Senior Rock Art Custodian Monty Hale at huge eastern Pilbara petroglyph site, with IFRAO President R. G. Bednarik

For ethnographic art, at least some levels of meaning remain accessible, and justify speculation about the function of such arts in the societies that produced them. In some rare cases one may be able to interview a producer or consumer of the art, in others there are established cultural traditions maintaining a level of knowledge about the art that permits some access to its meaning.

However, the fallibility of ethnographic constructs, the problems of reconciling the indigenous and ‘scientific’ (alien) conceptualisations of meaning, and the significant question of whether ethnographic interpretations of meaning can be relevant in the ‘text-free’ record of the distant past all need to be considered. All humans, including ethnographers, lack the ability to think outside of their severely limited intellectual and cognitive universe. Huchet (1991, 1992) has examined numerous expressed or implied ethnographic analogies in archaeological specialist literature, and has found that there are hardly any instances of rigorous logic in the examples he considered. This, admittedly, refers to inadequate rigour, and does not negate the principle itself. However, at the more fundamental level, one can question the entire underlying logic. Relational analogy involves a demonstration of causal mechanisms or other factors in the similarities between source and subject that determine the presence and interrelationships of perceived properties (Wylie 1985). In practice, this is hardly possible, and analogic propositions must not be based on a simplistic acceptance of inferred properties in both the source and subject data. That record is itself never totally reliable, it has been filtered by a variety of cultural and cognitive biases.

Archaeologists have tried in a number of ways to deal with the problems of using ethnographic analogy, but there is no consensus of how it would be achieved scientifically. The leading protagonist of the so-called New Archaeology, Binford (1967), suggests that documented analogies should form a basis for postulates as to the relationship between ‘archaeological forms’ and their behavioural context in the past. Deductive hypotheses drawn from such postulates would then be used to test the postulates. But there are numerous pitfalls in this kind of procedure.

First, there is the nature of the archaeological raw data itself. When we consider the amount of reliable dating we have for rock art currently, it seems quite inappropriate to pretend that we can speak of specific rock art traditions and relate them to archaeological entities. Secure age estimates are unavailable for nearly all rock art of the world. Most age claims in the literature are merely statements about degree of consensus that a certain perceived stylistic latitude within an art corpus refers to a particular technological pigeonhole in archaeological time. This does not constitute evidence, it is opinion. Perceived styles are not real styles (Conkey and Hastorf 1991), they are only what untutored alien ‘researchers’ lump together for the sake of creating order in disorder, in accordance with their own conditioned ways of experiencing reality. The mere fact that they seem unable to agree among themselves as to how to define styles demonstrates that these are not intrinsic properties. Styles perceived by archaeologists, be they of rock paintings or stone artefacts or any other archaeologically perceived object classes, are never more than styles perceived by archaeologists; they are ‘egofacts’ (Consens 2000). Some may well be valid, that is perfectly possible, but to prove this in a scientific fashion would be extremely difficult and would require a great deal more work than is being invested. Studies must be conducted without assuming a validity that has not as yet been demonstrated with any semblance of scientific method (Bednarik 1991a).

There are greater problems still with archaeological information. Not having been qualified taphonomically, such data prompt greatly distorted perceived realities. While the archaeologist has always made some allowance for these distortions by simple common sense, without systematic correction one must expect that any interpretation of the record will be substantially distorted (Bednarik 1992a, 1993a, 1994a).


The ethnographic record: what is it?

Much of the ethnographic record has not been collected under ideal conditions, but, just for the sake of the argument, we shall assume that it was all secured under the best of conditions possible. So we assume that alien researchers, who interviewed indigenous people, recorded the information offered accurately. It would seem that this record, then, should be objective. But is it really?

The communication between informant and recorder is always by means of translation. Even where the interviewer speaks the language of the people being studied, he or she is usually not very proficient in their language. He certainly has little or no linguistic access to those aspects of the culture that are avoided, or indeed taboo. His interpretation of what he does have access to will be affected by inadequate understanding of these inherent limitations. In many cases he uses a third party, an interpreter, and what he obtains is quite literally a biased interpretation, and not a rigorous account. He then interprets this interpretation in a way that makes sense in his own linguistic and cognitive framework. Moreover, it is well known today that extant traditional cultures do not permit outsiders access to all aspects of their metaphysical world. For instance, students of the oldest surviving culture on earth, that of the Australian Aborigines, have found that the knowledge bestowed on them is intentionally limited in several directions. Not only is it incomplete in the sexual sense, because of the strict gender divisions in cultural knowledge (there are restrictions according to the gender of both informant and interviewer), but also there are explanations within explanations, in the fashion of Russian dolls: upon opening one, there is always another one inside it. Thus interviewers, including myself, have found that the same informant used a different, more elaborate explanation for a phenomenon many years after he had given a simpler one to the same interviewer (‘But 25 years ago you didn’t know much!’). Or, as a senior Aboriginal custodian once remarked: ‘It took me 60 years to become an Aboriginal, so how could a young balanda (European) straight from university learn much in a matter of a few months or years!’.

