Apelles and the Birth of Illusionism: Ancient lessons in painting spatial depth

[Fig. 1] Man In Armor, Rembrandt van Rijn, Oil on canvas, 54” x 41”, 1655

Apelles and the Birth of Illusionism:

Ancient lessons in painting spatial depth

[This article appears in an edited form in Lessons in Classical Painting, by Juliette Aristides.]

During the early Middle Ages surfaces were not simply painted. That is, if painters wanted to create the look of gold they didn't just create a painted facsimile, they applied real gold to the surface. They reasoned; why bother imitating one material with another? In the 14th century, for example, Cennino Cennini recommended that if you wanted to paint wool you should scuff up the surface of your panel with a wooden block so that it would feel wooly. Artists were not interested in using paint as a tool to produce the mimetic illusion of gold the way Rembrandt did years later.

With the advent of Humanism, artists of the late Middle Ages such as Rembrandt ignored their immediate predecessors and began looking instead to the time of the ancient Greeks for inspiration. They viewed it as a time when nature itself was a guiding principle, and believed that what Proclus had said two thousand years earlier was still true: “Space is nothing other than the finest light.” Painters felt that if they could relearn how to capture the elusive effects of light, they too could command space.

Rembrandt’s stunning 1665 portrait of Alexander the Great, Man in Armor [Fig. 1], may seem relatively modern to our eyes but it employs techniques that had been invented in ancient Greece some two thousand years previously by a painter called Apelles. During Apelles’ own lifetime, his fame and reputation were unrivaled, and he soon became the favorite portraitist of Alexander the Great. So it is fitting, then, that when Rembrandt paid deferential homage to Apelles’ ancient techniques, he chose as his subject the same man; the 22 year-old Macedonian conqueror of half the globe.

[Fig. 2] The mosaic from the Galla Placidia mausoleum. Created around 450 AD, this mosaic captures the transitional style between earlier naturalism and medieval symbolism. Still, certain stylized methods were carried forward, including Apelles' line, highlighted in red for clarity (the original used gold mosaic tiles).

Apelles had understood what painters of the Middle Ages seemed to have forgotten, but his lesson was deceptively simple: Value (light and dark, essentially) is the best painterly means of producing the illusion of depth. Put another way, Apelles’ big discovery was that black recedes and white advances, or, hollows are dark and ridges are light. The term highlight indelibly carries with it this sense that what is light in color is also high in relief – a notion that did not exist in concrete form before Apelles.

Apelles' addition of a white line along the leading edge of a plane pushed space forward towards the viewer. Despite being based in observation, this was the advent of a painterly convention, a visual trick, and a lesson in perception that Rembrandt would have viscerally understood. It's one that has lasted until this day, and while it may seem blindingly obvious to us now, it’s worth taking a closer look to fully appreciate what Apelles was trying to tell us [Fig. 2].

[Fig. 3] Without Apelles' highlighted edge, the “cube” (A) on the left can be read spatially in two ways, while (B) and (C) cannot.

The Line of Apelles

Without a highlighted edge (Apelles’ so-called “line”), spatial depth in paintings can be ambiguous; form can appear raised or recessed depending on how one looks at it. The addition of Apelles’ lit edge collapses this ambiguity, and establishes forms as either raised or recessed – but not both [Fig. 3].

This was a discovery that owed its origin to developments in theatrical scene paintings of ancient Greece during the Golden Age, around 500 BCE, when artists first created the illusion of raised or recessed panels on a painted backdrop known as the skene. Scenic painters knew that the gleam on a raised edge destroys any spatial ambiguity as to the direction of the three-dimensional relief that might arise from a shifting light source [Fig. 4].


Ancient Greeks saw the world differently to medieval Europeans, and it showed in their painting. Broadly speaking, while medieval artists had subjugated the world of the senses in the service of higher ideals (in the mold of Plato), subsequent Renaissance painters re-awakened to the world of Aristotle - where knowledge could be acquired by apprehending the world via the sensory apparatus; where light and surface were worthy of study.

