119 COSTS AND BENEFITS IN TECHNOLOGICAL DECISION MAKING UNDER VARIABLE CONDITIONS: EXAMPLES FROM THE LATE PLEISTOCENE IN SOUTHERN AFRICA Alex MACKAY School of Archaeology and Anthropology, The Australian National University Ben MARWICK Department of Anthropology, University of Washington Abstract: The issue of technological time costs as applied to the manufacture of flaked stone artefacts is considered. Assuming a positive correlation exists between technological cost and improvements in resource capture, it is shown that the viability of costly technologies is constrained by the abundance of resources in a landscape such that more costly technologies would be likely to be pursued in resource-poor landscapes. This outcome mirrors the results of past assessments of ethnographic data concerning the relationship between subsistence risk and technological complexity. These hypothetical and ethnographic models are then compared to archaeological changes in technological costs at three sites occupied through the late Pleistocene in southern Africa. It is shown that while there is agreement in some respects, there are also times where archaeological outcomes differ dramatically from expectations. The results are taken to suggest that while costly technologies are generally pursued under conditions of increasing global cold, peak cold conditions at the height of Marine Isotope Stages 4 and 2 encouraged a reversion to least-cost technological systems. This may reflect a switch in the focus of optimisation from resource return rates to maximisation of early resource acquisition and/or maximisation of number of subsistence encounters. INTRODUCTION Major changes in the study of stone artefacts in the last half century have stemmed from a shift in emphasis from the description of technological variability to consideration of its underlying causes (Hiscock and Clarkson 2000; Odell 2000). As part of this shift, increased attention has come to be focused on the advantages conveyed by different technological systems under varying conditions. It is assumed that, when confronted with a new set of circumstances, the most beneficial options from the range of available technological alternatives will be preferentially selected (Bamforth and Bleed 1997; Kuhn 1995). Long-term trends in technological variation can thus be explained in terms of the differential persistence of more beneficial over less beneficial systems. A crucial part of the identification of which technological option is most beneficial is the evaluation of the costs incurred by each option. Though a component of early works in the field (eg., Torrence 1983), one aspect of technological decision making which has recently been neglected is the idea of technological costs. Specifically, while archaeologists tend to consider and speculate on the benefits of different technological systems under different contexts of deployment, they less often consider the relative costs of pursuing one given technological strategy over another. Intuitively such consideration would seem important. It is well documented that certain kinds of artefacts take longer to produce than others (Ugan et al. 2003). Furthermore, there are almost certainly costs involved in the acquisition of materials for the manufacture of artefacts, and these will vary depending on the materials selected. Finally, if there were no costs involved in the manufacture of technological items then there would be little incentive to pursue anything other than complex systems, assuming that complexity correlates positively with utility. In reality, however, complex technologies are not ubiquitous in either the archaeological or ethnographic records. This paper considers the nature of technological costs as they might have applied to stone artefacts. Several avenues of potential cost are explored and modelled, providing general criteria for assessing the cost of a given system. The paper then considers the changing costs of technological systems as they were deployed at three late Pleistocene sites in southern Africa. Changes in costs are considered in relation to changes in coarse climatic indicators. The results of this comparison suggest that there is a general relationship between environment and the technological costs borne by tool-using groups. The identified outcomes are not entirely consistent with predictions from existing models, requiring some consideration of the potential causes of variance. TECHNOLOGICAL COSTS Technological strategies featuring stone artefacts involve a complex interplay between costs and benefits (Bamforth 1986, 1991; Bamforth and Bleed 1997; Bleed 1986; Bousman 2005; Clarkson 2007; Hiscock 2006; Kuhn 1994, 1995; Torrence 1983, 1989). Technological costs chiefly arise as a result of three related problems. First, sources of stone suited to artefact manufacture are not ubiquitous in landscapes. Second, sources of stone will not always occur where and when tasks requiring stone artefacts occur. Third, that stone artefacts are rapidly depleted during use and resharpening. In order to maintain a supply of stone artefacts foragers must outlay time on the acquisition and manufacture of stone artefacts and stone artefact-making materials (Kuhn 1995). The