Very simply, the explanations given to ethnographers are commensurate with a researcher’s perceived competence and trustworthiness. This applies in any part of the world. For instance, a rock art motif may have many meanings, beginning from a very simple level. This is rather like an explanation a contemporary urban parent would give to a small child. Once it had become older, a more advanced explanation is considered appropriate, and so on. In indigenous or tribal societies, a great deal of knowledge is of restricted access. It may be of a secret or a sacred nature, and there may be levels of severity. In many such societies, serious breaches of sacred matters are traditionally punishable by death, and may be considered more serious crimes than murder, for instance. It is therefore inconceivable that information at the level of sacred knowledge would be passed on to uninitiated alien researchers, simply to satisfy their curiosity. Hence we can be certain that all the published ethnographic evidence of such metaphysical knowledge of tribal people is of the type given to people of inadequate understanding of the society in question. It should be obvious that this mechanism would have contributed to a simplification of ethnographic accounts: not only did the informants regularly observe the restrictions of tribal laws, they would have often felt obliged to simplify interpretations for untutored outsiders. At times they may have been forced to deceive their questioners in order to protect sacred knowledge. Ethnographers, often naively unaware of such factors as multiple meanings, may base their professional reputation and standing on their ‘findings’, and they may not be too willing to admit these severe limitations of their accounts. But that is understandable and we have to make allowances for such limitations.

The separation of religious and profane matters is not as clear in indigenous societies as it is in urban industrial societies. Since indigenous societies possess various levels of restrictions on metaphysical knowledge, these restrictions also affect the nature of the information that can be provided to uninitiated outsiders, such as researchers: one cannot explain to the interviewer every aspect of what appears to be, on the face of it, an economic and thus presumably profane aspect of the culture, because some of the meanings involved in a comprehensive explanation involve information of a secret nature. Hence even under the best possible conditions, the researcher will only obtain fragmentary, simplistic and even misleading explanations.

None of this is to negate that occasionally a non-indigenous researcher, through life-long endeavour, does reach a high level of understanding about the indigenous culture he partakes of. However, by the time he reaches such a state, he will have lost his desire to disseminate his knowledge ‘academically’, having learned that academic enquiry into sacred aspects of a society leads to its destruction — and having realised that he can celebrate his level of understanding without seeking the approval of his peers, through indiscriminate dissemination. Such a scholar, having risen above the need to prove himself to his peers, has no desire to inflict damage on the fabric of the culture of his teachers. Once again, knowledge ‘of the highest degree’, as it is regarded in Australia, will not become available for academic consumption.

Now we come to the main problem. In analogical interpretations of rock art we rely on two types of record: the source of the analogy (ethnographically ‘determined’ ‘meanings’, such as those just discussed) and the subject of the analogy (the empirical data acquired about a corpus of rock art). Archaeological data are severely distorted by many taphonomic processes and it is clear that the discipline has not as yet found systematic means of addressing this issue. A remedy is possible and some measures have been suggested (Bednarik 1994a), but this requires a sophisticated epistemology not realisable in the immediate future. The discipline is not capable of absorbing the required changes in a short time. We also know that most rock art is undated, or has only been attributed to some archaeological pigeonhole on the basis of inadequate assessment. Again, this is a shortcoming that can be addressed by improved research methodology and a great deal of hard work. Now we are trying to relate these biased and distorted archaeological data to the typically incomplete ethnographic interpretations that may be no more appropriate than children’s stories. We would be naive to expect much of such a procedure, which cannot be proposed or supported by anyone who has seriously considered the issues. Yet this is precisely what ‘archaeo-ethnographic interpretation’ of rock art is.

Nevertheless, there are of course many instances where ethnographic information has been used profitably in rock art research. This is particularly the case when it demonstrates the futility of seeking to explain meanings in the complete absence of ethnography. A good example is the work of Michaelis (1981), who followed up Titiev’s (1937) account of a journey of Pueblo Indians to collect salt. Titiev reports the practice of placing clan signs at the Willowsprings petroglyph site, and Michaelis found that nearly half of the known clan emblems of the Hopi occurred at that site. If the role of the petroglyphs had not been known ethnographically, researchers would have no means whatsoever of determining the meaning of the very diverse but also repetitive motifs. Similarly, if we were intergalactic visitors and found many rock art motifs with Christian crosses placed over them, we could not explain them. If we did discover the role of the cross, we would probably presume that it stands for death or resurrection. Its real role, iconoclasm, can only be known to us because we have some knowledge of the historical circumstances. It should be clear enough that for most rock art traditions, we certainly lack the necessary understanding of context, culture and ideology. In fact, most rock art cannot even be attributed to a specific cultural tradition at this stage.

There is one alternative way to determine the ethnographic meaning of rock art, but so far it has not been applied in any systematic fashion. In certain parts of the world, rock inscriptions occur together with rock art, and in many cases these can be directly related to the rock art: for instance by their spatial arrangement, by their content, or by colorimetric age estimation. When these rock inscriptions can be read or deciphered they can provide sound ethnographic information about the purpose, meaning or context of the rock art in question. This potential remains to be explored by rock art scientists, with the help of epigraphy.

R. G. Bednarik, 2001