art historians, the general idea that vision has a history had another major and in some ways even more influential proponent in the context of the turn of the Century German/Austrian art historical tradition, namely, Alois Riegl. Riegl’s main guiding principle in The Late Roman Art Industry was that the way people perceived the world in ancient times is radically different from the way we perceive now. More precisely, he argues that ancient people saw only “individual self-contained shapes”—and this explains some crucial features of their representational visual art.42 But even within the ancient era, he hypothesizes that the Egyptians perceived the world differently from the way the Greeks did and the Greeks differently from the way the Romans, especially the late Romans, did.

Perhaps the real influence of the history of vision claim was provided by the application of this general and abstract idea to the question of modernity—an idea that is present in the work of a very diverse group of thinkers: Charles Baudelaire, Georg Simmel, Friedrich Nietzsche, Laszlo Moholy-Nagy, Siegfried Kracauer, Lev Malevich, to mention just a few.43 But it was Walter Benjamin who made this application of the history of vision claim most explicitly and most influentially. Inspired by Riegl’s general claims about the history of vision,44 he says:

During long periods of history, the mode of human sense perception changes with humanity’s entire mode of existence. The manner in which human sense perception is organized, the medium in which it is accomplished, is determined not only by nature but by historical circumstances as well.45

Alois Riegl’s influence is clear in these passages and especially in the ways in which Benjamin applies these general ideas to specific periods in art history: “The period of migration in which the late Roman art industry and the Vienna Genesis came into being, had not only a different art, but also a different perception from classical times.”46

Benjamin’s main interest, however, is not the late Roman art industry, but the change in art and perception that happened at the end of the 19^th and the beginning of the 20^th Century. And Benjamin’s claim is that modernity is a change in “sensorium” and this idea has become one of the guiding principles of theorizing about modernity. For Benjamin, this change in “sensorium” was brought about by technological changes. Not just the changes in the streetscape around us: the speeding cars and the skyscrapers that Malevich and Baudelaire like to emphasize, but also the technological changes in art itself. As Régis Debray summarizes, “Photography has changed our perception of space, and the cinema our perception of time (via montage).”47

Still, the application of the idea of the history of vision does not stop, and does not begin,  with modernity. As Jonathan Crary argues at length, the general mode of perception may have undergone some important form of change already in the first half of the 19^th century.48 And theorists of postmodernism rely on the principle of the history of vision as much as theorists of modernity do. Frederic Jameson, for example, argues that postmodernism offers “a whole new Utopian realm of the senses.”49 The premise all these arguments share is that history, and art history, can be understood, at least partially, as the history of perception. This assumption is so deeply ingrained in much of the discourse on 19^th and 20^th century art and culture and in (at least some branches of) art history and aesthetics that it has been taken for granted without further discussion. As Whitney Davis summarized, “according to visual-culture studies, it is true prima facie that vision has a cultural history.”50 And one of the guiding ideas of post-formalism has become the claim that vision (and imagination) has an (art) history.51

How does the concept of defamiliarization (understood as distributed attention) fit into this framework? The concept of visual attention has not played a significant role in the history of vision debate, not even in its post-formalist version that puts the emphasis not on vision, but other Sehformen, especially imagination. One exception is André Malraux, who famously wrote that: “The creation of every great art is inseparable from [the] metamorphosis in the manner of seeing, which does not properly belong in the realm of vision, but of attention.”52 But if vision has a history, we can reasonably assume that visual attention (understood as part of perception) also has a history. In fact, given the rich empirical literature on cross-cultural variations in visual attention,53 the default assumption would need to be that attention is not a cultural universal—so we should expect not just cross-cultural, but also historical variations.

It should be clear from the discussion of defamiliarization that the kind of distributed attention I have talked about played an important role in the early 1900s and probably also some time before that. But it is not clear that, say, 15^th century observers exercised distributed attention in the same way as early 20^th century observers did. And even if they did, it is not clear that looking at pictures or artworks in general was considered to be the appropriate and even recommended context for such defamiliarized vision and distributed attention.

The present paper is not the place to outline such historical project, but it is worth noting that none of the “visual skills” Michael Baxandall talks about in his thorough study of how 15^th century Italian humanist observers looked at pictures implies (or is even consistent with) the kind of distributed attention that is entailed by defamiliarized vision.54 If this is so (and I won’t pretend that I’ve established this historical claim), then distributed attention does have a history.55 The history of (distributed) attention would be an important step towards understanding the history of vision and, as a result, should be of special interest to post-formalist art history.56

If many writers have described the many technical variations and contextual deployments of the magic lantern, her book takes a more contrary, deconstructive tack: how has the magic lantern not so much illuminated, than actually structured, the Western modern subject? Instead of dealing with the magic lantern as an object in separate fields of art, film or pop-cultural history, she finds its core operative mechanisms articulated in philosophy and science. Some readers will find this book’s methodology confusing—where is the art? the history?—but the payoff lies exactly in its Visual Cultural flouting of disciplinary boundaries. Not only a media-archaeological take on magic lanterns, the book also contributes to the ongoing decolonization of psychoanalysis.1

Casid’s argument is that magic lanterns are not simply mechanisms of transparency and reason, but work in more complex ways. As a bright light is projected across a distance, and a flat surface is lit up, a corresponding disavowal—of gender, race, or colonial imbrication—occurs in the self. A key component of her argument is the Freudian concept of “projection,” which involves a process of repudiation. At the very moment we project or cast a thought or emotion, a simultaneous and corresponding denial or rejection occurs, to which we are oblivious. The magic lantern, then, is at best a paradoxical object, at worst a ruse: seemingly offering up reason and clarity, it casts both audience and lantern manipulator into shadow. At times Casid’s book called to mind Hubert Damisch’s The Origin of Perspective, with its deft tracing back of a Lacanian split in the Cartesian subject to the origin of perspective.2 Perspective, camera obscuras, or magic lanterns were models of vision, in the way that they functioned schematically in the mind.

The first two chapters were, for me, revelatory, even stunning. Chapter one connects projection devices to what she calls “paranoid reason.” Going through a familiar list of names in the history of the camera obscura or magic lantern—della Porta, Kepler, ‘s Gravesande, Kircher, among others—Casid provides a rich counter-history of such “magical” devices. Many of us are familiar with the camera obscura as a metaphor of human consciousness; she argues, contrarily, that “the production of the phantom subject of rational vision depended on the vigilant demonstration of and subjective internalization of a paranoid version of projection.” (30) So if, in the old manner of looking at magic lanterns, the monster in the projected image alluded to a purifying of vision, a purging of sin, for Casid, the monster allegorized the split “in” the projecting subject. One externalized what one didn’t or couldn’t admit in oneself. (One important move she makes is to combine camera obscura and magic lantern historiographical traditions, which allows her to move beyond Jonathan Crary’s now-tired thesis of the long, 17th to 19th-century genesis of the disciplinary subject of vision.3) Chapter two is about “the extent to which an ‘Asia,’ ‘Africa,’ or ‘America’ failed to function as a blank screen that stayed safely ‘over there’ because of the ways the device was so fused and infused with what it mediated.” (31) Projection here also refers to the way that Europe geographically projected its power onto ‘its’ colonies. The chapter explores several curious cases in which projection devices were brought to far-flung places, whether the 1760s Obeah trials in Jamaica or Houdin’s 1856 trip to Algeria. Reason, the chapter argues, wasn’t formed in the West and then disseminated out to the colonies. No—the subject of European reason began with the prior disavowal of the colonial Other. If, at first, projective devices seemed to be brought to the colonies to dispel non-Western magic and proffer the superiority of Western science and reason, what emerges is a more fascinating story in which one form of magic faced off with another. Such moments in her book are forcefully decolonial, and recall the work of Michael Taussig.4

Chapter three considers the psychoanalytic process of introjection—the taking-in of objects into the self or psyche. Hence it is often thought of as the opposite of projection. From Freud’s Civilization and Its Discontents (1930), she salvages the fact that introjection was “a machine of social and political discipline.” (128) “[The] use of the magic lantern to produce images of phantoms, devils, and monsters endeavored to dissociate,” she writes, “through a complex manipulation of techniques of mimicry, the demonstration of the way in which images are cast into the dark room of the mind from the fear of introjection as demonic infection or false belief .” (129) Let play these dazzling images of monsters, in other words, to keep at bay a more unsettling spiritual contamination and invasion. Of the Tipu Tiger case, she shows how this “hybrid” tiger-organ “turns inside out the boundaries of Indian and European, top and bottom, masculine and feminine” as a transculturating “heterogeneous machine” (139)—a strikingly expansive take on the seemingly inward phenomenon of introjection.5 If chapter four, on early photography as a kind of “shadow writing,” especially in the work of William Henry Fox Talbot, feels a bit familiar, perhaps this is because of the intense reexamination Talbot’s work has received by historians of photography.6 Chapter five delves into Newton’s prism experiment, which he famously called “the crucial experiment,” and uncovers what she calls a “prismatic politics”—a pluralist politics of race and gender—via Frank Oz’s 1900 children’s novel The Wonderful Wizard of Oz. Casid’s explication of Newton’s experiment is convincing, and her treatment of Oz innovative and refreshing.

Projection in psychoanalysis refers to a form of defense whereby unwanted feelings are displaced onto another person, where they appear as a threat from the external world. So, for example, feelings of guilt on the part of the colonizer are transferred onto the colonized, who are taught that they are responsible for their own fraught condition. Racists project their own faults onto the group that they revile, occluding the fear and anxiety in themselves.7 Projection had a flurry of meanings in psychology and psychiatry in the late-19th century; out of this Freud forged a “strong,” synthesizing definition that took projection to refer to the displacement of thought or affect onto another person, and the corresponding disavowal that occurs in the self. (Seen especially in cases of paranoia, such as judge Daniel Paul Schreber.) Projection, always-already defensive, harks back to an early time in the psyche’s history. The strong definition, then, involves the constitution of the ego. But Freud also toyed with a “weak” definition: when projection is generalized to refer to all manner of throwing ideas onto people and things.8 We see this in Freud’s long quote from The Psychopathology of Everyday Life (quoted on Casid, 14), where he points out how primitive people misguidedly projected their ideas of religion onto the natural world, naïvely seeing inanimate objects as imbued with life and agency. These primitives were hampered by their own false consciousness, and refused to recognize the truth or reality of the object world. At these moments, Freud pulled away from the strength of his own insight—the strong definition of projection—and bought into the late-19th century Tylorian misunderstanding of animism, which was itself premised on a long mechanistic understanding of matter as “dead.”9 In other words, Freud became a victim of projection. Casid’s argument would have been strengthened if she had clarified this confusion in Freud’s own writing (cf. 39-40).

While the production of projected image and rational subject is expertly done, the book, I think, lacks an adequate account of the phenomenology of viewing projected imagery. Casid mostly deals with the spatial dynamics of projection, and not enough with time.10 It is easy to conceive of the instantaneity of projected imagery—the flash on the wall—but was such imagery experienced durationally? Was the power and effect of projective imagery cognitive rather than affective? The challenge, of course, is to describe the time of the ego. Teasing out the nuances of temporality may also help explain why so many contemporary artists who use projected imagery are eager to exploit the seam between old and new media or technology.11

While Casid’s argument certainly involves the construction of models of vision that are prior to local circumstance, the question begs as to when they are deployed in actual contexts: do all image projections entail disavowal? Did projection simply feed the education of spectacle—did it always come down, or back, to discipline? Viewers ‘caught’ in the projected image (as in figures 6, 17, 19) may have been desiring something other than what Casid suggests. We know that 18th- and 19th-century viewers often wanted to interact with the projected image, and imagine themselves ‘in’ the scene: compare the way viewers, in front of a mirror, inhabited Jacques-Louis David’s Sabines.12 Sculptures viewed by torchlight exerted a strange fascination—perhaps a form of virtuality we no longer have access to.13 These ambulatory viewers may have been trying to retrieve what was disavowed in themselves; introducing a third term between light source and screen may have been a way to break out of the destructive dyad of the Lacanian Imaginary.14

I hesitated at Casid’s phrases “projecting the projection” and “recasting” (e.g. 7, 33, 41), by which she means the reversal of the projection onto the projectionist, “sustaining an analytic reckoning with those abjected and opaque aspects of the subject” (14). If projection is premised on a founding disavowal, can such conscious reversal ever bring the subject back into light? I also wonder if bringing back the cast out or disavowed simply repeats the regressive logic of projection (e.g. 29). Perhaps a better analogy for psychoanalytic insight isn’t the shining of light onto darkness; it is to come to terms with the constitutive dimension of shadow. Shadow allows the presentness and presence of light. Casid’s “prismatic” politics seems to draw on a particular ecstatic reading of Deleuze and Guattari (cf. 86-7)—but I remain skeptical.15 So acutely deconstructive of identity politics in its first half, it would be a shame if the book ended up in a simple rainbow coalition politics in its second.16

Though she works hard to uncover the politics of projection, Casid ends up, I think, making projection’s relation to the subject mechanistic. Because if all image projections do this one thing to every beholder—subject one to paranoid reason—they are doing something, rather than meaning something, to us.17 With art that formally adopts image projection, one cannot know in advance, cannot discover, if the projection is paranoiac (rather than ordinary)—just as one cannot guarantee whether the work will count, as art. This is also to say that not every camera obscura or magic lantern suffers from projection, in the psychoanalytic sense. No doubt the artists discussed by Casid who adopt magic lantern techniques (181-94, 236-44) were shrewd and reflexive enough to unbind structures of reason, to turn projection on itself. If, indeed, the dialectic between illumination and darkness were built into artistic intention—and intention can arrive in the making of a work—this would have resulted in a more complex work of art. But rather than assume that the merging or confusion of body and world (e.g. “the real, enfleshed encounter with the palpable stuff of illusion,” 241) will result in a more capacious or critical viewer, it is the point of intention to figure the paradoxically non-programmatic constitution of the subject in the world.

By essentially offering a media-archaeological pre-history of the Kantian distinction between phenomena and noumena, appearances (Erscheinung) and the thing-in-itself18, Casid’s book bravely asks how modern history would look if reason is seen as the effect of a prior irrationality. Projection, Freud believed, played a crucial role in constituting the boundary between the ego’s inside and outside. To some degree, we are all projecting creatures ourselves.

Needless to say, here I cannot review the many formalisms in art history (let alone artwriting more broadly defined) by proceeding text by text and writer by writer, even if I were competent to do so. In addition to Wölfflin, and speaking only of writing published in his lifetime, one would have to deal with texts by Aloïs Riegl, Wilhelm Worringer, Clive Bell, Roger Fry, Albert Barnes, Hans Sedlmayr, Henri Focillon, and Clement Greenberg among others, not to speak of practices on the part of artists such as Wassily Kandinsky and Hans Hofmann and art teachers such as Denman Ross and Jay Hambidge—a hugely diverse group. I cannot attempt even the  most minimal exegesis that would be needed. Selective overviews are readily available.2

Instead I will offer an analytic commentary on the claims of art-historical formalism as an analytic construction in art history, and possibly as a tactical methodological construction. In turn this will enable me to situate “post-formalism” as a philosophy of art history today.

I take a special interest in Wölfflin’s formalism because it integrated the crucial post-formalist manoeuver as I will identify it: the form-making capacity of human intuition—especially if intuition simply is form making—must be historicized. Stated most simply, form is a history. Of course, there have been several “post-formalisms.” For example, art critics and art historians who reacted specifically against Greenberg’s modernism might be called post-formalists, and sometimes they have been. (Indeed, one might even argue that Kant inflected his formalism—a transcendental psychology of the form-making capacity of human intuition—with a post-formalist psychology of the form that is constituted specifically in aesthetic judgment of works of fine art.) In the present essay I take the view that post-formalism in art history became possible at the very moment of the principal inception of formalism in the work of Wölfflin. I will be concerned mostly to say how this possibility has been realized in recent years.

§1. Formalism and Post-Formalism. The overall gist is this. Post-formalism as I will describe it attempts to shift ground from the history of artworks to the history of visual imaging and imagining—what Wölfflin in 1915 in Kunstgeschichtliche Grundbegriffe called Sehformen (a neologism often translated as “ways of seeing” or “modes of vision”) and in 1923 revised to Formen der Vorstellungsbild (“modes of imagination” in the English translation of 1932, but probably better rendered as “forms of imaginative appearance” or even “forms of representationality”).3

This shift as I will describe it also involves a movement from an aestheticist “formalisticism” (though this is not a term used in the current self-description of art history) to formalism proper, a sub-Kantian psychology of intuition—of the “forming” activity of sensibility. Formalisticism effects a partial reification of what I will call “formality” (the apparent configuratedness of material things) as “form” in a specialized art-historical sense: form is a unity organized materially (and visibly apparent) in man-made things on the basis of essential habits of intuition, notably (for Wölfflin) sensitivity to “rhythm,” though Wölfflin also believed that some of these habits of intuition vary historically.4  (The things in question, of course, need not be artworks, and in Kant’s account of aesthetic judgment they need not be man-made things at all; still, works of fine art have sometimes been deemed to be impossible without form.) What many writers usually denote by “formalism” in art history is a wholesale reification—full-bore formalisticism in my sense. It puts form “in” the things, converting it from an aspect seen (or formality) to an attribute made. Instead of speaking of the formality of the thing for us as we see it, then, we come—in the formalist conversion forced by the reification of formalisticism—to speak of being aware of its form. (I expand on this briefly in §2 below.) It follows analytically that “post-formalism” in art history and artwriting would certainly try to get beyond formalisticism. Whether and how it gets beyond transcendental formalism is another question.

Given its post-formalisticism, post-formalism (whether post-Kantian or not) should be able to state the relation between artworks and Sehformen. Or at least it sets out to do so. The crucial point here, in my view, is that seeing, visual imaging, does not just have a history an sich or “in itself”—the history of the “optische Schichten,” or “strata of vision,” that Wölfflin said it was the first-order task of art history to uncover. (I will call this “vision historicism”: “Das Sehen an sich hat seine Geschichte und die Aufdeckung dieser ‘optischen Schichten’ muss als die elementarste Aufgabe der Kunstgeschichte betrachtet werden.”5) Imaging also has an art history, which I have tried to describe as the “successions” and “recursions” of “formality,” “pictoriality,” and other aspects of things made specifically to be seen in the seeing that sees them. In a sense, then, post-formalism is also pre-formalism—a grounding of the history of artworks in the art history of seeing. I will address this matter in the middle sections of the essay.

