Reflections on Art

Scientists debate whether the old masters used optical aids

Like a defense lawyer in court, David G. Stork was eager to know whether his closing argument was winning over his audience. Would a jury vote to convict? Stork asked the group assembled at NASA Goddard Space Flight Center early this month. None of the of the 100 or so people in the Greenbelt, Md.–facility raised a hand–just the response that Stork, chief scientist of Ricoh Innovations in Menlo Park, Calif., was hoping for.

RENAISSANCE SNAPSHOT? The top of the octagonal motif on the table rug is blurry, suggesting to David Hockney that Lorenzo Lotto’s “Husband and Wife” was painted with images projected by a mirror in the early 1500s. Art Resources
ANALYZE THIS. Details of the chandelier and mirror in Jan van Eyck’s “Portrait of Giovanni Arnolfini and His Wife” (1434) are used as evidence both for and against the painter’s use of optical instruments. National Gallery of London
WHAT’S YOUR ANGLE? These lines, presumably parallel when superimposed on a real chandelier, fail to meet at a single vanishing point. Their misalignment indicates that a major feature of Jan van Eyck’s painting isn’t in perspective and so wasn’t created with mirrors, says David Stork. National Gallery of London

Stork is no lawyer, but he definitely has a group of people to defend. An investigator of pattern recognition and an amateur artist, he’s on a mission to scientifically disprove the assertion by renowned British-born artist David Hockney that many of Europe’s greatest artists of the 15th and 16th centuries secretly used mirrors or lenses to project traceable images onto their canvases and thereby achieve the arresting realism of their paintings.

First publicized in a New Yorker article in January 2000, Hockney’s proposal jolted the art world and has received wide attention from the media and general public. Although Hockney recoils at the suggestion, many people interpret his hypothesis as an accusation that the old masters cheated.

The theory “touches some very raw nerves as to what we think art and artists are about,” says art historian and Hockney colleague Martin Kemp of Oxford University in England. Optical aids such as slide projectors are in widespread use today among painters, but art lovers typically revere the superb realism that Renaissance masters achieved without the apparent use of such devices.

In addition to the mass media coverage of his radical proposal, Hockney has made a documentary film on the topic and even published a book, Secret Knowledge: Rediscovering the Lost Techniques of the Old Masters (2001, Thames & Hudson, Viking Studio).

That the Goddard audience was wary of Hockney’s thesis was particularly gratifying to Stork. That’s because a year ago, art aficionado and Hockney collaborator Charles M. Falco, an optics professor at the University of Arizona in Tucson, had presented the scientific case for Hockney’s proposal from that same Goddard rostrum. At the end of his talk, Falco, too, had polled the audience to see whether it was with or against Hockney. By a show of hands, he says, 97 out of 100 listeners had declared themselves convinced of the Hockney hypothesis.

Falco had volunteered his scientific services to the artist after reading the New Yorker article. The optics specialist developed nearly all of the scientific evidence supporting Hockney’s original idea that many Renaissance paintings are simply too precise to have been done strictly by eye.

“I had the optics training to add the scientific information to what [Hockney] had seen,” Falco explains. “It took both [visual and scientific evidence] to make a compelling argument.” For instance, Falco used the sizes of objects and people in the paintings to calculate diameters, focal lengths, and other characteristics of lenses and mirrors that might have been used to project those forms.

In scientific circles, Falco’s ideas have been warmly received, except by a few vocal critics such as Stork and Christopher W. Tyler of the Smith Kettlewell Eye Institute in San Francisco. Tyler calls the idea that optics were used in the 1500s and earlier “just storytelling.” Particularly galling, he and other critics say, is the absence of any clear evidence from that time that the optical devices available could produce the kind of images Hockney claims the masters used.

In turn, Falco derides the objections raised by Stork and Tyler as unworthy of scholarly debate. Their criticisms are “an anomaly,” Falco says.

In recent talks and publications, Hockney, Falco, Stork, and Tyler have taken an especially close look at a few Renaissance artworks. Much of the discussion has focused on two paintings: “Husband and Wife” by Lorenzo Lotto and “Portrait of Giovanni Arnolfini and His Wife” by Jan van Eyck. Depending on the analyst’s point of view, even the same observations regarding those paintings lead to startlingly different conclusions.

Called on the carpet

Hockney and Falco have dubbed Lotto’s painting the Rosetta stone of their “opticality” theory. Completed in approximately 1525, the painting depicts a man and a woman seated at a table covered by a small oriental rug.

What Hockney and Falco consider so telling about this masterpiece is the curious distortion of a foreshortened, octagonal pattern in the front-center portion of the rug. The octagon is formed by a kind of train-track motif that jogs around a flower-blossom design. Hockney noticed that the octagon becomes indistinct–like an out-of-focus portion of a photograph–as it recedes from the viewer.

The octagon’s blurring is just the sort of distortion that someone might see in a projection of the rug by a concave mirror, Falco contends. Moreover, it’s a visual effect that an artist looking with his eyes alone wouldn’t see because human eyes automatically refocus as they range over a scene.

Falco has computed, down to the millimeter, just how Lotto might have used a mirror. To reproduce the complex details of the rug, the painter would have positioned a mirror roughly a meter and a half from the edge of the table at which his models were sitting, Falco calculates. Then, by placing himself to the side and about halfway between models and mirror, Lotto could have bounced a traceable image directly onto his easel.

Apparent mistakes by the artist in rendering the image in proper perspective strengthen that supposition, says Falco. According to the principles of geometric perspective, parallel lines appear to converge at a single point in space, known as the vanishing point, as they recede from the viewer. However, a short way back from the edge of the table depicted in “Husband and Wife,” some of the presumably parallel lines of the train-track motif converge to a vanishing point different from that of other parallel lines in the motif.

