A visit to the Cave of Lascaux in southwest France is an arresting experience. Colourful and often realistically posed images of bulls, deer, horses and many other animals cover large areas of Lascaux’s extensive passageways and chambers, making it by far the most prolifically decorated of all known caves from the Upper Palaeolithic, the period between 35 000 and 10 000 years ago. The style of these powerful images looks strongly familiar to Western eyes, so much so that the late French archaeologist Andre Leroi-Gourhan suggested that they represent ‘the origins of Western art’, painted as they were some 17 000 years ago, at the height of the Magdalenian period.
Yet this is clearly not the case. Representational painting and engraving all but disappeared at the end of the Magdalenian, in the so-called Azilian era. It was replaced by schematic images and patterns, a different mode of expression. Many of the techniques that had been applied in Lascaux, such as perspective and a sense of movement, had to be reinvented in Western art, during the Renaissance.
A closer look at the walls of Lascaux reveals the geometric patterns, such as grids, lines of dots and nested curves, scattered among the animal images and in some cases superimposed on them. Such ‘signs’ are not a normal accompaniment of most Western art, but are a constant component of Upper Palaeolithic art. They are also one its greatest enigmas. For the late Abbe Breuil, France’s greatest prehistorian, these geometric patterns represented hunting paraphernalia, including traps, snares, even weapons. This fitted in with his interpretation of the art as ‘hunting magic’. Leroi-Gourhan included them in his notion of a structural duality in the Upper Palaeolithic art. Dots and strokes were male signs, he said, while ovals, triangles and quadrangles were females signs.
Advertisement
Recently a South African archaeologist, David Lewis-Williams, has suggested that neither interpretation is correct. They are, he says, images plucked from a mind in the state of hallucination, a sure sign of shamanistic (medicine man) art. His argument is based on a study of San (Bushman) art, in southern Africa, and on a neuropsychological model that might be basic to much human image-making in hunter-gatherer societies, including those of the Upper Palaeolithic.
Lewis-Williams developed the neuropsychological model with his colleague Thomas Dowson. The shamanistic interpretation goes too far for some scholars. Lewis-Williams’s thesis ‘is very important in many ways’, says Margaret Conkey, an archaeologist at the University of California, Berkeley. ‘But I’d shy away from any kind of monolithic explanation, anything that claims to explain everything.’
Nevertheless, Lewis-Williams and Dowson’s thesis is receiving a cautious welcome in continental Europe. At a major symposium held last year in France to commemorate the 50th anniversary of the discovery of Lascaux, Lewis-Williams was the only speaker invited who did not come from France or Spain. ‘We are taking what he says very seriously,’ says Jean-Philippe Rigaud, director of Prehistoric Antiquities of the Aquitaine, Bordeaux.
Lewis-Williams’s thesis is as surprising as it is seductive. Prehistoric rock art is relatively common in southern Africa, where it goes back as far as 27 000 years, a full 10 millennia before Lascaux, at the Appollo-11 cave in Namibia. In many ways, the African art is much more schematic and contains more human figures than the European art, but it includes the same kinds of geometric signs. San art is a recent part of this body of images in southern Africa, and for a long time it was considered to be principally narrative and decorative, with a small element of mythological depiction.
Lewis-Williams’s interest in the interpretation of southern African prehistoric art goes back several decades, and he remembers some of the abortive approaches to trying to understand the art. ‘We were pretty naive,’ he now laments, recalling an idea from the late 1960s that quantification was the answer. ‘We thought that all we had to do was count various features of the paintings, and some sort of pattern would emerge and explain itself to us.’ Lewis-Williams and his colleagues had not realised that it was very difficult to know exactly what was significant enough to be counted and what could be ignored. ‘We discovered later on, when we learned what the art was really about, that many of the key elements in the paintings simply weren’t on our lists,’ he says. ‘Things like the bending forward posture, or the arms-back posture. It didn’t occur to us that these were important, but now we know better.’
