Perspective as Symbolic Form 
by Erwin Panofsky, translated by Christoper Wood.
Zone, 196 pp., £20.50, January 1992, 0 942299 52 3
Show More
The Language of Art History 
edited by Salim Kemal and Ivan Gaskell.
Cambridge, 245 pp., £32.50, December 1991, 9780521353847
Show More
Show More

Kant’s characteristic philosophical strategy – ingenious, original, and by his own assessment, revolutionary – consisted in transferring to the mind, as among its organising principles, a great many of the features heretofore ascribed to the objective order of the world. Causality, for example, rather than some bonding agency, linking event with event under the laws of nature, was instead a defining structure of the way we organise experience: it would not be experience were it not causally ordered. And space, rather than some vast container in which the furniture of the universe is stowed, is instead a form of perception – an innate a priori scheme through which bodies present themselves to the senses as coordinated. The laws of this mode of organisation arc given by geometry, which Kant had no grounds for believing other than Euclidian; and indeed, well after the advent of non-Euclidian geometries, it was still widely maintained that Euclid’s defines the structure of spatial experience for minds such as ours. Kant was interested in human nature only in the most universal terms, and it was his assumption that our spatial intuitions must be largely invariant inasmuch as we are all built the same. The organising principles of the form-giving mind must be the same from period to period of history, and culturally all of a piece.

There is very little if any discussion of pictorial space in Kant’s writing, as there is very little discussion in his aesthetic writings of art as such, since his main concern was with the judgment of beauty, whatever its occasion. My sense is that the beauty of a picture would, in his view, be derivative from the beauty of the subject of the picture, so a beautiful picture of a woman would ipso facto be a picture of a beautiful woman. Thus the picture would be treated as a pure transparency, with no features of its own, and the organisation of pictorial space would simply be continuous with the organisation of objects in ordinary space. The picture, in effect, would be like a window through which we see the world much as we would see it were we standing outside. And indeed, this specified pretty much the agenda of the pictorial artist, whose task consisted in creating what the psychologist Julian Hochberg calls a ‘surrogate’: ‘To prepare a surface that reflects to the observer’s eye the same proximal stimulus pattern as does the real scene.’ The picture may thus be construed as a section through the visual pyramid, and if, with Leonardo, we imagine this as a pane of glass, the artistic success consists in marking the glass’s surface in such a way that there is no perceptual difference to be registered between what we sec on it and what we would have seen through it antecedent to its being so marked: the skill of the artist consists in disguising itself. When the skill is maximal, there is accordingly no distinction to be identified in perception between looking at a picture and looking at what it is a picture of.

Very little of the pictorial production of humankind lives up to the ideal of the modified pane of glass, and indeed most of it deviates sufficiently from anything that could be taken as an immediate view on reality for one to question whether it really is an ideal widely subscribed to across artistic traditions. Had Kant been presented with a Chinese painting in isometric perspective, or a Byzantine mosaic with no perspective to speak of, or a Tlingit painting of a bear which presents all sides including the inside of the bear simultaneously across the surface, his response might very well have been that the artists had not as yet hit upon the truth of spatial representation – and their art would thus have been governed by an ideal they had yet to learn to actualise. This view of artistic diversity would have been pretty much the standard Western view of non-Western painting until at least the late 1880s, when Van Gogh and Gauguin turned their back upon their own tradition, which of course brought upon them the immediate charge that they did not know how to paint. It was a 20th-century contribution to recognise that there is no single path of progress in the visual arts, but instead an open disjunction of artistic ideals.

The idea of the symbolic form, originated by the neo-Kantian philosopher, Ernst Cassirer, belongs to this pluralistic reading of things. The notion of form is clearly akin to Kant’s notion of form, as something given by us to experience. But rather than being invariant from period to period and place to place, the symbolic forms are historically indexed: they define the way in which peoples in diverse eras and cultures organise their worlds. They are symbolic because they confer meaning on experience, which is not an especially Kantian notion; and they specify what it is to live in a culture, inasmuch as the culture itself consists in a complex of inter-related symbolic systems: language, myth, religion, science, art and ethics. Erwin Panofsky’s stunning essay was in a sense a long and learned footnote to Cassirer’s monumental The Philosophy of Symbolic Forms. And though Panofsky has a vivid sense of historical otherness and difference where Kant would have seen invariance and universality, his enterprise, like his colleague Cassirer’s, remains essentially in the Kantian spirit: it was to explain, in terms of the symbolic forms of art, the way the components of an artwork are structured by what we might term a ‘cultural a priori’. It was his extraordinary idea that the linear perspective discovered – or invented – early in the 15th century by Brunelleschi, given theoretical formulations by Alberto and Piero, and put into painterly practice by Masaccio and others, is not so much optical as symbolic, a case where ‘spiritual meaning is attached to a concrete, material sign.’ As such, it is a key to Renaissance culture, since there are other ways of handling space pictorially, each of which specifies what we might term a distinct pictorial world. Renaissance painting is only one such world.

