Summary essay for Chinese art history

1. Chinese bronze tripod of the type li. Height 15 cm. 15th or 14th century bc. Asian Art Museum of San Francisco.

[103]

In 1938 Max Mallowan excavated a temple of the fourth millennium bc at Tell Brak in northern Mesopotamia. The temple contained thousands of small stone plaques with prominent eyes (Fig. 2), and eye-like designs were con- spicuous also in the decoration of its walls and on the altar reconstructed in Fig- ure 3, so Mallowan called it the Eye Temple. But he could only speculate as to the meaning of the eyes. Thirty years later he commented that the interpretation of the eyed plaques “must in the absence of written records remain uncertain, and indeed the enigma of these unique objects is one that has long exercised the ingenuity of scholars. There is some satisfaction in reflecting that the true answer may never be known.”1 This is a sound conclusion, I believe, one that I am happy to endorse, but Mallowan’s satisfaction with it has not been shared by scholars with temperaments less philosophical than his. In the present paper I should like to examine some of the explanations that have been proposed over the years for designs like these.

As Mallowan himself was well aware, designs dominated by a pair of eyes are by no means confined to ancient Mesopotamia. The pot illustrated in Figure 4 is from Denmark, third millennium bc. Figure 5 shows something from prehistoric Britain, a carved stone cylinder from Yorkshire, early second millennium bc.

Interpreting Prehistoric Designs

2. Alabaster plaques from the Eye Temple at Tell Brak, northern Mesopotamia. First half 4th millennium bc. (For this date, suggested by recent excavations at Brak, see Akkermans & Schwartz 2003, pp.  186, 199.) The temple was rebuilt four times. The plaques belong to the second rebuilding, the altar (Fig. 3) to the fourth.

3. Reconstruction painting of the altar in the Eye Temple at Tell Brak. Mid 4th millennium bc.

[104] gombrich among the egyptians and other essays in the history of art

In 1957 the British prehistorian O. G. S. Crawford proposed an iconographic interpretation for these and all the other paired-eye images he knew of. In a book called The Eye Goddess he suggested that they all represent a fertility goddess: he suggested that the cult of the goddess originated with the first agricultural soci- eties in Mesopotamia, where Mallowan found the first images of her, and that it spread to Europe along with the agricultural way of life. A map he drew to indicate the spread of the cult is reproduced in Figure 6.2

Mallowan’s excavation report had already speculated that the Eye Temple was dedicated to a Mother Goddess who was symbolized by a pair of eyes.3 In extending the goddess’s domain beyond Mesopotamia, Crawford was making at least three assumptions. To believe that all paired-eye images represent a single goddess of Mesopotamian origin, he had to believe that the paired-eye design was invented once and only once in human history, that it had only one meaning, and that when it traveled, it traveled with that meaning permanently attached to it. He thought that the spread of agriculture would naturally take with it a design that meant the goddess of agriculture, but that without some such vital meaning

4. Pottery bowl from Svinø, Denmark. Height 12 cm. Passage Grave culture, 3rd millennium bc. National Museum, Copenhagen.

5. Carved chalk cylinder, the largest of three (the “Folkton Drums”) found in a Copper Age burial at Folkton, East Riding, Yorkshire. Diameter 14.8 cm. First half 2nd millennium bc. British Museum.

6. The Westward and Northward Migrations of the Faces, from O. G. S. Crawford, The Eye Goddess (1957).

interpreting prehistoric designs [105]

the design would not have traveled. He seems in fact to have believed that the only possible reason for any design to travel or even to exist is meaning, so he took it for granted that designs with prominent eyes have the same meaning wherever we encounter them. He appears to have made these assumptions more or less uncon- sciously. If he had formulated them explicitly to himself he might have realized that they cannot be correct.

To judge by the reviews of his book, prehistorians had misgivings about Crawford’s theory already in 1957. A prehistorian today would object that images dominated by eyes are found in places far beyond the reach of diffusion from Mesopotamia. The pot in Figure 7 is from pre-Columbian Brazil, from the lower Amazon region. The object in Figure 8, dating from the third millennium bc, is a Neolithic jade from the Pacific coast of China. As far as I know, no one has suggested that either of these designs came from Mesopotamia, and if paired-eye designs do not necessarily imply a connection with Mesopotamia, then Craw- ford’s theory dissolves. If the eyes in Figures 7 and 8 do not represent a Meso- potamian goddess, then the designs from Yorkshire and Denmark need not be a Mesopotamian goddess either. And of course isolated or exaggeratedly promi- nent eyes are not confined to prehistoric artifacts. We find them also, for example, on a Greek pot from about 500 bc (Fig. 9), a Chinese bronze from about 1400 bc (Fig. 1), and a pottery cup from medieval Nubia (Fig. 10).

Goddesses and fertility cults, though they still have devotees,4 were never really the explanation of choice for things like these. The standard way of accounting for paired-eye designs has long been to say that they had an apotropaic function, in other words that they were meant to avert evil. This explanation is usually offered without further comment, as though it needed no justification. One scholar cites it to explain the taotie motif on Chinese bronzes; another, equally without textual

7. Painted pot. Height 48 cm. Marajoara cul- ture, c.ad 500–1500, Marajó Island (Amazon estuary), Brazil. Museu Barbier-Mueller Art Precolombí, Barcelona.

