The alchemists of light

Nothing is visible without light; nothing is visible without a transparent medium; nothing is visible without colour; nothing is visible without distance; nothing is visible without mechanism."

These formulations by the great French painter, Nicholas Poussin, stand entirely within a long-standing optical tradition deriving from Aristotle and very much correspond to the commonsense view of colour as we see it. Objects are actually coloured (so it seems) and under the right conditions of illumination we use our eyes to see the coloured forms in the space in front of us. The painter concerned to present an illusion of reality aspires to arrange pigments on a flat surface to evoke this coloured array.

The Aristotelian tradition of colour, which prevailed largely unchallenged until Newton's experimentum crucis with his prism in 1666 and refused to die even in the 19th century, took the colours as real entities, as generally standing in a sequence between the poles of white and black (or lightness and darkness) and, in the most developed theories, as compounded from the polar qualities in set ratios like musical notes. By the early 17th century the most advanced ideas stated that there were three simple colours between white and black - yellow, red and blue - which in paired combinations would give the secondary hues of orange, purple and green. Three main ways of combining colours were recognised: by direct mixture of pigments; by the overlaying of translucent colours; and by the combination of small patches of juxtaposed colour. All this worked well with the painter's practice, not least because important aspects of the formulations had been built upon the accumulated wisdom of generations of painters.

It is one of the many paradoxes of colour that Newton's key recognition of the compounding of white light from the colours of the spectrum and his definition of the seven primary colours gave endless trouble to those artists and theorists who tried to make sense of the colour-inducing stimuli of the Newtonian theory and the behaviour of material pigments. The persistent failure to recognise the quite different processes of additive mixing with coloured lights and subtractive mixture with pigments did not help, but even after the clear definition of the differences by Helmholtz in the 19th century, the goal of painting light, or "painting with light" seemed to be as elusive as ever, if it was to be accomplished in terms compatible with colour science.

However, given the huge efforts of modern science in the fields of physics, physiology and psychology, and the sustained enquiries of philosophers, linguisticians and anthropologists, surely we ought to know better. The present volume, comprising eight papers, based on lectures given in 1993 at Darwin College, Cambridge, brings together leading specialists to review the main findings of the disciplines most centrally concerned with colour in its artistic, linguistic, cultural and scientific senses.

Unfortunately, the common theme which emerges most prominently is how far we are from being able to pin down any of the incredibly slippery businesses of seeing, naming and representing colour. Each of the successive chapters either opens or concludes with assertions of the inconclusive or relative nature of our understanding. Trevor Lamb and Jane Bourriau, in their editorial introduction, conclude that "the principle of colour is in the mind of the beholder, yet colours as we know them are a product of our language and culture". David Bomberg in his overview of the history of pictorial colour ends by quoting Paolo Pino, the Venetian author of the 16th century to the effect that colour is "the true alchemy of painting". Speaking as a painter, Bridget Riley tells us that "just because there is no guiding principle, no firm conceptual basis on which a tradition of colour painting can be reliably founded, this means that each individual artistic sensibility has a chance to discover unique means of expression". Malcolm Longair, expresses his relief to be writing about the concrete subject of the physics of light, since "every other aspect of light and colour is quite horribly complicated"; yet even his physics rests on the uneasy assumption that light simultaneously exhibits the properties of both waves and particles, depending upon how the analyses are conducted.

In his essay on colour mechanisms in the eye, Dennis Baylor recognises that "great advances have been made in unravelling the eye's mechanisms for colour reception, yet fascinating questions remain. How for example does a single cone cell decide which pigment to express?" From the standpoint of an experimental psychologist, John Mollon asserts that "there is no fixed relation between the spectral composition of lights and the hue that we perceive at a local point in the scene I In other words, we judge colours by the company they keep." When Peter Parks introduces his topic of colour in nature, he warns that "the observer must realise that what he or she sees may not be what others see, and therefore every one of us can rightly consider that what we see of colour is something very private". The art historian John Gage disconcertingly begins his account of colour and culture by declaring that "a series of discussions of colour from different perspectives soon lands us in serious conceptual difficulties. I want to start by proposing that most of the authors I are not really dealing with 'colour' at all: they are concerned with radiant stimuli in light or with the physiological processing of these stimuli by the eye." And he cautions in his conclusion that "the instability of colour-perceptions I should give pause to those many ethnographers and semioticians who have been tempted to speak confidently of colour-meanings and preferences in many cultures." Finally, among the conclusions to his review of colour naming, Jack Lyons states that "Colours I as we know them are the product of language under the influence of culture."

All this seems very much grist to the relativists' mill, and suggests that any analysis of colour is condemned to founder in a morass of perceptual instability, cultural incompatibility and linguistic difference. It is self-evident that I cannot prove that the sensation of a bright sunny glow which I gain when I see a bright yellow is the same as yours. And some cultures do not seem to have a word for yellow at all. Indeed, many languages do not even have a word for "colour" as a concept. We may be led to wonder what would have happened to Newton's theory had he been working in a culture with a different mode of colour naming - or whether he could have made any progress at all without the European heritage of colour terms. Yet if the essays are read as a whole, the various authors' ritualistic declarations of the lack of defined and agreed criteria seem less compelling. There is a general consensus that our seeing of colours is highly context-specific, and that the ways that different cultures construct colour language varies considerably, but there seems to be a robustness in our ability to make fine distinctions between many varieties of what we call hue, tone (or value) and saturation (or intensity), to match them when required in given circumstances, and to name them more or less coherently when culturally needed. Lyons's qualified support for Brent Berlin and Paul Kay's influential hypothesis that all natural languages have between two and 11 basic colour terms is particularly telling. Demonstrating this robustness is less easy than the verification of the slippery relativities, but it lies implicitly behind all our abilities to communicate effectively within and across cultures about what is seen and described. We all seem to have evolved a system of receptors for stimuli which we encode as red, green and blue, potentially to do the same kind of job in making some reasonably stable sense of what lies out there. It is how we describe that job, and what parts of it become the focus of directed concern, that is culturally determined.

The great merit of the present collection of papers is that the standard of clarity is high, and the current state of play is summarised in a way that may be recommended for any student wishing to gain a sane overall briefing on the main modes of colour analysis, past and present. The main problem arises from the very wide brief each specialist has been required to tackle, which results in an inevitable tendency for the reviews to become very compressed and summary. In the more technical of the papers, the result is a telegraphic quality which could well leave the non-specialist struggling. There is also a sense in which the very reasonableness of the ecumenical tone of the essays leaves the book short of really powerful argument, and there is little sign of stimulating polemic. Somehow the excitement of colour in science and art never quite emerges in the way that it can with a volume produced by a single, opinionated author. But in an area in which conflicting opinions are legion, perhaps we should at this stage be grateful for sober good sense.

Martin Kemp is professor of the history of art, University of Oxford .