ABSTRACT
What we know about early hominids at Olduvai (and elsewhere) is tightly linked to the process and structure of making inferences. Ideas about hominid activity are based on a hierarchy of conditional inferences, which corresponds to the levels of analysis mentioned in Chapter 1. Thus, for example, the stone cache interpretation is one of several ideas we can consider to explain why hominids carried resources to particular spots on the landscape. Yet this is what might be called a third-order inference, which depends on a set of prior interpretations. The latter include taphonomic (first-order) inferences (e.g., that the bones and artifacts indeed represent behaviorally transported remains and that hominids played a role in this) and second-order inferences concerning the specific activities of hominids at these sites and the ecological milieu (e.g., carnivore activities). There is also a conditional, or interdependent, set of inferences that pertains to site taphonomy alone. To distinguish agents of bone transport as “primarily hominid” and “primarily carnivore” depends on accurate recognition of stone tool and carnivore damage, contextual inferences based on the accumulations of stone artifacts and bone concentrations made by carnivores, as well as other lines of evidence and interpretation. On the one hand, some inferences are directly drawn from controlled experiments and “first principles.” Winnowing of small objects by water flow is one example, and marks on bone surfaces can also be related to basic differences between, say, the effect of stone versus enamel on bone. In contrast, other inferences are based on multiple lines of data and on theoretical ideas or recurrent observations from one or more disciplines that offer an interpretation of those data. The idea that the relative frequency of carnivores in a faunal assemblage corresponds to the degree of interference competition over carcasses is an example of this type of inference. The validity of higher order inferences relies upon an intersecting web of lower order inferences of various kinds. The latter serve as the foundation for interpreting hominid activities and paleoecology.
