Setting all the theoretical objections aside, it is quite reasonable to ask, as Smith does, whether archaeological remains can be quantified with sufficient precision to demonstrate that patterns are more consistent with foraging theory than some alternative. It is equally reasonable to ask, as he also does, whether enough can be known about prehistoric environments to generate the rankings and estimates of abundance needed to model diet breadth. I share Smith's concerns about the quality of the archaeological record, but if excavated faunal and floral assemblages will not sustain inferences as to the relative importance of various foods, more than foraging theory is at risk; a substantial fraction of any archaeological interpretation rests on knowing these things. Fortunately, we do not need to establish consumption in accordance with the diet breadth model at every site, any more than I need to be able to radiocarbon date every Rose Spring projectile point I find. Indeed, taken alone, individual assemblages, even sizeable ones, may not be especially revealing of economic decision making. Many assemblages, including small ones, in combination often speak more clearly to these patterns. These, in combination with knowledge of fundamental ecological relationships, can provide the information needed to construct and test plausible models of diet breadth. Waguespack and Surovell (2003), for example, employ the relationship between body size and density in mammals (smaller body = higher density) to estimate the frequency with which different prey would have been available to Clovis hunters, and compare this to the observed frequency of those prey in Clovis faunal assemblages. Because large prey are more commonly represented in these assemblages than they should be, Waguespack and Surovell conclude Clovis foraging was highly selective, implying relatively narrow diet breadth and relatively abundant resources. This body size to density relationship is sufficiently robust to serve as a general source of OFT prediction. It predicts almost perfectly, for instance, the relative frequency of Ache kills, suggesting that the Ache take these species upon encounter as would be predicted by the diet breadth model, when diet breadth is quite broad (Figure 14.1).
Smith's real message seems to be one of caution: HBE has substantial promise but theories can be deceptive and archaeologists, like all scholars, can be prone to foolishness. To keep us from mischief, he advocates rigid adherence to protocol of scientific methodology that involves, among other things, comparing the performance of HBE with equally plausible theoretical alternatives. The problem here, and with the idea of multiple working hypotheses in general, is that when it comes to foraging behavior it is hard to know what the alternatives are or should look like. Indeed, that is precisely the problem. I am drawn to foraging theory because it generates hypotheses with clear implications for the archaeological record. I would very much like to explore ways of thinking about prehistoric human behavior in other than strictly economic and evolutionary ways, perhaps with reference to the deep-lying mental structures that interest Levi-Strauss (1964), who talks of decision making weighed in terms of "good thinking" rather than "good eating." But "good thinking" does not lead me to predictions about the archaeological record in the same way that
FIGURE 14.1. Density of mammals taken by Ache hunters predicts the frequency with which they are taken. Density estimated from body size (cf. Waguespack and Surovell 2003) using average body size (Hawkes et al. 1982: Table 1). Number of individuals taken is as given by Hawkes et al. (1982: Table 1).
"good eating" does, when I construe that with reference to foraging theory. Evolutionary psychology offers some possibilities in this regard but not with the precision of OFT. My more fundamental problem with the method of multiple working hypotheses is its suggestion that I should spend time developing plausible alternatives. In my view the responsibility for that falls squarely on those who doubt the hypothesis I'm working on; it keeps me busy enough as it is. The alternatives in which I am mainly interested are the null hypothesis and the "middle range" behaviors acting on the patterns on which my attention is mainly centered, for instance butchering and density-mediated destruction affecting the composition of faunal assemblages. These provide complication enough to justify my choice of HBE theory as the simplest, most parsimonious account of human behavior.
Was this article helpful?