[106] gombrich among the egyptians and other essays in the history of art

authority, cites it to explain the eyes on the prows of ancient Greek ships depicted in vase paintings.5 But to me it seems problematic in the extreme. The scholars who invoke it are telling us: “The maker of this object thought that putting a pair of eyes on it would prevent bad things from happening.” How do they know what the maker of the object thought?6

And how can it be that they offer the same apotropaic explanation for objects from widely different times and places? Surely we must ask ourselves whether this explanation is any better than Crawford’s. When Crawford said that eyes everywhere stand for a fertility goddess, he understood that he was proposing a symbolism that could have become geographically widespread only by spreading from one place to another. If we suppose instead that eyes are always a protective device, is that not also an idea that would have had to spread from one place to another? If so, then the apotropaic explanation suffers shipwreck on the same rock that sank Crawford’s theory: the motif of paired eyes occurs in cultures too far apart to be connected by diffusion.

In his 1979 book The Sense of Order Ernst Gombrich tried to construct an argument for universal apotropaic eyes that would escape this objection. He understood the problem: if the association between a motif and an idea is purely conventional, then the association cannot be universal unless it has diffused uni- versally. But for reasons I can only guess at, Gombrich was unshakably convinced that belief in the protective power of paired eyes is universal.7 Or rather—and the distinction is important—he was convinced that people all over the world, including you and me, feel protected by paired eyes, whether we consciously believe in their protective power or not. When the testimony of native informants does not mention protection, Gombrich just postulates “an unconscious sym- bolism unknown to the craftsmen themselves.”8 In other words, he argues that the association between eyes and protection is not a matter of convention but a universal of human psychology. It is not iconography, it is instinct. According to Gombrich, we have an innate, automatic response to eyes: we feel protected by them. Hence that curious phrase “unconscious symbolism”: unconscious fears are a part of our psychological makeup, Gombrich says, and images of eyes relieve those fears.

8a. Jade cong (tube with round inner and square outer perimeter) from Qingpu Fuquanshan, Shanghai, China. Height 5.6 cm. Neolithic, Liangzhu culture, mid 3rd millennium bc. Shanghai Museum.

8b. Detail of corner.

interpreting prehistoric designs [107]

These are claims about human psychology. Should we believe them? What evidence can Gombrich call upon to establish that I am full of unconscious fears, and that the sight of paired eyes comforts me? There is in fact a useful scientific literature on reactions to eyes and eye images,9 but Gombrich does not cite it. Instead he cites folklore and aligns himself with an older generation of art his- torians, people like Emanuel Loewy, who explained not just eye images but all decoration as an attempt to ward off malevolent influences.10

Whatever we may think of this view of the psychology of ornament (and I have to admit that I find it hard to take seriously), we must not allow it to distract us from the real problem here—Crawford’s problem, the problem of the archae- ologist who digs up something with a pair of staring eyes on it: the problem is to understand why these images occur all over the world. It is not a difficult problem, in fact the solution is rather obvious. We are indeed dealing with a universal of human psychology, but it has nothing to do with unconscious fears or feelings of protection.

Let me propose a thought experiment. Suppose that next to Figure 10 there was another illustration exactly like it, but with no eyes on the cup. Which cup would you be looking at? You would be looking at the cup that is looking at you. Evolution has programmed us to respond to eyes with attention. Eyes give us vital information: they tell us where other people or animals are directing their atten- tion. We are acutely alert to eyes because they furnish our best clue to attitudes and intentions that are sometimes matters of life and death to us.11

What the designs in Figures 1 and 10 have in common is that they stare at you. Because of that, they have power over your attention. They also have power over your imagination. In Figure 10, while you are consciously focused on the pair of eyes, your peripheral vision takes the vertical and horizontal lines for a nose and a mouth. If the eyes were erased, you would not see a nose or a mouth on the cup, but given a convincing pair of eyes, your imagination will try to interpret anything nearby as part of a face.

In Figure 1 it is hard to look away from the eyes, and hard not to feel that the whole object is somehow alive. That so simple a motif can be so mesmerizing is

9. Attic black-figure eye cup. Diameter 36.4 cm. c.520 bc.

[108] gombrich among the egyptians and other essays in the history of art

a little uncanny, and much the same animating effect can be obtained even from eyes much less carefully drawn than these. We are so primed to search our sur- roundings for eyes that might be watching us that an artist can make us see a pair of eyes just by drawing a pair of circles. And we will accept almost any squiggle between or below those eyes as a nose or a mouth.

The people in Detroit responsible for Figures 11 and 12 no doubt knew per- fectly well that they were designing cars with monstrous faces on the front. The faces were not intended to avert evil or to calm unconscious fears, they were meant to sell cars; they make the whole car look alive, full of energy, ready to eat up the road. Gombrich was well aware of this effect, but he called it “protec­ tive animation”12—why “protective”? Why must animation be protective? These Buicks are not examples of protective magic, nor is there anything in them that we could reasonably call symbolic meaning.13 Their chrome trim is a means of enlivening things, giving them visual interest.14

To help me summarize my argument so far let me introduce a convenient bit of shorthand: I will use the term “universal motif ” to mean a motif that occurs independently in unrelated cultures. A universal motif, in other words, is a motif that has been invented more than once. Then my argument goes as follows: Gom- brich says that a universal motif cannot be explained by an attached meaning; and I agree. He says that such a motif can only be explained by our innate reaction to it, and again I agree. But in the case of eyes he identifies that innate reaction as a feeling of release from primal anxieties, and I disagree. The innate reaction, the reaction built into us by natural selection, is alertness. For that reaction to be prejudiced by any fixed emotional coloring—comfort or fear or anything else— would surely be maladaptive, and natural selection would not have programmed it into us. Our only biologically programmed reaction to eyes is attentiveness.