The challenge today is that the art history of seeing (such as it is) is confronted by an expansive foundationalist neurology of seeing that is merely aesthetic, and in a sense aestheticist within its domain, or formalist—a visual neuropsychology that is now sometimes called “neuroaesthetics.” Post-formalism might best be defined, then, as post-neuroaestheticist. (In addressing this side of the story, the present essay follows on my “Neurovisuality,” published in nonsite in 2011.6)

Wölfflin tackled a similar problem in the terms of the psychophysiology of his own day, as did some of the other early formalists already mentioned. (Some were expert psychologists, and formalism is a psychological theory—a method of “virtual historical psychology,” as I have put it elsewhere.7)

But he did not quite solve it because of his tendency toward formalisticism, and the circularity of the analysis that resulted: the form of artworks (identified in the reification of formalisticism) became Wölfflin’s evidence for the historicity of vision (an open question for formalism in psychology) at the same time as the historicity of vision explained the form of artworks. Post-formalism seeks to avoid this circularity. It tries to state a historical relation and therefore in my terms a recursive relation between the form of artworks and the historicity of vision, at least so far as we limit ourselves (rather artificially) to formality or formal aspects. I will deal with this matter in the final section of the essay.

§2. Formalism Proper and Formalisticism. In Kant’s system and in the Kantian tradition—what we might call formalism proper—form belongs to intuition. As Kant expressed the point in the First Critique, form can be defined as “the manner [die Art] in which we are affected by  objects”—die Art as distinct, that is, from our knowledge of “content” let alone our empirical contact with the real matter of the objects. As Robert B. Pippin has put the point in Kant’s Theory of Form, in Kant’s doctrine “form is inextricably linked with the knowing subject”; it is the condition of the possibility of our sensory awareness of anything.8

Responding to the First Critique, in 1787 Johann Georg Schlosser sharply criticized what he called Kant’s “Formgebungmanufaktur,” his “pedantic” invention of a forming capacity of the mind—what we might call Kant’s “mere formalism” in transcendental psychology, or formalism proper. In 1796 Kant responded equally sharply and reiterated his primary claim: “If the thing is an object of the senses,” he insisted, “so its form is in intuition (as an appearance).”9  Responding to the First Critique, in 1787 Johann Georg Schlosser sharply criticized what he called Kant’s “Formgebungmanufaktur,” his “pedantic” invention of a forming capacity of the mind—what we might call Kant’s “mere formalism” in transcendental psychology, or formalism proper. In 1796 Kant responded equally sharply and reiterated his primary claim: “If the thing is an object of the senses,” he insisted, “so its form is in intuition (as an appearance).”10

In the terms of the aspect psychology that I have adopted in A General Theory of Visual Culture, constitutive form (when converted from space and time to color, shape, etc.—to color, shape, etc., as constituted in space and time in intuition) is “formality”: the apparent configuratedness (the “appearance” of configuratedness) in things. Needless to say, some of these things have been made specifically to be seen: they have been colored and shaped by someone with a “forming” sensibility who has produced them for the prospect of a “formality”—seeming form—that they might afford to sight. But it is crucial to preserve the theoretical sense in which this relay is a complex recursion—the succession of formality in the visibility of things in historical visualities.11

In the reification, form in this sense—Kantian forms-in-intuition; formal aspects (“formality”) as described in aspect psychology—gets transferred or translocated to the object, as if we perceive—receive or pick up—its “form.” The apparent configuratedness of an object as we constitute it becomes our seeing of its configuration or formedness. Many commentators on art history and criticism use the term “formalism” to designate the reification (or objectification). But it might better be called “formalisticism” (as I will do here) because formalism need not objectify form. It is “formalistic”—but not inherently “formalist” in the transcendental-psychological sense—to proceed analytically as if form subsists primordially in the object, even if it was put there (as one might say) by the intuitional activities of its maker and beholders in visualizing or visibilizing it.

It should be noted, however, that formalistic reifications are not wholly unwarranted in art history. At least, they should not be unexpected; there is a credible reason for them. In its methodological self-invention as a second-order archaeology, art history conjures a real thing (i.e., “form”) to dig up for display to its “looking” when the first-order object (i.e., “the manner in which objects affect us”) cannot be proffered as any (properties of a) physical thing. Formalisticism literally creates a material object as objectified form(ality). Philosophical tradition has enabled art history to trade on this shuttle, as Kant himself had done. As Pippin has put it, “should the object of knowledge be an object of the senses, its ‘form’ is not a property of the object, but is the intuiting activity of the subject. Nevertheless, by his association here with the tradition, Kant admits that knowledge of this form will reveal something like the ‘essence’ of the object—for Kant the universal conditions necessary for it to be an object at all (space and time).”  By “association with the tradition,” Pippin specifically means Kant’s starting-points in Aristotelian and scholastic philosophies. To use the very words with which Kant opened the remarks of 1796 that I have already quoted “[a] the essence of things consists in their form (forma dat esse rei, as is said by the Scholastics) [b] insofar as this thing might be known through reason.”13  As I understand the matter, the passage from (a) to (b) qualified the tradition in Kant’s special philosophical terms. Marking the beginning of his formalist transcendental psychology, it introduced what Kant called his “Copernican revolution” in philosophy. But (b) is not intelligible without (a).

Now let us suppose that art history (not to speak of art criticism) understands its primary objects to be material works of art and perhaps, in an extension, items of visual and material culture that need not be identified specifically as works of art. (Recall that Wölfflin said that these things are not the theoretical object of the project of art history, at least considered in themselves; the “elementary datum” of art-historical inquiry is their optische Schichten. In principle this is vision historicism in formalism proper.) In turn it is obvious that art history at that level or in that register can (and perhaps must) proceed archaeologically from (1) form “in” the object to (2) “form” as intuited (“seen”) by the beholder-historian to (3) “form” attributed to the intuiting activity of the historical maker, though psychologically the order—the business of Kant’s transcendental psychology and other formalisms proper—is partly inverse to this (i.e., [3] to [1] to [2]).

In other words, the reification might have a tactically valid methodological status: formalisticism functions as a forensic archaeological method for an art history framed theoretically as a history of forming activity in sensibility. The problem arises when the methodological reification (that is, the forensics of the recursions of formality) is treated as primary foundation—gets ontologized. This risks mischaracterizing the intuiting activity of the subject (“forming” the object in space and time) as the intuiting of objective form. More exactly for my purposes, it risks conflating them in such a way that the “elementarste Aufgabe” of art history—the historicity of forming, and its specific historicality as it were—becomes hard to identify, and perhaps will be cut out of the story entirely. In §4 below I will return to this matter in the case of Wölfflin’s formalism.

There is another reason for art-historical formalisticism aside from its status as a forensic method of formalism proper. The contiguity between the hugely general terms of Kantian formalism (laid out in the First Critique) and the highly specialized terms of Kantian aesthetic judgment in constituting the “perfected ideals of beauty in the fine arts” (spelled out in one short section of the Third Critique) may have motivated some formalisticist reifications of art as an empirical object specifically made to affect us in its form, that is, as a putatively aesthetic object. But the contiguity does not fully justify the reification. Kant did not simply identify form and artwork as the subjective and the objective faces of sensibility. If anything, the perfected ideal of beauty in art is form reformed—detached from any interest we have in the empirical existence of the object. And this reformed form (the specifically “normative” form) is not “in” the object that is an ideal artwork, just as it was not “in” the images of things that we find appealing from which the ideal artwork is built. It is in sensibility. Still, because the reformed form—ideal art—is not a mere emanation of immediate subjective intuition (instead it is a product of “subjective universality”) it has some claim to be regarded as an objective form afforded to the subject for his or her intuition. Perhaps this model of the history of the forming of artworks as idealizations can be reconciled analytically with the history of their optische Schichten. After all, Kant admitted that his model of the perfection of ideals of beauty in the fine arts invoked an “optical analogy” of the intuitive superimposition of visual images of appealing and attractive things. Taken literally, then, the consolidation of perfected ideals of beauty in an artwork simply is the historical constitution of an optische Schicht as a particular aesthetic horizon. But it will be difficult, maybe impossible, to make this case by indulging the formalistic reification. 

§3. Post-Formalism Proper. In the sense that I use the term here, the term “post-formalism” first cropped up (for me anyway) in David Summers’s book Real Spaces, published in 2003—a post-formalist world art history (as he called it) partly intended to describe the “rise of Western modernism.”15  Real Spaces is not only art-historically post-formalist. Summers not only tried to get beyond the formalist reification in art-historical formalism, or what I have been calling formalisticism. It is also philosophically post-formalist, that is, post-Kantian. Summers did not want to start from the forms of intuition (i.e., space and time) as described by transcendental psychology, or to be required always to return to them as the ground. Even Martin Heidegger had done so, though he thought that he had analytically managed to constitute “original time”—over and against space—as the privileged or primary route of the understanding (and its final existential constraint). In turn, Ernst Cassirer severely criticized this view—what he took to be extreme onesidedness. Partly in this light, Summers hopes to redescribe space (one of the “two basic pillars” of the understanding, as Cassirer insisted after Kant and against Heidegger) in terms of what he calls “real space,” that is, the geometrical-optical organization and sociocultural architectonics of topographical “place” and the man-made configurations set up in it, notably pictures.16 He takes his analysis to break decisively from art histories that reify form, thereby opening new art histories—new lineages, for example, of the “rise of Western modernism” that has been treated so often by art historians.

Art historians have usually connected Western pictorial naturalism to the forms of Classical Greek sculptural contrapposto and pictorial construction of fictive depth. In one of the most innovative of his new histories, Summers partly derived it instead from planar constructions of ancient Egyptian depiction. As he put it, “Egyptian painters and sculptors made choices that were to establish the basis of Western naturalism . . . accomplished by the development of planarity into the virtual dimension, with consequences reaching to the present day.”17 I want to emphasize the striking novelty of this art-historical claim. Ancient Egyptian depiction has sometimes been said to constitute “the origins of Western art.” But it has been far more usual (largely under the influence of Hegel’s aesthetics) formally to contrast Egyptian “symbolic” or “conceptual” procedures in picture making with the naturalism of Classical Greek art as the formal “birth of Western art,” that is, of one of its characteristic formalities in visual culture.18

According to Greenberg, for example, “of all the great traditions of pictorial naturalism, only the Greco-Roman and the Western can be said to be sculpturally oriented. They alone have made full use of the sculptural means of light and shade to obtain an illusion of volume on a flat surface.  And both these traditions arrived at so-called scientific perspective only because a thoroughgoing illusion of volume required a consistent illusion of the kind of space in which volume was possible.”19  “Sculptural means of light and shade,” “illusion of volume on a flat surface,” and “illusion of the space in which volume is possible” are overly formalistic—reifications of formality. We can convert them (post-formalistically) to modeling in light, flat surface seen as voluminous, and virtual volume seen in space under specified optical, geometric, and architectonic conditions in real space; Summers set out a demonstration in terms of the “axis of direct observation” that he took to be maintained in Egyptian “planarity.” If we do this—I will not rehearse Summers’s careful analysis here—we can see how the Egyptians produced illusion of volume set it in a space in which it is possible. To be sure, this virtual space was organized on the metricized “virtual coordinate plane” that Summers has identified in Egyptian depiction, not in “scientific perspective” in the sense that Greenberg meant. Still, “metric naturalism” and optical naturalism can now be described more accurately relative to each other, and no longer as mere opposites; Summers’s explications of “relief space” in Classical Greek pictorialism, of the “optical plane” in medieval pictoriality, and of “painter’s perspective” in Italian Renaissance pictoriality benefit from—depend on—his antecedent identification of the virtual coordinate plane in Egyptian pictoriality.20  Any phenomenal, technological, and sociohistorical continuities between metric naturalism and optical naturalism can be explored archaeologically (for example, by investigating the transfer of technologies for constructing proportions on the plane from Saite Egypt to Archaic Greece). They need not remain juxtaposed as polarized à la Greenberg, who pushed each formalistic category of immanent form to its “full,” “thoroughgoing,” and most “consistent” self-realization. Real Spaces exemplifies Summers’s training in the traditions of archaeological Strukturforschung, which I share: my first undergraduate teacher at Harvard, G. M. A. Hanfmann, was a student of Friedrich Matz, the best mid-twentieth-century exponent of Strukturanalyse, which he imbibed from Sedlmayr and Guido von Kaschnitz-Weinberg. Strukturanalyse might be the first post-formalism in art history.21 But Summers’s teacher at Yale, George Kubler, was right to say that Matz and others practiced a kind of Weltanschauungsgeschichte—“idealist archaeology” as Kubler called it. That is, they tended to derive the optische Schichten from preexisting Weltbilder, or at least from a “central pattern of sensibility” to be found among poets and artists living at the same time. This was more or less empty. Analytically Weltbilder simply are the optische Schichten.

By contrast, Kubler saw his overall project in examining “some of the morphological problems of duration in series and sequence” as making good on a question that art historians had put aside “when [they] turned away from ‘mere formalism’ to the historical reconstruction of symbolic complexes”—that is, to iconology. His scare quotes imply, it seems to me, that for him “mere formalism” had certain analytic advantages, at least when it was reconstituted as archaeological method. It had advantages, that is, if one could shake off its tendencies to universalize about  sensibility (as in Strukturforschung) and to indulge formalisticism in the ontology of (art)making—to treat form as a feature of the artifact, often unique, rather than as an emergent morphological boundary of its artifact-type (or “form class”), in which each artifact is “formed” as much under constraints of its serial position as in virtue of its maker’s sensibility.22

Following Kubler, Summers works inversely to “idealist archaeology”: a worldpicture emerges historically within an optical stratum in human imaging, or as it. (The two most general strata identified in Real Spaces, “planarity” and “virtuality,” would seem to be modifications on the second of the five polarities Wölfflin had set out in Kunstgeschichtliche Grundbegriffe to describe the realization of form, namely, plane [or surface] and recession [or depth].) This is a history in imaging, not—or not essentially—a history of ideas or beliefs or worldviews that precede and determine it. As Wölfflin had already urged, excavating this history is the primary task of art history. And Summers did so in a “post-formalist” way, as he saw it, because for him the serial making of assemblages (or environments) of things in real spaces in history is the elementary (quasi-Kublerian) datum of our archaeology, not the form of the artwork as put into it by a spatializing sensibility said to precede the agent’s experience in the world and especially the agent’s experience of socially shaped topography—of particular cultural “places” in “real space.” Stated another way, Summers wants to “world” form (that is, put it in the world) whether or not his “world art history” succeeds in each and every one of its genealogies and chronologies of artmaking considered globally.23

Of course, Summers’s Real Spaces is not the only example of recent post-formalism, though it is notable because it explicitly described itself as such. By now many teachers of art history must have gotten used to hearing students call themselves “post-formalists,” though many of these teachers (including me) probably described themselves as “anti-formalists” when they were students. When I ask my students what they mean, some cite Summers. Others cite Hans Belting’s Bild-Anthropologie of 2001: it proposed Entwürfe für eine Bildwissenschaft said by him to be an “anthropology.” There is considerable overlap between Belting’s and Summers’s descriptions of the functions of images (both writers deal extensively, for example, with substitutive effigies) and in turn with Horst Bredekamp’s recent Theorie des Bildakts, also markedly post-formalist. I will not rehearse these projects here, however. Suffice it to say that the art-historical attainments of post-formalism are now quite clear. In each case—Summers, Belting, Bredekamp—we find the art historian identifying affiliations or even historical connections and cultural interactions between productions that formalism had sometimes overlooked—could only overlook when formalistic.

Certainly Belting and Bredekamp proceed post-formalistically. In Belting’s account of the force of effigies, what counts is the functional substitutability of material effigy and prototype, even if there is no visible “formal” congruence between the visible features of the former object (a thing with “form”) and the latter (which may not be visible at all). It is not so much that form is  objectified in the effigy, though effigies have formality. The prototype is objectified when the effigy functions. Belting insists that he does not identify images with pictorial artifacts; indeed, he identifies images primarily with the body and mental imaging. This could allow formalism proper. But it should be “anthropological.” The forms of intuition do not precede bodily awareness in the social field. In anthropology, intuition simply is the body’s imagemaking in the social field.24

If I have any objection to Belting’s Bild-Anthropologie, it is simply that his historical anthropology might not be radical enough to displace transcendental psychology. Of course, Belting’s anthropology situates the transcendental activity of intuition within the bounds of historical cultures that he differentiates one from the next. But this anthropology tends to reify each cultural tradition as a priori collective intuition within its bounds—a “common pattern of sensibility” as Kubler put it in describing the idealist archaeology of Strukturforschung. Indeed, it objectifies that intuition as supposed cultural tradition. Therefore it does not quite issue in a radically historical analytics of imagemaking.

