A New Chronological Framework for Prehistoric Southeast Asia, Based on a Bayesian Model from Ban Non Wat

Article excerpt

Introduction

As Movius observed of the European Upper Palaeolithic, 'Without ... a [chronological] framework the over-all picture becomes confused and, in certain instances, almost meaningless. Time alone is the lens that can throw it into focus' (Movius 1960: 355). The passage of time is equally vital for a proper understanding of the prehistoric sequence in Southeast Asia. While the cultural sequence is agreed by most scholars, its timing is not. The ancestors of the first rice farmers in Southeast Asia probably lived in the Yangtze Valley to the north (Liu et al. 2007), and spread south, via the coast and the major rivers, to enter the broad riverine plains of Southeast Asia. They brought their Austro-Asiatic languages, and a way of life that centred on settled village communities incorporating widespread exchange in exotica, a sophisticated ceramic industry, weaving, and a mortuary tradition that involved both extended inhumation and interment in lidded jars. This Neolithic settlement phase was followed by the adoption of copper-base metallurgy, in which copper and tin were alloyed from the earliest known contexts. The transition into the Iron Age has not been precisely dated, but it is known that early states were forming by the fourth to fifth centuries AD. The timing and the degree to which Iron Age communities developed social and technological sophistication prior to the rise of early states is poorly documented: Noen U-Loke is the only extensively-excavated Iron Age site in Thailand to be published (Higham, C.F.W. et al. 2007).

We do not know when the first farmers reached Southeast Asia and there remains a basic uncertainty over the date for the inception of copper-base metallurgy in Southeast Asia. This has generated a lack of understanding of the social changes that occurred with the early Bronze Age. As Muhly (1988: 16) stressed 20 years ago in a dictum still true, 'In all other corners of the Bronze Age world ... we find the introduction of bronze technology associated with a complex of social, political and economic developments that mark the rise of the state. Only in Southeast Asia ... do these developments seem to be missing.' One of the objectives of our recent excavations at Ban Non War has been to open an area large enough to identify just those variables Muhly describes.

In retrospect, the causes of controversies over chronology are readily understood (Solheim 1968; 1970; Bayard 1972, 1979; Gorman & Charoenwongsa 1976; Bayard & Charoenwongsa 1983; Higham 1983; Loofs-Wissowa 1983). Radiocarbon determinations have virtually all been derived from charcoal, with its problems of 'old wood'. Only very rarely has the species of tree been specified, a practice that needs to be addressed in future dating programmes. No recognition was given to the unreliability of mixed samples (Ashmore 1999). In many cases, the relationship between a charcoal sample and the event being dated was unreliable. Major cultural changes, such as the beginning of copper-base metallurgy, have been dated on the basis of only a handful of determinations. When a sample of dates was available, the construction of the site's chronology followed procedures now shown to be importantly wrong (Bayliss et al. 2007: 9).

Resolving this situation first requires a prehistoric site with a cultural sequence spanning the early Neolithic to the end of the Iron Age. Such sites are very rare in Southeast Asia. Phases within such a site would need to be ordered in terms of a relative chronology, and we would then require a sufficient number of radiocarbon determinations, preferably generated on the basis of samples with no inbuilt age, to provide dates for the successive cultural phases identified. Armed with such a series of dates we could apply the refinement of the Bayesian approach as outlined by Bayliss et al. (2007). The Bayesian method is able to provide us with quantitative, probabilistic estimates of archaeological events through a combination of calibrated radiocarbon likelihoods and given archaeological information, for example, the sequence of phases within a site's sequence (see Buck et al. …