Did shamans always play the drum| Tracking down prehistoric shamanism in Central Asia

Historical records from different times and parts of Asia clearly show that shamanism is an archaic belief system of Asian peoples. Although shamanic beliefs and practices have been documented in different parts of Asia, many studies still consider Siberia to be the land of their most classic forms1. Regardless of the discussion about whether the term ‘most classic’ is appropriate to characterise Siberian shamanism (e.g., Rozwadowski 2008a), there is no question that shamanism in Siberia is an ancient tradition. To pinpoint its ‘birth’ or identify the people in whose culture it originated is, surely, a highly problematic question, and one can even doubt that it is resolvable at all. However, if we consider Siberian/Central Asian shamanism as a cultural tradition, I believe that the answer to the question of its antiquity should not be reduced to ’very ancient’, but should actually be investigated. In this paper, I attempt to show that such an investigation is possible and I outline several lines of enquiry into the issue of the existence of shamanism in the past. To do so, I refer to the border area of southern Siberia and Central Asia, where I carried out research on this issue for several years. However, the region to which I refer should not be considered diagnostic for the question, as its choice resulted from various coincidences. Other areas surely may appear to be equally significant for the problem, but until we have access to data from adjacent regions this case study may serve as an introduction to the issue.


Introduction
Historical records from different times and parts of Asia clearly show that shamanism is an archaic belief system of Asian peoples.Although shamanic beliefs and practices have been documented in different parts of Asia, many studies still consider Siberia to be the land of their most classic forms 1 .Regardless of the discussion about whether the term 'most classic' is appropriate to characterise Siberian shamanism (e.g., Rozwadowski 2008a), there is no question that shamanism in Siberia is an ancient tradition.To pinpoint its 'birth' or identify the people in whose culture it originated is, surely, a highly problematic question, and one can even doubt that it is resolvable at all.However, if we consider Siberian/Central Asian shamanism as a cultural tradition, I believe that the answer to the question of its antiquity should not be reduced to 'very ancient', but should actually be investigated.In this paper, I attempt to show that such an investigation is possi-ble and I outline several lines of enquiry into the issue of the existence of shamanism in the past.To do so, I refer to the border area of southern Siberia and Central Asia, where I carried out research on this issue for several years.However, the region to which I refer should not be considered diagnostic for the question, as its choice resulted from various coincidences.Other areas surely may appear to be equally significant for the problem, but until we have access to data from adjacent regions this case study may serve as an introduction to the issue.turn of the 19 th century.Rich descriptions and thorough analyses of Siberian shamanism were possible because shamanism was still practised when pioneer ethnographers entered this mysterious and forgotten land.Historically, shamanism in Siberia enjoyed a rather vigorous tradition, which was also little influenced by other cultures and religions (like by Buddhism, Lamaism, in southern parts of the region).Until the Russian conquest in the 17 th century the majority of Siberia more or less remained 'untouched'.Central Asia evidently has a more complex history.The inclusion of the southern parts of the region into the Persian Empire already in the middle of the first millennium BC introduced Zarathustrian ideas.At the turn of the Christian era, Buddhism and Christianity appeared here.By the 1 st centuries of this era, Manichaeism had developed and, since the 7 th century, Islam has spread into Central Asia, dominating the history of these lands ever since.Although this does not concern the whole area of Central Asia (the steppe remained an area of local traditions), it is clear that if shamanism existed here in ancient times it must have undergone more significant transformations than in Siberia.In fact, it has long been recognised that shamanism function here in the guise of Islam (especially Sufism) and other eastern religions.However, its antiquity remains unknown.Was it just with the influx of Turkish tribes from the north that shamanic ideas also came to Central Asia and were incorporated into the local religions?In this essay, I introduce new archaeological data, particularly on rock art, which in combination with linguistic sources suggest a significant antiquity of shamanism, or aspects of shamanic culture, in Central Asia, which predate the appearance of the Turks in the region.

