Australia: The Land Where Time Began

A biography of the Australian continent 

Jimede 2, Western Arnhem Land – point production

In models of prehistoric Settlement and Territoriality, and economy inferred patterns of production and morphological variation have a central place. In this paper Peter Hiscock re-analyses the Jimede 2 assemblage that was excavated at Kakadu by Carmel Schrire which provides the basis for redescribing the nature of the production of points in Western Arnhem Land.

A key strategy in our attempts to recast our understanding of the prehistory of Australia has been the reanalysis of artefact assemblages for more than a decade. It has been repeatedly shown by such analyses of lithic artefacts that few technological insights were provided, while simultaneously allowing new and sophisticated models of the manufacture of artefacts and land use to be tested. There are a number of other well-known sites that have produced assemblages that have been redescribed such as Burkes Cave (Shiner et al., 2007), Ingaladdi as well as nearby sites (Cundy, 1990; Clarkson, 2002a, 2206, 2007), Puritjarra (Law, 2005, 2009), Puntutjarpa (Hiscock & Veth, 1991), Lake Mungo (Hiscock & Allen, 2000; Allen & Holdaway, 2009), Mussel Shelter (Hiscock & Attenbrow, 1998) and Capertee 3 (Hiscock & Attenbrow, 2002, 2003, 2004, 2005). The technological re-examinations of them have been fundamental in improving our comprehension of ancient tool manufacture, as the development of explanatory models of the variability and nature of prehistoric technologies in this continent these original interpretations have been important.

The basis for extensive debate over the nature, timing and causes of technological change in the region, and the construction of models, that were influential, of spatial and chronological technological change in Australia was a series of assemblages from Western Arnhem Land excavated in the 1960s and 1970s (Hiscock, 1999, 2009). Jimede 2 (also called Jimeri 2 and Tymede II elsewhere), a cave that has been excavated (Schrire, 1964-1965) was one of the key sites in the production of archaeological interpretations with regard to human occupation of Western Arnhem Land. This is a deposit spanning much of the Holocene, with the occupation beginning prior to 7,000 BP. The later assemblage was characterised as point-dominated (Schrire, 1982: 245), with her typological classifications being 38 % of the flaked implements as points and 34 % as fragments, some of which had come from broken points (N=502). It has been argued that there were 2 different types of points (Schrire, 1982: 246), each having been produced by a different manufacturing process, and that in any particular landscape setting the abundance of points indicated either a function that was distinctive to a season, or an identity that was distinctive for the occupants (White, 1967a; 1967b, 1971; White & Peterson, 1969; Schrire, 1972).

There are a number of possible explanations for the varying numbers of points that have been suggested. It was originally hypothesised (white, 1967b) that differences in the assemblages between the lowlands and the escarpment/uplands resulted from co-existence on the long term of 2 cultural groups, each of which occupied a different part of the landscape. Subsequently an alternative interpretation was offered by White, that a single group of people moved between the lowlands and the highlands according to the season, with points being used more frequently in the part of their country they occupied in the wet season, the uplands, where they occupied sites such as Jimede 2.

As the magnitude and the continuous nature of spatial variations of assemblages was recognised in later work these models were rejected: a lowland site rich in points was found by Allen & Barton (n.d.), while differences in lowland assemblages were documented (Brockwell, 1989, 1996). In the 1980s seasonal models were still proposed, though they did not offer plausible explanations for the extreme range of differences observed in lithic assemblages. An economic model of tool production in Western Arnhem Land has been proposed to explain the lithic variability of the region (Hiscock, 2009). According to this model the differences between assemblages can be explained in terms of a change in technological practices at the different landscape positions, which was a response to material cost, and a major process that produced variation was differential lithic material use, and the extending of flake tool use life by the use of additional retouching to maintain their edge. Procurement economics resulted in the effect on the abundance of points and form in Western Arnhem Land based on a number of observations and inferences.