[Fig. 4] In the left panel, the image reads as small raised rectangles, while on the right, it looks as though the rectangles are recessed and it’s the background grid that’s raised. They are however the same image; the one on the right is just flipped upside down. It’s an illusion caused by the fact that we evolved to orient ourselves by the sun, and it’s much less likely that the sun is shining from below than it is for the rectangles to have gone from raised to recessed, so we believe the latter. 

[Fig. 4a] Above, the lessons of Figs. 3 and 4 have been combined into a rendering of the capital letter E. In (A), the letter can be read spatially as either recessed grooves lit from below or raised relief lit from above. In (B), the application of Apelles’ white line forces us to read the letter as raised relief, lit from the top left.

Not surprisingly, this intellectual precedent had cultural significance in the ancient world stretching back even further than the Greeks. “In Old Kingdom Egyptian tombs, the figure of the deceased had to be recognizable so the Ka could find its proper habitation,” said the American historian Daniel Boorstin, explaining the cultural focus on verisimilitude that clearly informed artistic production in Hellenized Egypt [Fig. 5]. Apelles was the product of a culture that valued observational reality, and when this method of inquiry was rediscovered in the 16th century, so was he.

If we compare the ancient Egyptian paintings known as the Fayum mummy portraits (named after their place of discovery) with Rembrandt’s Portrait of a Bearded Man [Fig. 6] painted almost two thousand years later, we can see very little difference. The lighting is practically identical – from the accentuated bright spot on the tip of the nose and the sparkle of light in the eye to the hint of reflected light on the right edge of the cheek.

And then of course, there is Apelles’ line running down the bridge of the nose. Rembrandt knew, as had the Greeks, that placing your brightest values next to your darkest draws our attention. The human eye is inescapably drawn to high contrast. In a field of white sheep, it's hard not to stare at the black one. Echoing Apelles, Rembrant merely exaggerated the transition from light to dark into what was to become his signature style; a schizophrenic drama of bright illumination on the one hand and brooding darkness on the other.

[Fig. 6] Fayum mummy portrait, C. 100 BCE (left) Portrait of a Bearded Man with Wide Brimmed Hat (detail), Rembrandt Van Rijn, 1633  (right)

Contrast as a Cue to Spatial Depth

Not only does the addition of highlights naturally draw our attention, it contains valuable information regarding spatial depth. Physiologically speaking, if the visual data received by the eye are either highly contrasted, have concentrated detail (or both, as in the eyes and nose in Figs. 5 and 6), we interpret them as being closer to us than data that have less contrast and less detail.

There is an unavoidable psychological response to fine-painted detail at play here. Humans discern spatial information from surface texture such that when our eyes make small jumps (called saccades) from point to point while looking at a rough surface, our brain picks up the message that that surface is near to us. Rough texture is a depth cue understood by the brain to signify proximity. Thus, objects with sharp light-to-dark transitions or strong background contrast are interpreted by the brain as being close to us, while objects with smooth light-to-dark transitions or weak background contrast are interpreted as being far away. No doubt, there was an evolutionary advantage to being immediately able to tell whether that charging wildebeest was near or far. Depth cues manipulated by artists trace a psychological lineage all the way back to early hominids hunting the plains and avoiding danger [Fig. 7].

[Fig. 7] Here, Munsell’s value scale (1-9) has been superimposed in rows upon a graduated background. Known as the Bartleson-Brenneman effect, values along the red diagonal axis appear to be further away because they contrast less with the background, while values along the green axis appear closer because they contrast more with the background.

Value Constancy

Furthermore, we interpret our world not through absolutes of value, but via relative differences. A white floor tile seen in an interior by Vermeer, for example, can be instantly recognized as "white" by the brain despite being nowhere near the white end of Munsell's value chart in terms of its actual value [Fig. 7a].

Form and space are interpreted not through absolutes of value but through the relationship of one value to another. What's more, even in a darkened room we can still read depth and form despite all the values being dark and relatively similar in an absolute sense [Fig. 8]. In a room that's almost completely dark we can still spot that white floor tile, reminding us that perception happens in the mind and not the eye.

The actual value of the floor tile outlined in red (A), when plotted against Munsell's value chart, is surprisingly dark despite us knowing that it is "white."