For his part, Bredekamp collates objects that formalistically would not be affiliated as iterations of the same kind of artifact. Rather, they replicate images. To explain this history Bredekamp would appeal as much to the agent’s visual experience (in complex interaction with fantasies, desires, beliefs, and concepts) as to formal sensibility. Indeed, the agent’s formal sensibility  simply is this history in his or her visual experience. Bredekamp’s history of Charles Darwin’s diagrammatic visualizations of transmutation exemplifies this theory of the action of images: set down in Darwin’s notebooks of the mid 1830s, the diagrams (according to Bredekamp’s archaeology of the replications) relayed images dating to earlier periods in Darwin’s experience and in the experience of other agents whose imagemaking he encountered historically. According to Bredekamp, then, the image is not so much formed in intuition as primary; rather, it is formed in the relay of images. Better, the historical account of the relay of images simply is an analysis of intuition as essentially historical.25

If I have any objection to Bredekamp’s Bildakttheorie, it is simply that he still requires formalisticism to do some of his historical work. Bredekamp sometimes takes mere morphological similarity in the form of images (regardless of their diverse historical locations in the Sein und Zeit and the Zeit und Ort of the agents) to be the evidence forsituating them as relays of images, as replications. It turns out, then, that the radically historical genealogy that supposedly situates form historically is simply an ordinary formalisticist history of reified form. We’re back where we started. To be sure, in Bredekamp’s history of images, the transcendental psychology of forming intuition is limited to—bounded by—a history of a mind that forms. He offers no “anthropology,” let alone reifications of “collective sensibility.” Still, it is not enough to say that a human mind is held together merely in virtue of the consistency and continuity of its forming activity. This assumes the consequent, and a more radical history of mind might radically dispute the very idea of any such coherence. Images have their effect on us—their “power”—not only because they replicate the form of images that we have assimilated (let alone produced as objects). They also affect us precisely because the form of the image has not already been integrated. (If it has been integrated, “we’ve seen it all before.”) This history needs to be excavated beyond purely morphological (merely formalisticist) descriptions of the images. Again we find that a radically post-formalistic history of imagemaking remains to be written.

Both of these kinds of Bildwissenschaft offer philosophically sophisticated responses to art-historical formalisticism and to formalism proper; whether or not they fully succeed, they seek to address Formen der Vorstellungsbildung nonformalistically. They are post-formalist to the degree that vision—or bodily awareness more generally—has been historicized, or more exactly in the degree to which they can offer a historical analysis of intuition as such. Needless to say, then, they operate in conceptual proximity to psychology and even to neurology and evolutionary biology, though they are not mere applications of the formalisms that can be found there. If anything, in fact, they would seem to demand in the end that we apply art history to psychoneurological formalism. I will turn to this very question in the final section of this essay (§5).

§4. Formalism and the “Wölfflinian method.” A historical question intervenes, however—a loose end in my account so far. What about Wölfflin’s formalism, his history of Formen der Vorstellungsbild, in relation to his formalistic method? Just as postmodernism in art criticism partly involved the rejection of Greenberg’s aesthetics, one might take post-formalism in art history to involve the rejection of “Wölfflinian method.” I refer, of course, to the comparative juxtapositions of photographs or other illustrations of artifacts that Wölfflin used to clarify our comprehension of the optische Schichten. Still, this way of construing post-formalism could mislead us. As I have already suggested, formalisticism may have certain tactical methodological values. And it is not formalism anyway.

Juxtaposition was ubiquitous in Wölfflin’s array of methods. But it was not essential. Wölfflin’s theory did not require it. To expose the Vorstellungsbild of Albrecht Dürer (when the artist was working on his master engravings of 1513-14) to readers in 1926 only required that Wölfflin make a single presentation: an illustration not of the artwork per se, such as a photograph of Dürer’s Ritter, Tod, und Teufel of 1513, but a visual demonstration for us of the artist’s optische Schicht in making it, which Wölfflin included in his later editions of Die Kunst Albrecht Dürers, first published in 1905. (In this demonstration Wölfflin removed all depicted figures and landscape except for the figure of the Knight himself, hoping to dramatize—to visibilize—its linear profile silhouette.) For the theoretical comparison in Wölfflin’s art history is not only between two different pictures constituted in the same or different optical strata in history. It is also between two imagings of the same picture, one clarified—relative to the other—by recognizing the optical stratum to which the picture historically belongs. This phenomenal clarification can be helped by the juxtapositions. But more generally it requires “formal analysis”—specification of the ways in which the artwork exemplifies Vorstellungsbildung, however the analysis is carried out. (Perhaps it is a specification in written discourse and perhaps it is a simulation in visual diagram or model, or perhaps both, as in Wölfflin’s formal analysis of Dürer’s Ritter, Tod, und Teufel.) As Wölfflin’s student Panofsky recognized, what really counts is the distance one travels between seeing the picture outside the horizons of the optische Schicht in which it was constituted historically and considering it within them—moving from our optical stratum (or Daseinserfahrung as Panofsky put it with a nod to Heidegger) to the picture’s visuality. As Panofsky put it in 1938, “this is rational archaeological analysis at times as meticulously exact, comprehensive, and involved as any physical or astronomical research.”26

Panofsky thought that Wölfflin was overly confident about this transfer—about the warrants that enable art historians to shuttle among optische Schichten of different pictorial styles. (For Wölfflin these warrants included experimental psychophysiologies of rhythm in optical stimuli, or in responses to them.) Panofsky condemned “pseudo-formalism,” as he called it when stating his objections in English in 1939 in rewriting a major programmatic statement published in German in 1932: formalism that proceeds from Daseinserfahrung, our imaging of the object given our Sein und Zeit (our being, and especially our aesthetic valuation, in our existentially limited historical time), and never breaks out of it—never reaches the optische Schicht of the object in its proper Zeit und Ort, its historical time and place. He repeatedly instanced the way in which Wölfflin had overlooked the historically particular aesthetic theory—an idiosyncratic canon of proportions—by which Dürer had created visible rhythm in his picture. Wölfflin thought that we can just see the rhythm, or see it with a little help from formalism in its presentation of the artwork analytically clarified to reveal the form imposed by Dürer in imaging the picture. But Panofsky insisted that this analysis was not helping us to see Dürer’s rhythm (as configurational) by way of the artist’s own theory of proportions. That theory had not been art-historically identified until 1915 and by Panofsky himself (or so he supposed—there had been other proposals). Wölfflin’s formal analysis, then, was pseudo-formalism: objectified Daseinserfahrung passed off as analysis—even visual presentation—of “’what we see’” when we see Dürer’s form. Presumably he intended his scare quotes here to designate the tendentious formalistic reification, as if “form” is something that we can see in the object (or can see). As formalism or formal analysis, Wölfflin’s illustration was an analytic simulation of the artist’s forming activity—activity of intuition—in making his picture. But it risked being interpreted as an actual visual presentation of the form—giving us the form the picture can be seen to have if we look through some kind of non- or paraformal visual material, thereby overlooking much of the picture and its symbolism.

As I have argued in A General Theory of Visual Culture, pseudo-formalism in Panofsky’s sense—what Richard Wollheim has called “Manifest Formalism”—generates an infinite regress. Treated as visual presentation of form-as-visible, Wölfflin’s simulation nonetheless is simply another “formed” object (maybe an artwork in its own right) to be looked at in terms of its own form, if we continue to extend the fallacy. By the terms of the formalistic (or Manifest Formalist) analysis, there would have to be another formal analysis (discovering and presenting the form of the formal analysis . . .), and so on without end.

Here we must be careful, I realize. It is possible that one of the intellectual claims of formalism (as formalisticism)—one of its critical resources—is that it accepts the logical possibility of its endlessness, perhaps as an argument (or at least a belief) about the sensuous inexhaustibility of art, perhaps as a philosophy of the irreconcilability of human knowing in its discursive and nondiscursive registers, or perhaps as a practical sociology of varied human interactions when  we show art to one another or tell one another about it. Here again we find that formalisticism might be a valid tactical method of formalism proper—even its necessary social and discursive forum (formalisticism is the pragmatics of transcendental formalism) and to an extent the evidence for it (without formalistically defined objects—even as objects of disagreement—one might doubt the very existence of a forming subject). If we drop the formalism proper, however, the formalisticism can be abandoned. Indeed, it need never arise. Rather than having to ask how the artwork “looks,” or “’what we see’,” we can ask different questions. Above all (and if we stick somewhat artificially to the register of visual and visible aspects) we can ask what it looks like. To the extent that we suppose the form of an artifact is constituted in essential primary movements of intuition, the question of how an artifact looks can receive a partly transcendental answer. Indeed, it must receive such an answer if there is no material account of historical variation in the forming capacity as such. But the question of what it is like can never receive anything but a historical answer.28

As late as 1951, Panofsky continued to criticize Wölfflin, at least for American readers. And his influence in America (given the strong pragmatist orientation of art history in the United States) reinforced worries about the supposed subjective origins, scientistic appeals, and transhistorical claims of Wölfflin’s formalism, regardless of its pedagogical appeal as integrated (for example) into the “Fogg Method” of training the art historian’s “eye.” In Panofsky’s wake, then, many North American art historians combined formalism with iconology—an unstable blend of transcendental psychology, critical phenomenology of art, and art-historical positivism that eventually imploded, though not without severe pressure exerted by critics who considered themselves to be anti-formalists.

Should we call Panofsky a post-formalist, then? For my purposes, probably not. Panofsky’s objection to pseudo-formalism in the “formal analysis” of artworks—whether or not it was the “Wölfflinian method” of comparative juxtaposition—was limited to its pre-iconological application without requisite iconological correction, as in Wölfflin’s simulation of Dürer’s Ritter, Tod, und Teufel. There is little evidence that Panofsky wanted to jettison formalism proper, or that he had a philosophical vocabulary in which to do so—at least one that he was willing to accept as a historian and a humanist. And there is considerable evidence that he took iconology to be the historical and humanistic application of formalism proper—that is, to be a culturally particular account of the interaction of “concept” (or symbol) with sensibility in the pathways of transcendental deduction, as Cassirer’s neo-Kantian “philosophy of symbolic form” had proposed. (Of course, Panofsky admitted—insisted—that freely created aesthetic values vary from one human group to another; but both Winckelmann and Kant had said the same thing, and in part as the very motivation for art-historical archaeology conceived as correction of aesthetic judgment.) In my terms, then, the dispute between Wölfflin and Panofsky was mostly about formalisticism in Wölfflin’s occasionally unqualified Manifest Formalism, such as his simulation of Dürer’s forming sensibility in making Ritter, Tod, und Teufel. Panofsky could have had other formalist fish to fry, such as Fry or Barnes. But perhaps it was self-evident to him that their formalisms were historically oriented even though they did not deploy the “rational archaeological analysis” that he recommended.

The “New Art History” of the 1980s also criticized “Wölfflinian method.” For example, in his Rethinking Art History of 1989 Donald Preziosi said that it set up art history as a technological “Panopticon.” On this view, the discipline of art history surveys an archive of illustrations of visual and material culture from around the world—an archive created by the Olympian gaze of missionaries, ethnologists, collectors, curators, and art historians, usually colonial-imperialist and maybe racist. (Of course, the art-history Panopticon also involved photography, expositions, museums, art handbooks, and other institutions beyond the one specifically associated with Wölfflinian pedagogy.)29

In 1989 art history undeniably needed postcolonial rethinking, and for some readers at the time Preziosi’s critique was decisive. But it has lost force in post-formalism, especially in so-called “object biography” (such as Richard Davis’s Lives of Indian Images of 1997), in transcultural art history (such as Barry Flood’s Objects of Translation of 2009), and in world art history (such as Summers’s Real Spaces). Post-formalists can treat the results of nonformalistic comparison not so much as a Eurocentric Panopticon as a postcolonial kaleidoscope. Indeed, object biography, transcultural art history, and world art history would be unthinkable if they could not undertake nonformalistic comparisons. Thus we have already found Summers—to stick to my main example of post-formalism—comparing Egyptian metric naturalism and Classical Greek optical naturalism in terms of the axes of observation adopted relative to the virtual coordinate plane, however one contextualizes the continuities (or not) between these optische Schichten.

§5. Post-Formalism as Art History Proper. In excavating the optische Schichten in which artworks—that is, drawings, paintings, sculptures, and so on—are constituted, and to return to my starting point, post-formalist art history calls for histories of the aesthetic orders and structures (as it were the “art”) of human vision, of imaging and envisioning, that is, of its active imaginative force whether or not any actual historical artwork was (or is) in vision or in view. The optical appearance of visual artworks—the supposed object of Wöfflinian formalism—is becoming less important analytically than the configuring force of imaging, regardless of what is imaged: an artwork; another kind of artifact; a person; a state of affairs in nature. Stated most dramatically, then, in a post-formalist history of imaging it would be perfectly possible for an art historian not to write about artworks or “objects” at all, at least if they are taken to be the primary object of study or the basic unit of analysis—als die elementarste Aufgabe der Kunstgeschichte, as Wölfflin specifically said they are not (situating him as a pre-post-formalist or as the very first post-formalist avant la lettre). In no central theoretical respect would this compromise the post-formalist’s identity as an art historian—as a historian of imaging as artful. In Summers’s analysis of planarity in ancient Egyptian depiction, what counts is the optical “axis of direct observation,” “completion” of volumes, and virtualization of the “coordinate plane”—parameters in imaging and productions of imaging. Any depiction constituted within these parameters of imaging will be so produced, even if “formally” it is a perspective projection. (In this case, an ancient Egyptian beholder accommodated to the optische Schicht of planarity would likely see it—optically “form” it—to be optically incorrect in specifying the real size of the objects depicted on the coordinate plane, or, alternately, would try to “form” it in such a way as to see their social and symbolic status in the size they are depicted to have.) Anything that we say art-historically about the objects—say about the colors of the painting used to “sculpt” virtual volumes in relation to the depicted space—must follow from the primary relations of their optische Schicht.

Of course, the question of “The Object” remains. Artifacts such as artworks do not have the same “materiality” as imaging, even if visual perception should be described as aesthetically ordered. But the aesthetic force of visual art is constituted in imaging it: in seeing it, or in imagining its visibility—aspects it might have when seen. Therefore the question of the art object lies within the questions of imaging as aesthetic. To repeat, then, in a post-formalist art history it is possible—sometimes desirable and maybe essential—not to write about any particular objects at all. This isn’t news in Strukturforschung after Wölfflin or in the archaeology of art (conceived, for example, on Kublerian lines). It is unnerving mostly for formalists mired in formalisticism. Having put the formality in the artwork or object, obviously they have to start with that thing. “Formal analysis” of an artwork, “close looking” at objects, an “iconic turn” to their “presence”: all of these court formalistic fallacy—formalism rampant.

It will not escape notice, however, that I have come close to depending on the same circularity that bedeviled Wölfflin. Sometimes it derailed his art-historical practice: as we have seen, his formal analysis of Dürer’s engraving—supposedly reporting empirically on the historical order of Dürer’s formal sensibility—was mocked by Panofsky as the very nadir of “pseudo-formalism.” How do we excavate the optische Schichten—the visual strata of art as historical—without “formally analyzing” the configuration of artworks? Isn’t imaging, including the constitution of formality, just a black box? Isn’t artistic object-form its manifest correlate? If optical strata are the objects of analysis, aren’t artworks our data?

Yes, artworks—and the wider field of artifacts—are data. But no, we do not use their formality—their form for us—as evidence for the optische Schicht in which they were constituted. Rather, we look at what people in the past did with the things, what they used them to do, in order to infer the network of aspects that the things had for them—aspects not limited to formality but including pictoriality and visible style. We look especially at how they replicated things: which features they chose to preserve, what they allowed to vary. Perspectival effects were not replicated in Egyptian depiction even though they would have to be ubiquitous in imaging the very same pictures. (Summers’s diagrammatic analysis of the virtual coordinate plane in ancient Egyptian “metric naturalism” adopts a natural visual perspective on the depiction in order to illustrate the phenomenon; given specified architectural conditions, anyone—including the Egyptians—could see the picture at this visual angle. But this is not pseudo-formalism because it isn’t showing “what we see”—that is, what the Egyptians saw when beholding these pictures. Rather, it virtualizes the real-spatial parameters of the optische Schicht of Egyptian pictoriality.) We can therefore infer that the optische Schicht within which its pictorial formality was constituted—the apparent configuratedness of such pictures—was not perspectival. On this virtual coordinate plane, “depth” does not mean diminution. One did not and could not use the virtuality to tell him how far away the depicted objects are from the plane of the format, though he could use it to show that they are separated in space—even to show how “big” they are relative to each other.30

Moreover, imaging is not really a black box. Vision science and perceptual psychology bring a mass of anatomical, experimental, and clinical data to bear on vision treated as active configuring of information in light reflected into the eyes, as if Wölfflin’s Sehen an sich were a painter painting a picture. This metaphor has been fully exploited (indeed analytically integrated) in Semir Zeki’s formalist neuroaesthetics, though the “Painter,” of course, is the human visual brain and the “Painting” is the world it sees. Because I have commented on this elsewhere, here I can go straight to the implications for post-formalist art history.31

As I see it, post-formalist art history is post-the-formalism of neuroaesthetics, or it should be. The mass of neuroaesthetic data was not collected in terms of Wölfflin’s theory that vision has a history. In fact, much of it was collected in terms of an opposite hypothesis, namely, that the processing of reflected light in the visual brain can be treated as a historical invariant (barring neuropathology) even though the things it makes to be seen are historically variable—formally multiform. If we extend Zeki’s metaphor, one Painter—the brain—paints all the paintings that have ever been made—that ever will be made for a very long time.

How does the theory of optische Schichten square with this? Wölfflinian vision- historicism, if it is accepted at all, would commit post-formalist art history to the metaphor that there have been many different Painters Painting Pictures—many neurologically real “visions.” And why?

Because literally there have been many different real painters painting different real paintings in the world—people making things to be seen, to be used visually, including pictures and artworks. When the Painter (vision) sees these things, and finds out what can be done with them or how they can be used visually, it will—if adaptive and intelligent—Paint differently. Indeed, it must Paint differently (in greater or lesser measure) in order to find out what can be done with these things or how they can be used visually, especially if the things were made by painters (and Painters) other than himself or herself. Of course, when it Paints differently he or she will paint different real paintings to be seen by other painters (and Painters). The historical cycle will spiral on. New optical strata will be laid down in the accumulated repertory of the Painter that each real painter is (and that each Painter has), and as the art history of his or her seeing: Das Sehen an sich hat seine Kunstgeschichte.

Indeed, we might derive the historicity of vision from mere variance in the replication of pictures that accrue in the visible world and demand to be used pictorially in ways that seem to be commensurate not only with their apparent configuration as we see it but also with their historicity—that is, with our historical awareness that we haven not yet seen what they can show us. It is possible that vision has an art history because pictures can be historical for us: erupting in and rupturing our visual field, our Painting, they create its optical strata. Seeing them as such, we are asked—maybe required—to see things anew.