Making no progress with the quantification approach, Lewis-Williams and his colleague Patricia Vinnicombe decided to look to ethnography for inspiration, specifically to the traditional stories of the Kalahari San. The last of the images here were painted about 120 years ago, at the fringes of recent history. Although the Kalahari San were 1500 kilo-metres from the sites of the prehistoric images they had been studying, Lewis-Williams and Vinnicombe hoped to draw some insight from their art.
Prominent among San images was the eland, just as it was over large parts, but not all, of southern Africa. In what way was the eland important? Immersion in Kalahari ethnography and anthropological theory suggested that the eland was a symbol with multiple meanings. The eland turned up in all kinds of contexts, in the girl’s first puberty ritual, the boy’s first kill, in marriage ritual, in rainmaking, in healing rituals, everywhere. But Lewis-Williams made his first real discovery as he assembled his dissertation, eventually published as a book, Believing and Seeing.
‘When I was doing the book, I could find only a few pictures of the girl’s puberty dance, only a few of the boy’s first kill, the marriage ceremony, depictions of events in people’s lives,’ Lewis-Williams explains. ‘But when I got to the shamanistic chapters, rainmaking and so on, there were more pictures than I could accommodate. It was one of those cases where the data challenged the theory.’ The art was not principally narrative and decorative, as people had supposed. It was dominated by shamanism. But the next step of discovery was decisive, for it set in train a most unlikely but highly productive line of investigation.
‘I used to take copies of the pictures home with me, prop them up on the mantlepiece, and contemplate them for days, hoping I’d see something in them that would give me a clue,’ remembers Lewis-Williams. ‘Then, one day, I was looking at a picture in which there was a dying eland and a man apparently holding its tail. The man had hoofs, like the eland; his hair was standing out, like the eland’s hair; his legs were crossed, in imitation of the eland’s legs. I suddenly saw that the dying eland was a metaphor for the dying medicine man. Shamans are said to ‘die’ when they enter the spirit world through trance. And the dying eland is a source of potency. The penny dropped, and I thought, how could I have been so dumb not to see it before?’
Not only were the images produced in a shamanistic context, but they were its very essence. They were not meant to depict the shaman carrying out rainmaking, healing or whatever his speciality was. They were a product of the trance that the shaman enters to perform these functions.
Shamans may induce a trance by various techniques, including drugs and hyperventilation, but it is almost always in the context of rhythmic singing, dancing and clapping by groups of women. As trance deepens, the shamans begin to tremble, their arms and whole body vigorously vibrating. They may also bend over as if in pain, and need the support of a staff. They are ‘dying’, visiting the spirit world beyond this world. They are also hallucinating. Later, the shamans of the past painted the images.
It was the workings of the mind during this altered state of consciousness, this hallucination period, that intrigued Lewis-Williams. So he turned to the neuropsychological literature. There he learnt of the stages of trance, and of the nature of hallucination that the trance produces. Research stretching back over half a century had identified the occurrence of so-called entoptic phenomena, shimmering, incandescent, moving patterns produced within the visual system in early trance states. These entoptic images take geometric forms, such as grids, parallel lines, zigzags, dots, spirals, nested sets of curves and filigrees.
In deeper trance states, these geometric images may be construed as real images, depending on the state of mind and cultural background of the individual involved. As one researcher put it: ‘Thus the same ambiguous round shape on initial perceptual representation can be ‘illusioned’ into an orange (if the subject is hungry), a breast (if he is in a state of heightened sexual drive), a cup of water (if he is thirsty), or an anarchist’s bomb (if he is hostile or fearful).’ The San commonly see nested curves as a honeycomb, some-thing that is extremely important to these people as honey is a great delicacy.
Moving into a yet deeper stage of trance is often accompanied, according to laboratory reports, by an experience of a vortex or rotating tunnel that seems to surround the subject. The external world is progressively excluded and the inner world grows more florid. Iconic images may appear on the walls of the vortex, often imposed on a lattice of squares, like television screens. Frequently there is a mixture of iconic and geometric forms. Experienced shamans are able to plunge rapidly into deep trance, where they manipulate the imagery according to the needs of the situation. Their experience of it, however, is of a world they have come briefly to inhabit; not a world of their own making, but a spirit world they are privileged to visit.
