Shadows and Enlightenment 
by Michael Baxandall.
Yale, 192 pp., £19.95, June 1995, 0 300 05979 5
Show More
Show More

Powdering one’s nose is a strategy for controlling the effects of light. The powder changes the reflectivity of the surface of the skin. Oily skin acts as a mirror which bounces light off at an angle equal to the angle at which it arrived; powdered skin is matte and reflects light in all directions. Powdering thus turns a bright, specular highlight (the kind you get on glass or polished metal, which moves across the surface of the object as you change your point of view) into a diffuse one. For an artist, getting to know the nature of highlights, grasping the fact, for example, that the end of a nose or a cheek, even a brown or black nose or cheek, will often be the brightest part of the face – brighter than the ‘white’ of the eye, which is, in fact, usually grey and shadowed by the eye socket – is one step on the path leading from naive to ‘realistic’ representations of faces.

Shadows and shading in pictures represent the variations in the light reflected from objects which, on reaching the eye, become the raw material of perception. Although we are good at using make-up and camouflage in order to manipulate appearance, it takes strenuous effort (what amounts to an act of suppression) to become aware of just how the perceptual material relates to such stratagems. Painters (well, some kinds of painters) must, as Michael Baxandall puts it, ‘backtrack down the channels of perception, undoing the integration of features that is higher perception’s achievement, pushing right back down to the early visual modules of brightness, colour and the rest’.

To be interested in Shadows and Enlightenment you must be interested in this kind of thing, interested, in particular, in looking at shadows. Baxandall says his book is ‘coloured by being an offshoot of work-in-progress on problems of visual attention in 18th-century thought, in modern thought and in the art of painting’. It raises very basic questions. Do we look at marks on the canvas, or the things the marks represent, or both? And if we look at both, in what sense can we be said to give our attention to, say, a patch of pink paint and, at the same time, to the Virgin’s cheek?

It helps, in reading Baxandall’s book, to be aware of the extent to which his examples from computer and cognitive science, as well as from the words and work of painters, help to define a central paradox: that if we pay close attention to variables like shadows, which are, in the ordinary way of things, perceived inattentively, our attention itself gets in the way of the process we are trying to understand. We have, in Baxandall’s words,

no clear idea of specifically in-attentive perception as a productive complement to attentive perception – attention this time in the sense of a directed, focused and constructive scrutiny in some reciprocal relation to consciousness ... ‘Attention’ effectively disables itself as a concept by reducing the ‘in-attentive’ to a negative or absence of something, rather than the active, determined and structured field in which consciousness plays.

Baxandall’s book, which is admirably, indeed severely, focused (itself unforgiving of any straying of attention), could be seen as a long series of footnotes to that sentence. It is not set out as a linked argument, for his purpose is to characterise how we see, represent and interpret shadows rather than explain them in relation to a single theory, or even to a single definition of what shadows are. And it is only partly about how ‘we’ in general see shadows: the concerted attention his 18th-century Enlightenment painters and thinkers gave to shadow theory made them, at the very least, better informed than we are and, in various ways, more (or at least differently) attentive to them. Some of Baxandall’s chapters would not be out of place in Scientific American, others are essays in intellectual history and art theory. Passages of introspection, in which he describes his own spells of shadow-watching, offer a verbal equivalent of visual backtracking.

Among the things one learns about, or is reminded of are: Locke’s epistemology, and his distinction between what is seen (the visual array) and what is known (that there is a table in the room); the arguments about whether the derivation of the second from the first is a matter of analysis or of empirical experience; the fascination which Molineux’s hypothesis – that a newly-sighted person would have no understanding of the visual array – and its semi-confirmation in William Cheselden’s medical account of the case of a boy operated on for cataracts, had for writers in the 18th century (the boy’s relation to the seen world being memorably encapsulated in the story of his being unsure if it was the dog or the cat he was looking at, feeling it to make sure and saying: ‘So puss, I shall know you another time’); and the taxonomy of shadows.

