Acrimonious and sometimes bitter debate has marked the evolution of the understanding of rock art in Southern Africa. The central figure in this debate since the 1980's has been David Lewis-Williams, who is presently about to retire from the Rock Art Institute at Wits University. In a stream of publications and books, starting in the early 1970's, he established himself as the leading theoretician of rock art studies in Southern Africa. Although the seeds of his argument were present in his earlier works, it was only in the mid- 1980's that he began to propound forcibly what was initially called the 'trance,' and what he now calls the 'shamanistic' hypothesis. Thus rock art is understood to be much more than, although also consisting of, images of experiences within trance, painted later by the medicine man or shaman.

In a recent paper (1998) he argues that:

"The making of San rock paintings was essentially (or principally) associated with a range of shamanistic beliefs, rituals and experiences and was situated within a tiered shamanistic cosmology and complex social relations. The images comprise symbols of supernatural potency (e.g., paintings of eland), images of trance dances, 'fragments ' of trance dances (e.g., single figures in the arms-back posture), 'processed' (recollected and formalised ) visions (e.g., the capture of a rain-animal), transformed shamans (including the so-called therianthropes ), monsters and beings encountered in the spirit world, activities performed in the spirit world (e.g., fighting off malevolent spirits of the dead) … and many images that in a range of ways, show the interdigitating of the spirit realm with the material world."

" The main focus of San art was the building up, through generations of painters, of a cumulative manifestation of the spirit world, its fear, pain, beings, creatures and power, and , importantly its pervasive infiltration of the material world."

Lewis-Williams has fought two sets of theoretical battles. The first battle was waged against the adherents of the view that art was a depiction of everyday life and experience. The second battle is still being waged against a new generation of theorists who argue that rock art reflects all San mythology, not only the trance ritual and who argue that other meanings (like the negotiation of gender within San society) are also embedded within the art.

As Lewis -Williams became more convinced of the shamanistic hypothesis he took on the proponents of the view that rock art was merely the record of everyday life of the San. This was a conflict fought through the 1980's. In his view the 'bitterness' of these exchanges was partly due to differing views of the Bushman/San - one describing the San art as childlike and simple , while his own describing a San cosmology which was sophisticated and filled with religious feeling.

However the bitterness might also have been fuelled by the sense that the professional academic archaeologists were elbowing aside the amateurs (people who were not professional archaeologists or academics, but who had nevertheless devoted their lives to the recording and writing about rock art. Lewis-Williams in fact suggested that the time of the layman was over. These amateurs were writers like Woodhouse, Battiss, Lee, and Wilcox. They had continued the tradition of Stow, one of the first recorders of South African rock art in the 1860's, who had believed that the rock art was a record of "everyday life'. And they initially resisted the suggestion that the majority of rock art was an expression of trance experience. According to Wilcox, for instance, rock art was created for the pleasure of the creation process as well as " the recording of events important to the individual or group, and to illustrate myth." While accepting that some art could arise from trance experience of medicine men, he felt that only a small proportion of the art could be explained by this hypothesis. And while believing that some art might reflect myth, the amateurs used fragments of myth in a rather haphazard way in making sense of certain images. Thus in a chapter called 'Mythology'in the book Art on the Rocks, Lee and Woodhouse wrote that when no sense can be made of images, there is "the temptation to toss it into the pigeon hole of mythology. This is what we have done with these 'animals'." Generally, the model of the amateurs was to say, 'these images can tell us something of San/Bushman life. Here we see hunters, here women with digging sticks, here San/Bushman fighting with Sotho invaders. From these paintings we can get a sense of their life.'

Lewis-Williams started from the opposite premise. His argument was that we had to go back to the ethnography (the recorded mythology, folklore, and accounts of ritual and life, given verbatim to recorders in the late 1870's.) Southern Africa was especially fortunate in that there was a rich record of material collected by W.Bleek, L.Lloyd and Orpen. This material came from the Xam Bushmen of the area north of Cape Town (collected by Bleek and Lloyd), and from the Bushmen of the Maluti mountains by Orpen. Lewis-Williams also began in the 1960's from an empiricist (proof through data collection) position. Although he later discarded this position it was extremely useful in highlighting certain issues. Prior to Vinnicombe, Pager and Lewis-Williams, who all adopted this approach in the 1960's and then early 1970's, the approach was to find the new and unusual painting. However Lewis-Williams, Pager and Vinnicombe, conducted detailed and exhaustive recording of all the paintings of certain areas. Their detailed recording of all content within their respective research areas raised interesting questions. Why was it that the almost 50% of all antelope painted in certain areas were eland? Why were wildebeest painted so infrequently when early travellers had written of the large numbers of wildebeest who had lived in these areas like the Drakensberg? Also, archeological evidence through excavations of bone remains showed that the bulk of animal food eaten by the San, was small buck and animals like tortoise, rather than eland. These statistical anomalies seemed to sugggest that the Eland had some symbolic significance in the San cosmology and that the paintings therefore probably indicated a religious or spiritual purpose, rather than a mere reflection of what was seen in daily life by the San. Lewis-Williams then turned to the 1870's ethnography of the Southern San as well as modern studies of the Kalahari bushmen. He argued that there did seem to be a pan-cognitive system (certain basic beliefs) which were held across geographical space and at least stretching back a few hundred years if not much earlier. A central piece of the puzzle of understanding was material recorded by C.Orpen in the 1870's.

The painting from Melikane that Qing commented upon in the 1870's. His comments are still being scrutinised and debated.