Shaman, kam and baksy: antiquity through language
There is no single word for shaman across Siberia and Central Asia.Almost every culture has its own name for this spiritual practitioner, and even in a single language group the names for shaman differ regionally.This is also the case among the Turkic peoples (Alekseyev 1984), who constitute a crucial ethnic stratum of Central Asia.The most eastern branch of the Siberian Turks, the Yakuts or Sakha, call shamans oyun. 2 In southern Siberia, mainly in the Altai and Sayan region, therefore in the close proximity to Central Asia, the most common term is kam (kham/qam), from which the term for the shamanic ritual known as kamlanie is derived. 3In written form, the word kam first appeared in the Uygur Kudatku Bilik written in 1069 (Chadwick, Zhirmunsky 1969.235).Then, it is also noted in the Codex cumanicus of 1303 and in the writings of the Persian historian Rashid ad-Din Fadlallah Hamadani in 1302, with reference to the shamans of the Mongols (Laufer 1917.369).In fact, Potapov writes (1978.7)that the term kam as the term for shaman among the Yenisey Kirghiz is recorded earlier in Chinese chronicles from the Tang dynasty (618-906).Towards the close of the first millennium BC, the Chinese chronicles mention also wu, often interpreted as shamans (Boileau 2002), amongst a group of the Huns, a people closely related to the Turkish tradition (Potapov 1978.9).However, in Central Asia (Kazakhstan, Kyrgyzstan, Uzbekistan, Turkmenistan, Tajikistan), a shaman is known as baksy or bakshy/bakhsy (amongst the Kazakhs, Kirghiz, Uygurs, Uzbeks, but also among Tajiks in Afghanistan - Centlivres et al. 1971) 4 .Various etymologies of baksy have been proposed.These are usually derived from the old Turkish bak (bakmak) 'to look carefully', 'to watch', or from the Chinese word boshi (bag-∏i) meaning 'teacher', which could have come into the Turkish languages via Mongolian.It is worth mentioning, however, that an Indian genesis cannot be excluded, from the Sanskrit term bhikshu for a Buddhist monk (Żerańska-Kominek, Lebeuf 1997.39).
From a functional perspective, baksy and kam are analogous personages.The main function of a baksy is healing based upon negotiations in the spirit world, although he/she may have other abilities, such as changing the weather, prophecy and finding lost objects.Since illness is caused by a bad spirit which seizes the soul of the sick person, the core of the baksy's ritual activity concerns negotiations with the spirits.Sometimes it is not merely negotiation, but a struggle.Baksys use helping spirits to perform these ceremonies.Just as in kam's ceremonies, the baksy's healing power reaches its climax in a trance state.The following Kazakh baksy healing ceremony was recorded in the 19 th century (Levshin 1832.62-63): "The baksy takes the kobyz and begins to sing to its accompaniment.When his voice becomes louder, the movements of his body became all the more violent.He tries harder and harder; the sweat flows down his face and foam appears on his lips.He throws down his kobyz, and suddenly stands up, jumps up, shaking his head and start shouting in an affected voice, calling on the spirits, waving his hands (in this way, he calls some of the spirits to himself, and chases others away).In the end, having lost his strength, pale faced, with bloodshot eyes, he falls upon the rug, giving off wild cries, falling silent, as though dead." This account corresponds closely to the kamlanie of Siberian shamans.As in the Siberian tradition, a baksy is called by spirits to fulfil the shaman's role and the shamanic gift is usually transferred within families.From an ideological point of view, baksys also share the pan-shamanic belief in the threefold vertical structure of the world.In Kazakh shamanism, a threefold horizontal perception of the world is also discernible: this is a characteristic feature of northern shamanic ideologies, in which the river is the most typical element of 'passing over' (Kośko 2002.20).
The similarities I have pointed out may actually suggest that shamanism was brought to Central Asia by Turkish peoples spreading out from their Altaic cradle in the middle of the first millennium AD.However, taking into account that the expansion of Turks into Central Asia coincided with the expansion of Islam, and the fact that analyses of the culture of Central Asian peoples suggest that shamanism in Central Asia can be viewed as a pre-Islamic cultural feature (Basilov 1976.149-157;1992;Baylieva 1972;Jettmar 1986.289-295;Snesarev 1973.37-46;Snesarev, Basilov 1975;Sukhareva 1960.41-58),these all point to a possible conclusion about its ancient existence.Perhaps we may assume earlier than the arrival of historical Turks.
