Analogical Evidence and Shamanism in Archaeological Interpretation: South African and European Palaeolithic Rock Art

Abstract

Rock art studies have been strongly reliant on ethnography in recent decades. Since the 1970s, the (re)turn to ethnography has been considered short of a paradigmatic change, and it has indeed stirred a lot of theoretical discussion in the very under-theorized field of rock art research. The ethnographic turn has been mainly built around shamanism, very loosely defined here as the causal association that researchers establish between shamanic practices and rock art, and from which explanations have been sought. The application of this approach has changed through time, depending on 1) the archaeological context in which it was to be applied, 2) the use of additional sources of evidence (namely, neuropsychology), 3) the role of shamanism as a hypothesis or as an established fact. As a hypothesis it has been built on the basis of three different kinds of analogies: ethnographic, formal and uniformitarian. This paper addresses the shifting character of shamanism in South African and European Palaeolithic rock art studies, seeking to contribute at least in part to a broader reflection on the nature of analogical reasoning and its implications.

ACKNOWLEDGEMENTS

The core of this paper was written while I was a MEC/Fulbright postdoctoral scholar at UC Berkeley, 2007–2008. I thank Meg Conkey, Juan Vicent, Antonio Gilman and Leonardo García Sanjuán for their reading and comments. I warmly thank Janette Deacon and Maribel Martínez Navarrete for their clever and useful suggestions. Marcos Llobera thoroughly revised the paper and the English, improving it a great deal, for which I remain really thankful. Carmen Pérez, Teresa Rodríguez and Alex Suárez gave me a hand with pictures. Primitiva Bueno and Rodrigo de Balbín also provided bibliography. Reviewers of a previous version pointed out its weaknesses. Any mistake remains my own.

Notes

1 Shamanic practices or shamanic religion have been a widely debated topic in broader anthropological and historical contexts (see, e.g., Hamayon 1993, 1995, 1998, Kehoe 1996, 2000). Their relationship with the counter-culture movements of the 1960s has been pointed out (which, for instance, led some critics to see a link between drug use in the present and hypotheses about drug use in the past leading to the making of rock art (Meighan 1982:226)). Also the romanticized vision of the past turned out by shamanism has been criticized (Hamayon 2001:3). More problematically, it is generally assumed that there exist varying criteria behind almost every use of the concept.

2 This paper is mostly focused on shamanism as it was presented and popularized by the work of J.D. Lewis-Williams. In his important paper of 1982, however, he did not use the term shamanism, instead referring to medicine men, trancers and healers (interestingly, two of his commentators assumed the shamanistic nature of these practices and even the term). ‘Shamanism’ would not be used as a label by Lewis-Williams until several years later (J. Deacon pers. comm.).

3 Examples of which would be the idea of a membrane as an interface between two worlds and the use of hallucinogenic substances by shamans. The assumption of a separation of the sacred and profane spheres has also been represented as a Western construction (e.g. Bahn 1997, 2001) that should not be attributed to past peoples.

4 This would be due to its failure to address the issue of cultural change over time, which promotes a view of southern African rock art as static and universal (McCall 2007:225).

5 Ethnographic analogies can ultimately be either relational analogies, based on understanding the causal relationships between the variables that can be observed, or formal analogies, based on a straightforward comparison of some aspect of form or observable characteristics (Wylie 1985). But I will use the concept ‘ethnographic analogy’ to stress the source of the information employed in the analogy.

6 A middle-range theory (MRT) ‘is widely regarded as a useful means by which archaeologists can reconstruct human behaviour from a materialist and rationalist perspective. As defined by Binford … this approach postulates that, if a strong correlation can be established between specific aspects of behaviour and specific aspects of material culture in the modern (actual) world, the presence of such material culture in the archaeological record allows the archaeologist to assume that the associated behaviours also were present in the past. The bridging or warranting argument, which establishes the relevance of the modern generalization for inferring past behaviour, is that of uniformitarianism. It is assumed that the correlation applies to all situations in which human cognitive and behavioural capacities were the same as those of modern human beings. The goal of middle-range theory is not to explain human behaviour but to infer it from material remains recovered from archaeological contexts’ (Trigger 1995:450).

Or, an ‘MRT is a body of theory (or methodology) which bridges the gap between our observation of the archaeological record and our interpretation of it as a cultural system – it provides the middle ground between unreflective observation and general theory … Thus an MRT might have very general cross-cultural applicability, but its role in interpreting the past is as a bridge or methodological tool and is quite distinct from general social theory … . In themselves … middle range theories are as theoretical as any other theory and no more or less general than “general” social theory’ (Lucas 2001:184–185).

7 These representations can relatively easily be interpreted as shamans undertaking a spiritual transformation, based on the ethnographic records. But the relevance given to these depictions has been considered a bias of shamanism, because they are in a minority (Bahn 2001:76) and could also be interpreted as disguises used by hunters, as spirits or as other mythical creatures (Solomon 1998, 2001, 2008).

8 Originally, shamanic practices did not imply altered states of consciousness (Hamayon 2001, Bahn 2001). Trance became a key component of shamanism for South African rock art researchers and their particular contribution to a definition of shamanic religion.

9 ‘Since the publication of “Signs of all times”, researchers from various parts of the world have taken up the model as a tool for approaching the understanding of rock art. It is in this globalization of the model that potential dangers are found … a further potential danger of the use of the model is the tendency to treat the entoptic component as the final explanation. Once it is realized that an art depicts entoptics, some people feel that there is little more to say. “Abstract” images, like painted and engraved entoptics, do, however, have cultural values and meanings’ (Blundell 1998:9–10).

10 But see Kosso (1991:626), who argued that middle-range theory and hermeneutic approaches have a ‘common structure’ and ‘are fundamentally the same method’.

11 It has been noticed that such a focus on the religious character of the art can contribute to regarding its makers as fundamentally different (Bahn 1997:64; see also Klein et al. 2001:229).

12 It has been noticed that the turn to ethnography ‘has at least two implications: first, it mandates that interpreters (and especially the followers of shamanism) confront explicitly the rather gnarly if not longstanding issue of the nature, role, and limits of ethnographic analogy in archaeological reasoning and interpretation; and second, it asks how rock art researchers can develop interpretive frameworks that draw from the rich bodies of theory – of representation, social agency, performance, semiotics and structuration, feminist theory, and so forth – that are unlikely to be considered in (sometimes problematic) ethnographies, but may be particularly well suited to the nature of rock art as a cultural phenomenon. Furthermore, what is painfully missing or limited in many of these approaches is the archaeology; that is, the use of archaeological materials and contexts, where available, to provide additional lines of evidence into understanding the (presumably dynamic and changing) social contexts within which shamanism, ritual, and trance, for example, would have been embedded’ (Conkey 1997:170).