Hallucination in this deepest of the trance stages can become quite bizarre, with human and animal forms combining, forming chimeras or part animal, part human therianthropes. One early experimenter reported the transformation of a human head into a cat’s head. Another described his experience this way: ‘I thought I was a fox, and instantly I was transformed into that animal. I could distinctly feel myself a fox, could see my long ears and bushy tail, and by a sort of introversion felt that my complete anatomy was that of a fox.’
From these and other details of the neuropsychological literature, Lewis-Williams and Dowson developed what they call their three-stage model, which corresponds to the three, increasingly deep, trance stages. The model included an identification of specific entoptic images and of their perception and manipulation. The model was meant to predict what images would look like, in a general sense, if indeed they had been generated as the result of a trance state. The first thing that was necessary was to test the model against known shamanistic art.
‘It had seemed apparent that San art had indeed been shamanistic art, and we were able to support this idea when we interviewed the old lady in the Transkei,’ says Lewis-Williams. ‘She was the daughter of a shaman, and she told us how she had seen the images done, how people used the images, turning to face them, touching them to receive potency. It all seemed to us to be about as firm a confirmation as we could hope for that San art was shamanistic.’
It was important that Lewis-Williams and Dowson could test their model against known shamanistic art, otherwise they might have found themselves trapped in circular arguments. The presence in San art of the six specified entoptic forms, the mix of geometric and iconic images, and the importance of therianthropes are strongly indicative of shamanism, say the South African researchers: ‘We take (it) . . . to be an initial confirmation of the validity of our model.’
What of Upper Palaeolithic art? Interpretation of its ‘meaning,’ a highly culturally bound concept, has moved through two major stages and is in the infancy of its third. The first stage, championed by the Abbe Breuil, was that of sympathetic hunting magic, beginning early this century and continuing until the early 1960s, when the Abbe died. The second stage focused on how images of different types of animals are distributed in different parts of the caves, and the association of images with one limb or the other of a sexual duality. This was the view of Leroi-Gourhan and Arlette Laming-Emperaire. By the time Leroi-Gourhan died in 1986, his theory had already begun to weaken. No single authority has since emerged to dominate the field, and no single explanation dominates. The very diversity of interpretations offered today suggests that the images themselves may have had many different meanings for the people of the Upper Palaeolithic. It was into this new intellectual setting that Lewis-Williams and Dowson ventured when they decided to apply their neuropsychological model to Upper Palaeolithic art.
‘The neuropsychological model orders and fits Upper Palaeolithic art as well as it does San art,’ says Lewis-Williams. ‘The ‘fit’ is by no means simple; it is in fact highly complex, and this increases confidence in the conclusion that it does not result from chance.’ For instance, all six entoptic forms are to be found in Upper Palaeolithic art, although this does not account for all geometric forms in the caves. The presence of the forms associated with altered states of consciousness was, however, an encouraging sign, an apparent confirmation of stage one hallucination and associated images.
Stage two, the construal of entoptic forms, proved more difficult to identify. Nevertheless, Lewis-Williams and Dowson suggest that the exaggerated curved ibex horns with a zigzag outer margin in the Pyrenean cave of Niaux might qualify. So, too, might similarly exaggerated horns of ibex on a stone lamp from the cave of La Mouthe in the Dordogne. The South African researchers also cite paintings and engravings of mammoths in Rouffignac, also in the Dordogne, in which tusks of various animals appear as nested curves. These, and other, suggestions are tentative, admit Lewis-Williams and Dowson, but they ‘expect examples as convincing as the San ones to come to light’.
Stage three, the most complex of all, provides stronger evidence. Looking for combinations of iconic and geometric images, Lewis-Williams and Dowson cite the famous horses of Peche Merle, in which red and black dots were placed within the black outlines of these animals. With the more enigmatic images of therianthropes, there are strong parallels between San and Upper Palaeolithic art. Often in the past, such images had been interpreted as hunters or shamans wearing masks. This does not explain the common occurrence of therianthropes with hoofs as feet. It was once suggested that therianthropes were the product of a ‘primitive mentality’ that had ‘failed to establish definitive bounda-ries between humans and animals’.