As is commonly known, the word shaman (saman) comes from the Tungus language.5However, its po-sition in this Siberian tongue is not very clear.As long ago as the 18 th century (Shirokogoroff 1935. 270), the suggestion of its Indian inspiration appeared, based on the Sanskrit word śramana.This parallel was found attractive, and it was argued that this term had passed to eastern Asia from India through Central Asia and China (the Chinese sha men is a transcription of the Pali samana) with the spread of Buddhism (Shirokogoroff 1935.276-287).Although this theory was criticised already in the mid-19 th century (Laufer 1917), recent studies have tried to verify this 'Indian path' and extend the date of possible borrowing back almost to the previous two millennia.An analysis undertaken by Sidorov (1997) allows him to suppose that it occurred at the time of the eastward movement of the early Indo-Europeans, whose main migratory waves reached India and Iran in the mid second millennium BC.Thus the sphere of Indian influence should be understood as having been far wider than previously believed -in Indo-European, or at least Indo-Iranian, terms.In fact, studies of the Indo-Iranian tradition demonstrate that it possessed practices for altering consciousness, although there is no agreement as to whether the term shamanism is appropriate to Indo-Iranian culture (and I will refer to this issue later in the text).
The antiquity of the shaman's function should not thus provoke reservations.Historical sources related to Central Asia allow us to claim that the history of shamanism in this region stretches back two thousand years, and the 'Indian path' adds another two millennia to its antiquity.However, even if the term 'shaman' appeared in Siberia 4000 years ago, it does not necessarily mean that the introduction of this term was accompanied by a prepackaged ideological and ritual context.In the following part of this essay, I try to show that archaeological sources, particularly rock art, permit us to suggest that aspects of shamanic culture could have already existed in Central Asia 4000 years ago.

Prehistory, ethnography and visions: antiquity through rock art
Identification of the shamanistic context of rock art requires not only a thorough acquaintance with the material, but also a theoretical perspective (Hoppál 1992;Rozwadowski 2009;in press).This is especially important with regard to prehistoric art, for which we have no direct indications to identify which images were connected with shamanistic practices or beliefs.For many years, the primary approach to reconstructing ancient beliefs was to think through ethnographic analogies.This reasoning has been a subject to strong criticism, particularly since the middle of the 20 th century, when structuralism questioned the usefulness of ethnographic parallels for elucidating ancient meanings.But it should be remembered that such criticism was a natural reaction to the decontextualised applications of such parallels to very ancient cultures, between which the distances of time and space were particularly large, as in the case of the Palaeolithic cave art which was viewed through the prism of Aboriginal Australians.New studies of the last three decades show, however, that ethnographic contexts can be of profound significance for elucidating the meaning and function of rock art (which often appears not to be of such great antiquity as parietal paintings in Western Europe).Studies in Africa (e.g., Lewis- Williams 2002), America (e.g., Whitley 2000) or Australia (Layton 1992) demonstrate that a fresh look at local ethnographies can open new interpretative insights into local rock art.A similar situation can be postulated in Central Asia, where rock art imagery belongs to both prehistoric and historical periods, including the quite recent past.It is significant to note that shamanic themes can easily be found, particularly in this recent rock art (of the last few centuries) in Southern Siberia (Kubarev 2002).This proves that the tradition of making images on rocks, at least to some extent and in given contexts, was somehow related to shamanism.The question, then, is if such a relation can also be identified in much older art?
