David Hockney believes that the only way the past masters achieved near photographic quality in some paintings was through optical trickery. Martin Kemp reports.

It is so realistic that it looks like a photograph." This comment from an unsophisticated observer looking at a naturalistic picture by a Dutch master of the 17th century or at one of Canaletto's Venetian views is guaranteed to arouse the scorn of an expert. But what are we to think when a draughtsman, painter, scene-designer, maker of photo-works, and inquirer into representation of the sophistication of David Hockney says essentially the same thing of a range of painted images from the Renaissance to the 19th century?

More precisely, what was I to think in June 1999 when Andrew Dempsey, formerly of the Arts Council and Hayward Gallery, suggested that I talk to Hockney about his use of a camera lucida to draw portraits and his idea that historically painters had used optical drawing devices far more widely than anyone had previously supposed?

When I visited Hockney's London studio during one of his visits from California, I found that he was undertaking a form of visual inquiry into the art of the past - through his inimitable approach to looking and picturing - that was venturing into territory that no art historian had previously dared enter.

ÖйúAƬ

In practical terms, he was experimenting with a camera lucida to make drawn portraits. The camera lucida, patented by William Hyde Wollaston in 1806, is a prism-based apparatus that relies on refraction and internal reflection to allow the draughtsman simultaneously to see the subject and the drawing surface. The effect, once we learn to use the device, is as if the image is actually hovering on the surface. Hockney was inspired in his experimentation by his observation that some early 19th-century drawn portraits by J. A. D. Ingres - then on display in New York and London - showed clear signs of the use of optical projection. As I had come to recognise from my research, the line that results from the tracking of an optically projected contour is quite distinct from that of drawings that are directly "eyeballed" (Hockney's term) from the subject. Once one has an "eye" for the telltale signs, they are relatively easy to spot.

As one of the portrait subjects of the artist's unrelenting scrutiny, between passing clouds of cigarette smoke, I was surprised by the way that he was using the camera lucida. Rather than meticulous and laboured tracing of the projected image, Hockney used it for a form of summary charting, rapidly laying down the key locations in the sitter's facial map, before the expression froze and the features sagged. After some 90 seconds or so of charting, the rest of the drawing, taking an hour or more, was laid in by normal "eyeballing".

ÖйúAƬ

This method opened the possibility that the use of optical devices might not be limited to those artists who paint meticulously "photographic" images and could be extended to masters of rapid brushwork and alla prima virtuosity, such as Frans Hals.

This was not the end of the surprises. From the springboard of the Ingres observations, Hockney was diving into the great pools of naturalistic image-making in the European tradition, from the Renaissance to the 19th century. Driven by an impetuous curiosity and unbounded enthusiasm, he simultaneously began to build his own beguiling devices, most particularly camera obscuras - "dark chambers" with convex lenses or concave mirrors - and to re-examine the visual evidence within historical paintings about how the painters might have achieved their often miraculous ends.

He could not believe that some pictorial effects, such as creased draperies bearing elaborately foreshortened patterns, were simply "eye-balled" without recourse to optical assistance. In an exchange of correspondence, published in part in his latest book, we began to debate the relative roles of the geometry of linear perspective and optical projection in the formation of such images as the globe in Hans Holbein's The Ambassadors and Dutch interiors of the 17th century.

Whereas I looked for evidence in the historian's conventional places - artists' treatises, inventories of studios, accounts of artists at work, surviving instruments and so on - Hockney's evidence was drawn from the visual qualities of the pictures themselves, combined with his painter's question, "how could I do that?" and the strange visual magic of the images formed in his own "cameras". In his definition of visual qualities, he was led to create the kind of startling and telling juxtapositions that are now largely confined to the distant era of "art appreciation". A bit of Ingres drapery is juxtaposed with a traced egg-whisk by Warhol, while a Byzantine mosaic of Christ is set beside a Van Gogh portrait. The two latter images frame his great chronological wall of reproduced images in the California studio.

It would be too simple to say that I acted as the pedantic keeper of the historical conscience in this enterprise, while he dashed creatively into daring realms of historical supposition - but there is something in this characterisation. Hockney's studios in Los Angeles and London became laboratories of visual research and experimental modelling, involving a loose team of diverse talents, not least those of David Graves who chased up historical texts and helped construct the viewing chambers. Charles Falco provided detailed know-how about the nature of lens-based images and how to determine the parameters of the "camera" needed to make a particular picture. Philip Steadman contributed his experimental knowledge of Johannes Vermeer and the camera obscura.

Unexpected artists entered the fray. Caravaggio, master of dramatic narratives painted with no apparent drawing, became a kind of film director, or rather a photographic collagist, assembling his deeply shadowed tableaux from figures projected in strongly foreshortened poses. The concave mirror, which offered greater flexibility and was less dependent on a large "dark chamber" than the lens, began to play a dominant role. All this continues, and in a sense the book is an intermediate staging post on a great journey.

ÖйúAƬ

Where do I stand as Hockney's claims have become ever more extensive? As he knows, I don't think optical projection was used in all the cases he adduces. I still accord linear-perspective constructions a conspicuous role, particularly for artists who regarded the mastery of geometry as part of their claim to be intellectuals, and I tend to give greater scope to the ability of great painters to "fake" the quality of a camera image through visual memory and incredible acts of painterly sleight of hand. But Hockney has utterly transformed the scope and basis of the debate about means-ends in naturalistic art for those prepared to cast aside prejudices about artistic "cheating". Some cases seem to me to be cast iron, particularly from the 17th century onwards. The suggestion that Jan Van Eyck and his fellow Dutch masters of painted light utilised optical devices is about as convincing as it could be without written evidence that continues to provide the cautious historian with succour.

But to conclude by picking out niggly "nos" and qualified "yeses" seems less important than to acknowledge that Hockney's bigger picture brings home to us how the mainstream of western image-making after the invention of photography went not so much into the specialised practice of "art" but into photography itself, into film and television, and now into all the forms of graphic realism produced by computers.

ÖйúAƬ

In a sense, the predominant tone of western painting from Renaissance Italy and Flanders to 19th-century naturalism was indeed "photographic', if we are granted the licence to use the term retrospectively to denote illusions that aim at some kind of optical matching of light-borne images and the disposition of pigments on a flat surface. In the face of the predominant form of academic art history that seems scared of hard looking, my view is that we should bring on as many David Hockneys as possible.

Martin Kemp is professor of art history, Oxford University. David Hockney's Secret Knowledge: Rediscovering the Lost Techniques of the Old Masters is published by Thames and Hudson, ?35.00.

THE DIALOGUE CONTINUES

Extract from letter to Martin Kemp - May 7 2001

Dear Martin,

A few days ago, we made a drawing using optical projections of figures like the composition of [Caravaggio's] Supper at Emmaus . It is amazingly simple to do. You do not begin with a tableau of four people. Not necessary, not even a large table. Every figure actually occupies the same space in the studio; it is the canvas that is moved around, forwards and backwards (difference only about two feet) and upwards and downwards - also not that much difference.

The actors move only on a straight line from the lens - away or closer - for changes of scale. It is ingeniously simple, and I think the only way optics can have been used for it. It also explains the scale of the hands and even the different scales of the still lives on the table. It would not be possible to set four people at the table this way and photograph them. It wouldn't look like that. It is composed in his head and on the canvas.

This is very exciting, as exciting as discovering the mirror lens. One draws with optical projections - indeed, it is really like Photoshop. If the young users of Photoshop knew all this, they could really make new-looking pictures, although paint is better than printing ink...

ÖйúAƬ

David H.