ABSTRACT
In 1996, two of us (A.G. and I.S.) sent ten samples from the early Iron Age sequence at Tel Dor to be dated by the Radiocarbon Dating Laboratory, Weizmann Institute of Science, Rehovot, then headed by Israel Carmi. Shortly afterwards Israel Finkelstein’s first paper advocating a ‘low chronology’ for the Iron Age in Israel (1996) was published. The Dor dates turned out to be low, even somewhat lower than Finkelstein’s suggested chronology. This was totally unexpected and in fact somewhat startling, but we decided not to trouble our minds, and sent eleven more samples to the Radiocarbon Dating Laboratory at Rehovot and one to Beta Analytic inc., Miami, FL. The results were the same. These dates (Gilboa and Sharon 2001; Sharon 2001; the stratigraphy and ceramic composition of the Dor sequence is presented in Gilboa and Sharon 2003) were the first empirical evidence demonstrating that the low chronology cannot be simply brushed away. Still, we were skeptical. Such a chronological shift, with its tremendous implications, cannot be based on 22 dates from a single site. Rather than publish these results and be done, three of us (I.S., A.G., E.B.) proposed a much wider study. It was our conviction that in order to construct a
statistically viable data-base we would have to consider hundreds of dates from meaningful contexts in many sites, measured by different labs and methods. A general solution to the chronological problem meant that rather than dating individual contexts, or (supposed) historical events, what we needed was to date the (beginning and end of) chronological horizons. Such horizons would first need to be defined, chiefly by ceramic considerations. Furthermore, to overcome the menacing chaine of pitfalls-problems of archaeological contexts, ceramic seriation, sample preservation, laboratory procedures and statistical modeling-a concerted effort by a multidisciplinary team would be needed. We approached all the excavators who, to our knowledge, possessed organics from such contexts as are described above. With the remarkably unstinting collaboration of all of them we collected more than 100 samples from 21 sites in Israel, from which more than 400 measurements have been produced, most of them already available. Other aspects, mainly bearing on characterization of the raw material for analysis, still await their completion. All the samples were dated at the Radiocarbon Dating Laboratory at Rehovot, and 22 were double-checked at the Arizona Accelerator Mass Spectrometry (AMS) Laboratory, Tucson. Calibration of dates employs the INTCAL98 radiocarbon calibration curve (Stuiver et al., 1998). The initially projected length of the project was four years, and the cost of some 400 radiocarbon determinations considerable. The initial Israel Science Foundation grant we were awarded for this purpose covered, however, only two years and half of the costs. It is only recently that additional funding has been obtained that will enable us to complete the project as planned. A laboratory intercomparison based on 22 of the samples (Boaretto et al. 2005) has shown that there are no systematic differences between the participating laboratories and methods. This intercomparison assures the accuracy of the 400 radiocarbon dates produced. Meanwhile, there have been several developments in 14C dating on the Levantine Iron Age scene. Chiefly, the publication of another substantial database-32 dates from Tel Rehov, measured at the Centre for Isotope Research, University of Groningen-purported to support the ‘high’ chronology, and refute the ‘low’ one (Bruins, van der Plicht, and Mazar 2003). The discrepancy between the Dor data set and that of Rehov was explained by the Rehov team by claiming that in the 1990s the Rehovot lab was experiencing difficulties, which caused its dates to be consistently too low (Mazar 2004: 31-34). Although these allegations were not supported by any explicit reasoning, we re-measured the ‘suspect’ Dor sequence, by extended replication and intercomparison. The results of these investigations are presented here, together with the methodologies used to analyze the partial data sets we have examined so far and by which we propose to investigate the full data set once all the dates are measured, and once the checks that are still pending are completed. The focus of this paper is on methodology. But while the case studies presented here are shown mainly for the purpose of illustrating the methods used, and the results should be considered provisional, the direction in which they are all pointing is clear. Archaeological Considerations
As a result of his own set of radiocarbon dates, A. Mazar recently proposed what he terms the ‘extended conventional chronology’ (e.g. Bruins, van der Plicht, and Mazar 2003: 318; Mazar 2004: 30-31). This ‘extension’ incorporates most of the 10th and 9th centuries BCE within Iron IIA. This, in itself, is quite a generous concession towards the low chronology, which has somehow passed practically unnoticed. Thus, at least one consensus (or near consensus) seems to have formed in the last two years: that the ceramic culture termed in Israel Iron Age IIA encompasses at least the first 70 years or so of the 9th century BCE. The debate, therefore, centers on the beginning of the sequence and may be narrowed down to two questions:
1. How do we define, stratigraphically and artifactually, the transition between the cultural entities we term Iron Age I and II in many different sites covering a large geographical area? What transition are we actually after? Do we seek to date the late Iron I destructions (such as Megiddo VIA)? Or the transition between the Iron Age I and the Iron Age II ceramic assemblages? (Does this latter transition bear any meaning at all for issues of state formation, emerging identities, etc.?) Or do we aim at dating the transition from nonmonumental to monumental stages at sites attributable to Israelite entities? All these are not necessarily co-terminus nor contemporaneous.