The oldest expressions of the rock art in Central Asia (the rock paintings discovered in southern Uzbekistan and Tajikistan) may date to the Mesolithic, although their precise age is not clear (Rozwadowski, Huzanazrov 1999).We find no explicit shamanistic indications in these paintings.Possible shamanistic motifs appear only in the rock art of the Bronze Age (2 nd millennia BC).At this time, human figures appear more often in rock art imagery, suggesting that a group of people or individuals were perceived as holding special positions in local social contexts.Some of these images, discovered in the Altai Mountains of southern Siberia, and south-east Kazakhstan, Uzbekistan, and Kyrgyzstan, are distinguished by attributes for which a shamanistic context seems to be likely.The painted and engraved ima-ges of the 'ox-headed', 'bird-headed' or wolf-headed (Fig. 1) zoo-anthropomorphic figures discovered in stone burial chambers on the bank of the Karakol River in the Russian Altai (Kubarev 1988;2002) and other parts of southern Siberia (Esin 2009) are the most spectacular.Archaeologically dated to the early 2 nd millennium BC, they may be the oldest well-dated images of shamans in this part of Asia.These humans are associated with birds (a common metaphor for shamanic flight) or are simply transformed into birds of prey.Their bodies are covered by something that resembles well-known historical shamanic ritual costume (Fig. 2).Equally intriguing are the petroglyphs in the Chu-Ili and Karatau Mountains in Kazakhstan.Here we find human figures dressed in costumes, some of which seem to imitate the hide of a horse-like animal. 6Their heads are decorated (in some cases there can be masks) and they hold staffs, which could be hardly explained in purely utilitarian terms (Fig. 3).The staffs are often crosier-shaped.All of these attributes -headdress, dress, and staffare characteristic attributes of historical shamans.If we refer to ethnography, however, it can easily be noted that this set of attributes is lacking the most characteristic object of Siberian shamanism -the drum.
6 Shamans in modern Kazakhstan and adjacent areas do not wear special ritual costumes.Some people, however, still remember old-time baksys who wore special headgear, such as feathers, wolf heads, fox skins, or metal helmets (Arik 1999.14).

Fig. 1. Wolf-headed anthropomorphic figure; note also bird aspects in the form of feathers covering the head. Painted in red on the grave stone slab, Karakol, Altai in southern Siberia; height approximately 40cm (photo A. Rozwadowski).
In fact, the drum is commonly seen as most essential object of Siberian shamanic culture.However, it is surprising to realise that rock images of drums in Siberia are relatively rare and mostly associated with recent historic periods.When we consider strictly prehistoric art, the situation becomes complicated.
To some extent, of course, it results from the problem of dating rock images.However, more then one century of rock art research in this part of Asia (Rozwadowski, Lymer 2012) gives us tools to differentiate at least between recent images and those connected with Iron Age nomads (which are characterised by unique formal features known as 'animal style') and the earlier imagery of the Bronze Age or Neolithic (Rozwadowski 2004).On the basis of this research we can conclude that there is evident lack of images of drums in the Bronze Age and possibly also in the Iron Age rock art (see also Rozwadowski 2012b).However, even if the drum is a recent object in the tradition of Siberian shamanism, it is still important for our discussion, particularly from the perspective of its symbolism, which could be more archaic, being inherited from a 'pre-drumming' tradition.
First of all, the drum in Siberian shamanism is perceived as an animated object, precisely as an animal.The shaman's act of drumming is thus equated to riding the drum-animal.During the ritual, the drum is transformed into a horse, deer, or camel (in Altaic steppe cultures).Analogous symbolism refers to the shamanic staff.It is also symbolically perceived as an animal, being a beast on which the shaman journeys to the other world.The same can be said of shamanic costume -it imitates an animal, because the shaman is entering the world of animals.This shared symbolic code, common to different material attributes, suggests it can be considered as a core symbolic code of more ancient shamanic culture.The question of the antiquity of shamanism should then be discussed more in terms of identifying the symbolic code of shamanism, than merely seeking ancient images of contemporary shamanism (Rozwadowski 2009.243-281;2012a).Therefore, although drums are absent from prehistoric rock art, the essential idea of the human-animal/object-animal metamorphosis, as encoded in the drum, can still be recognised in prehistoric art (Rozwadowski 2009.249-275).