That’s easily explained, Falco argues, because a mirror wouldn’t be able to project the whole rug pattern in focus at once. He calculates that Lotto twice would have had to shift the mirror a bit further back from the table in order to refocus on details of the rug closer to the background figures. Each time, if the artist misaligned the mirror even slightly, the vanishing point would shift.

There’s a simpler explanation, claims Tyler, who finds a hodgepodge of vanishing points in Lotto’s painting. This multiplicity of perspectives suggests that Lotto painted freehand, not worrying about getting the perspective exact, as he would have done were he using an optical aid, Tyler says.

In another experiment to test whether the rug’s octagon was drawn in proper perspective, Tyler scanned the painting into a computer and used Photoshop, a popular image-processing software program, to digitally reorient the rug-covered tabletop to look as if it were seen from directly above instead of at an angle. With the perspective removed, the octagon appeared asymmetric.

Tyler says this suggests that Lotto botched the perspective in the first place–further evidence that the rug was painted by eye instead of reproduced optically.

Tyler unveiled his findings last October at the annual meeting of the Society for Literature and Science in Pasadena, Calif. He also has posted an online report of his investigation (

In Falco’s view, however, Tyler “completely blew the analysis.” In his own Photoshop experiment, Falco treated the octagon as if it were composed of several pieces. That’s because each time Lotto would have moved the mirror to bring a portion of the octagon into focus, he also would have automatically changed the magnification of the projected image, the Arizona scientist explains. Indeed, if the octagon’s pieces are each resized adequately, “the octagon fits the entire Lotto tablecloth pattern to plus or minus a percent. It’s quite stunning,” Falco says.

Even granting the thesis that Lotto used a concave mirror to paint “Husband and Wife,” a mystery would remain: Why, ultimately, would Lotto have left a blurry section of the octagon in the painting? Falco and Hockney conclude that the refocusing steps led to several separately magnified octagon pieces that Lotto couldn’t completely reconcile. So the artist fudged the details as best he could to minimize the visual impact.

To Tyler, given that the octagon is such a noticeable feature of the painting, the refocusing explanation is farfetched. Had Lotto enjoyed the benefit of a projection mirror, he would have arranged his set-up to capture that figure correctly, Tyler argues. “The octagon was the one thing he would have needed the lens for, so he would get it right,” he says. A better explanation, Tyler suggests, is that Lotto painted the rug freehand without fretting too much about the details.

All that glitters

If it’s controversial to claim that Lorenzo Lotto used optical aids, then it’s even more so to say the same of Jan van Eyck, a Flemish artist who was painting a century before Lotto. Nonetheless, in his book, Hockney argues that van Eyck may have used a concave mirror to create his famous 1434 portrait of the local merchant Giovanni Arnolfini and his wife.

In the book, in television interviews, and on Web pages, Hockney and Falco have argued that van Eyck would have needed a mirror or lens to render so faithfully the chandelier depicted in the portrait.

To Hockney’s eye, the chandelier is in “perfect perspective,” he told a 60 Minutes interviewer. However, by identifying points on the chandelier that lie along parallel lines and then plotting those lines by computer over a reproduction of the chandelier image, Stork has determined that the lines don’t merge to the expected vanishing points. “Wham! What a mess,” he declares, showing a slide of van Eyck’s chandelier covered with a jumble of colored lines. “It’s in terrible perspective,” he says. And that, Stork argues, is evidence that van Eyck didn’t use a mirror.

The painted chandelier’s arms also fail another test. Were the arms identical, lines drawn through the same features of each arm should meet at the chandelier’s axis of rotation. That’s not the case with the couple’s chandelier, Stork finds.

Both Hockney and Falco also have drawn attention to a convex mirror that appears in the painting on the back wall of the room in which the couple stands. Had van Eyck flipped over that mirror, they suggest, the painter could have created the optical equipment required to make the painting.

Yet Stork determined the probable focal length of a concave mirror made by inverting and silvering the convex mirror shown in the painting. For instance, by computer-correcting the reflection painted, he found that the focal length of the mirror would have been too short to project images of Arnolfini and his wife and other features of the painting. This mismatch, he says, militates against the Hockney-Falco claim that the painter relied on optical aids.

Stork presented his challenges to the chandelier and mirror claims in his talk at Goddard, which also ranged over several other paintings, and other criticisms of the Hockney-Falco thesis.

For instance, to explain the sudden blossoming in the early 1400s of a painting style that’s almost photorealism, Stork recommends alternative explanations to optical technology. For example, the advent of oil paints and the control they offered artists are well documented in the historical record.

Not that any of these arguments are convincing the father of the opticality theory. “Mr. Stork doesn’t really change my view at all,” says Hockney.

Even so, Stork has scored some points. Hockney-supporter Kemp rates Stork’s chandelier deconstruction as “pretty convincing.” So does H. John Wood, lead optical engineer for the Hubble Space Telescope, who heard Stork’s argument at Goddard. Both Wood and Kemp say they still find the Hockney-Falco theory persuasive but not convincing beyond the reasonable shadow of a doubt.

With his hypothesis now supported and challenged by technical analyses, Hockney muses about whether science can ever settle the issue. After all, he arrived at his hunch about optics by looking at paintings and, later, tested it by trying out optical devices in his own art.

“The only way you’d know [whether those devices were used] is by looking at pictures,” he says. To him, the photographic style that started showing up in the early 15th century is a dead giveaway that the old masters used optical technology.

Now that he has sparked an art-history debate among scientists, Hockney is retreating from the fray. “I’ve gone back to painting now,” Hockney says. “I’m bored with the optical view of the world.”


If you have a comment on this article that you would like considered for publication in Science News, send it to Please include your name and location.

More Stories from Science News on Physics