There was a general co-variance of the abundance of quartzite and points across the landscape in a way consistent with quartzite points being transported from sources in the uplands.

The manner in which the ratio of bifacial to unifacial points varies around the landscape is consistent with reduction of points being more intensive away from the uplands, with foragers reducing the cost and inconvenience of obtaining points by maintaining the points.

Bifacial points were often converted to other forms of tool, possibly for different functions,   as maintenance of tools was extended, such as “bifacial ovals” that have been reported by Schrire (Hiscock, 2009).

Based on these observations it was hypothesised by the model that composition of assemblages across the landscape was explicable in terms of the economics of procurement: as knappers rationed, recycled and substituted artefacts as a response to the varying cost of replacement varied in each location.

According to Hiscock the benefit of this approach is that it is possible to explain much of the geographical variation among assemblages by understanding how technological behaviours are sensitive to economic contexts of the manufacture and use of artefacts, and to explain the persistence of those geographical patterns through time and the existence of temporal changes in the extent of implement reduction (Hiscock, 1999, 2009).

As foragers adjusted their technology to suit changed conditions of tool use in step with the evolution of landscape and climate over time  , but geographical differences in the maintenance and tool production cost persisted as these are largely a reflection of distances to sources of lithic raw materials that are not much changed.

In the recent model assemblage differences presented of the economics of reduction, recycling and the procurement of raw material (Hiscock, 2009) were based to a large extent on the characterisation of assemblages by earlier researchers, and including the point production description previously published for Jimede 2. There is therefore the potential for technological examination of     Jimede 2 to enhance understanding of point reduction and variation.

In this paper Hiscock presents a reanalysis of the points recovered from Jimede 2 which tests and refines models of the point manufacturing processes at Jimede 2 already existing, which has implications for the way in which models of economy and land use can be framed for Western Arnhem Land. The results of the reanalysis illustrate how the reanalysis of old assemblages is able to yield significant new information.

Jimede 2 – previous analyses

Shirere’s analysis of points from Jimede 2 was part of a long debate over the stone implements from northern Australia. The relationship between bifacially and unifacially flaked points recovered from assemblages has been debated for more than 70 years. According to one model, the divergence model, the 2 forms of points were manufactured by different methods, therefore representing the end point of 2 different manufacturing sequences. The divergence model was argued to be the best description of the point diversity at Jimede 2 (Schrire, 1982). The diversity of the points were depicted by the alternative “sequence” model as a continuum from unifacial specimens that had been subjected to limited retouch, with the result that they presented different forms of point as merely different stages in the manufacturing process. It has previously been argued by Hiscock that the best description of points recovered from Jimede 2 was the “sequence” model, basing her argument on the superimposition of scars on 48 specimens, which at this site indicated that most bifacial points from this site displayed the same order of retouching as unifacial points, with the initial retouching onto the dorsal face and ventral retouching occurring subsequently (Hiscock, 1994). Hiscock concluded that, though a range of retouching patterns were visible on the points recovered from Jimede 2, bifacial points from northern Australia were generally reduced more extensively than, and transformed from, unifacial points. Hiscock subsequently used this conclusion in the testing of models of land use in the region, using the ratio of bifacial to unifacial points as 1 measure of the reduction of points, and, therefore as an expression of geographical variation in the cost of accessing replacement material and the extent of the maintenance of points.

Hiscock’s interpretation of the variability of the assemblage was based to a large extent on the suggestion that the majority of points had manufacturing histories that were similar, that could be presented as a linear sequence, though a range of point manufacturing and recycling procedures was acknowledged. According to this image bifacial points are presented as being reduced more than unifacial points, and “bifacial ovals” as being more reduced than bifacial points, in a series that is diverse but directional. The other elements of this scenario that result from interpretations of Schrire’s classification and description, in particular the proposition that all bifacial points were reduced more than all unifacial points had not been determined by examination and examinations of the assemblage, though the idea that bifacial points were initially unifacial points has been demonstrated empirically for the points recovered from Jimede 2 (Hiscock, 1994). Consequently, the next step in developing a detailed understanding of the production of points at Jimede 2, and evaluating the economic model of variation in the assemblage is to technologically reanalyse the assemblage. Just such a reanalysis is presented in this paper.