It must be stressed that it is no part of Panofsky’s thesis, though he may well have been tempted in that direction, that people under different complexes of symbolic forms perceive space differently. Our perceptual system is genetically-coded, and DNA is impermeable to cultural differences. The ancients would have been subject to the same optical illusions as we are, which is part of the reason the senses come under philosophical suspicion in Plato. Ancient sculptors enlarged the heads of figures meant to be placed on high columns, in order that they should look normal seen from the ground; ancient architects induced a subtle curvature into a line of columns so that it would look straight. Kant’s philosophy of perception would not require symbolic emendation were we to live under the starry heavens without language or culture, matters to which he paid no great heed. What Panofsky supposed is that people under different symbolic systems represent the world differently in their art. The mistake would consist in supposing that it was part of their undertaking to represent the world exactly as they perceived it. That became a cultural ideal chiefly in the Renaissance. The fallacy is to export the Renaissance ideal to other cultures and to suppose, since the aim of pictorial art is always to imitate perception, that the artists of different cultures faithfully replicate the world as they perceived it, and that they perceived it precisely the way their pictures show it. So the Chinese must live in a world of isometric perspective and the Tlingit in a forest of two-dimensional bears. In truth, if their pictures show the world as they perceived it, and they represented it faithfully, their pictures would look exactly like our pictures (this is a generalisation on what Gibson called the El Greco Fallacy). So the task for us is to see what, other than modifying the metaphoric pane of glass, artists in different traditions were doing in their art. And this is to ask for the symbolic meaning of, among other things, their various ways of organising pictorial surfaces.

Die Perspektive als ‘Symbolische Form’ was first published in the Vorträge of the Warburg Library in 1927-8, and it is something of a mystery, in view of the frequency with which it is referred to and Panofsky’s importance as scholar and thinker, that it has not until now existed in an integral English version. It is a crucial text, and the translation makes it as clear as is consistent with Panofsky’s own sometimes confused formulations what the thesis really amounts to. Panofsky in fact wavers between two views of Renaissance perspective as symbolic form, which amount to two deeply different views of art-historical reality. The first view holds that there is, in any given period, what he terms (in another work) a ‘unifying principle which underlies and explains both the visible event and its intelligible significance, and which determines the form in which the visible event takes place’. His name for the discipline which investigates such underlying principles was ‘iconology’. Linear perspective, in particular, is a visual expression of the underlying principle of the culture of the Renaissance, which forms a kind of totality – forms what Continental philosophers call a ‘totalising totality’ in which everything is brought into a kind of unity with everything else. There have been and are many such totalising totalities. The ancients had their perspective, which expressed the underlying principle of their culture, and in connection with it the systematic perspective central to the Renaissance was strictly unthinkable – ‘as unthinkable for antique philosophers as it was unimaginable for antique artists’. The two perspectival orderings are, as it has become fashionable to say in the history of science, incommensurable. ‘Antique perspective is thus the expression of a specific and fundamentally unmodern view of space.’

The other view regards ancient and Renaissance perspective as stages in a progressive history, in which Renaissance artists solved the problems the ancients grasped, but did not know how to handle. ‘When work on certain artistic problems has advanced so far that further work in the same direction, proceeding from the same premises, appears unlikely to bear fruit, the result is often a recoil, or perhaps better, a reversal of direction.’ For Panofsky, the Middle Ages was ‘the greatest of these “recoils” ’, was a ‘turning back to apparently more primitive modes of representation’ which lay ‘the groundwork for a creative re-engagement with older problems.’ It was a matter of reculer pour mieux sauter. But this connects ancient, Medieval and Renaissance painting into a single, heroic narrative.