By “visuality,” art historians mean socially constructed ways of seeing, Sehformen as Heinrich Wölfflin called them, often shaped in interaction with styles of art, depiction, and built form (often called “visual culture”).1 To take Wölfflin’s most famous example, we might compare what he called the “linear” mode of seeing exemplified by Lorenzo di Credi’s Venus (c. 1490), with the “painterly” mode exemplified by Gerard Ter Borch’s Concert (c. 1657). Both modes of seeing are also modes of painting, perhaps primarily modes of painting—a point to which I will return. In A General Theory of Visual Culture, I have argued that an interaction between vision and the visible features of visual culture—notably the recursion of “pictoriality” in vision and as vision—constitutes visuality. I also noted, however, that the recursions of this interaction are not well understood analytically, let alone neuropsychologically, as operations of retinal proprioception and subsequent processing in the visual cortex. In A General Theory of Visual Culture I ventured no speculations, then, about the neural manifestation of visuality, if any. My proposals were meant to be compatible with any neuroanatomical or neuropsychological model of imaging and visual knowing, especially with models that assume the plasticity and the pluripotency of the visual brain. In this essay, however, I push further. Should a general theory of visual culture be accommodated to a general science of vision, and vice versa? If so, how should the equation be stated?2

I. Vision, Visuality, and Visual Culture

Wölfflin understood Sehformen to be cultural styles particular to a time and place (such as the fifteenth and sixteenth centuries in western Europe in the case of Lorenzo di Credi and Gerard Ter Borch) and natural routes or relays of vision, that is, ways in which human beings are visually sensitive to contour, shape, and so on. The ways, in fact, in which all human beings are visually sensitive: in removing the background in Dürer’s engraving of Knight, Death, and the Devil of 1513 (the figures of Death and the Devil, the Knight’s dog, and the landscape) to expose the outline silhouette of the Knight, to visibilize it, Wölfflin meant to show how his readers—that is, people today—can still see the primary rhythmic configuration of a pictorial artwork made five hundred years ago. Wölfflin’s contemporaries appreciated that this investigation of expressive form in art was compatible with psychophysiological investigations of the human awareness of “rhythm” and other aesthetic orders in artworks and many other visible things. But other art historians have not fully followed him in that direction. Responding directly to Wölfflin’s treatment of Dürer’s engraving, Erwin Panofsky argued that we must understand the artist’s explicit theories of beauty and proportions (the subject of Panofsky’s earliest art-historical researches in writing his doctoral thesis) in order to interpret the allegory, that is, the Protestant (specifically Lutheran) visuality within which Dürer conceived the image—what Panofsky called “iconology.”3

Later “social historians” of art have proposed that the visual skills needed to interpret pictures are coordinated as a “period eye,” to use Michael Baxandall’s version of E. H. Gombrich’s general ethological theory (based on evolutionary psychology) of “perceptual readiness” or “mental set.” Such socially constituted visuality might be found, for example, in a fifteenth-century Florentine merchant’s ability to judge the size, volume, weight, and mass of things in space (and therefore value and cost—especially value and cost), a visual skill cultivated specifically in his tasks and purposes in commerce or banking and brought by him to the work of making sense of simulations of volume-shape produced in painter’s perspective at the time. To my mind, Baxandall may have erred in restricting explanation of Sehformen to “social history,” reducing the general ethology of perceptual readiness as conceived by Gombrich (his teacher) to a sociology of perceptual readiness that might be too limited. Regardless, the point here is that Baxandall’s notion of “period eye” has been widely accepted as a material explication of Wölfflin’s Sehformen—as “the social history of pictorial style.”4

II. The Hypothesis of Neurovisuality

In sum, visuality as conceived from Wölfflin (or before) to Baxandall (and beyond) seems to involve historical variations and specializations—variations and specializations that might be described by terms like Sehformen or “period eye.” As I would like to put it, human beings succeed to visuality when they recognize the forms of likeness that things have in a particular historical form of life—the visible and invisible aspects that things come to have in a network of analogies constituted in that form of life. Nonetheless, vision does not always and wholly succeed to visuality. Things remain visible to people outside the visuality within which they were intentionally produced, though what is visible in an artifact in this context (or what is visible about it) may differ from what is visible in the context of visuality. By the same token, people can succeed to many visualities, though both Wölfflin and Panofsky were somewhat uncertain (on different grounds) about just how far it is possible to do so when we are dealing with visualities constituted in the past and accessible to us only in things made to be visible within them that happen to have survived into our own visual world.

It is precisely for these reasons that the relations and recursions—logical, neurological, psychological, or sociological—between vision and visuality are not easy to state. I do not share Dürer’s Protestant faith, let alone his familiarity with figurations analogous to his depiction of a Christian Knight on a dread journey in the wilderness of temptation and imminent death (at least according to Panofsky’s interpretation of the iconography of Dürer’s engraving of 1513). But I can see the rhythm and balance of his engraving of the Knight, or more exactly I can see it when I am confronted with the visual replication of it in Wölfflin’s illustration, where it has been visibilized for my seeing of rhythm and balance in the engraving. (The illustration functions, then, as an autonomous artwork, though one produced by Wölfflin rather than Dürer.)

The visibility of rhythm and balance in the engraving, if any, is presumably a matter of processing in the visual brain—a material question in visual neuropsychology. Of course, Wölfflin could not invoke the results of later twentieth-century neuroscience. But he assimilated the psychophysiology of his own day, and during his own lifetime he could read experimental psychology that could have been pertinent for his art-historical phenomenology, notably Gestalt psychology. (For the sake of economy I set aside the way in which Gestaltists might have responded to art-historical formalist phenomenology, not to speak of the incorporation of both art-historical formalism and Gestalt psychology in the modern arts addressed both by art history and by psychology.) At the time other art historians made explicit use of psychophysiological concepts.5

But what about visuality? Should we suppose that historical periods or phases of human picture-making—Sehformen or “period eyes”—depend on neural circuitries unique to the human populations that used those pictures visually or, to use my terms, visually understood their forms of likeness in a form of life? (In Panofsky’s iconology, remember, we can only use a picture that was made in the distant past or in a different culture in discursive ways; we cannot fully use it visually in the way that its makers did.)

To designate the neural manifestation of the succession to visuality (if any) in natural vision, I will use the term “neurovisuality”: the neural specificity, if any, of vision-in-visual-culture, or visuality. The neologism is a convenient way to designate (hypothetical) successions to visuality in neural circuits that are laid down for such recursions in natural history and by such recursions in social life. They are laid down, then, as visual culture and in visual culture. (Still, if visual culture is nothing but neurovisuality then it may not be necessary to ontologize it: it might suffice to describe natural-historical recursions of vision in human society.) The term echoes (and in certain circumstances it overlaps) two other terms: “neuroaesthetics” and “neuroarthistory.” But it is not quite the same concept, as we will see.

The hypothesis of neurovisuality—that there might be a neural correlate or even causation to the visual succession to visuality—is unpopular among art historians, if it is entertained at all. Still, if it is correct it makes a powerful (and in some arenas a decisive) contribution to vision science from ophthalmology or psychiatry to lighting design or the engineering of human-computer interfaces (HCI)—a specifically art-historical contribution. For this reason it can be unpopular in vision science too.

III. The Hypothesis of Neurovisuality in Vision Science and Art History

To investigate neurovisuality in my sense, experimental neuropsychology must join forces with art history (and vice versa) in full measure. Why? And how?

On the one hand, laboratory or clinical experiments on visual processing in living human subjects must address visual brains that are already acculturated in historical visualities (that is, “visual culture”). But it may be difficult to detect this parameter experimentally without historical perspectives. And by definition these cannot readily be built in to any laboratory or clinical experiment.

It will not do simply to administer questionnaires and tests to populations of undergraduates in psychology courses in American colleges. In fact, it probably will not do simply to administer questionnaires and tests to any range of human populations of any kind anywhere on earth today. If art history and the historical study of visual culture are correct—correct, that is, as an account of the history of art and visual culture whether or not explicated as neurovisuality—then all living human populations possess visuality. But in theory not all neurologically possible visualities are represented in any range of living human populations on earth today. Therefore it is logically impossible fully to specify the neural correlates (if any) of visuality specifically as a question of the neuropsychological capacities and adaptations of the visual brain by way of any experiment on living human subjects. More exactly, neuropsychological experiments with living human subjects can identify a (neuro)visuality—namely, the (neuro)visuality of the subject(s) who have been observed experimentally. But this might be no more helpful than the art-historical or anthropological evidence it purports to explain. Indeed, art-historical or anthropological evidence may suggest that populations in the past or in other cultures have (had) neurovisualities different from the ones observed.

On the other hand, however, art history must be just as unsatisfyingly partial as experimental neuropsychology, though for inverse reasons. Art historians writing in the main lines of formalism and historicism from Wölfflin (and before) to Baxandall (and beyond) have usually addressed human beings in visuality without any direct experimental access to the experience of many of the people they purport to survey, and whose configurations and figurations they claim to interpret, even if such access is available. Since its consolidation in the late eighteenth century, to be sure, professional art history has had many opportunities for such experiment, notably among living human populations who have made and used the visual arts. But usually it has eschewed laboratory or clinical evidence about these experiences unless there has been a special art-historical reason—usually a biographical reason—to investigate them (for example, in the case of artists who have suffered damage to the visual brain). And it has disdained psychological and sociological questionnaires and tests as well as protocols in the pedagogy of art even when they could be salient. (Since the late nineteenth century, then, art history has rarely included in its professional purview such writers on configuration in the visual arts as Guido Hauck, Denman Ross, Hans Prinzhorn, Henry Schaefer-Simmern, Erle Loran, Rudolf Arnheim, Roy Schafer, Rhoda Kellogg, Margaret Hagen, or John Willats, regardless of their influence in institutions of art making in modern society—in classrooms, lawcourts, hospitals, and so on—and even when they have made valid contributions to the resolution of specifically art-historical problems.6)

For this very reason, art historians can overlook neural causalities that might operate outside visuality—causalities that might explain why pictures or artworks can retain their visual “power” or “agency” (aesthetic or otherwise) far beyond their original contexts of making in a particular historical visuality, that is, why they can be globally transmitted between historical visualities despite tenuous material connection between the social groups or visual cultures in question. In this regard it is not surprising that neuropsychological research has been most visible in art history specifically in the frame of “world art studies,” within which it subsists (a bit uneasily) with comparative anthropologies of art and histories of intercultural interaction in visual culture.7

At the same time, art historians may have missed a chance to enrich their historical understanding of visuality (Sehformen or “period eye”) by entertaining the hypothesis of neurovisuality. Often they have eschewed responsibility for investigating ontogenetic succession into visuality—neurological, psychological, or sociological. They have left that question to developmental psychology, projective testing, the pedagogy of art, art therapy, and other professions, where the thesis of neurovisuality sometimes has been stated too strongly or simply taken for granted.8 For example, children’s drawing, depiction, or art-making has typically interested art historians only when artists or visual cultures of the past have been interested in it, despite the obvious sense in which the very notion of Sehformen or “period eye” demands anthropological or experimental study of visual acculturation in history, that is, in the experience of human subjects being integrated over time into visualities. If children learning to draw or to interpret pictures are not instances of this historical process of succession-to-visuality, it is very hard to know what possibly could be an example of it.

In sum, neither neuropsychology nor art history is especially well placed (or has taken itself to be well placed) to address the hypothesis of neurovisuality, whether the reasons are theoretical, methodological, or ideological (most likely a mixture). Overall, vision science has approached visuality, if it has done so at all, by way of vision. Art history has approached vision, if it has, by way of visuality. But I have already noted that vision and visuality do not fully intersect, despite their essential recursion in ordinary human visual experience in history; what is visible is partly specific to each domain. Therefore we cannot predict which direction to follow analytically. From vision to visuality? From visuality to vision? Or both together, as I suggest?

In the remainder, I proceed on two levels. On one level, I will draw on research in three areas of vision science (physiological neuroaesthetics, computational psychology, and evolutionary aesthetics) in order to point out where neurovisuality might be active as a recursion in neural circuitry (with emphasis on the “might”). This is analytic: an attempt to get as clear as possible about concepts and arguments. On the other level, I will mention possible neurovisualities that have been suggested in the historical record of human art (namely, in artistic modernism, in making virtual coordinate space, and in adjustments to the vertical in built form). This is provisional and speculative.

IV. Neurovisuality and Physiological Neuroaesthetics

Though it has been defined in many ways, in general “neuroaesthetics” studies the neural correlates (if any) of aesthetic experience, and perhaps the identity of certain aesthetic experiences and certain material states of the visual brain. For my purposes in this essay, its most appealing general thesis has been well stated by the neurophysiologist Semir Zeki in his Inner Vision, published in 1999.9

The thesis has two parts. One part says that human visual imaging, or seeing, is intrinsically aesthetic: it actively configures the visual image (including any pictures presented to natural vision), like a painter painting a picture. (Zeki deploys this metaphor for all it is worth, and we might want to take it quite literally in a theory of neurovisuality—a theory of how the history of art, such as the history of painting as an art, may be “wired into” the visual brain.) Seeing, we might say, has “image structure” (to appropriate John Kulvicki’s term) in the way that the painting might have “formal structure.”10 To quote Zeki: “the brain . . . is no mere passive chronicler of the external physical reality but an active participant in generating the visual image, according to its own rules and programs. This is the very role that artists have attributed to art, and the role that some philosophers have wished that painting could have.”11

This is a familiar psychological claim (or at least the first sentence is). It is even a philosophical claim, recently reexamined by philosophers such as Alva Noë and Dominic McIver Lopes.12 Its genealogy might (or can) include Kant’s doctrine of the transcendental imagination, Ernst Cassirer’s phenomenology of knowledge, and Nelson Goodman’s constructivist psychology. (Indeed, Goodman’s Structure of Appearance and Languages of Art provide powerful analytical resources for mapping “formal structure” onto “image structure” and vice versa.) Art historians do not usually endorse it explicitly. But they are often comfortable with it: as Zeki says, it attributes processes and functions to the human brain that are often attributed to artworks or pictorial representations (and possibly have been caused by imaging these artifacts). Indeed, this part of the thesis may simply be one way of stating the thesis of visuality—that human visual perception or seeing has active modes and forms, phases and styles—without building in any particular causal explanation, that is, without appealing, say, to visual skilling in carrying out social tasks and fulfilling social purposes (as in Baxandall’s “social history of pictorial style”) or alternately to habituation in environments of visual affordance (as in the “neuroarthistory” mentioned in a later section of this essay).

The other part of the thesis derives specifically from neurophysiology. Active visual-aesthetic configuration of the visible world has its own “rules and programs” in the brain, as Zeki puts it, whether we refer to electrochemical activity in the brain that materially accompanies our experience of a painting as an art or refer to activity that produces it in the sense that if we were to stimulate or simulate the activity in the brain—let’s say by giving the subject an “art pill” or an “art injection”—we would induce just that aesthetic experience of just that painting. Soon enough, artists will begin to administer art pills, if they do not already do so, and the neurophysiological claim in aesthetics will have come full circle. For according to the strongest version of neuroaesthetics any artwork simply is an art pill or an art injection. Why?

First, there is evidence for the cortical localization and functional specialization of neural-aesthetic activity, notably the “firing” (or heightened electrical activity) of specific cell-complexes in relation to stimuli to which they are specifically adapted or for which they are “selective.” For example, in primate vision some neurons only fire when stimulated by light within certain wavelengths—the red range, say. One can also identify cells that fire in response to a vertical bar moving only from left to right in the visual field. Unresponsive to movement in the other direction, the cells can be said to be “directionally selective.”13 There are many similar findings. Hundreds of technical publications in primate neurophysiology and neuropsychology report them in detail.

For my purposes in evaluating the general thesis of neuroaesthetics, it should be possible to integrate these results—many dozens of them, many overlapped or intersecting—to give a full account of the neural identity of a given artwork or other item of visual culture, that is, of its subjective visual-aesthetic effect, of its “look” or “feel” and its “power” or “agency.” If that subjective visual-aesthetic effect includes the beholder’s awareness of aspects of the painting constituted in visuality—awareness of forms of likeness of the artifact that visibilize it in a particular way in a historical form of life—then the neuroaesthetics is de facto a neuroaesthetics of neurovisuality though it need not be strictly limited to it. (As an example of visibilizing forms of likeness, I have discussed the visibility of a building built to have a “color scheme” in a certain historical form of life, that is, in visuality. Outside the visuality, the colors of the building may be visible, but the “scheme” of the colors may not be. Within the visuality the visibility of the colors is specifically the visibility of the color scheme in the building.14)

At the moment, as already noted, most neuroaesthetic research does not explicitly entertain the hypothesis of neurovisuality. Indeed, it may prefer to skip over it to move directly to putatively panhuman visual aesthetics—to supposed invariants in visual processing. At the moment, then, it is not easy to say which neuroaesthetic findings, if any, speak specifically to neurovisuality, if any such exists, though cross-cultural or comparative experimental research may be useful if carefully framed in terms of the hypothesis. Still, neuroaesthetics can assert as a point of its principle or general theory that the whole of the artwork (including its analogical aspectivity in visuality if any) is intelligible to vision in virtue of neural wiring, whether or not that wiring was neurally laid down in recursive interactions with visual culture (that is, as neurovisuality). On this account, then, neuroaesthetics would fully replace the art-historical analysis of the look or the feel of the painting as an art (including its aspectivity in visuality). Or at least it would fully replace it if neuroaesthetics does encompasses neurovisuality.

It hardly needs to be said, of course, that the neuroaesthetic analysis (if fully realized) probably would not look much like current art-historical or art-critical descriptions of pictorial artworks, that is, like the present-day advanced “artwriting” that putatively attends both to the (neuro)intentional order of the artwork and to the (neuro)aesthetic responses of its beholder/critic as well as to the ambiguities and uncertainties within both of these horizons. The neuroaesthetic analysis would be a read-out of the neuroelectrical activity specifically correlate with the perception of the painting as an artwork. Of course, some neural “firing” will not be specific to the visual perception and processing of just that painting as an artwork and only that painting. But the particular hierarchy, sequence, and recursion of firings—the overall pattern of neuroelectrical activity—will be specific to just those paintings that have that particular aesthetic order, whether one or many.