Why then do eye designs occur all over the world? Because the power to com- pel attention is something that artists can use. And it is a power that an artist can turn to any purpose he wishes. Try to imagine what the advertisement in Figure 13 would look like if the cat’s eyes were closed. Having got your attention with a pair of staring eyes, the designer can attach whatever message his purposes require. The motif is universal, but the uses to which it is put are infinitely various. In other words, we are dealing with a universal motif that has a multiplicity of local meanings. It can mean one thing in a Mesopotamian temple, something else on a Chinese jade, and on a Buick, nothing at all.15 The meaning it carries in a given context is specific to that context.

Now if that is so—if the motif of paired eyes is a universal motif with local meanings—then two important corollaries follow. First, the meaning that attaches to the motif in one context cannot tell us anything about the meanings assigned to it in other contexts. We cannot hope to get at meanings through cross-cultural comparison or ethnographic analogy. And second, we do not need to know meanings to understand why the motif exists. It exists because artists all over the world have discovered that our perceptual system has an inbuilt reaction to it that they can use.

The same two conclusions apply, it seems to me, to any universal motif, any motif that has arisen more than once in unrelated cultures. First, anywhere a universal motif does have a symbolic meaning, that meaning is locally assigned

10. Cup from Nubia. Christian period, c. ad 550–1500. Aswan Museum.

interpreting prehistoric designs [109]

to it. And second, repeated invention of the motif can only be explained by the response it evokes independent of meaning.

Figures 14 and 15 show a motif that is just as ubiquitous as paired eyes. Ignore the element in the center of each bowl; the design that runs around it is a continu- ous series of spirals joined by diagonal arms. This is the running spiral, a motif that occurs worldwide, sometimes in completely unrelated cultures, from the Neo- lithic onward—one scholar calls the Neolithic “the age of the spiral.”16 The pot- tery bowl in Figure 14 is from pre-Columbian New Mexico, eleventh century ad. The gold one in Figure 15 is from Minoan Crete, seventeenth century bc. The two objects are about three thousand years apart. An article or indeed a book could be written about the history of the running spiral, about the theme and the variations that have been wrought upon it. One of the simplest variations is seen in Figure 16 on a pot from Bosnia, fifth millennium bc. We have encountered much the same design already on the pot from the lower Amazon region shown in Figure 7.

The running spiral has been independently invented, probably many times, and endlessly reproduced and varied. Why? Meaning cannot be the answer. Sup- pose I were to declare that the person who invented the running spiral in prehis- toric New Mexico was thinking about a symbolic meaning. What feature of the spiral would I have explained? What feature of it would have been different if the artist had been thinking about a different meaning? What feature is different when another artist with other ideas reinvents it? Perhaps we should speak of universal motifs as “discovered” rather than “invented.” The reason for the running spiral’s repeated invention—or rather, for its popularity with its many discoverers—can 13. Advertisement.

11. 1949 Buick Roadmaster.

12. 1951 Buick XP-300.

[110] gombrich among the egyptians and other essays in the history of art

only lie in our perceptual system’s response to the formal properties of the motif: its combination of periodic symmetry and dynamic movement, its suitability for decorating continuous curved surfaces like pots.17 In short, people use it because they like the way it looks. I have suggested that paired eyes are sometimes used purely to enliven (think of those Buicks); I would not be surprised if this were always the raison­d’être of running spirals. To say exactly why the perceptual sys- tem favors this particular configuration of curves would require a detailed under- standing of its workings that I  at least do not have. But as Gombrich himself pointed out many times, artists are in the business of probing the mechanisms of our perceptual system. They look for configurations that tickle our perception in ways they can use, and among the configurations they have found effective are running spirals and paired eyes.

Let me suggest an analogy here with universals in music. Ethnomusicologists nowadays despair of finding very much that is common to all musical cultures, but one universal they seem to agree on is the association of music with the super- natural; in the words of Bruno Nettl, “All known cultures accompany religious activity with music.”18 Notice that he did not say “All known cultures use music to convey sacred messages.” Music’s association with the sacred arises from its power to affect psychological states. Of course it sometimes does carry religious mes- sages: the words of the Catholic liturgy can be set to music and sung. But it does not have to. Nor does it have any necessary connection with the sacred; it can serve other functions too. I would propose that all the same observations apply to visual designs. They affect psychological states; that is why they exist. They may or may not have messages attached to them. The theologian who declares eye contact with holy images to be a vital act of religious communication is making a theological gloss on a psychological fact. Running spirals no more have an intrin- sic meaning than the C-major scale does. Running spirals and paired eyes and musical scales are ways of exploiting the mechanisms of our perceptual system.19

14. Painted bowl from the Swarts Ruin, south- western New Mexico. Mimbres culture, Clas- sic period, c. ad 1000–1130. Peabody Museum, Harvard University.