Once we have learned to recognise the difference between shading (for example, the darkening in ambient light of the surface of a cheek as it curves away from you), self-shadow (the dark side of a half-moon) and cast shadow (a hand making a rabbit on the wall by lamplight), we are already some way to understanding the basis of the rules painting manuals offer – for instance that the shadow a figure casts on the ground will be darker than the self-shadow (to which it is attached at the feet) because the self-shadowed side of the figure receives light reflected back from the ground. Geometrical regularities in the way curved surfaces shade off suggest that the brain may use such regularities to simplify the process of deriving information about the shape of objects from information about their luminance. The book also deals with the problems which arise when you try to teach a computer to recognise an object. Shadows are both givers of information (about profiles, convexities, concavities and textures) and something very close to visual ‘noise’ (the camouflage of dappled shade). They can remove ambiguities (drawings in car manuals use shadows to distinguish grooves from ridges), but as work on computer systems has shown, deriving an account of a three-dimensional environment from a logical/geometrical analysis of a two-dimensional array demands such a formidable amount of calculation it seems very unlikely that the human brain gets to understand the world in this strictly mathematical way.

Two interlocked sets of questions arise. The first centres on the phenomenology of shadows: how are they made? What colour are they? How are they affected by light reflected from other surfaces? What is the nature of the shadow edge? These questions are both scientific – concerned with the physics of light – and artistic: concerned with the practice of representation. They, and the shadow-watching they encouraged, were apparently extraordinarily attractive to some 18th-century minds.

The second set of questions concerns the psychology and physiology of perception. Do we see by a bottom-up process – taking the raw data and analysing them to produce information about the world – or do we work top down: applying schemata of objects and choosing the ones which fit best? Are we physiologically better set up to notice some things than others? (Does the retina itself have preferences – do we, for example, see edges better than fields?) Are there aspects of the visual array which, because they remain constant under changing light conditions, are relied on for spatial information?

Mid-18th-century French studies of shadows – their colour, attenuation, reflection, refraction and so on – had their beginnings in Leonardo’s account, and were pursued by both artists and scientists. Sciography – essentially, the rules of linear perspective applied to cast shadow – had, by then, become an applied craft, most fully developed in the training of architects and engineers. Natural philosophers and painters were noticing, and trying to give orderly accounts of, or recipes for, the same phenomena. Baxandall describes the engraver Cochin’s model of shadow intensity as

elegant, with its two counter-scales of force and the moving point of dominance between them. [Cochin was explaining why shadows in the near-middle-distance, not those closest to, or furthest from, the viewer are darkest.] It has the virtue of avoiding any crude linear sense of simple degradation by distance. It builds in a number of physically describable peculiarities of the behaviour of light. It sustains a clear sense of the distinction between objective and observed shadow. But the particular virtue is its neat variable, ‘a certain line set at some distance into the picture where the shadows are strongest and darkest’, on either side of which ‘they diminish in force, whether coming forward to the foreground or moving back behind.’

Cochin’s account, involving testable hypotheses, is very like a scientific theory.

Why do we have an appetite for representations? Why would we rather look at a peach painted by Chardin than at a real peach? Why would we (at least sometimes) rather look at a Velázquez portrait of an unknown, very plain person than at a real, pretty human being? And why at a Chardin or a Velázquez rather than paintings of the same subjects by other painters?

Reading Baxandall’s book puts up hares of this sort, and suggests ways to chase them by emphasising that perception is an active, interpretative faculty, not a passive, receptive one, so making sense of the idea that looking is, in itself, pleasurable. Once we accept that paintings are a kind of visual gymnasium, art can become for some people what ball games are for others: an arena in which skills can be explored, whether by players or by viewers, and pushed beyond mere usefulness. The difficulty of striking a snooker ball so that it goes just where the player wants it to is not unlike the difficulty of truly representing the shades of white in the feathers of a dead goose; and to share some of the snooker player’s pleasurable sensations when watching a truly struck ball on a television screen is not unlike the enjoyment we get in an art gallery observing the painter’s recapitulation of his acts of perception. Art and sport are doubtless closer than followers of only one of them would like to admit. If it is ever possible to map the parts of the brain which participate in aesthetic response I would be surprised if the parts which respond to painting did not overlap with those which respond to pool.