This phenomenon is called value constancy: it’s our innate tendency to group and hold the perception of an object despite changes in visual data. Thus understood, representational painting is no longer the strict transcription of optical data but becomes, according to historian Ernst Gombrich, a “system of notification;” a set of agreed-upon conventions that may have little to do with observational reality - we paint what we know, not necessarily what we see. Apelles knew this better than any Middle Age painter.

Equally, Rembrandt knew that he didn’t have to pepper his whole canvas with strong blacks and whites in order to render form. He knew that by manipulating value constancy, contrast and value, he could create a more convincing illusion of spatial depth and surface texture (such as gold) than by simply applying real gold to canvas. Understanding the way the mind perceives the world is the key to Rembrandt’s metallic sheen in Man in Armor, and he valued it among his most prized magic tricks.

It's important to note here that learning by observing reality does not mean painting what you see in front of you. This seems counterintuitive at first, but what was important about Apelles' discovery was not that it conformed to what the eye sees, it's that it conformed to how the mind perceives. Aristotle, as the father of the scientific method, knew that insights into human perception came from direct observation of nature. It's up to the representational artist to learn how to manipulate this psychology into illusionism.

Talk of "mimesis" and "observation" may fool us into thinking that we are learning how to see the truth of the world. The greatest illusion of the artist is to convince the viewer that she is seeing reality. All of this is nothing but painterly smoke and mirrors. Tricks of the trade. All that we in the West are seeing when we look at a Rembrandt, for example, is a set of Western cultural conventions; artistic norms that may have a basis in shared psychology but are temporally and culturally rooted. Indeed, there's nothing less realistic about the Art of the Middle Ages than that of the Renaissance, or any other time for that matter.

[Fig. 8] This classic illusion demonstrates “value constancy." Even with all the lights turned down (or in shadow, as above) we can still recognize a “white” square on the floor, although in absolute terms the square is far from white: A and B are the same value.

Metallic Sheen

Rembrandt's ability to capture the effect of metallic sheen was a hard-won technical victory in what was, at the time, a march towards mimesis - the ability to capture the optical effects of light bouncing off the material substance of the world around us - and it owed much to Apelles. Indeed, there is a direct line from the ancient Greek's early innovations that carries all the way through the Renaissance to inform representational painters to this day. But when it came to rendering metal in paint, Rembrandt took Apelles' lesson one step further.

Rendering highlights is not the same as rendering reflection. With the further inclusion of specular highlights to the effects of modeling, much information about the nature of the surface texture can be suggested. Metallic surfaces (particularly curved surfaces such as Alexander the Great's armor or the lip of a goblet) tend to reflect the intensity of the light source – usually the sun. With the exception of bronze and gold, these metallic surfaces reflect little if any local color.

Unfortunately for the painter, white pigment has an upper limit as far as its ability to depict specular highlights – it simply will never be as bright as the sun – so Rembrandt had to cheat when painting his armor. He knew that if he simply painted everything else darker, then his lights would automatically look lighter. Because we intuitively know what we’re looking at thanks to the gestalt principle of value constancy, he knew that he could artificially lower the observed values of the overall scene without compromising spatial depth, and that even on a darkened canvas we’d still be able to read form and surface texture. The only way to make your specular highlights seem as bright as the sun is to darken everything else.

This trick would simultaneously enable him to create a higher relative jump in value to his specular highlights – a necessity when it comes to painting metallic sheen [Fig. 9]. This observation holds true when we consider Man in Armor: precisely because the overall value is relatively dark and indistinct, we know that we are looking at the shiniest metal when we see the bright reflection on Alexander’s helmet and breastplate.

Perhaps, in the end, old people are not simply bad at computers and unpredictable in traffic. Maybe they do have something to teach us.

[Fig. 9] The Rape of Prosperine, (detail) Rembrandt van Rijn. In order for Rembrandt to depict specular light on metal he lowered his surrounding values, enabling the highlights to pop. His specular highlight (f) is almost 4 value points from its next closest value, perceptually implying that we are looking at metallic sheen. And yet, notably, when plotted against the Munsell value chart, it’s clear that even his highest highlights (f) are still relatively low-key (7 on the Munsell chart). Rembrandt still has room to spare at the top end of the value chart if he needed it.