However, tracing shamanic themes in rock art through ethnographic analogy (which is understood here more in terms of identifying symbolic code than material objects alone) does not exhaust all the possibilities of identifying shamanism in prehistoric rock art.As is commonly accepted, trance is crucial to shamanic practice, a psycho-mental state in which the shaman is able to fulfil his or her ritual functions, be it curing or changing the weather.As has extensively been demonstrated (Siegel, West 1975), the experience of trance is very complex, but some reactions to altered states of consciousness appear to be shared cross-culturally.This does not mean that each person while in a trance will have exactly the same experiences, but that these experiences will fall within a predictable range of possibilities.From the perspective of art, the important issue in this regard involves visual and somatic hallucinations and their influence on visual depictions.7 Among visual hallucinations, two groups of images may be distinguished: those based on an individual's memory, derived from any number of phenomena tied to that individual's life; and images generated by the central nervous system.This second kind of imagery is universal and can be perceived regardless of cultural context.'Entoptic images', as these are most frequently labelled, are light patterns in geometric shapes in the forms of ovals, crosses, chains or zigzags.They appear as either individual images or may be superimposed on hallucinations evoked from personal memory (in deeper states of trance).A hallucinating person commonly attempts to interpret these geometric images, trying to make sense of them in terms of imagery in his or her memory (e.g., Lewis-Williams, Dowson 1988).
These universal hallucinatory phenomena are particularly important for our discussion.In Central Asia, there is a unique class of rock images in the form of anthropomorphic figures, presumably dated to the Bronze Age.These most spectacular and, at the same time, most enigmatic rock images in this region are characterized by rather schematic bodies, but their most exceptional feature is the elaborate forms of their heads, most often in the form of one or more concentric circles, either filled or surrounded by point-dots and short lines radiating outwards.These features, which at first glance resemble a schematic image of a radiating sun, informed the name 'sun-headed' or 'sun gods' which was attached to these petroglyphs found at different sites in Central Asia (Kadyrbayev, Mariyashev 1977;Maksimova et al. 1958;Kuzmina 1986;Martynov 1988;Martynov et al. 1992;Davis-Kimball, Martynov 1993).However, as I have demonstrated elsewhere (Rozwadowski 2001;2003;2008b), their interpretation as images of solar gods raises questions.
First, it is difficult to accept the argument that a solar cult was the predominant aspect of Bronze Age culture in Central Asia (as has commonly been argued by proponents of this interpretation, mostly be reference to Indo-Iranian mythology).Second, circular shapes do not necessarily denote the sun (and not all of the strange heads are actually sunlike forms).Instead, the circular forms of the 'heads' and the concentrations of dots packed around and inside surprisingly correspond to universally visualised oval and radial entoptic images, which have been observed during experimental studies (Fig. 4).
Interpretation of these petroglyphs in terms of graphic expressions of trance visions raises new question.Even if one agrees on this interpretation, the problem remains: should it be accepted as evidence of shamanism?The answer is not straightforward.
Trance is rightly considered crucial to shamanic practice, but should trance always point to shamanism?Should, for example, Islamic Sufi trance experiences be interpreted in terms of shamanism?These questions are pertinent to our problem.The tranceinspired rock images most probably belong to the Bronze Age, the period during which the main migrations of Indo-European peoples took place on the steppes of Eurasia; in Central Asia, these were Indo-Iranians.These peoples were characterised by a distinct class of priests, and the Vedic and Avestan texts offer rich data on their ritual life.The Vedas are dated linguistically to the turn of the 2 nd and 1 st millennia BC, but whether the picture of the culture described in these texts is also of such antiquity is not clear.We know that they appeared in written form only in the 6 th century BC (Mallory 1991.37).
We cannot be precisely sure how ritual life looked in the Bronze Age and earlier periods.One fact, however, is important in this context.It concerns the sacred ambrosia soma, a core ritual drink known both in the Vedas (soma) and Avesta (haoma), hence assumed to have a crucial role in Indo-Iranian cultures.Drinking soma enabled the priests to enter the world of the gods, become 'immortal' for a short time, and lose all the limitations associated with human nature.Soma gave them insight into a sphere of reality that was inaccessible to normal human senses.As a source of sacred knowledge, it enabled priests to ensure social, moral and cosmic order (defined in the Vedic writings as rta).There is no doubt that soma was a real concoction.