Lewis-Williams and Dow-son acknowledge that during certain rituals shamans of the Stone Age may have used animal adornments, but argue that the overall nature of therianthropic images is strongly consistent with stage three trance. This holds good for paintings and engravings in Upper Palaeolithic art, as it does for San art.
Therianthropes represent a small but arresting proportion of Upper Palaeolithic images. The most famous example is the so-called sorcerer in the cave of Les Trois Freres, in the French Pyrenees. Deep underground, in a cramped cavern, the sorcerer dominates the space. Denis Vialou, of the Institute of Human Palaeontology in Paris, has studied the cave in detail, and describes the image like this: ‘The body is uncertain, but is some kind of large animal. The hind legs are human, until above the knees. The tail is some kind of canid, a wolf or a fox. The front legs are abnormal, with human-like hands. The face is a bird’s face, odd, with deer’s antlers.’ Unusually for Upper Palaeolithic images, the sorcerer is staring directly out of the wall, a full-face stare that transfixes the spectator.
Below the sorcerer are several heavily engraved panels, a riot of animal figures with no apparent order, no pattern. In the midst of all this is another human/animal figure, again with human hind legs. Human hind legs on animals are common in Upper Palaeolithic art, as are hoofs on otherwise human figures. This therianthrope is standing upright, with a bison’s body and the head of a bison, with horns but a somewhat human face. The front legs are odd, in the same way as the sorcerer’s forelimbs. This individual is holding something that might be a bow or a musical instrument. ‘Directly in front of this image is an animal,’ explains Vialou. ‘It has reindeer hind legs and rear end, showing female sex prominently displayed, the only one known in Upper Palaeolithic art. The rest of the body is bison, the head turned, looking back over its shoulder at the first individual. Something special is going on between these individuals, I’m sure of that.’
Something similar is to be seen in Lascaux. The very first beast in the stampede in the Hall of Bulls is an enigma. Known as the Unicorn – wrongly, because it has two very straight horns – this beast has a swollen body on thick limbs, and a head of no known animal. There are six circular markings on the body, and the partial outline of a horse. Look at the head again, squint, and the profile snaps into that of a bearded man. It is a curious image, one that Lewis-Williams and Dowson believe fits very well into the kind of therianthropes hallucinated during stage three trance.
In arguing that Upper Palaeolithic art conforms to the three-stage neuropsycho-logical model, and is therefore shamanistic, the South African researchers do not claim to have explained the meaning of the images. ‘Meaning is always culturally bound,’ say Lewis-Williams and Dowson. ‘What we are pointing to are neurological mechanisms that underlie shamanistic art, wherever it is produced. How people construe and manipulate entoptic images, and what kind of iconic images they depict, all this will be influenced by the cultural context.’ It is like working with a given palate of paints, from which any desired image may be constructed. The meaning of the images remains elusive, but knowing the art is shamanistic – if indeed it is – at least offers a more secure foundation from which to analyse it.
If Upper Palaeolithic art is shamanistic then, Lewis-Williams and Dowson suggest, there might be a clue to the origins of image making, the origin of the notion that two-dimensional lines can represent three-dimensional objects. Prehistorians and psychologists have speculated on this issue for decades, wondering about the mental processes that would be required to provide the key innovation. One idea, recently revised, is that the production of representational images grew out of nonrepresentational marks, that it was a progressive process, a maturing of cognitive abilities and insights. The suggestion that the art is shamanistic makes this unnecessary. Image-making, suggests Lewis-Williams, might have derived directly from the experience of hallucination itself.
For instance, some of the studies on hallucinations report that the images, both geometric and iconic, often appear to exist on surfaces, as if projected onto the wall or ceiling. ‘Pictures painted before your imagination,’ commented one observer. In a more naturalistic context, shamans often perceive their hallucinations as if they were on rock surfaces. This is natural, because the rock surface represents an interface between the real world and the spirit world, a passageway between the two. ‘They see the images as having been put there by the spirits, and in painting them, the shamans say they are simply touching and marking what already exists,’ explains Lewis-Williams. ‘The first depictions were therefore not representational images in the way you or I think of them. They were fixed mental images of another world.’