15. Gold bowl from the Aegina treasure. Diam- eter 9.7 cm. Minoan, c.17th century bc. British Museum.

interpreting prehistoric designs [111]

The point I wish to establish is that in prehistoric times, as in any other time, there existed motifs that were found attractive for reasons unconnected with iconographic meaning, motifs whose meanings, if any, were assigned to them by convention and therefore cannot be deduced from them. As a last illustration let me mention the whole class of geometric patterns that have emerged repeatedly from simple plaiting and weaving techniques. Patterns that move diagonally are effortlessly natural in these techniques. Complex though it may seem, the pattern of interlocked T’s on a modern basket from Brazil (Fig.  17) is something that basketmakers weaving two colors at right angles have certainly invented more than once, for we see it also on an ancient Chinese basket (Fig. 18).20 According to ethnographers, the makers of the Brazilian pattern say that it depicts a local animal and has a symbolic meaning. If so, the pattern came first; animals and

16. Pottery bowl from Nebo, Bosnia. First half 5th millennium bc.

17. Sieve from the upper Rio Negro region, Brazil. Collected 1993. Diameter 51 cm. Museu Para ense Emilio Goeldi, Belém, Brazil.

18a. Lacquered bamboo basket from Jiangling Jiudian tomb 410, Hubei province, China. 18.9 by 24.6 cm. 4th or 3rd century bc.

18b. Drawing of basket.

[112] gombrich among the egyptians and other essays in the history of art

19. Grave at Wujin Sidun, Jiangsu province, China. Neolithic, Liangzhu culture, second half 3rd millennium bc. The grave contained about 120 jades, including disks, ax blades, and cong (tubes with round inner and square outer perimeter). Most of the cong have much taller proportions than the examples in Figs.  8, 20, and 21. The tallest is 36 cm high.

meanings were attached to it later. They had nothing to do with its invention, and they certainly could not be read out of it.21

Let me return now to the paired-eye motifs on Chinese Neolithic jades. Third millennium burials on the east coast of China have yielded hundreds of jades like the one in Figure 8; the grave shown in Figure 19 contained thirty-three of them. Invariably carved with paired-eye designs, these peculiar objects are the defining artifact of the Neolithic culture that produced them, but their origin and use are a mystery (perforated from top to bottom, they cannot be containers). On most of them the carving is limited to simple eye designs on the corners (Fig. 8). But the very exceptional object illustrated in Figure 20, discovered in 1986, is carved on the sides as well, and the design on the side is unusually complicated (Fig. 20b). It seems to show the feathered headdress, trapezoidal face, and shoulders of a man

interpreting prehistoric designs [113]

whose bent arms reach down to touch the eyes of a crouching animal. The animal has a bulbous nose, a fanged mouth, and clawed forelegs.

When the jade of Figure 20 was unearthed, its combination of a man and an animal was immediately hailed as providing a prototype for the simple paired-eye designs long familiar from other jades. In Figure 8 the round eyes with horizontal prongs seem to stand for the feathered man, the big oval eyes joined by a strip for the animal; the same abbreviations can in fact be seen in Figure 20a, on the cor- ners of the object. But having related the familiar design in Figure 8 to the excep- tional design in Figure 20b, are we any closer to having a meaning for it? And what should we make of the jade shown in Figure 21? Here we find only the eyes of the feathered man—twice. Does the man-plus-man design mean something different from the man-plus-animal design? I see no way of answering this question. I do not believe that we have the slightest clue what any of these designs meant to the people who made them, and I doubt that we ever will.22

Nevertheless, needless to say, there have been many attempts at symbolic interpretation of these jades. Invariably they rely either on ethnographic anal- ogy or on the assumption that the symbolism is obvious and open to inspection. The argument from ethnographic analogy says for example that in a certain nine- teenth century Siberian tribe, an animal companion helps a shaman take ecstatic voyages, so the jades must show a shaman with his animal helper. But this is only to assume that all images of men riding on animals are depictions of shamans with their animal helpers, which is nonsense.23

In its proper place ethnographic analogy is indispensable. When we guess that a certain piece of chipped stone from a Palaeolithic site was a tool used to skin car- casses, we are implicitly or explicitly making an analogy with similar stones that we know were used for skinning carcasses in more recent societies. But ethno- graphic analogy does not supply proofs, it only suggests possibilities. And while it is invaluable when our problem is the function of a tool, it is useless when our

20a. Jade cong from Yuhang Fanshan, Zhe jiang province, China. Height 8.8 cm. Neolithic, Liangzhu culture, first half 3rd millennium bc.

20b. Detail of design on side.

[114] gombrich among the egyptians and other essays in the history of art

problem is symbolic meaning, because—as ethnography reminds us—the pos- sibilities for meaning are limitless. Eyes on the prow of a boat do not have to mean protection; tourists in modern Bali are told that eyes help a boat find its way. Analogy with other cultures would never have told us what the ethnographers tell us about the basket in Figure 17; what other culture would see an anaconda there? I  see no role for ethnographic analogy in iconography; cross-cultural compari- sons cannot help us recover beliefs.24

As for the notion that prehistoric symbolism is so straightforward that a mod- ern observer can interpret it by inspection, this I find even less defensible. We would not take seriously the suggestion that our own symbols could be under- stood by an uninformed observer. What would such an observer tell us about the lion and unicorn in the Queen’s royal arms? What would he tell us about the image of a man nailed to a cross? (An ancient Roman observer might have told us that it means “justice.”) We do not believe that our symbols are self- explanatory. Why should we suppose that the symbols of prehistoric people are? Why should we expect the design in Figure 20b to be any more intelligible to us than the design of a man on a cross would be to a Neolithic Chinese observer?