It also goes without saying that at the moment most art historians as well as art critics and other artwriters probably believe that this read-out could not possibly be as informative as an extended discursive description of individual works. (In artwriting, one does routinely acknowledge that words—or at least words of discursive analysis, explanation, and interpretation—tend to fail us in dealing with images, or some images at any rate. As T. J. Clark suggests in narrating his own sustained attempt to engage a particular single painting as a work of art, it may only be “the physical, literal, dumb act of receiving the array on the retina [that] will satisfy the mind,” or perhaps, as he also suggests in the same book, the writing of a poem that in some way analogizes aspects of the painting or thoughts and feelings one might have in relation to it, direct or indirect. Still, the words proliferate: Clark’s meditation on Nicolas Poussin’s Landscape with a Man Killed by a Snake of 1648 as an aesthetic object for him—and a specifically visual one, as he repeatedly insists—is nearly two hundred and fifty pages long.15) In principle, however, the neuoraesthetic read-out and the extended discursive description can be fully translated into one another. Taken on their own terms, they are different representations of what we can see (or of what is seen by us) when apprehending the painting as an artwork. Our choice of the representation likely has as much to do with our disciplinary inheritances and professional territories as with any point of putative aesthetic principle. We can pursue them or not, or approve them or not, in light of what they allow us to do with our experience of the painting as an art, and to do about it, if anything. In some contexts we might reasonably prefer the extended discursive description (say, in describing the artwork to someone who has not seen it—one of the original and foundational roles of art criticism since the eighteenth century). In other contexts we might prefer the read-out (say, in determining whether the artwork has any causal role to play in arousing fear or stimulating aggression—one of the clinical arenas in which neurophysiological aesthetics might have a place). In practice, most art-historical description already is partly critical ekphrasis and partly read-out—an analytic report of patterns in the data, that is, the observed aesthetic effects and attributed intentional orders of artworks and other items of visual culture.16

In fact, there may be art-historically salient uses of neuroaesthetic read-outs that would not be possible on the basis of extended discursive descriptions of individual works. The read-outs might make it analytically possible, for example, to identify patterns of similarity and difference in vast series of artifacts, pictures, or artworks—patterns of similarity and difference founded in their functions and meanings—that are invisible to the naked eyes, or to our naked eyes. (This point was accepted in “structural research” [Strukturforschung] in art history and archaeology in the earlier part of the twentieth century, which pioneered typological, seriational, and statistical methods of formal analysis.17) In the theory of visuality that I have proposed to defend, two morphologically indiscernible artifacts or artworks can be distinguished analytically—and may be distinguished in seeing, in visual use—in virtue of the different forms of likeness they relay (a mere multicolored building, say, as distinct from a color-schemed building). In principle, the neuroaesthetic read-out should capture this distinction, in turn enabling us to “excavate” complex networks of forms of likeness that may be inaccessible to introspection or ethnography—not visible, that is, either to the historical actors precisely because the forms of likeness in question are the very grounds of the visibility of visual culture for them or to the art historians situated outside this visuality.

Art history is a long way away—probably decades away—from fully actualizing research of this kind. Instead, and partly for this reason, it remains deeply bound to visualist and formalist prejudices—its claim of responsibility to “what we see” in and as individual art-objects—even when many historical questions about visual culture are questions neither of visibility nor of sensuous form and cannot be answered by focusing on individual art-objects. As I would like to put it, they are questions not (or not only) of how an artwork or artifact looks but (also) questions of what it is like. Still, we can predict that some art historians in the future may pursue a post-visualist and post-formalist phenomenology (paradoxical as that might sound) by way of neuroaesthetic investigation and its potential to identify—empirically to track—the forms of likeness that constitute visuality in the replication of series of artifacts, pictures, or artworks.18

Admittedly it may be premature to judge these matters. Here we need to stick to the question of neurovisuality in neuroaesthetics. As noted, neuroaesthetics is probably only relevant to art history, and art history to neuroaesthetics, if the hypothesis of neurovisuality is correct. Where exactly does it enter the story analytically?

Overall, cells in one region of the visual brain, known as V4, activate when colored arrays are viewed, that is, when color is neurally constituted. V5 activates when the subject engages moving affordances, maybe by moving. In Zeki’s summary diagram of the confluence of color and motion processing in vision, then, a multicolored “Mondrian stimulus” of the kind introduced by Edwin H. Land engages V4, while a black-and-white televisual or digital-animated stimulus engages V5.19 To be sure, configurations made by Piet Mondrian himself involved calibrations of figure, ground, edge, and frame that were not configured in Land’s “Mondrian stimulus.”20 In turn, then, we might suspect that painters like Mondrian manipulated the routines of natural color vision (or other aesthetics of the visible) in their painting as an art, that is, in Mondrian’s case, in making his versions of the “Mondrian stimulus.” Neuroaesthetics not only makes this assumption. As noted, it tries to justify it neurophysiologically—to show how the subjective effects of artworks are produced in stimulating the “rules and programs” of the brain in ways that can be identified neurophysiologically.

Still, this raises the question of possible neurovisuality. Does Mondrian’s painting—does any artifact, artwork, or picture—prod the brain to reintegrate its circuitry? We know that reintegration can occur when the brain compensates for loss of particular functions due to damage or disease; some of the research results reported by Zeki and other neurophysiologists and neuropsychologists writing about aesthetics are partly derived from clinical-medical studies of brain damage in major stroke or other injury.21 Presumably artworks, unlike strokes, do not cause the brain to lose anything. But do they cause it to gain something? Something to be gained by way of the artwork?

One single painting by Mondrian likely cannot prod the brain to reintegrate the confluence of color and motion processing in V4 and V5. But the painting was generated, some art historians might want to say, in a historical visuality in which the configurative orchestrations of Mondrian and many other artists and other image makers (making productions in many domains of visual culture in many techniques and media) demanded the confluence of V4 and V5 in certain ways that maybe could only be stimulated by the integration relayed in just such works. The objective sensuous conditions of modernity, it has often been said, became its new subjective modern sensations as reprojected in its contemporary arts—a recursion debated by Walter Benjamin, Georg Simmel, and many others. Right or wrong, many artists and art critics and some art historians have treated certain modernisms in the arts as neurovisualities, though when their claim for the arts in question is stated this way they may disavow it.

To be sure, many critical and social historians of modernisms in modern art are sceptical of the claim. Certainly it can be treated critically as a historical formation in visual culture—an aesthetic theory—in its own right, like the correlate theories of nervous energy and transformation.22 As a positive hermeneutic assumption, it is most common (if often still latent as a material neurological claim) in aesthetics of modernism committed to strong claims for the aesthetics of modernism—for its material power to “alter perception,” as one sometimes hears, or to “affect the senses in new ways.” To quote Stephen Kern, “From around 1880 to the outbreak of World War I a series of sweeping changes in technology and culture created distinctive new modes of thinking about and experiencing time and space. Technological innovations including the telephone, wireless telegraph, x-ray, cinema, bicycle, automobile, and airplane established the material foundation for this reorientation; independent cultural developments such as the stream-of-consciousness novel, psychoanalysis, Cubism, and the theory of relativity shaped consciousness directly. The result was a transformation of the dimensions of life and thought.”23 To be fair, the proponents of the aesthetics in question usually blend a historicist approach to past ideologies of something like neurovisuality with their own unstated theories of neurovisuality. The latter may be universalist, referring to panhuman sensations constituted in or by aesthetic experience of the arts, or more specifically historicist or visual-culturalist in theoretical definition, referring to neural integrations in the experience of certain particular historical arts.24

I cannot possibly decide the empirical point, though deciding it is surely one of the outstanding projects for neurophysiology and art history. Instead I want to make a strictly analytic point. In Zeki’s illustration of the confluence of color and motion processing in vision, supposedly the beholder of the Mondrian stimulus is neurally constituting color in V4: the diagram shows part of his cortex “lighting up” at the reflected luminance of the green patch in front of him. (In Land’s experiments, and in his theory of “retinex” or the processing of retinal proprioception in the cortex, the question is the similarity or difference of that green patch, qua green, to other patches perceived to have the same or different colors.) But he is not seeing a Mondrian, as already noted. And not only because the pattern is not in Mondrian’s style or by Mondrian. Even if it were a Mondrian, the beholder depicted in Zeki’s diagram stands too close to it to see almost everything else: bounded rectangular planes of discrete but uniform colors partly occluding one another at different apparent distances from the imaging point and regardless of the real variance of luminance across the painting. If the beholder is to see those attributes of the painting (painting them in vision), that is, if he is to see a Mondrian, then he has to see the configuration integrated by Mondrian in those very actions—succeed to the colors, planes, and virtual depth configured by Mondrian to be visible to him in seeing the painting. In other words, when we look at things that have been actively configured for our seeing as actively configuring what is seen we aestheticize twice over or in a feedback loop, redoubling the aesthetic momentum of seeing: we paint the painting painted for our painting of it—repaint it. This recursion or redoubling is a necessary condition for neurovisuality even though it may not be sufficient for it.

I want, then, to supplement Zeki’s point that “the brain is an active participant in generating the visual image, according to its own rules and programs—the very role that artists have attributed to art and the role that some philosophers have wished that painting could have.” In some contexts, art (or other things made specifically to be visible to us) may be an active participant in generating the visual image, according to its own rules and programs—the very role that philosophers have attributed to the brain, and the role that some artists have wished that vision could have. In modern human lifeworlds, we are always in such contexts. Where might we look to find this recursive redoubling of actively-configuring vision by way of the actively-configured visual culture that it sees? In the next two sections I explore a possible example.

V. Neurovisuality and Visual Computation

Consider our visual inspection of a common natural thing. In his book of 1982, Vision, David Marr illustrated an early stage in the “computation” of information transmitted in reflected light to visual proprioception: a leaf on one stalk hanging in front of a leaf on another stalk, or as he put it an “image of two leaves.” Marr’s illustration models information at (or in) photoreception—what he called the “gray level” of the “image” (color has not yet been computed at this stage of visual processing). The table accompanying the simulation of the “image” at photoreception assigns numerical values to the discrete units of the image (not exactly pixels, they are discrete incidents of photoreception): the measurable “intensity value” of luminance at these locales in the image. But where in the numbers can we find the distinct edge of the leaf, or the boundary or gap between the leaves, that we can see with just a bit of work? As Marr wrote, “there is not a sufficient intensity change everywhere along the edge . . . to allow for its complete recovery from intensity values alone, yet we have no trouble perceiving the leaves correctly.”25 Marr explicated this observation in a complex model (what he called a “representational framework for vision”) of the hierarchical computation of information in reflected light sequenced into three “representational stages” beyond the early and primitive gray-level image that “represents intensity” (modeled in his illustration of the “image of two leaves”): what he designated (1) the “primal sketch,” (2) the “2½-D sketch” (a “viewer-centered coordinate frame” in which the edge is getting computed in the discontinuities in the surface and in “distance from the viewer”), and (3) the “3-D model representation” (an “object-centered coordinate frame” in which the edge gets fully computed as a spatial distance in depth).

Marr’s model is complex, and he stated it in several ways. To use his own words, “The overall framework . . . divides the derivation of shape information from images into three representational stages . . . (1) the representation of properties of the two-dimensional image, such as intensity changes and local two-dimensional geometry [i.e., primal sketch]; (2) the representation of properties of the visible surfaces in a viewer-centered coordinate system, such as surface orientation, distance from the viewer, and discontinuities in these quantities; surface reflectance; and some coarse description of the prevailing illumination [i.e., 2½-D sketch]; and (3) an object-centered representation of the three-dimensional structure and of the organization of the viewed shape, together with some description of its surface properties [i.e., 3-D model representation].” Marr took the primary theoretical problem to be the transition from (2) to (3) above. Given the information encoded in the image (by stereopsis, shading, texture, contours, or visual motion), the sequence of early visual computations from the image to the primal sketch and the 2½-D sketch (i.e., from image to [1] to [2] above) is “unsuitable for recognition tasks” (that is, insufficient in identifying what the shape is the shape of) because it depends critically on the vantage point. As Marr put it, “[i]t must be remembered that coordinates referring to the line of sight are not very useful to the viewer” precisely because “one must continually allow for the angle of the line of sight, .  . . a difficulty that is compounded by the effects of eye movements.” Therefore the crucial final step of visual computation (i.e., from [2] to [3] above) likely “consists of transforming the viewer-centered surface description into a representation of the three-dimensional shape and spatial arrangement of an object that does not depend upon the direction from which the object is being viewed.” Seeing, we might say, is always subject to the visual angle; it is always coordinated in natural visual perspective. Nonetheless it extracts a recognition of an object by computing a representation away from the vantage point or as it were around it. Momentarily we will see why this theoretical problem (perhaps an artifact of Marr’s own theoretical conceptualizations) might have a historical solution.26

Most important, we see the edge or gap between the two leaves more or less easily, Marr thought, because of what he called “consistency-maintaining processes in the 2½-D sketch.”27 Still, we can see the edge in several ways. (The image in Marr’s sense is ambiguous.) I may see the leaf on the lefthand side of the image as “behind” the leaf or leaves on the righthand side of the image in virtue of the apparent continuity of its contour with the stretch of leaf to the right of the middle stretch of leaf—one of Marr’s “consistency-maintaining processes.” I see lefthand and righthand stretches of leaf, then, as the same leaf—recognizing the leaf behind the middle leaf. At the bottom of the leaf in “back,” this continuity seems readily computable; the curve connecting them is smooth. At the top, however, no smooth continuity can be constituted. Moreover, there is a stark contrast in intensity values along both connecting curves (“bottom” and “top” in the image), smooth or not. Given this, we might compute the leftmost stretch of leaf as a third leaf arising from a stalk that is not visible—a leaf abutting the middle leaf, lying in the same plane across the line of sight. The boundary or “discontinuity in the surface” (supposedly seen as the one edge of the leaf “in front,” silhouetted as contour against the leaf “in back”) is where these leaves are touching along both their edges—not separated by the distance from the vantage point that can be computed at any point on the edge as a quantity on the Z-axis. (The image is two-dimensional; it encodes information on the X and Y axes of the plane in a viewer-centered coordinate frame. As yet there is no Z-axis plane—no “third dimension”—coordinate with the X and Y axes.)

Whether there really is a third leaf can probably be resolved by moving around the objects in real space. (I say “probably” because it might be theoretically possible to design objects in real space that always remain ambiguous about their spatial relations in three dimensions at any and all possible visual vantage points, even if the ambiguity changes its visual particulars from vantage point to vantage point, each having its own ambiguity—and even if succeeding ambiguities partly resolve preceding ones.) But assuming no motion of the vantage point, in Marr’s model the viewer-centered representation of the image feeds into the object-centered one, and helps constitute. And that representation could have two three-dimensional renditions computed from the intensity values of the image. One is objectively correct and the other is incorrect. To get us to the one and not the other, visual processing according to Marr prefers to integrate the intensity contrasts and primal (dis)continuities represented in the 2½-D sketch in order to give “rough depth” and “distance from the viewer”—thus two leaves, not three. This is the very idea of Marr’s two-and-a-half-dimensional representation. And a strange animal it is: for some critics, too strange quite literally by half, an internally contradictory sleight-of-hand in Marr’s model of three-dimensional visualization.28

But in a recursion in neurovisuality, the 2½-D representation in visual computation might be perfectly viable. As commentators on Marr’s model of the representational framework of vision have pointed out, it is a widely-used technology of orthographic projection, usually designated as axonometric projection in “paraline drawing.”29 Equally important, it is a common way in which two-dimensional renderings on the plane three-dimensionalize when the visual angle at vantage point constructs what David Summers has called a “virtual coordinate plane,” that is, a Z-axis-plane visibly coordinate with the plane of the X and Y axes—the “plane of the format” in Summers’s account (equivalent to the “picture plane” if the picture in question is constructed in linear perspective) and the plane of the “image” in Marr’s model.30 If this rule and program of depiction loops into the rules and programs of vision, then Marr’s problem could take care of itself in the modern human case—that is, in a neurovisuality. The Z-axis-plane would seem to be rarely if ever encountered as a visible plane in nature outside pictures. (Where, for example, is the flat “floor” in the arboreal space of the forest in which the leaves will likely be seen in natural primate vision?) It only becomes fully visible in virtual pictorial spaces. Hypothetically, once the brain computes visually (or in Marr’s terms once it represents the image) by way of this plane, if it does, it can spatialize intensity contrasts, surface (dis)continuities, and so on, in terms of “rough depth” and “distance from the viewer,” just as Marr requires of the 2½-D sketch as one of the sequential representations of the gray-level image in visual processing.

Indeed, it might be possible to eliminate this mysterious stage in the overall representational framework, that is, to move directly from primal sketch to 3-D model representation. After all, the main theoretical job of the 2½-D sketch as Marr conceived it is to take the representation of the image from a viewer-centered coordinate frame (insufficient for object-recognition) to an object-centered coordinate frame (in which the shape is readily recognizable as the shape of a particular object-volume). But this can be done on the virtual coordinate plane when it is visibly coordinate at right angles with the upright plane of the format or the picture—something visible only in artificial architectural spaces architectonically configured to have just such a coordination. For this coordination sets up a what might be called a “box” or “quadration” that situates things virtually in the three-dimensional coordinate space they inhabit; regardless of visual angle (that is, of viewer-centered coordinates), we can see the relation between the depicted object and the coordinate frame, that is, the intersection of the plane of the format (or the picture plane in linear-perspective constructions) and the virtual coordinate plane (depicted or not as an optical plane in the picture). We can see the object, in other words, in what might be called “virtual coordinate space.” When fully generalized or extended to its “notional abstraction,” this virtual space might be called “metaoptical,” to use Summers’s terms, because the virtual coordinate space visibilized in pictorial architectonics becomes notionally identical with infinite homogeneous three-dimensional coordinate space (or Cartesian point space).31 But any virtual coordinate space can be sufficient to three-dimensionalize a depicted object in natural visual perspective—that is, in natural visual perspectives on pictures in architectonics in which a virtual coordinate plane appears. In a recursion of neurovisuality, it might be that natural visual perspective enfolds this pictorial effect in its representations—its sequential and hierarchical recomputations—of the primordial image.