Chardin’s still lifes, painted in the controlled environment of the studio, have none of the smooth, anonymous finish which became common in 19th-century academic painting – what Baxandall calls the ‘licked trickery of trompe l’oeil’. But they do have an extraordinary verisimilitude, which seems to come from a highly ‘backtracked’ account of what the artist has seen. Their force was recognised at the time he painted them. But there were all sorts of other subjects – grander in terms of scale and social prestige – which could never have been put on canvas in a Chardin-like way. Some of the most highly regarded paintings in 18th-century France were hybrids of a particular kind, as Baxendall explains in a very convincing account of portraits by Rigaud and Largillière (‘entertainments, and still very enjoyable’):

Their representation of optical facts is a representation of the pictorial representation of optical facts. For instance when they paint the degeneration of shadow with distance, which so interested Cochin, they paint it in inverted commas ... such pictures, being representations of painting, are committed to testing our perceptual tolerances ... for instance the painting of the microshadow of the surface of silk and velvet is a very old trick, here with a delicate balance ... between our perceiving it as silk-reflected light and our perceiving it as differently thinned, differently edged application of white or white-lightened pigment in marks of different morphologies, scales and frequencies ... They declare themselves as conventionalised not just through the standard set-up – standard poses, angled heads, cute dogs, still-life accents, rich fabrics and so on – but in the conventionality of the lighting topics handled: mixed lighting, heads and hands studio lit against dramatically illuminated ambience, with arbitrarily highlighted leafy sprigs, and emphatic aerial perspective. Perceptual tolerance is tested first by alerting us to light with these cues, isolating them by differentiating their levels of pictorial specificity or styles of detail, and then denying any systematic consistency among them.

The kinds of unnatural lighting we are willing to accommodate in paintings can be matched with the kinds we accept in photographs and films: to see an outdoor scene being shot with huge are lights and reflectors applying ‘studio light’ to the actors who take up the foreground of the image is to see (in perceptual terms) the photographic equivalent of a Largilliére portrait in the making.

If Baxandall’s book could be said to have a hero it is Chardin. The Young Draughtsman provides Baxandall with a gamut of shadow types. To start with, the image includes another image (the chalk drawing of a figure lit from the side); there is the figure of the young draughtsman, ‘modelled against a mid-toned ground in the three-tone scheme of academic chiaroscuro’; and there is his cast shadow, without which we would be at a loss to locate the wall which forms the spatial background. Here Baxandall identifies a ‘corner effect’ (of the kind which makes us see three straight lines meeting as edges of a three-dimensional object): a line which ‘keeps twanging faintly, now and then, and reminding us that attentive vision and inattentive vision may give conflicting knowledge’. And there is – Baxandall’s description of this brings the physical presence of Chardin’s painting most clearly into view – the matter of ‘microtexture’:

As for microtexture or shadow, there is almost none in the direct representation; Chardin has effectively appropriated it for the actual facture, for his own performance ... he learned to speak with extreme, demanding quietness – small pictures one has to approach close to make out, banal subject-matters with little independent interest ... And he has the nerve to represent, disguised as a canvas against the wall on the right, a picture of abstracted texture, pure Chardinian facture ... a picture of a picture of nothing at all ... By some odd conflation in the perceptual process, the rich and subtle real texture of the paint surface cues us to supply a sense of texture to the fictive surfaces. The release of perceptual energy involved in this flashover seems enjoyable in itself, but it also lends vitality to the fiction.

Baxandall describes what goes on in one kind of picture, produced in one place and in one period, but his account is so suggestive that one wishes to pre-empt him in the wider investigation of visual attention which he promises to give us. It is doubtless better to wait. But there is no need to go without the ideas he has supplied. They answer, for example, questions about abstract pictures. When ‘what is it of?’ is replaced by the question, ‘How does it work?’ Chardin and Rothko are found to have something in common. The ‘faint twang’ Baxandall hears in Chardin has a lot in common with whatever sound it is the blurred edges of Rothko’s rectangles make. More generally, his book suggests why painting easily holds attention and goes on giving pleasure when its subject-matter seems unimportant and its content minimal.

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