I have to confess to a cynical suspicion that iconography without texts flour- ishes because it is safe: prehistoric symbols can be interpreted without fear of contradiction. Would an interpreter who confidently explains eyes on prehis- toric objects be equally ready to tell us the meaning of eyes on an Egyptian coffin (Fig. 22)? I doubt it. There is too much danger that a literate Egyptologist might actually know what the eyes mean. I doubt that anyone would claim the ability to penetrate Egyptian iconography without help from texts.

But let me suppress my cynicism and ask how these interpreters would them- selves justify what they do. As far as I can see, they have only one justification

to offer. They assume that prehistoric symbolism, unlike ours, is open to inspec- tion, because prehistoric people, unlike us, were simple thinkers who could only use symbols in simple, straightforward ways. Because primitive man was not capable of anything beyond the most obvious associa- tions, the color red always means death. (Or perhaps it always means life.) The thinking of primitives is so simple as to be transpar- ent to us.25

I hope that I do not need to argue against this assumption; to make it explicit ought to be sufficient to explode it. But perhaps I can clarify the issue by quoting a sentence from a recent book on prehistory. The sentence reads: “The classic definition of a symbol is ‘something that stands for something else’—the colour red for blood, for instance, which then might be general- ized as a symbol for . . .” I invite the reader

21. Jade cong from Yuhang Yaoshan, Zhejiang province, China. Height 7.9 cm. Neolithic, Liangzhu culture, early 3rd millennium bc.

interpreting prehistoric designs [115]

to finish the sentence: “which then might be generalized as a symbol for . . .” For what? If the thinking of primitives were transparent, then we should have no dif- ficulty in agreeing on how the sentence should end, but in fact I doubt that we would all independently end it with the same word. The color red has an obvi- ous association with blood, but that is not the same thing as having a predictable symbolic meaning. The association, no matter how obvious, puts no constraint whatsoever on symbolic use of the color.26

Palaeolithic specialists studying the emergence of our species from its primate ancestors are concerned above all with the emergence of distinctively human cog- nitive capacities. One of those capacities is the ability to use arbitrary symbols, most notably, the words of a language. Upper Palaeolithic cave painters undoubt- edly had language; is it likely that they lacked the ability to use arbitrary visual symbols? Certainly late prehistoric people had that ability. In several parts of the world late prehistoric people invented writing systems, and writing systems are an extremely sophisticated use of visual symbols.27

I would in fact see a close analogy between deciphering the symbolism of Neolithic designs and deciphering a script in an unknown language. Until the Egyptian hieroglyphic script was connected with the Coptic language, attempts to decipher it amounted to little more than free association with the pictorial content of the signs, and since most of the signs in a piece of Egyptian writing do not mean what they depict, the result was fantasy.28

This brings me to a conclusion that nobody but Max Mallowan could be happy with. If the makers of the designs I have been discussing had an ability to use arbitrary symbols that was not significantly different from our own—and I see no reason to doubt this—then unless their symbols occur in a context that is far more detailed and informative than anything I have encountered, the pros- pects for iconography without texts look grim. The only consolation I can offer is that we do have access to those features of running spirals, eye motifs, and other prehistoric designs that depend not on arbitrary convention but on the way the perceptual system works. If we find prehistoric designs striking or beautiful, that is not likely to be a coincidence irrelevant to understanding why they exist. It is more likely to be a manifestation of our biological identity with their makers.

22. Painted wooden coffin. Length 200 cm. Egypt, Middle Kingdom, 12th or 13th dyn- asty (first half 2nd millennium bc). British Museum.

[116] gombrich among the egyptians and other essays in the history of art

notes Lecture delivered on 3 June 2005 at the colloquium “Ico-

nography without Texts,” Warburg Institute, London. I should like to thank John Ellington, Stephen Houston, Susan Naquin, and Kyle Steinke for helpful comments on drafts of this paper, and John Baines, Jerrold Cooper, Rebecca Farbstein, and James Trilling for answering ques- tions and suggesting references.

1. Mallowan 1965, p. 48. Mallowan’s excavation report (Mal- lowan 1947) reviews a wide range of explanations, for instance that the plaques might depict the god of the temple or the worshippers who presented them to the god (p. 209), or that they might display “benevolent and protective eyes, intended to bring good fortune to those that deposited them” (p. 151); see also pp. 32–5, 150–59, and 198–210.

2. Crawford 1957. 3. Mallowan 1947, p. 153. 4. E.g. Gimbutas 1989. 5. Watson 1977, p.  50; Kirk 1949, pp.  132–3. Deonna 1965,

pp.  183–9, gives a long list of ancient cultures with eye designs that he takes to be apotropaic.

6. The interpretation of paired eyes as an apotropaic device seems to have been inspired not by any ancient textual source but by superstitions prevalent in the Mediterra- nean region in recent times, where belief that the gaze has power to harm, i.e. belief in the “evil eye,” is said to be widespread. (Notice that modern believers in the evil eye do not necessarily rely on images of eyes for protection against it. Most of the amulets illustrated in Maloney 1976 do not incorporate eyes.) As far as I can tell, the notion that eyes in ancient designs have an apotropaic function is only a careless bit of ethnographic analogy, a guess that has hardened into a “fact” through lazy repetition in the scholarly literature.