I am uncertain to what degree these two theories are co-tenable. The philosophy of symbolic forms vacillates between a proto-Kuhnian and a proto-Foucauldian theory of the history of art. On the one theory, spatial representation is an expression among many of a period’s underlying principle. Thus, speaking of systematic pictorial space, as in Alberti or Piero, Panofsky writes: ‘This perspectival achievement is nothing other than a concrete expression of a contemporary advance in epistemology or natural philosophy.’ Byzantine space, by contrast, was the artistic correlative of the ‘space of contemporary philosophy ... the metaphysics of light of pagan and Christian Neoplatonism’. Or again: ‘The space of Giotto and Duccio corresponded to the transitional, high Scholastic view of space.’ All this seems to imply that each period is a kind of symbolic whole, in which art, philosophy, architecture, and whatever else gives meaning to signs, are, as it were, transforms of one another in different media. But then it seems to me you can’t also say that the history of art is a struggle with the same problems through time, until they at last get solved: then the history of philosophy or of architecture would not belong to distinct, autonomous periods, strung like beads on the string of time, with no rapport with one another. It would have been valuable had Panofsky used some non-Western examples, if he really wanted to argue for periods as integral totalities. The Jesuit painter Giuseppe Castiglione introduced perspective into China in 1729, with mixed results. One writer said: ‘The ancients lacked perspective method, and when it is used so skilfully as here, one regrets that the ancients had not seen it.’ But the Emperor intuited that it was not something for which Chinese artists could have a use: occidental painting, he observed with great wisdom, while resembling, only resembles, whereas painting may be judged by quite different criteria.

One question, the answer to which depends on which view of history is true, is whether, granting perspective is a symbolic form in the Renaissance, it is a different symbolic form in every culture which uses space non-perspectivally. Were the ancients conscious of space as a problem – or were they interested in representing figures, set in whatever space enabled them to do so without that space having some meaning of its own? Panofsky did say, in another context, that symbolic forms are ‘above the sphere of conscious volition’. But linear perspective was actually invented, consciously. Space had never been treated quite that way before. But does that mean that non-perspectival space was also invented, and bore its own meaning? No one knows the answer to this, but an art history which does not reflect on such matters is by comparison with Panofsky’s robust Kunstwissenschaft a fairly pallid discipline. His is a rich and stimulating book, and we are all in the debt of Christopher Wood and Zone Books.

I cannot say the same for the rather inaccurately named Language of Art History, which is merely an anthology of philosophical writings on art, together with an example, certainly admirable, by Richard Schiff of what Michael Baxandall has called ‘inferential art criticism’. Baxandall is represented by an attenuated version of a 1979 paper called ‘The Language of Art Criticism’, which would have been a marginally preferable title for the compendium. What Baxandall wrestles with is the adequacy of words to describe visual art. Given that we use words to describe the visual world, this ought not to be a terribly vexing problem. It would be no problem at all if pictures were as transparent as Kant’s theory tacitly holds them to be, for the properties of a visual work would reduce then to the visual properties of the subject of the work. But pictures are in some way discontinuous with the rest of the visual world in that we may want to describe them in their own terms – for example, as Panofsky does (of a mosaic from San Vitale), as ‘a space merely excised by the picture’s edge beginning to give way to the principle of a surface bounded by the picture’s edge’, or (of Van Fyck’s Virgin in the Church): ‘The beginning of the space no longer coincides with the border of the picture.’ Were this to be the subject of the collected essays in this volume, it would be of very great value. Art-critical language is not like what the Ancients called ekphrasis – a matter of finding verbal equivalents for pictures, which was a rhetorical exercise – but rather a way of using words to make salient certain features of pictures which merit attention. Curiously, Baxandall finds this practice almost unintelligible, chiefly on the grounds that language is linear and pictures are not, by contrast with literary texts. So, as the bewildered metaphysician might ask, ‘how can the linear possibly address the non-linear?’ My sense is that the issue of linearity is a sign of some mental cramp from which Baxandall suffers not at all in his practice, whatever problems it raises for him in theory: a good dose of old-fashioned philosophical analysis would be quite the best antidote for ‘How ... possibly?’ questions that seem to make impossible what everyone can do. Schiff, in his contribution, demonstrates the way ‘touch’ applies to the paintings of Cézanne, but the rest of the papers do little to address the language of art criticism or art history. I would regard this as a most inauspicious inaugural volume in a series meant to examine the relationship of philosophy to the arts.

Send Letters To:

The Editor
London Review of Books,
28 Little Russell Street
London, WC1A 2HN

Please include name, address, and a telephone number.

Read anywhere with the London Review of Books app, available now from the App Store for Apple devices, Google Play for Android devices and Amazon for your Kindle Fire.

Sign up to our newsletter

For highlights from the latest issue, our archive and the blog, as well as news, events and exclusive promotions.

Newsletter Preferences