Rock Shelters as Indicators of Patterns of Mobility in the Inland Pilbara

Australia: The Land Where Time Began

Rock Shelters as Indicators of Patterns of Mobility in the Inland Pilbara

Consulting archaeologists in the Pilbara often use a land-use model depicting rock shelters as ephemeral and atypical. In this paper Bird & Rhoads report a study in which they evaluate assemblages that have been excavated from rock shelters in the Chichester Range, comparing them against characteristics that have been predicted for sites that were used by small highly mobile groups on a short-term basis. It is shown by the results that rock shelters do not clearly conform to the model. When assemblages from selected shelters were compared with neighbouring surface sites rock shelters have proved to likely be among the most complex and diverse known sites in the study area.

The proliferation of early dates make the Pilbara significant to the discussion of the issues of the peopling of the continent, the occupation of the arid zone, and mark it as a possible refugium during the Last Glacial Maximum (LGM). As a result of commercial consulting archaeology there has been little attempt to synthesise this large body of primary data. Most of the data is still to be published in a vast array of consulting reports. Therefore, consultants work mostly within a generalised regional research agenda rather narrowly, and necessarily, focusing on issues of antiquity and the distribution of the sites.

Rock shelters that have been excavated provide most of the published archaeological evidence from the Pilbara. However, surface sites dominate most of the regional archaeological record, as is the case for much of Australia, ranging from small scatters to extensive sites that contain thousands of artefacts. The ethnographic evidence of the use of rock shelters is sparse and anecdotal, what there is suggests rock shelters were not generally the camping places that were preferred, being used mainly for shelter during bad weather or for storing food and personal equipment (Brown, 1987: 17; Clarke, 1983; cf. Veth, 1993: 77). As a result it has been argued (Ryan & Morse, 2009) that rock shelters distort the interpretation of the archaeological landscape of the Pilbara as they are atypical.

According to Bird & Rhoads it is certainly true that rock shelters and caves are often focused on because they produce evidence that surface sites cannot. It has long been a challenge for archaeologists to find an analytical framework for the surface archaeological record (Holdaway et al, 1998; Rossignol & Wandsnider, 1992). There is also a problem of integrating the records of rock shelters and surface sites that formed over a wide diversity of timescales, and the Pilbara is not an exception. Here, a model of land use for classifying sites in terms of the size and mobility of groups that was originally formulated for the Western Desert is often used to interpret surface sites (Ryan & Morse, 2009; Veth, 1993). Bird & Rhoads say the model has been evaluated very much, particularly for its application outside the area it was originally designed for.

In this paper Bird & Rhoads argue that if rock shelters are indeed ephemeral sites that were used only rarely for short periods by small groups, then a test of the land-use model for the Pilbara is provided by the artefact assemblages in them. After a discussion of the archaeological correlates of the model, Bird & Rhoads draw on data from excavated rock shelters and surface sites in the Chichester Range to test if the assemblages from the rock shelters conform to the expectation that they are the result of ephemeral occupation within the context of the regional archaeological record.

Conclusion

Bird & Rhoads began this paper by testing the suggestion that rock shelters in the Pilbara were ephemeral sites. The land-use model, which is widely used by consulting archaeologists in the region, sets out the characteristics that are expected for each site. These were applied to rock shelters from the Cloudbreak-Christmas Creek study area that had been dated to the Holocene. The analysis by Bird & Rhoads suggests that rock shelters should not be regarded as a single uniform site type. They showed considerable variability in the Cloudbreak-Christmas Creek study area. It was found that most of the assemblages that had been excavated that are discussed here do not clearly conform to the criteria that are predicted for ephemeral sites. In terms of complexity and diversity, at least some assemblages from rock shelters appear to be comparable with nearby surface scatters. Bird & Rhoads suggest it is likely that throughout the Pilbara region it is that rock shelters will display comparable variation.

There is a long tradition among archaeologists of borrowing explanatory models from each other and even other disciplines. However, it is important to carefully examine their underlying assumptions, and the degree to which they are applicable outside the original context. It is suggested by this analysis that, though a useful starting point is provided by desert land-use models, it is possibly they may not be transferrable to the Pilbara, as it has diverse topography and major river systems.

Though rock shelters may not provide a complete picture of life in the past, they are still important to archaeology for very good reasons (Frankel, 1991: 57). Whether rock shelters or surface scatters are more representative of the archaeological record of the Pilbara is not, according to Bird & Rhoads, the issue. When building regional interpretive frameworks, the integration of different strands of evidence at very different temporal and spatial scales is critical. Bird & Rhoads say it is essential to understand rock shelters, surface scatters as well as other types of material evidence in terms of their relationships within the natural and cultural landscape.

Sources & Further reading