Should dates trump context? Evaluation of the Cave 7 skeletal assemblage radiocarbon dates

Abstract

“Massacre” was the accepted interpretation for a prehistoric skeletal assemblage of around 90 individuals from Cave 7 in the SE Utah of the North American Southwest since the 1890s. Coltrain and others (Journal of Archaeological Science 39, 2220–2230) rejected this account based on a dispersed suite of AMS radiocarbon dates on purified bone collagen from the interred individuals. Since dates from skeletons exhibiting perimortem damage were scattered across some 400 radiocarbon years, Coltrain et al. argued for multiple interments of victims of violence and related kin spread across several centuries. The temporal placement of interment events in Cave 7 clearly cannot be known independent of radiocarbon assays but such assays should not be privileged above contextual information about which individuals were interred together unless verified by an independent dating laboratory. Only by ignoring important information about burial context in the 1893 field record can the dates of Coltrain et al. be accepted as accurate estimates of time of death. We redated residual collagen from 11 of the Cave 7 individuals because of significant contextual anomalies with some of the prior results. The new assays combined with contextual evidence demonstrate that some of Coltrain et al.’s dates are either too old or too young; the dates are neither sufficiently accurate (true estimates of sample age) nor sufficiently precise (small confidence intervals) to refute a single-event massacre or to confirm multiple interment events in Cave 7. Nonetheless, dates that meet these criteria disclose at least two interment events, a large one of mostly adult males, many exhibiting perimortem damage (a massacre assemblage), and a small one consisting of an adult female with three children. A third interment event of a few adult females and child is possible based primarily on context since the date distribution in this case overlaps substantially with that of the massacre victims; context also hints at other interment events. Aside from chronology we clarify the number of Basketmaker individuals interred at the site and the incidence of perimortem violence. A single-event mass killing continues to be the most likely interpretation for around 58 individuals, mostly adult males (at least 35) but also adult females and children. This incident occurred between cal. AD 20–80 and doubtless had a significant social impact at the time because of its scale, reverberating throughout the early farming communities of the Southwest.