VI. Neurovisuality and Evolutionary-Ecological Aesthetics

The history of this recursion, if it has occurred, requires further analysis. The philosopher Marx Wartofsky urged the radical thesis that the “dimensionality of visual space” in visual perception (human abilities to apprehend and interpret ambient space as three-dimensonally coordinate and possibly as infinitely extended in a correlate metaoptical coordinate space) was a consequence of the “pictorialization of visual space” acquired through “practices and conventions of pictorial representation” that had been developed in Classical Greek and Renaissance Italian styles of depiction, notably the invention of linear perspective projections.32 This proposal would seem to follow from Wartofksy’s more general historicist thesis that “modes of our visual cognition change with changes in the modes of our pictorial representations” and specifically that “canonical styles of representing the seen world change . . . and introduce transformations of vision.”33 Wartofsky himself did not invoke the biohistoricist hypothesis of neurovisuality in my sense because he wanted to say that human vision has a history that “goes beyond the biological evolution of the hominid visual system and is part of that activity of self-creation and self-transformation which we call cultural evolution.”34 But if we set aside this reification of cultural evolution, Wartofsky’s thesis is compatible with the hypothesis of neurovisuality, and in the end it may require it. What Wartofsky calls “self-creation and self-transformation” simply is part of the biological evolution of primate vision once any feedbacks were introduced into it by cultural activity—by using sticks, stones, teeth, and bones to do and to make things and especially to depict things. (This brute culture still needs to succeed to what I have called “culturality”—awareness of the forms of likeness of artifacts produced in cultural activity—in order for it to constitute a visuality in my sense. Brute or “osteodontokeratic” culture can be found among monkeys and great apes and it is bioculturally characteristic of hominid species in the human line. But the visuality of culture—and disjunct visual cultures—may be a more recent development in Homo.35)

Restated as a historical identification of a supposed neurovisuality constituted in interaction with painter’s perspective, Wartofsky’s “dimensionality of visual space” may still seem untenable. As Arthur Danto has complained, there can hardly have been substantial neural evolution “in the bare six hundred years from Giotto to Ingres,” or at least any evolution of the scope that Wartofsky’s thesis would seem to entail.36 For all art-historical intents and purposes, Danto would treat human vision as invariant since the Upper Paleolithic period. Indeed, he has argued that art historians categorically depend on treating vision as an invariant, for otherwise their identifications of pictorial styles (even as “ways of seeing” or “period eyes”) cannot be possible in the first place.37 Still, we should reserve judgment about the pace of biocultural evolution and therefore about the rate at which any recursions of neurovisuality could be disseminated.

In this regard, Wartofsky’s art-historical chronology might be way off even if his historicism can (perhaps must) be accommodated to evolutionary histories. Painter’s perspective in the Italian Renaissance was not the first place, or even the most important place, in which the “[three-]dimensionality of visual space” could be constituted pictorially. When visibilized, the virtual coordinate plane naturally three-dimensionalizes and spatializes depicted objects—“perspectives” them—without using the particular techniques of painter’s perspective. As Summers has argued, it can be found in ancient Egyptian pictures made at the beginning of the third millennium BCE. In fact, he has suggested that it was metrically organized in Egyptian pictorializations—that the plane was segmented in bands of equal width on the plane or of “depth” in the resulting virtual coordinate space. As he writes, “Egyptian painters and sculptors made choices that were to establish the basis of Western metric naturalism . . . by the development of planarity into the virtual dimension [i.e., the virtual coordinate plane], with consequences reaching to the present day.” Summers treats the history of these long-term consequences as an art history—a history of successive styles and technologies of pictorial representation building on the visibilization of the virtual coordinate plane in Egyptian pictoriality, culminating in painter’s perspective in the Italian Renaissance before its metaoptical generalization in Cartesian point space.38 Still, Summers does not always undertake to show how the technology was transferred from one locale of visuality to the next—for example, from ancient Egypt to archaic Greece—despite an extensive discussion of the conceptual origins of painter’s perspective in Arabic optics. I do not think that he would endorse the hypothesis of neurovisuality as I have tried to state it; his world art history depends on a strictly psychological (and Kantian) conception of the contribution of understanding (or the transcendental logic) to sensibility and its intuition of the sensible manifold (transcendental aesthetic) in a kind of transcendental deduction, or what he calls “abstraction to the notional.” But the hypothesis of neurovisuality could help enable him to explain how the consolidation of the virtual coordinate plane was disseminated, ramified, and preserved in human visualities and pictorialities, some of which had little material intercultural interaction.

In this regard, the virtual space constructed in painter’s perspective is a particular historical variant of spatiality constructed on the virtual coordinate plane; as Summers puts it neatly, “the orthogonals in a perspective construction are ‘really’ parallel lines perpendicular to the baseline, and all modules marked off by transversals in the grid are ‘really’ perpendicular to them.”39 These parallel lines can be constructed on the virtual coordinate plane, whether or not they are treated as modules. The ancient Egyptian examples given by Summers may be questionable. But it seems likely that some Greek architectural sculpture (such as the frieze of the Parthenon at Athens) was organized by explicitly planning the spatial relationships of depicted objects and figures on the virtual coordinate plane before laying them out as pictures on the plane of the format and carving them back to the secondary plane(s), even if these plans did not have to divide the virtual coordinate plane into bands of equal width in which the figures were located and even if the figures were not reduced in size on the plane of the format proportionate to their location in depth (as in perspective projection).40

Indeed, the virtual coordinate plane could well be found in any visual culture in which pictures are installed on flat upright surfaces axially organized as perpendicular to the line of sight (though permitting the oblique axis that naturally visibilizes the plane under certain visual angles allowed or enabled by the architectonics). It could be found, that is to say, in virtually every large-scale or monumental architecture known to us in complex societies, in which planed planks of wood, dressed blocks of stone, and other flat, rigid materials have been used to construct what has been called a “carpentered” visual world of right-angled architectures or “cuboid” architectonic spaces into which the virtual spaces of pictures were integrated. Often associated specifically with the rise of state-level civilizations in the ancient Near East, the Nile Valley, and elsewhere, this monumental architecture and the pictorial architectonics correlated with it is both very ancient (Neolithic) and practically worldwide in distribution. Many different kinds of visual and virtual spaces and many different kinds of pictorial effects can be constructed within it. But all of them can dimensionalize visual space in the correlate pictorial architectonics; within well-defined limits of the visual angle, all of them permit the visibilization of an artificial or man-made visual plane coordinate with the plane of the image in a retinocentric or viewer-centered frame.

The notion of a carpentered (visual) world—of a visual world that has been carpentered—has a number of ancestors. But it achieved classic formulation in famous studies by a trio of psychologists and anthropologists, Marshall H. Segal, Donald T. Campbell, and Melville J. Herskovits, who examined cross-cultural responses to the Müller-Lyer Illusion, exploring whether perceivers sensitive to (visually sensitized in) effects of plane recession in carpentered architectural environments (such as receding walls at right angles to the groundplane and increasingly distant from the vantage point) would transfer this sensitivity or skill to spontaneous interpretation of the two-dimensional graphic stimulus.41 The thesis is usually stated as a strictly psychological hypothesis about cultural relativity in visual perception, not a thesis about a historically evolved neurovisuality that is activated in contexts of visual affordance similar to the ones in which the neural circuits were laid down. Though not mutually exclusive, these hypotheses might be related analytically in a variety of ways. Onians has emphasized that the original study depended on an overly narrow (and ideologically shaped) hypothesis—namely, that perception can be influenced by culture—rather than a “broader” one, one that he says “goes further”—namely, that perception is influenced (neurally shaped) by environment, “whether cultural or natural.”42 But the more encompassing hypothesis must perforce include the narrower one. And the latter may require a special theory of particular recursions in vision: that is, of neurovisuality.

Still, problems abound. Many statements of the thesis as tested by the perception of geometric illusions (1) construe as possible effects of a carpentered world what could equally be described as effects of a “perspectived world”; (2) overlook possible culturalized interpretations of the illusion other than the “carpentered” ones, and in general downplay what I have called “radical pictoriality”43; (3) attribute carpentered worlds as a matter of real architectures only to urban and/or Western civilizations, often modern ones; and (4) understand carpentered worlds overly literally as a matter of visual spaces in real architectures as opposed to visual-virtual spaces in man-made architectonics that might include natural vistas, landscapes, buildings, pictures, and other artifacts organized axially, or in terms of what Summers calls “paths” and “centers,” and in terms of frontality and planarity (in the case of pictures installed in such architectonics). These limitations on statements of the thesis and cross-cultural experiments designed to test it mean that its cultural reach and historical depth, let alone its neurological implications, remain uncertain. For immediate purposes, I note here that the original hypothesis did not take the virtual coordinate plane in Summers’s sense to be responsible for the interpretations of the geometric illusions preferred by urbanized subjects, even if the plane is visually constituted or pictorially constructed only in carpentered architectonics. Rather, it took the interpretations to be steered or partly caused by certain effects of perspectival recession in carpentered worlds. Though related, these visual calibrations are analytically distinct. The virtual coordinate plane in Egyptian pictorialization did not lead to perspectival treatment of depicted objects on the plane of the format—no recession and no diminunition—even if such phenomena were perfectly visible in natural visual perspective in ancient Egyptian architectures.

Before Summers, perhaps the most sustained visual-psychological discussion of the standpoints, sightlines, planes, and spaces constructed in the monumental architectures of the ancient world was attempted by the architectural theorist Sigfried Giedion. Giedion emphasized the verticals and “verticalization” constructed in the monumental dressed-stone architecture of Egypt and Sumer, which he contrasted with the “multi-orientationality” of Paleolithic cave painting and other rock arts.44 “Adjustment to the vertical” is not quite the same thing, however, as visibilization of the virtual coordinate plane, though it may be correlated insofar as the virtual coordinate plane becomes visible in a picture organized on the plane of a format perpendicular to the observer’s groundplane (notionally flat) and to his line of sight (when axialized to address the picture head-on). And anyway I would locate the virtual coordinate plane not only in monumental architectures of the ancient Near East, Egypt, and the Mediterranean world. It may visibilize in “megalithic” or large-scale post-and-lintel construction in Neolithic and Bronze-Age Europe and elsewhere. Probably it had independent visibilizations in rectangular architectonics and pictorial planarity elsewhere in the world. Despite the contributions of Giedion, Summers, and others, archaeologies of ancient visual-virtual spaces have yet to be compiled and compared in these terms.45

Here, and once again, I want to make an analytic point instead of an empirical one. The question of the virtual coordinate plane as possible neurovisuality inevitably brings us to evolutionary-ecological aesthetics: how aesthetic experience in Zeki’s sense, or seeing that actively configures what it sees, has been shaped in natural selection in natural (wild) and artificial (cultural) environments, however “second nature” for their inhabitants.46 The logical synthesis of neuroaesthetics and evolutionary-ecological aesthetics might be called “neuroarthistory.” It has been so called in an ambitious book by John Onians, published in 2007. Onians urges that the neural routines of vision, including the visual routines of making artifacts, pictures, and artworks, are evolutionary products of visual adaptation to distinct ecologies of visual affordance, such as seasonal illuminations of topography, typical local geology, and the growth or behavior of regional flora and fauna. These environments have their own histories: climatological; geomorphological; botanical and zoological; and, of course, hominid and human. To quote one of Onians’s formulations, vision (and visuality if any) can be most economically and comprehensively understood in terms of “automatic responses generated by neural networks whose configuration has become more or less permanent as the result of frequent exposure to particular features of our environment.”47

Neuroarthistory in Onians’s sense need not invoke visual adaptation to visual-cultural environments as the explanation of neural-visual rules and programs—of the “automatic responses” of the brain habituated in “frequent exposure to particular features of our environment.” Neuroarthistory does not assume neurovisuality, though it is open to finding it. In recent work on the Aurignacian painted cave of Chauvet, for example, Onians suggests that the painters configured a simulation of certain things they saw in a natural ecology to which they were long adapted. If we adopt the strongest version of his theory, they could only do so. If this is correct, it might suffice to say that cave-painting (that is, this cave painting) merely stimulated the routines of vision already habituated to the ecology it pictured. No recursion of neurovisuality has occured.48 I set aside the question whether Chauvet or any painted cave in the Upper Paleolithic period in Franco-Cantabria had any kind of “frequent exposure” to viewing in a population—exposure frequent enough to habituate vision, as Onians’s theory requires—or why the vision of those who did make and see it, however restricted their number, might have been preferred or “selected for” in the biosocial reproduction of the population if whatever could be seen in or about the painting by these special people could be seen equally well in the extra-pictorial world by anyone else in the population. In the most economical statement of Onians’s theory, those people who did see the paintings in the cave could see and use them visually as presentations of things to which their “neural networks” were already habituated or disposed, with or without pictorial mediation.

Still, the paintings in Chauvet (on Onians’s account of them) may be special cases of trompes l’oeil. Arguably pictoriality as such has little or no causal role to play in the visual perception and interpretation of such works. Indeed, one can defend the radical thesis that we should sharply distinguish “illusionistic imaging” from depiction—that trompes l’oeil are not pictures.49 On this account, a trompe l’oeil might be able to shape vision neurally if it is frequently encountered as a particular feature of a habitual environment so long as it is not a mere duplication of just those things that can be visible just as well outside its illusion (if it is, then its causal role in shaping vision is diminished or eliminated). But it would not do so as a picture visible as such. It would bring unnatural or artificial things into being in the visual world—things such as the flat groundplanes of virtual coordinate planes—simply in virtue of the illusion.

The recursions of neurovisuality, if any, do not require that they be recursions cycled though the perception of pictures specifically. Nevertheless, pictorial recursions would be dramatic instances of neurovisuality precisely because most pictures envision some aspects of a world that are not otherwise visible in the world (or about it) in any way to which the visual brain could have become environmentally habituated. And they do so as visible pictoriality. The difficulty lies in determining when neural adaptation occurs in relation to such depiction—neurovirtuality. Even more fundamentally, the difficulty lies in determining when pictoriality as such becomes visible. When in neuroarthistory does neurovisuality specifically coordinate as neurovirtuality in a pictorial ecology?

VII. Neurovisuality Now

Over the long term, it would be surprising if human vision were not adapted to virtualities generated specifically in picture-making and cognate activities (including the making of “artificial memory systems” and “compu-notations”), that is, if its circuitries were not reintegrated in having played “image games” for more than forty millennia in (and as) different visual cultures.50 For all the reasons I have noted, however, it will be hard to detect these hypothetical successions and recursions in the past. And for this very reason, my suggestions have been analytical, not empirical. As remarked at the outset, they are intended to explore whether a general theory of visual culture can or should logically be accommodated to a general science of vision.

Of course, it is one thing to suppose that the human species is partly disjunct from ancestral species—that it is a historical variation in the hominid line—partly in virtue of its species-wide neurovisuality in picture-making activities or in the production and material management of particular aesthetic sensations. Clark, a historian of modern art, has been happy to identify the “species-defining” nature of the “form” that he thinks emerged in tool-making in Solutrean cultures of the Upper Paleolithic period—presumably an order of manual, visual, and mnemic order that (he thinks) indexes speciation.51 And it is another thing altogether to suppose that in neurovisuality the human species is continuing to vary—to diversify in ways that could lead to speciation. Clark would hardly suppose, I would guess, that artistic modernism is species-defining, whatever else it defines. Modernist rhetorics of the creation of a “New Spirit” (or a “New Man” or a “New Woman”) in the regimes of modern sensation and its reintegrations and reprojections in modern visual and material cultures have rarely gone that far. If they have, the proximity of dubious eugenics—of sociogenetic engineering by means of art and architecture—is obvious. If “creative evolution” occurs at all, presumably it is not limited to the evolution of late-modern men and women.

It is striking, then, that the rhetoric of present-day new media routinely asserts their power in terms of neurovisuality: the rules and programs of the software (sometimes conceived as having a self-autonomizing epiprogrammatic capacity) are getting wired up in those of the wetware, the human brain. Maybe this is just a metaphor for addiction or other processes of habituation and acculturation that do not amount to neural reintegration, or could be said to cause it. The wetware voraciously consumes the software in part because it has produced it specifically to satisfy cravings for imaging not so readily released from human image banks by other technologies or by natural visual perspectives unmodified by man-made extensions into virtuality—virtuality expressly tailored pictorially to be stimulating, often sexually. But in other times and places, in any time and place, it might well be that the then-contemporary “new media” exerted a similar grip or had a similar efficacy—that there is nothing neurally new about the effects in question. (Certainly the extraordinary expenditure of human energy in the Paleolithic visual cultures has long been taken to betoken a “creative explosion” in human media and even in human mind—a world-historical threshold.52) New-media enthusiasts point to new forms of community and subjectivity that emerge from personal and social interaction in new media. But there does not yet seem to be any clear evidence for infraspecific variation, let alone interspecific variation, driven by new-media adaptations of neurovisuality. Is it too soon to tell? Or simply unlikely?

How can we address this contradiction? In conclusion, I notice the analytic equation with regard specifically to the vexing empirical question. The many “old media” can be investigated in many ways, from ophthalmology to art history. As we have seen, however, recursions of neurovisuality, if any, are likely to be invisible analytically, at least in part. They are a “black box”: we cannot make the relevant neuropsychological experiments; comparisons and controls are inaccessible; evidence is indirect, read out of uncooperative fossils that present many forensic difficulties and must be subject to intrusive interpretation, especially if they are pictures or artworks. In technosocial environments of the new media today, however, we would seem to have ideal scientific and hermeneutic conditions to investigate neurovisuality—for the very first time in history. For the first time, psychophysiological observations of the electrochemical activity of the human visual brain or other experimental and clinical investigations can be brought into direct relation with introspective, ethnographic, and sociological information about subjective experience in light of the historical thesis that interactions with visual-cultural media generate new neural circuitry. Which media and what effects produced in them stimulate neural reintegration and adaptation, if at all?