Orpen recorded statements of a bushman called Qing in viewing some rock paintings. Qing commented on four of these paintings, however one in particular , a painting from Medikane, is centrally important. The painting is of men with rhebuck heads. Qing commented that " they were men who had died and now lived in rivers, and were spoilt at the same time as the elands and by the dances of which you have seen paintings." Lewis-Williams brilliantly reinterpreted this statement as an extended metaphor of trance and the trance dance. Dying and living in the rivers was seen as metaphor for the experience of trance. Lewis-Williams therefore used various sources of information to come to his hypothesis. Primarily he used ethnographic material collected in the 1870's, but he also used material collected from the 1960's onwards by researchers of modern San in the Kalahari, especially their research on trance dances and trance experience. In addition he used modern psychological studies of experiences of trance. These were helpful in explaining some of the abstract geometric symbols, which Lewis-Williams argued were universal images which people had in the early stages of trance.

This hypothesis could then explain numerous details and images of the rock art:

• Nasal bleeding from numerous figures in the rock art which had never really been noticed, and if noticed never commented upon, was now seen as the nasal bleeding which literally occurs when the San medicine men go into trance.
• Images of underwater, death and flying are also understood as metaphors of trance.
• Images of halfmen/halfantelope (therianthropes) are understood as shamans undergoing transformation into animal form in states of trance. Shamans are reported as transforming into lions for instance, and going roaming through the night.
• Postures like arms stretched behind which is the actual posture that is taken up by the San shaman going into trance.
• Use of sticks, one in each hand, used to take the weight of the body when the shaman begins to feel physical pain in going into trance.
• The frequency of the eland which has a special symbolic richness in San culture and is understood as embodying a potency which is used by the shaman to go into the state of trance.

The hypothesis was taken up by numerous researchers and became the dominant model of explanation with the proponents of art as "reflection of everyday life" fighting a rather losing battle throughout the 1980's. The trance model began to be used for explaining Zimbabwe rock art (Huffman and Garlake) , the South West Cape (Manhire, Yates, and Parkington), The Free State (Ouzman), and the Cedarberg (Deacon).

However in the 1990's another shift in rock art research began to take place. Another generation of researchers emerged who were no longer just willing to apply Lewis-Williams' theory to different material.

Initially, willing to bow to the "shamanistic" explanantion before adding their additional explanation, these researchers' tone has become more strident and they are clearly less willing to submit to the final authority of Lewis-Williams. And once again the debate is becoming rather bitter and angry. Although the debate is filled with detail, all of which cannot be presented here, these are some of the central issues.

Pippa Skotnes, an artist and art historian has suggested that greater emphasis should be placed on composition, colour, and style in understanding the paintings.

Solomon and Parkington both argue that the rock art might well illustrate gender relations within San society.

Finally Anne Solomon has launched the most confrontational challenge to Lewis-Williams' thesis. Solomon uses the same sources as Lewis-Williams, the ethnography of the 1870's and the modern research on the Kalahari San. However the understanding she reaches is quite different. In her view Lewis-Williams' model relies too much on the ritual of the trance dance and not sufficient weight has been placed on the mythology of the San. She takes the famous statement made by Orpen's informant of the 1870's Qing, and reinterprets his statement concerning the painting of men with rhebok's heads. His statement was that they were " men who had died and now lived in rivers, and were spoilt at the same time as the elands, and by the dances of which you have seen paintings."

Her reinterpretation is that the 'spoiling' of the elands refers to the reference in San mythology of the second creation time when men became men, and animals became animals. Prior to this there was no real distinction. Thus the halfmen/halfanimals are ancestral San, who live in another realm, the realm of 'under the water'. Death and underwater can be seen to be equivalent. And consequently the therianthropes (halfmen/halfantelope) are understood as either the spirits of dead people, or spirits of the ancestral San, not necessarily as Lewis-Williams would have it, as expressions of animal possession in the state of trance.

In two articles published in The South African Archaeological Bulletin in 1998 and 1999, Lewis-Williams takes on his new challengers. His rebuttal tackles the criticisms raised.

Firstly he feels that he has been misunderstood in terms of the difference between shamanic and shamanistic. All art is not, as sometimes understood, merely the hallucinated visions of the San healers, (a view of rock art as shamanic), although some of it is. Rather, all the art is a reflection of a shamanistic vision of the world , which is a view of a tiered universe with access available from the material world into the spirit world through certain portals like the waterhole, or beyond the veil of the painted rock.

Thus certain images (of therianthropes for instance) could be the spirits of the dead, although Lewis-Williams convincingly argues that the fact that many of these therianthropes are associated with nasal bleeding, arms in a backward position and supporting sticks , all elements associated with the actual trance dance, suggest the greater likelihood that they reflect the transformation into animal form in trance.

Further, in Solomon's analysis of Qing's comments, she fails to adequately explain the phrase " and by the dances of which you have seen paintings". This is clearly, the trance dance to which Qing is alluding.

Lewis-Williams believes that while different meanings might be intertwined in the art, he does believe that some meanings should be 'privileged' while others are peripheral. This is in response to Solomon's criticism that one set of meanings, shamanistic beliefs, should not be privileged over others, like gender, mythology or composition for instance.

This debate has not run its course, and is likely to lead to further heated interchanges. However, for the moment, the shamanistic hypothesis retains its central position in understanding San rock art.

References:
Lewis-Williams, J.D . 1999. 'Meaning' in southern African San rock art: another impasse? South African Archaeological Bulletin 170: 141-145.
Solomon, A . 1999. Meaning, models and minds: a reply to Lewis-Williams. South African Archaeological Bulletin 169: 51-60.