7. Gombrich 1979. See e.g. p.  259, where he speaks of “the equally universal belief in the apotropaic power of eyes or masks” as though it were established fact.

8. Gombrich 1979, p. 269. 9. For a brief summary with references see Ellsworth 1989.

The biological literature that deals with the evolution of eye spots (e.g. on the wings of moths or the tails of fish) is complex and less directly relevant but nevertheless instruc- tive in showing that no single explanation, for instance that eye spots deter predators, is generally applicable.

10. According to Gombrich (1979, p. 257), Loewy “proposed that this ‘apotropaic’ function could account for the pur- pose and origin of most if not all decorative motifs.” On p.  259 Gombrich momentarily distances himself from Loewy’s view, but the position he ends up taking seems indistinguishable from Loewy’s: he relates not just eye

motifs but all decoration to unconscious fears. Interlace, knots, and mazes, he tells us, confuse evil spirits, thus protecting us from them: “If it could be shown that there also exists a widespread disposition to attribute to hostile powers a dread of confusion and indeterminacy, we would come closer to an interpretation of the tangled dragons of Northern art” (p. 263; the “widespread disposition” intro- duced here in an “if ” clause is subsequently taken to be a fact). Imaginary animals, Gombrich says, provoke an ambivalent response, fear mixed with laughter, and this somehow (the point is not explained) enables them too to function protectively (p. 256). On the strength of these examples, along with that of supposedly apotropaic eyes, Gombrich concludes that all ornament arises from an unconscious urge to allay fears.

11. In an earlier essay Gombrich himself made just this point (Gombrich 1973, pp. 203–4): “there is an obvious survival value in recognizing the eyes, and even the direction of the gaze, of our fellow creatures. It is useful to know when and how we are being looked at if we want to respond adequately to the threat or invitation of another creature.” This entirely plausible analysis conflicts sharply with Sense of Order’s insistence on a specific and rigid response. The range of words we use for different ways of looking is itself testimony that eyes carry no fixed message: from glaring to ogling, from glimpsing to staring, from squinting to leer- ing; we can look tenderly, warily, angrily—the possibilities are endless.

12. Gombrich 1979, pp. 171, 261–3, 273. 13. Throughout the present paper I am using the terms “mean-

ing” and “symbolic meaning” in the narrow sense of a meaning attached by convention to a sign; a crucifix, for example, has a well-established conventional meaning in our culture. In the 1950s the Detroit designer may well have intended for the car buyer to be reminded of jet aircraft (the comparison was sometimes explicit in advertising: Heskett 1980, p.  179), but this is not the kind of associa- tion to which I would apply the term “symbolic mean- ing.” Broader uses of the term are discussed in Bagley 1993 and Halle 1998. A remark of Halverson’s is worth quoting (Halverson 1987, p. 82): “There are, of course, other mean- ings of ‘meaning’. One of the most elusive is imbedded in the very vague word ‘meaningful’, which seldom means ‘full of meaning’; in fact, it is seldom possible to extract any specific meaning from its use at all. It often expresses merely a suspicion, wish, or feeling that something might have specific meaning if only it could be discovered or articulated.”

14. Notice that the effect of animation in Figures 11–12 does not depend on any conscious awareness of resemblance to eyes or a face. The designers actually may not have wanted

interpreting prehistoric designs [117]

the buyer consciously to see a face on the front of the car—perhaps this could be confirmed in the archives of General Motors—since he might have thought that silly. Paul Wilson (1976, chap. 5, esp. pp. 200–202, 245) is acute and entertaining on the subject of these automobiles.

15. Greek cups of the type shown in Figure 9 should perhaps be classified with Buicks. John Boardman (2001, p.  277) does not ascribe iconographic meaning to the eyes on such cups; instead he points out that when the cup is tilted upward for drinking, a spectator sees the handles as ears and the foot as a mouth, the eyes thus having the effect of animating the entire vessel. Of course, even if some iconographic interpretation should be correct—for example that the eyes abbreviate a mask of Dionysus (Fer- rari 1986)—it would still be the psychological power of eyes that explained why the mask was abbreviated to a pair of eyes rather than, say, to a nose and mouth.

16. Trilling 2001, p. 94. 17. Hence, perhaps, its appearance in the Neolithic and not

earlier: it has no special aptness for the surfaces to which Palaeolithic artists applied designs.

18. Nettl 1983, p. 40. Perhaps this should be regarded as a uni- versal of religion.

19. A similar point is interestingly formulated in Spinden 1913, p. 48. [Note added in 2014: On archaeological evidence for universals in musical scales see Bagley 2005.]

20. To match up the two patterns it may help to focus on the light parts; try following a light zigzag line from top to bot- tom in each pattern.

21. Lucia van Velthem (in McEwan et al. 2001, p. 208; cf. Boas 1955, pp. 90–91) says that the pattern in Figure 17 contains

“figurative or iconic elements” that are taken from an ani- mal native to the Amazon region (an anaconda, she seems to believe) but that are hard to recognize because the weaving process has given them a “geometric and stylized” appearance. But if the pattern was arrived at in the Ama- zon region by “stylizing” an Amazonian animal, how was it arrived at in other places? Are we to suppose that pro- cesses of stylization which had different figurative starting points—an anaconda in South America, something else in China, something else again in other places—all con- verged on a single geometric pattern? This pattern arises from weaving, not from depicting.