A detailed description of the production of points has been provided for a region to the south (Clarkson, 2006, 2007), that clarifies the value and nature of further analyses of the Jimede 2 points. The analysis of Clarkson was based on an index of retouching that allowed him to evaluate the amount of reduction observed on individual specimens, independent of the retouching kinds that were employed. Several trends in the production of points were inferred by this approach (Clarkson, 2007: 102-112).

As production proceeded points became progressively smaller, varieties of bifacial points often being discarded when they were smaller than unifacial points.

As retouch proceeded, retouch expanded around the perimeter.

As reduction proceeded, the cross-sectional shape changed from wide and relatively thin to wider and relatively thicker, finally becoming lenticular and relatively thin in the later stages of bifacial reduction.

As reduction proceeded, butt trimming/thinning (the base of the point being retouched) became more frequent and pronounced, eventually leading to bases that were more curved.

As reduction continued, retouch to the ventral face was added, though in the majority of specimens the points began to be retouched on the dorsal surface only.

Only the larger points continued to be reduced to produce bifacial points, while the smaller points continued to be retouched unifacially, though the extent of their reduction was often less than that of bifacial points.

Hiscock suggests that             it is valuable to redescribe the points from Jimede 2 in such ways that the presence or absence of these trends are allowed to be established, which would thereby test his economic models concerning technology and land use.

Conclusion

A complex interaction between the size of flake blanks, the extent of retouching and the order in which retouching was applied to different surfaces. Many of the trends that were observed (Clarkson, 2007) in Wardaman country have also been established for Jimede 2, including the decrease of the dimension of points as the reduction continued, the greater reduction of larger blanks, changes to the cross-sectional shape, as well as the progression from unifacial dorsal retouch to bifacial retouch. The most significant finding is, however, the diversity of the patterns of superimposition scars that has been observed in the assemblage from Jimede 2 (Hiscock, 1994) is explicable within a single ramified production process, that began with the same retouching pattern of all specimens, though subsequently diversified as different decisions as to the location of retouch were made. There were a number of ways of working a point at moderate levels of reduction, which included unifacial, bifacial or a combination of both on different margins. This finding demonstrates that a simple sequential model according to which a unifacial succeeded by bifacial flaking is unable to account for the diversity of knapping options that are present within the production sequence. The results of this analysis have revealed that bifacial to unifacial point ratios could possibly provide inaccurate and simplistic images of the geographical variation in point reduction intensity, though the economic model that was recently formulated (Hiscock, 2009), and implication of mobility on a seasonal basis was not a factor involved in the shaping of lithic variability, is based on several measures of the reduction extent. For this reason, economic models will be tested better after detailed technological descriptions of the assemblages from western Arnhem Land, which enables studies to continue beyond typological classifications and measure directly the geographic differences in reduction and recycling that reflect the prehistoric economics of the region.

Further questions are raised by the recognition of a single ramified system of the production of points, concerning whether the different patterns of retouching are random or idiosyncratic or are a direct consequence of the characteristics of the blank, whether they are a reflection of tool preparation or maintenance with different functions of hafting procedures, and whether they display geographic variation reflecting economic factors. It remains to be explored whether similarities between factors creating point variation and those that are the basis for variation in other forms of tools. Hiscock suggests that further technological examination of Jimede 2 will help in resolving these questions, and the revision in our understanding of lithic variation in western Arnhem Land.

Sources & Further reading

1. Hiscock, Peter, 2011. Changing perspectives in Australian archaeology, part VI. Point production at Jimede 2, western Arnhem Land. Technical Reports of the Australian Museum, Online 23(6): 73–82.