In our own time, this question should provoke terrifying ethical and political excitements and doubts—inevitable excitements and essential doubts. But there is no way around that. And it may be just as well. Within visuality we might accept neurovisuality, or contest it—we must do so. For if one thing is certain it is that somewhere some people are now working very hard to bring neurovisuality into being.

Things remain visible to people outside the visuality within which they were intentionally produced, though what is visible in an artifact in this context (or what is visible about it) may differ from what is visible in the context of visuality. By the same token, people can succeed to many visualities, though both Wölfflin and Panofsky were somewhat uncertain (on different grounds) about just how far it is possible to do so when we are dealing with visualities constituted in the past and accessible to us only in things made to be visible within them that happen to have survived into our own visual world.

Whitney Davis elegantly lays out the relation of visuality to history in this passage, early in his impressive account of what he calls “neurovisuality.” As his references to Heinrich Wölfflin and Erwin Panofsky suggest, there are important ways in which the problems he elaborates are continuous with old problems in the field of art history. Crucially, in the current context, he addresses himself to the problematic notion that people can look at a work of art made in an earlier epoch and find that “what is visible” in those works is not what they were intended to make visible, and that what they were intended to make visible is no longer visible in them.

Davis takes for a concrete example a work of Albrecht Dürer’s, which both Wölfflin and Panofsky discuss. Davis glosses their efforts. By “removing the background in Dürer’s engraving of Knight, Death, and the Devil of 1513…to expose the outline silhouette of the Knight, to visibilize it, Wölfflin meant to show how his readers—that is, people today—can still see the primary rhythmic configuration of a pictorial artwork made five hundred years ago. That is, Wölfflin stands for the idea that we can still see Dürer’s linear rhythm after all these centuries. Panofsky stands for the contrary: “In Panofsky’s iconology, remember, we can only use a picture that was made in the distant past or in a different culture in discursive ways; we cannot fully use it visually in the way that its makers did.” We can talk about how a sixteenth-century Lutheran would have seen it, but we can’t see it (“use it visually”) ourselves.

But what is “it”? Neither Wölfflin nor Panofsky (nor Davis) would claim that we can’t see Dürer’s Knight, Death, and the Devil. The complicated relation between cultural position and vision is the topic they address. But they all admit that we can look at the work of art. Panofsky might be a pessimist and Wölfflin an optimist, but they are concerned about the uncertain prospect of our succeeding to visualities intended for historically remote beholders.

Our best chance seems to be that offered by Wölfflin, who removed the background from the print, thereby enabling the latter-day beholder to “see” the rhythm of Dürer’s linear work. But one might equally argue that Wölfflin is at least as pessimistic as Panofsky, because, in order to make it possible for us to succeed to the original work’s visuality, he had to show us “an autonomous artwork, …one produced by Wölfflin rather than Dürer.” We can “use it visually,” but “it” is a new work, one Wölfflin produced for beholders who were his contemporaries. Or, perhaps we should say that, what we can finally see (“use…visually”) is Dürer’s print, but we see it by looking at Wölfflin’s illustration. Either way, the object of our visualization is not in front of us. We can see Dürer’s print, just not by looking at it.

To say that (neuro)visuality defines itself in terms of the effects an object of visual culture produces in the viewing subject calls into question the place of the work of art or of visual culture in it. If Wölfflin’s altered image can afford me an equivalent for the experience of linear rhythm Dürer’s print afforded its original audience, then the “succession to visuality in neural circuits” does indeed take place crucially “in natural history” and “in social life,” but the specificity of such successions lies in the “neural correlate” or “causation” or at any rate in “the visual succession to visuality,” and not in the work of art or visual culture. Ultimately, this means we’re each of us his or her own Wölfflin: we experience neural correlates afforded by images, but the “succession to visuality” takes place “in neural circuits.” And looking at the work of art turns out to be irrelevant (if not inhibiting) to our success in seeing it.

Strange as it may sound, I think this is obviously right. Imagine looking at a badly damaged copy of Dürer’s Knight, Death, and the Devil. Imagine tears, stains, vandalism, etc., obscuring the image. Perhaps even an inscription applied to the work in, say, the nineteenth century. (To enhance the thought experiment, let’s imagine something inane written by some commentator whose understanding of Dürer’s project was evidently limited.) Indeed, let’s say it has been colored in.

An astute scholar of Dürer’s work will still be able, we hope, to succeed to visuality by doing what Wölfflin taught us to do—namely, by responding to what we know rather than what we’re looking at. This learned acculturation may, as I hope I rightly take Davis to speculate, become part of a recursive neurovisuality. That is just to say that our knowledge of history and our sense of what time’s toll looks like may become part of the way we see old works of art at the neural level, thanks (I gather) to the plasticity of our brains. Whether this hypothesis is true, and regardless of the extent of that truth, this will justify me in doing what I have always done, as an art historian: attempt to interpret works of art.

But Davis refuses this point. In fact, he notes regretfully that “art historians can overlook neural causalities that might operate outside visuality—causalities that might explain why pictures or artworks can retain their visual ‘power’ or ‘agency’ (aesthetic or otherwise) far beyond their original contexts of making in a particular historical visuality, that is, why they can be globally transmitted between historical visualities despite tenuous material connection between the social groups or visual cultures in question.” In other words, after convincing me that my knowledge should trump my vision, Davis tells me my vision should be counted on to transcend my knowledge. The object of my attention is now the object of my vision—the thing I’m looking at.

Can we have it both ways, or is this special pleading, now on behalf of interpretation, now on behalf of subjective experience?

One can imagine a way to reconcile the two commitments. Suppose those “pictures or artworks” that “retain their visual ‘power’ or ‘agency’…beyond their original contexts of making in a particular historical visuality” and can therefore “be globally transmitted between historical visualities despite tenuous material connection between the social groups or visual cultures in question” are precisely those pictures or artworks that do not happen to raise the historical problem we began by discussing—that works from remote milieux are liable to afford us different visualities from those they were intended to produce. Let’s say, they’re works for which an all-knowing Wölfflin would make illustrations (such as his altered Knight, Death and the Devil) that looked just like the original. If that is so, Davis has no contradiction to explain. But if that’s so, it turns out that interpretation always precedes vision (if only logically). Art history needs no neuroanything.

Whitney Davis Responds to Charles Palermo:

Charles Palermo’s response to my speculations about “neurovisuality” (let alone what I called “neurovirtuality”) is right on target. Is the concept needed at all? Or does it needlessly proliferate art-historical ontology (whether at the level of method or theory or both) beyond essential categories? I believe in Occam’s razor as much as the next hairy beast. To use Palermo’s turn of phrase, then, does the hypothesis of neurovisuality simply ratify us in doing what he always does—“attempt to interpret works of art”? (For myself, I’m involved in a quite different project—explaining pictures. But we can be indifferent to that difference, I think.) And in an obfuscatory language to boot? Or does it bring something to that very enterprise? To give the hypothesis of neurovisuality its best shot, the answer to his final question is that “interpretation always precedes vision” because of neurovisuality or as neurovisuality, not in spite of it or without it, as he implies. More exactly, what he calls “interpretation [that] always precedes vision” is neurovisuality, the neural representation of the “knowledge” that he might use as a beholder to make sense of the intentional order of the artwork or picture, just as the effect that a “visual object produces in the viewing subject” simply is neuroaesthetics (though I tried to show that neuroaesthetics without an account of neurovisuality likely cannot handle the subjective effects of beholding an artwork or a picture insofar as its culturally particular intention is salient).

In both logical and material terms there may be no ontological difference between interpretive looking in Palermo’s sense and neurovisuality in mine. As I said in my essay, our choice of analytic representation has little theoretical valence. It’s instrumental. In talking about visual artworks or pictures made to be visible, you can talk about interpreting their intentions or you can talk about seeing them. Claims advanced in either terminology are fully interconvertible if they are indeed claims about the intentional autonomization of autonomic processes of vision or proprioception of any other kind. I have no truck with (and I try not to trade in) essentialist claims for verbal ekphrasis as the only way to represent the (neuro)intentional order of the objects in view. (Palermo plays his cards close on this score: he just “interprets works of art.” But it’s possible that this means he just doesn’t want to learn to write differently.) If you don’t like the specialist or mandarin terms of the one, move to the other. My hunch is that his intentionalism should best understand itself as materialism, and therefore as an archaeology of successions and recursions in (for example) the visual brain, with the phylogenetic and ontogenetic histories that are entrained by that analysis.

At a rhetorical level, however, there may be a salient difference between interpretation in visuality (and of it) and the neurovisuality of interpretation (and in it)—though ironically it’s the difference that explains why neurological investigations and what seem to be ontologically more economical inquiries might well make epistemological common cause. Obvously interpretive “looking”—in a parallel to “reading”—is culturally bound, and historically particular. It confronts not only the problem of making sense of the productions of “historically remote beholders” (Palermo focuses on it as an art historian). It also confronts the question of people in one’s time and place who do not share one’s culture (this is the troubling case behind the scenes in this debate). Intentionalism has long been associated with exclusivistic hermeneutics—with having the right knowledge to look at things aright—and by circular appeal to intention it has long justified the teaching of the supposed knowledge that enables correct interpretation. (That’s why intentionalism has fallen into such bad repute.) The hypothesis of neurovisuality dodges this defect without denying the principle. If visuality is exclusive to those who can “look” and “read,” neurovisuality is open to anyone with a brain: intentionalism for everyone and anyone. Again, a point of intellectual tactics, not strategy.

I must clear up one misunderstanding, however. I did not mean to argue that art historians overlook “neural causalities that might operate outside visuality” because they ignore neurovisuality. That would indeed be a contradiction. Rather, I suggested that they might overlook one contribution of neuroaesthetics (distinct from neurovisuality), namely, that it might address such causalities. I’m agnostic about them as a point of science, but I’m interested in them as a matter of history. As I said, they could help explain the cross-cultural or transhistorical “power” and “agency” of artworks, pictures, and artifacts, possibly even their “interpretability”—one of the deepest unresolved questions in art history. Palermo notes a contradiction between neuroaesthetics and neurovisuality, between “subjective experience” and “interpretation” as represented neurally (and irrespective of their discursive representations in aesthetic and hermeneutic science). And so there is. But it’s not my contradiction. It’s the contradiction—or, more realistically, the uncertainty—between a neurology that does not find conventions and intentions wired or writ into the rules and programs of the brain (at least at some empirical level of recursion that could be modeled analytically) and one that does.

Art history may need no neuroanything. But Palermo may have missed the drift of my argument—my stated theme. Neuroeverything needs art history.

Einstein’s absence is most telling in the literature on cubism, the art on which he wrote prolifically and for which he is one of the most important early commentators–indeed, his chapter on cubism in his Die Kunst des 20. Jahrhunderts (1926), his most developed treatment of the subject up to then, had the distinction of being the longest, most intellectually ambitious text on this art to have yet appeared in either French or German.4 Yet that account, his revised and expanded versions of it in the second and third editions of the book (1928, 1931), his Georges Braque (1934), and his numerous other shorter writings on cubism and cubist artists are rarely noted in the literature on this art.5

Einstein’s marginal reception contrasts markedly with that of his friend, Daniel-Henry Kahnweiler, the transplanted German who was the Paris dealer of Braque and Picasso during their cubist phase and knew their work more intimately than anyone else. His book Der Weg zum Kubismus (The Rise of Cubism), which appeared in 1920 under the pseudonym Daniel Henry, is still taken seriously, still debated within cubist scholarship. The art historian Yve-Alain Bois has hailed Kahnweiler as the only early commentator “to give an intelligent account of cubism,” “a passionate critic” with a “fantastic eye,” “whose breadth we have only begun to appreciate.” Indeed, he has praised Kahnweiler’s theoretical account of cubism as “one that in many respects remains unequaled today.”6

What then of Einstein? Could we not say much the same of him? I believe we could and should. I say this even though I regard Kahnweiler’s account of the formal development of Braque’s and Picasso’s cubism as the more concrete, nuanced, and persuasive one, because more attentive to the evolving formal problematic of their painting. Einstein’s interpretation of cubism may be flawed as an explanatory model, yet it is the only early writing on this art that does justice to the radical implications of cubist representation, as painting that can fundamentally alter not only our conception of art but our intuition of the visual world, and in so doing alter our subjectivity. But here, rather than exploring the different interpretations in detail, I wish to focus on one issue that is central to both authors’ interpretations of cubism—the role played by memory in the act of viewing cubist pictures. On this question Einstein and Kahnweiler held diametrically opposed positions. Moreover–and this is my main interest–their respective positions correspond to successive phases in the developing neuroscientific understanding of the visual brain. Kahnweiler’s interpretation of cubism was shaped by the neuroscience of his day while, remarkably, Einstein’s account of seeing, as he believed it to be embodied in cubist paintings, anticipates by half a century a fundamental breakthrough in the neuroscientific understanding of vision.

The differences between Kahnweiler and Einstein take on added interest in light of the work of Semir Zeki, a neurobiologist who has done major research on the visual brain and has proposed a neurological explanation of cubism. He does this in his pathbreaking book, Inner Vision: An Exploration of Art and the Brain (1999), the first extensive attempt to apply the new neuroscience to the study of art. Moreover, Zeki offers an account of the functioning of the visual system, the parts of the brain engaged in processing data from the retina and producing our images of the world, that has certain striking affinities with Einstein’s account of cubist perception, even as it helps us more precisely to locate the radicalism of Einstein’s conception of art in relation to the current neuroscientific understanding of the visual brain. Yet there is irony here. Einstein’s theory of art, as we shall see, was founded on the idea that vision was an active process, long before neuroscientific research corroborated this fact. Cubism was the pictorial embodiment of that truth. Yet Zeki’s interpretation of cubism is closer is some respects to Kahnweiler, whose understanding of vision was based on the old neuroscience.

One may legitimately ask why this matters—what do we gain by relating Einstein to a scientific discourse that postdates his death by more than three decades? In my view, it is a matter less of what neuroscience can do for Einstein than what Einstein might offer to neuroscience. While Einstein anticipates current understanding of the visual brain, he also offers an important corrective to efforts by neuroscientists to apply their knowledge to art.7 The applications of the neuroscience on the visual brain to visual art have up to now been overwhelmingly concerned with ostensibly universal aspects of visual perception and aesthetic response; in general these neuroscientists have shown little interest in the specificity of art-historical cases, and this also remains largely true in the nascent fields of neuroaesthetics and neuroarthistory. I am less interested in ‘universal’ neurobiological responses to art than in understanding how historically specific artistic practices, and historically specific interpretive accounts of those practices, can be understood in neuroscientific terms and how in turn they might enrich neuroscientific research on the visual brain. Here I will argue that Einstein’s interpretation of cubism has notable implications for these emerging fields and productively complicates the current understanding of the relationship between artistic activity and the functioning of the visual brain. As such, his writing deserves consideration in the emergent discourses of neuroarthistory and neuroaesthetics.8

Kahnweiler’s early writing on cubism dates from his five-year exile in Bern during the First World War and its aftermath. When Germany declared war on France in August 1914, Kahnweiler and his wife were on holiday in Italy. As a German national he was barred from returning to France, and, at the invitation of his friend and client Hermann Rupf, he opted to sit out the war in neutral Switzerland. During his Swiss exile, forced into suspending his work as a dealer, Kahnweiler had dedicated himself to reading philosophy, psychology, and art history in an effort to “explain to myself and to others what had happened, what cubism was.”9 In 1915 he wrote a long theoretical essay, ”Der Gegenstand der Ästhetik“ (“The Object of Aesthetics”). In this text, which remained unpublished for more than a half-century, Kahnweiler treated cubism in the broadest historical trajectory as a decisive shift in the practice of plastic representation and of viewer response.10 His short book on cubism, Der Weg zum Kubismus, he developed from the last chapters of this manuscript. It was destined to become the classic early critical text on the movement.11

For Kahnweiler the fundamental problem facing Cubist painting was a strictly pictorial one that had emerged with impressionism: the conflict between Darstellung and Aufbau, between illusionistic representation and an increasingly autonomous pictorial structure (Kahnweiler, Rise of Cubism, 1). One of the fundamental tasks of painting, as he formulated it, was “to represent three dimensions and color on a flat surface, and to comprehend them in the unity of that surface” (7). In the first phase of cubism Braque and Picasso attempted to resolve this conflict by adapting three-dimensional objects to the painting surface through extreme distortions of form. Yet this discrepancy between the beholder’s generic memory images of such objects and their deformed pictorial representation was deeply disturbing. In the summer of 1910, writes Kahnweiler, Picasso found a solution to this conflict. It was then that he took “the decisive step that detached cubism from the previous language of painting. . . . He had pierced the closed form” (10).

Comparing Picasso’s Woman with Pears (fig. 1) with his The Guitarist (fig. 2) we can see the effects of this radical step. Even in the abstract, faceted forms of the woman we can still perceive closed contours; line and color are synthesized to create a sense of volume by means of modeling. In The Guitarist there are no closed forms; planes are suggested but not consistently defined; line and color now began to function as independent entities. As a result brushstroke, largely subordinated to modeling in the earlier work, becomes more individuated and texturally varied. This piercing of the closed form eliminated perplexing deformations of the motif—indeed, without the title, which Kahnweiler deemed essential for these more abstract works and which he himself usually provided, we would scarcely if at all be able to identify it. As Kahnweiler described it, following this step the subordination of parts to the unified pictorial structure “can take place without producing disturbing deformations, since the object in fact is no longer ‘present’ in the painting, that is, since it does not have the least resemblance to actuality” (12). At the same time, these highly abstract works came to include selected “real details,” as Kahnweiler called them—lettering, clay pipes, bottles, tassels, etc.— integrated into the structural whole (fig. 3) (11). Such details, augmented by the painting’s title, were, in his words, “a stimulus which carries with it memory images (Erinnerungsbilder). Combining the ‘real’ stimulus and the scheme of forms, these images construct the finished object in the mind. . . . There is no possibility of conflict here, and yet the subject ‘recognized’ in the painting is now ‘seen’ with an intensity of which no illusionistic art is capable” (12). Ultimately, for Kahnweiler the painting is reconciled with the known, familiar world as given.