22. A few scholars have argued that the motif on the jades is ancestral to the animal-like motif of Bronze Age art con- ventionally called taotie (see e.g. Li Xueqin 1993). Even if this were correct, the connection would shed no light on questions of meaning, since there is no agreement about what (if anything) the taotie means. But it is unlikely to be correct. The only feature cited as evidence that the two motifs are related is paired eyes, a feature which, as

I hope to have established in the present paper, could just as well have arisen independently; or, if not independent, could have been adopted by bronze casters quite igno- rant of or indifferent to the meanings that were attached to paired eyes a millennium earlier by the makers of the jades. Nothing resembling the elaborate form of the jade motif—a man atop a beast—ever occurs on bronzes; the bronze decoration never shows one creature sitting on top of another.

23. For a shamanistic interpretation of these jades (and of much else) see Chang 1989; Chang 1986 (epilogue); Chang 1994; and compare Bagley 1993. Chang’s interpre- tations do not rest on archaeological evidence but seem only to reflect his conviction that shamanism was the reli- gion of all prehistoric societies. This conviction licenses calling any artifact or design “shamanistic” without fur- nishing any evidence for shamanism in the society that produced it.

24. A special case ought to be acknowledged here. Archae- ologists investigating a prehistoric culture sometimes have available to them ethnographic accounts of a cul- ture descended from it. An archaeologist in the Ameri- can Southwest excavating a sixteenth century pueblo, for example, can interpret his finds in the light of ethnographic accounts of nineteenth century pueblo Indians. He may thus be led to interpret a circular underground chamber as a kiva used for ceremonies; and he might go further and attach certain beliefs to it (he might for instance suppose that a hole in its floor was intended to symbolize a tribe’s place of origin). Analogies of this kind can be very con- vincing, and no doubt they are often correct. Awkwardly, however, they depend on an assumption of unchanging beliefs that the archaeologist has no way of verifying. The same kind of reasoning, based only on similarity of form, might lead a European archaeologist to identify a Roman public hall as an early Christian church, or the Winged Victory of Samothrace as a Christian angel. In the example just given, did the beliefs of sixteenth century Indians coincide exactly with what a nineteenth century Indian told an ethnographer? If not, which nineteenth century beliefs can we safely project backward and which can we not? Might sixteenth century builders have had a different reason for putting a hole in the floor of a kiva? We have no way of answering these questions. For an illustration of the abuse all too common with this type of analogy, see a sha- manistic interpretation of pipe bowls from North Ameri- can Indian sites summarized by Hill (1994, pp. 87–9): here, surely, the assumption of beliefs uniform over time and space is pushed beyond credible limits. In the realm of ideas the changelessness of “traditional societies” is more a convenient assumption than a demonstrated fact (Boas

[118] gombrich among the egyptians and other essays in the history of art

1955, pp. 6–7). The merest glance at a society with a long historical record, our own for instance, makes it clear that the ideas attached to artifact types can change.

25. Allan (1993) claims just this ability to understand the thinking of “primitives” (a category which for her includes the literate Chinese of the Anyang civilization). If we remind ourselves of the intellectual achievements that we lump together under the term “the rise of civilization”— writing and metallurgy for a start—and then reflect that the people who achieved them were prehistoric, we will take a less condescending view of prehistoric thinking.

26. The sentence, quoted from Gosden 2003, p. 118, ends with the word “danger.” Arriving at London Heathrow a few days before the colloquium, I found the walls of the termi- nal placarded with advertisements for a bank that boasts an ability to bring local expertise to bear on international banking (“Never underestimate the importance of local knowledge”). One of them showed three identical squares of red labeled with three different captions: “China— Lucky”; “Britain—Danger”; “Indonesia—Brave.” (Gos- den is British.) In ancient Egypt the color red acquired some of its associations not because it is the color of blood but because it is the color of the desert (as contrasted with the irrigated black land: black on the face of the god Osiris signifies resurrection).

27. The various nonfigurative motifs of Upper Palaeolithic cave art (for which see Bahn 1998, pp. 180, 222–5), if they are symbolic, presumably qualify as arbitrary symbols (Mithen 1996, pp. 155–7, so interprets them). Merlin Don- ald (1991, chap. 8), overlooking the nonfigurative motifs (pp. 280, 285), believes that the cave painters were not able to use arbitrary visual symbols. He takes the invention of writing to signal the advent of the necessary cognitive skills (p. 287), indeed he interprets the failure of Palaeo- lithic people to invent writing as evidence that they lacked those skills (p. 279; it seems more likely that they lacked a use for writing). His further argument that the history of writing systems reflects a continuing development of cognitive abilities is vitiated by misconceptions about how writing systems function (see e.g. pp. 292–3 on “learning to read hieroglyphics out loud”).

28. Pope 1975.

interpreting prehistoric designs [119]

Akkermans & Schwartz 2003. Peter M. M. G. Akkermans and Glenn M. Schwartz. The Archaeology of Syria. Cambridge: Cambridge University Press, 2003.

Allan 1993. Sarah Allan. “Art and Mean- ing.” Whitfield 1993, pp. 9–33.

Bagley 1993. Robert Bagley. “Meaning and Explanation.” Archives of Asian Art 46 (1993), pp. 6–26. [Reprinted in the present volume.]