Descriptions of reactions induced by soma/haoma show clearly that it was an intoxicating potion with hallucinogenic properties (e.g., Keith 1922.104;Renou 1971.96;Wasson 1972;Stutley 1980.74;Elizarenkova 1989.445;Flattery, Schwartz 1989;Brockington 1990.14;cf. O'Flaherty 1981.119, 137) 8 .Its ingestion led to trance states, as demonstrated by the descriptions of these experiences: soma caused visual sensations of macroscopy and microscopy; it produced the impression of flight; and quite frequently a person identified himself/herself with a bird, feeling an unusual lightness and lack of any physical limits (Jurewicz 1995).Soma was conceptualised as a drink of immortality and to achieve this 'immortality' effect soma was intentionally drunk in ritual contexts.The prayers for immortality addressed to the god Soma by priests did not concern the wish to achieve such a state after death.Soma was drunk in order to "experience here and now, during one's lifetime, during the offering, a happiness similar to a post-mortal one" (Jurewicz 1995.98).
Thus it is reasonable to claim that soma as a hallucinogen played a central role in priestly rituals in Indo-Iranian cultures (i.e., the Bronze Age).The fact that in the ceremonial life of the Indo-Iranians one can also find other elements which echo shamanistic practices (Rozwadowski 2002) suggests three possibilities: we are dealing either with strong influences from archaic shamanistic Asian cultures, or the priestly function emerged from the shamanistic function of an institutionalised specialist in the field of communicating with the supernatural world, or that ceremonial life was a completely independent ritual context focused on visionary experiences.Given the current stage of knowledge, it would be risky to suggest that any of these solutions is correct.This ambiguity, however, has implications for our discussion.Assuming that 'sun-headed' petroglyphs are graphic expressions of visionary experiences, it is still difficult to say that all of them, and in every case, point to shamanism (entoptic visions could be stimulated by soma drinking -see Nicholson 2002).In some cases, however, this ambiguity may be moderated.An analysis of the 'sun-headed' images in the Tamgaly Valley in Kazakhstan can serve as an example.A thorough iconographic study of these petroglyphs shows that these images may represent shamans while in trance (Rozwadowski 2001;2008b).Several features support such an interpretation.First, in the row of small (dancing?)humans visible in the lower part of the composition with sun-headed figures are two men holding short crooks (Fig. 4).Analogous staffs one can be seen in the hands of other human images, which are dressed in horselike garb and associated with horses (see Fig. 3).Furthermore, some of these petroglyphs are evidently integrated with natural cracks in rocks (Fig. 3).If this integration was not accidental, at least in some examples, it had to be meaningful.I believe that popular beliefs in cracks as entrances to the spirit world, as noted in the tradition of local shamanism (Sagalayev 1992.55-56, 60-78;Tátar 1996;Rozwadowski 2012b.461-465),provide a reasonable explanation for this iconographic motif.It could then be seen as an additional argument in favor of shamanic overtones of the 'sun-headed' petroglyphs in Tamgaly.Some sun-headed petroglyphs can thus be explained through shamanism.Whether the same could be said about other similar petroglyphs at other sites in Central Asia is a question to be investigated.Simila- rity in form should not automatically point to semantic homogeneity; although it cannot be excluded, it needs to be demonstrated.
Making final conclusion concerning the issue of antiquity of shamanism in Central Asian/south Siberian border zone through the prism of rock art, we can note that both classes of images (those identified with shamanism through ethnographic parallels and those identified with visionary experiences) probably date to the Bronze Age.It is thus reasonable to believe that forms of shamanic symbolic code as well as practices which focused on visionary experiences already existed four thousand years ago in Central Asia and southern Siberia.This does not, of course, necessarily mean that shamanism originated at this time.We merely note that, at this juncture, reliable sources support the existence of shamans or aspects of visionary practice about four thousand years ago.

Fig. 2 .
Fig. 2. Human figure with animalistic features.Note bird-like claws instead of feet, protrusions from the body which strongly resemble historical shamans' coats, and the bird flying under the left arm of the human.Engraving on a grave slab at Karakol, Altai; height approximately 50cm (afterKubarev 1988.Fig.33).

Fig. 4 .
Fig. 4. Fragment of a large panel with petroglyphs in Tamgaly.The shapes of the heads of these anthropomorphic figures (height approximately 2 metres) resemble visionary patterns seen in trance states -A (afterHorowitz 1975.Fig.2).Copy by author.