In 1923, before Einstein had himself written anything of substance on cubism, he praised Kahnweiler as “the only one in Germany who described and explained cubism correctly.”12 Yet, in his long letter to him from that same year it becomes clear that Einstein’s views on what was at stake in cubism went far beyond Kahnweiler’s narrowly circumscribed aesthetic interests. “I have long known that the thing one calls ‘cubism’ goes far beyond painting,” he declared. “Cubism is tenable only if one creates equivalents in the mind. ”13 In these two sentences, Einstein opened up a wide gulf between his conception of cubism and Kahnweiler’s. First, that it goes beyond visual art; secondly, that ideally its impact will ultimately be a radical refiguring of the mental world of viewers and their intuition of visual phenomena. The experience of cubism is no mere theory, he insists, but leads to a “gradual modification of sensations,” in which “a person waxes and wanes . . . in the sensation of himself or his feeling for objects, in the harnessing of time,” so that what is represented is “the very history of the sensations, experiences brought close up, whose symptoms are at best so-called things” (140, 141). Kahnweiler had written of the cubists’ desire to glorify “the beauty of things” (1); for him cubism was about a new form of aesthetic pleasure; it was not about changing human subjects and their mental representations of self and world. He did not respond to Einstein’s letter, but we can easily imagine that it must have struck him as alien to his own conception of cubism.

Three years later, in Die Kunst des 20. Jahrhunderts, Einstein presented his own account of cubist painting. The most critical point about Einstein’s cubism is that, in contrast to Kahnweiler, he understood it as an “example of a subjective realism” rather than as an abstract art—realist, because as Einstein interprets it, cubism is based on “the direct experiences of the human subject”; it seeks to reconnect to a direct phenomenological perception of objects as experienced in time and space (63). What is represented is not, as in illusionistic art, ostensibly stable objects that exist apart from perceiving subjects, but rather our own unfolding subjective process of vision as we apprehend and mentally construct the volume of objects in moving toward and around them in space. In short, a cubist painting presents us with a synchronic image of a diachronic process, a process Einstein called das vorstellende Sehen, envisaging seeing (59). These different viewpoints, these discrete moments of envisaging an object or objects, rather than being abstracted into a single motif as in previous painting, are synchronically represented as planes and juxtaposed on the two-dimensional picture surface. What Einstein describes here fits the style of Braque’s and Picasso’s cubism of the period 1908 to early 1910; it is harder to reconcile with the subsequent phases of their cubism up through 1914, when their paintings often began as “pure paintings,” furnished only toward the end of the process with “attributes,” figurative details.14 It is noteworthy that he says nothing about the fundamental shift that came with the puncturing of the closed form later that year, which Kahnweiler regarded as the fundamental strategic shift in cubist painting. Yet it follows from Einstein’s interpretation of cubism that for him there was no problem with deformation, so there was no need of a solution—“We can speak of deformation,” he wrote, “only if we declare imitation to be the task of art.” Our normal habits of perception were deformed, but not objects, which were but “signals and symptoms” of our own visual and cognitive functions (63, 64).

Einstein’s interpretation of the significance of multiple viewpoints in cubist paintings, a focal point of most of the early commentary, was radically original within cubism criticism itself. For Kahnweiler this feature was motivated by the quest for plasticity in representing a three-dimensional body in space on a two-dimensional surface (Kahnweiler, Rise of Cubism, 11, 12). The most influential explanation, however, was that of Maurice Raynal, a friend of Picasso’s, which was widely embraced and repeated. According to Raynal, writing in a 1912, if art is a means of “augmenting knowledge,” he declared, “its function will only be served by painting forms as they are conceived in the mind.” This interpretation was directly related to multiple viewpoints. “We never see an object in all its dimensions at once.” “We must fill in the gaps. Conception gives us the means.”15 Yet despite their shared focus on multiple viewpoints, Raynal’s understanding of cubism is diametrically at odds with Einstein’s.16 For Raynal what is painted is the world as already known. The objects are seen from multiple viewpoints, but it is information-gathering, not a dynamic process of generating an object sequentially out of the viewing subject’s movements in space. For him and most other early cubist critics, including Kahnweiler, the object exists before the painting; for Einstein the painting is the product of the object’s genesis in the mind of the artist.

Yet it is on the issue of visual memory, crucial for both Einstein and Kahnweiler in their respective theorizations of cubism, that the most fundamental difference between them emerges. For Kahnweiler, as we have seen, Braque and Picasso solved the conflict between representation and structure by abandoning deformation of the object in favor of an abstract geometric grid into which realistic details and sometimes words were inserted as clues. These realistic details, aided by the picture’s title, triggered memory images in the viewer, memories of the familiar object world. For Einstein it is precisely such memory images that prevent us from seeing concretely and directly: “in reality,” he writes, “we do not at first see purely, with optical directness, but quickly associate a cumulative memory image with some known optical stimulus that obscures the genetic stimulus with a supposedly stable and comprehensive image. We conceal from ourselves that this memory image is a reconciliation of temporally as well as optically (qualitatively) distinct actions, and this image seems to endure because as something latent, mechanized, and rather unspecific . . . it is fused with visual function,” namely the present act of perception. Cubism disrupts these well worn mnemonic circuits. “We discover,” Einstein writes, “that the object is a nodal point of functions, the result also of subjective activity, that its rigidity is effected above all by linguistic habit and the desire of enabling extremely easy–i. e., conformist—actions; thus it is a matter of biological memory” (Einstein, Kunst des 20. Jahrhunderts, 58, 64). Cubism displaces and overwrites that memory by painting phenomenological optical experience. It is important to note the distinction between two kinds of memory—the biological memory formed by long-term perceptual experience, i.e. the kind that Kahnweiler has in mind when he refers to Erinnerungsbilder, and what I would call instantaneous short-term memory that accompanies a single event of perception as it unfolds in space and time. This latter form of memory is what Einstein refers to with the term simultané (63, 68).

The function of visual memory and its role in object recognition—the final step in the complex physiological and neurological process of seeing—is not only central to the differences between Einstein and Kahnweiler, it also figures in Semir Zeki’s treatment of cubism. He relates current knowledge of the visual brain to cubism, and finds this interpretation supported by some of the earliest interpretations (50-57).

As Zeki relates in his earlier book, A Vision of the Brain, until recently science understood vision as an essentially passive process; the image, it was thought, is received whole by the retina as though on a photographic plate,17 and is then transmitted through nerve fibers to the primary visual cortex at the back of the brain, and analyzed and interpreted by another, surrounding and distinct cortical area, which was known as the “association cortex.” There “received visual impressions are associated with previous visual impressions of a similar kind, resulting in recognition.”18 Kahnweiler was at least superficially familiar with this neuroscience on the visual brain. In a chapter of his Der Gegenstand der Ästhetik entitled “Das Sehen” (“Seeing”), he quoted from the recent book Gehirn und Auge, in which its author Robert Bing describes the primary visual cortex as a “Wahrnehmungszentrum,” a center of perception, which through the “Assoziationsfasern,” association fibers, connects to the occipital lobe, which contains “die optischen Vorstellungszentren,” the centers of mental representations, “in which the memory images (Erinnerungsbilder) for establishing the meaning of a seen object are stored.” Erinnerungsbilder, we recall, is the term Kahnweiler would use in Der Weg zum Kubismus, and Bing was almost certainly his source.19

Since the 1970s neuroscientists have discovered that the visual brain is much more complex than the old model of the primary visual cortex and an association cortex. Most importantly, they have now established that vision is an active rather than a passive process. The visual brain does not receive images from the retina, as was previously believed; rather the retina transmits to it unfiltered, unprocessed visual data from which the brain actively produces images, synthesized from stimuli processed by geographically and functionally distinct areas of the visual brain, in temporally discrete stages separated by milliseconds. The neurons in each area have the specialized function of responding to a different attribute of the visual scene, such as form, color, and motion. The primary visual cortex, V1, has been compared to a post office, where information gets sorted and distributed to other visual centers with specialized functions. V3 processes information related to orientation and, it is believed, dynamic form. V4 deals with hue discrimination and color constancy; V5 processes information on motion and stereoscopic depth.20 There is further specialization within these areas; some cells are sensitive to vertical lines, others to horizontal or diagonal lines in specific directions; other neurons respond to particular colors, and so forth. All of this data must be coordinated and matched with items in the cumulative catalogue of our visual memory, which is believed to be localized, at least in part, in the area anterior to V4.21

This process of object recognition is crucial to the cognitive function of the visual brain, for the information we receive from the visual domain is, in Zeki’s words, in “a continual state of flux” (Zeki, Inner Vision, 5). We see objects in different lighting conditions, which affect the appearance of their color, from different angles and distances, distorting their shape. This means that the visual brain has work to do if we are to make sense of what we see. Zeki explains: “Vision must . . . be an active process requiring the brain to discount the continual changes [of visual phenomena]and extract from them only that which is necessary for it to categorize objects.” The brain acquires “knowledge about the enduring and characteristic properties of the world; the brain is consequently interested only in the constant, non-changing, permanent and characteristic properties of objects and surfaces in the external world, those characteristic that will allow it to categorize objects”(6).

The major premise of Zeki’s book is that “the function of art and the function of the visual brain are one and the same or at least that the aims of art constitute an extension of the functions of the brain” (1). Art, like the brain, seeks “to represent the constant, lasting, essential and enduring features of objects, surfaces, faces, situations, and so on, and thus allow us to acquire knowledge, not only about the particular object, or face, or condition represented on the canvas but to generalise from that to many other objects and thus acquire knowledge about a wide category of objects or faces” (9-10). This quest for “the constant, lasting, essential and enduring features” of visual phenomena is central to Zeki’s interpretation of cubism. As he proposes, the cubists, like the visual brain, seek to grasp the “constant and essential elements” of the visual field, disregarding the fugitive accidents of appearance (50). Zeki really believes that objects have constancy and stability; for Einstein that is a convenient fiction of “biological memory,” born of the practical necessity of our functioning in our visual environment.

Zeki’s theorizing is not completely unmoored from history, however. He cites some of the earliest French commentary on cubism, which supports his conviction of the common function of the visual brain and of visual art. He cites statements by, among others, Picasso and Juan Gris and the critic Jacques Rivière, whom he quotes: “The true purpose of painting is to represent objects as they really are, that is to say differently from the way we see them. It tends always to give us their sensible essence, . . . this is why the image it forms does not resemble their appearance” (11). Zeki probably also found confirmation for this view in Kahnweiler, whose book on Juan Gris he cites. According to Kahnweiler, the cubists “strove to produce a complete image of the objects signified . . . which should be at the same time devoid of everything ephemeral and accidental, retaining only what was essential and permanent.” “Unsatisfied by the fortuities of a single visual impression, [cubism]endeavored to penetrate to the very essence of an object by representing it, not as it appeared on a given day at a given time, but as it exists ultimately composed in the memory.”22 Later in the chapter, Zeki again cites a passage by Rivière that fits neatly with a neurobiological perspective: “Contrary to what is believed,” writes Rivière, “sight is a successive sense; we have to combine many of its perceptions before we can know a single object well.”23 What the cubists were trying to do, Zeki now asserts, “was to try and mimic what the brain does. . . . They decided to depict all the different views and unite them on a single canvas, much as the brain unites what is seen from all different views” (51).

This may sound a lot like Einstein, but when Zeki turns to Picasso’s Man with a Violin (fig. 4), a funny thing happens: the painting is so abstract, with so many points of view “that the final result is only recognisable as a violin player through its title. A brain ignorant of that title can hardly construe that this is a violin player. The brain of course regularly views objects and people from different angles, but it is able to integrate those different views in an orderly way, allowing it to obtain knowledge about what it is viewing. The attempt by Cubism to mimic what the brain does was, in the neurobiological sense, a failure—an heroic failure perhaps, but a failure nonetheless.”24

Now what Zeki sees as a failure is precisely what Einstein, with a radically different understanding of cubism’s agenda, saw as cubism’s triumph. But before I elaborate this point let us, now that we are familiar with Zeki, revisit a passage from Die Kunst des 20. Jahrhunderts that I earlier quoted in part, in which Einstein describes what he calls cubism’s Gegenstandsgenetik, its genesis of the object. It is even closer to contemporary neuroscience’s understanding of the visual brain than the statements by Jacques Rivière cited by Zeki. Addressing the hermetic abstraction of cubist painting, Einstein writes:

One might perhaps object, this is not how we see in reality; yet in reality we also do not at first see purely, with optical directness, but quickly associate a cumulative memory image with some known optical stimulus that obscures the genetic stimulus with a supposedly stable and comprehensive image. We conceal from ourselves that this memory image is a reconciliation of temporally as well as optically (qualitatively) distinct actions, and this mental image seems stable because as something latent, mechanized, and rather unspecific —specificity comes from the linkage of the memory image [i.e., the “biological memory”]with the individual stimulus–it is fused with visual function. (Einstein, Kunst des 20. Jahrhunderts, 58)

This is a much more nuanced account of the visual process than we find in other writers on cubism. Moreover, Einstein’s emphasis on perception of an object as comprising “temporally as well as qualitatively distinct actions” anticipates current neuroscientific knowledge. Even if he doesn’t know of the functionally distinct regions of the visual brain, he seems to intuit that the image is a result of an active neural process unfolding in discrete stages. In Einstein’s words, “Cubism put an end to the laziness or fatigue of vision. Seeing had again become an active process” (Einstein, Die Kunst des 20. Jahrhunderts [1931], 107). This is what I meant by cubism’s triumph, as Einstein saw it. In his “Notes on Cubism” (1929) he illustrated Man with a Guitar, a work from the same year as Man with a Violin, which so perplexed Zeki.25 “It was,” Einstein writes there, “the cubists who undermined the object forever identical with itself, in other words they undermined memory, in which mental images are reconciled with one another. Their chief merit is having destroyed mnemonic images.” The cubist painting becomes “the distinguishing sign of the visually active human being, constructing his own universe and refusing to be the slave of given forms.”26 In other words, for Einstein cubism short-circuited the normal operations of the visual brain, allowing us to discover vision as a creative act to make the world new. Einstein summed it up concisely in his Die Kunst des 20. Jahrhunderts: “Painting or sculpture become necessary to a critique of visual intuition. . . . For the sensory is not some fixed, limited material that is altered only when interpreted through concepts; visual intuition and seeing change and exhaust themselves, and optical dissatisfaction forces such change; . . .what is at stake is not reproducing [Abbilden] but forming [Bilden]” (57).27

The notion that visual art precisely does not parallel the operations of the visual brain, but is rather continually unsettling and refiguring our construction of the visual world is not unique to Einstein—one finds it as early as 1876 in the writings of Conrad Fiedler. “Artistic activity begins,” wrote Fiedler, “when man finds himself fact to face with the visible world as with something immensely enigmatic; when, driven by an inner necessity and applying the powers of his mind, he grapples with the twisted mass of the visible world which presses in upon him and gives it creative form. . . . What art creates is the world, made by and for the artistic consciousness.” This is not a world, Fiedler insists, that existed prior to its realization through art: “What excites artistic activity is that which is as yet untouched by the human mind” (Fiedler, On Judging Works of Visual Art, 48-49). In that same year Stephane Mallarmé wrote of impressionism: “the eye should forget all else it has seen, and learn anew from the lesson before it, should abstract itself from memory, seeing only that which it looks upon, and that as for the first time.”28 This is consistent with how Claude Monet described his practice to Lilla Cabot Perry:

“When you go out to paint, try to forget what objects you have before you—a tree, a house, a field, or whatever. Merely think, here is a little square of blue, here an oblong of pink, here a streak of yellow, and paint it just as it looks to you . . . until it gives your own naive impression of the scene before you.”

He said he wished he had been born blind and then had suddenly gained his sight so that he could have begun to paint in this way without knowing what the objects were before him.29

It is noteworthy that the young Wassily Kandinsky, not yet a painter, experienced bafflement when he saw a Monet Haystack for the first time: “That it was a haystack, the catalogue informed me. I didn’t recognize it. I found this non-recognition painful. . . I had a dull feeling that the object was lacking in this picture.” And yet this strange picture “gripped me” revealing “the unsuspected power of the palette.”30 There are countless such examples.

In other words, effective visual art does not, as Zeki claims, parallel the operations of the visual brain, which always favor generalized repetition of the previously seen; rather art is continually unsettling and refiguring our construction of the visual world, working against the brain’s reproductive and classificatory operations. Vision has the potential for agency. As Einstein puts it: “In the act of looking we change man and the world.”31 He wrote these words with reference to cubism, and it is in his writings on cubism that we find his ideas most fully developed.

Current neuroscience, even as it acknowledges the ‘plasticity’ of the brain, its capacity to be structured by experience, treats vision as purely neurobiological. The situation described by the philosopher Marx Wartofsky nearly four decades ago is little changed today: “. . . the historical development of modes of perceptual action is not yet mapped into accounts of neurophysiological structure.”32 The leading neuroscientists who work on vision have shown no acknowledgment of the historicity of vision, or of how artworks and other visual representations might alter the structures and influence the cognitive activity of the visual brain. Yet, as Wartofsky eloquently argued, “Human vision is itself an artifact; with the advent of human culture the visual system breaks loose from its previous biological domain, and acquires a history; and . . . in this history, it is we who shape and transform the modes of visual praxis, of visual cognition and perception.”33 This has clear implications for art: ”with the development of representational practice, we come to see by means of the forms and styles of visual representation that we create; and . . . our modes of visual perception change with changes in these modes of representation“ (Wartofsky, “Sight, Symbol, and Society,” 28). This idea, that art had the capacity to shape human vision, was, as we have seen, a central tenet of Einstein’s writing.

As a young man, Einstein defined the task of art in these words: “Negation says nothing at all, and affirmation just as little. The artistic begins with the word ‘otherwise’.”34 It is this concept of art that Einstein can offer to neuroscience, and so enrich our understanding of the visual brain and the neurobiological foundations of human aesthetic response.