Bagley 2005. ———. “The Prehistory of Chinese Music Theory.” Proceed­ ings of the British Academy 131 (2005), pp. 41–90.

Bahn 1998. Paul Bahn. The Cambridge Illustrated History of Prehistoric Art. Cambridge: Cambridge University Press, 1998.

Boardman 2001. John Boardman. The History of Greek Vases. London: Thames and Hudson, 2001.

Boas 1955. Franz Boas. Primitive Art. New York: Dover Publications, 1955.

Chang 1986. K. C. Chang. The Archae­ ology of Ancient China. 4th ed. New Haven: Yale University Press, 1986.

Chang 1989. ———. “An Essay on Cong.” Orientations, June 1989, pp. 37–43.

Chang 1994. ———. “Ritual and Power.” In Robert E. Murowchick, ed., China, Ancient Culture, Modern Land (Nor- man, OK: University of Oklahoma Press, 1994), pp. 60–69.

Crawford 1957. O. G. S. Crawford. The Eye Goddess. London: Phoenix House, 1957.

Deonna 1965. Waldemar Deonna. Le symbolisme de l’oeil. Paris: Éditions de Boccard, 1965.

Donald 1991. Merlin Donald. Origins of the Modern Mind: Three Stages in the Evolution of Culture and Cognition. Cambridge, MA: Harvard University Press, 1991.

Ellsworth 1989. P.  C. Ellsworth. “Eyes.” In Erik Barnouw, ed., International Encyclopedia of Communications

(Oxford: Oxford University Press, 1989), vol. 2, pp. 131–4.

Ferrari 1986. Gloria Ferrari. “Eye-cup.” Revue archéologique 1 (1986), pp. 5–20.

Gimbutas 1989. Marija Gimbutas. The Language of the Goddess. San Fran- cisco: Harper & Row, 1989.

Gombrich 1973. E. H. Gombrich. “Illu- sion and Art.” In R. L. Gregory and E. H. Gombrich, eds., Illusion in Nature and Art (London: Duckworth, 1973), pp. 192–243.

Gombrich 1979. ———. The Sense of Order. Oxford: Phaidon, 1979.

Gosden 2003. Chris Gosden. Prehistory, A Very Short Introduction. Oxford: Oxford University Press, 2003.

Halle 1998. David Halle. “Material Artefacts, Symbolism, Sociologists and Archaeologists.” In Colin Ren- frew and Chris Scarre, eds., Cognition and Material Culture (Cambridge: McDonald Institute, 1998), pp. 51–9.

Halverson 1987. John Halverson. “Art for Art’s Sake in the Palaeolithic.” Cur­ rent Anthropology 28 (1987), pp. 63–89.

Heskett 1980. John Heskett. Industrial Design. London: Thames and Hud- son, 1980.

Hill 1994. James N. Hill. “Prehistoric Cognition and the Science of Archae- ology.” In Colin Renfrew and Ezra B. W. Zubrow, eds., The Ancient Mind: Elements of Cognitive Archaeology (Cambridge: Cambridge University Press, 1994), pp. 83–92.

Kirk 1949. G. S. Kirk. “Ships on Geomet- ric Vases.” Annual of the British School at Athens 44 (1949), pp. 93–153.

Li Xueqin 1993. Li Xueqin. “Liangzhu Culture and the Shang Dynasty Taotie Motif.” Whitfield 1993, pp. 56–66.

Mallowan 1947. M. E. L. Mallowan. “Excavations at Brak and Chagar Bazar.” Iraq 9 (1947), pp. 1–266.

Mallowan 1965. ———. Early Mesopo­ tamia and Iran. London: Thames and Hudson, 1965.

Maloney 1976. Clarence Maloney, ed. The Evil Eye. New York: Columbia University Press, 1976.

McEwan et al. 2001. Colin McEwan, Cristiana Barreto, and Eduardo Neves, eds. Unknown Amazon. London: Brit- ish Museum Press, 2001.

Mithen 1996. Steven Mithen. The Prehis­ tory of the Mind: The Cognitive Origins of Art, Religion and Science. London: Thames and Hudson, 1996.

Nettl 1983. Bruno Nettl. The Study of Ethnomusicology: Twenty­nine Issues and Concepts. Urbana: University of Illinois Press, 1983.

Pope 1975. Maurice Pope. The Story of Decipherment: From Egyptian Hiero­ glyphic to Linear B. London: Thames and Hudson, 1975. Rev. ed. 1999.

Spinden 1913. Herbert J. Spinden. A Study of Maya Art. Cambridge, MA: Peabody Museum, Harvard, 1913. Reprint, New York: Dover Publica- tions, 1975.

Trilling 2001. James Trilling. The Lan­ guage of Ornament. London: Thames and Hudson, 2001.

Watson 1977. William Watson. Ancient Chinese Bronzes. 2nd ed. London: Faber and Faber, 1977.

Whitfield 1993. Roderick Whitfield, ed. The Problem of Meaning in Early Chi­ nese Ritual Bronzes, Colloquies on Art & Archaeology in Asia No. 15. London: Percival David Foundation of Chinese Art, School of Oriental and African Studies, University of London, 1993.

Wilson 1976. Paul Wilson. Chrome Dreams: Automobile Styling Since 1893. Radnor, PA: Chilton Book Co., 1976.

works cited