Katharina Streit and Felix Höflmayer NEAR EASTERN ARCHAEOLOGY 79.4 (2016) 233 I n the June 2016 issue of Near Eastern Archaeology, Michele D. Stillinger, James W. Hardin, Joshua M. Fein- berg, and Jeffrey A. Blakely published a paper entitled “Archaeomagnetism as a Complementary Dating Technique to Address the Iron Age Chronology Debate in the Levant” (Stillinger et al. 2016). The article addresses a key issue in the Iron Age chronology debate, namely, long date ranges of cali- brated radiocarbon determinations caused by the considerably flat shape of the calibration curve in the tenth century b.c.e. (fig. 1). As a result, radiocarbon dating frequently does not provide the precision desired by archaeologists for answering chronological questions, such as the Iron Age I–IIA transition. Distinguishing between an early or late tenth-century b.c.e. date based on single calibrated radiocarbon determinations is close to impossible without the aid of Bayesian analysis (fig. 2). In their article, Stillinger et al. present preliminary results from the archaeomagnetic studies at Khirbet Summeily and claim that archaeomagnetism could be used to reine the current text- or radiocarbon-based chronologies of the Iron Age Levant. Unfortunately, their approach is hampered by circular reasoning, which calls into question the potential of the proposed method for the Iron Age chronology debate. his debate revolves around the comparison of two indepen- dent strains of absolute chronological information: the histori- cal–biblical chronology on the one hand, and the radiocarbon dating of archaeological contexts on the other. While some have claimed that radiocarbon dating challenges the traditional his- torical–biblical assessment and are in favor of a signiicantly lower absolute chronology (e.g., Sharon et al. 2007; Finkelstein and Piasetzky 2010a, 2010b), other scholars have argued that radiocarbon dating might be in agreement with the traditional chronology (e.g., Bruins, van der Plicht, and Mazar 2003; Mazar and Bronk Ramsey 2008; Mazar 2011; Garinkel et al. 2015). As the discrepancy between the diferent chronological models is only about half a century, the distinct shape of the radiocarbon calibration curve and the resulting long calibration ranges ham- per a simple decision as to which chronological model should be preferred. As Stillinger et al. rightly point out (98), statistical methods, such as Bayesian analysis (incorporating stratigraphic information into the calibration process of radiocarbon determi- nations and reducing the calibrated date ranges; Bronk Ramsey 2009), directly depend on the prior information or assumption(s) that underlies the chosen model. he radiocarbon model of the early Iron IIA site of Khirbet Qeiyafa demonstrates the power of Bayesian analysis for the Iron Age (see Garinkel et al. 2015 for details). While it is not possible to decide whether the site dates to the eleventh or tenth century B.c.e. based on single calibrated radiocarbon determinations (ig. 3), the archaeological observa- tion that all samples derive from a single destruction layer (with some samples maybe also being residual), adds additional in- formation to the radiocarbon model. his single-phase Bayesian model restricts the calibration ranges based on statistic grounds and thus calculates the destruction of the city into the early tenth century b.c.e. at latest (ig. 4). Nevertheless, an independent dat- ing method to corroborate the results of radiocarbon dating and Bayesian modelling would be highly welcome for the current de- bate on the Iron Age chronology. Stillinger et al. argue that archaeomagnetism could have the potential to solve the current debate, to reine the Iron Age chro- nology, and to make it more robust. he method itself is based on the paleomagnetic signature trapped in materials such as pot- tery, mud brick, slag, or lithic material during a heating process. his signature is not constant over time, but instead depends on the luctuations of the earth’s magnetic ield. he geomagnetic signal does not contain any chronological information per se, but has to be compared against an archaeomagnetic reference curve, which is built on a securely dated sequence of measure- ments deriving for example from laminated lake sediments or cave speleothems (the archaeomagnetic reference curve). he archaeomagnetic reference curve needs to be absolutely dated in order to be meaningful for the archaeomagnetic dating method, for example, employing varve- or speleothem growth layer counting, U-series dating, radiocarbon dating, or a com- bination of them (91–92). Such reference curves, for example, ARCH3k.1e, CALS10k.1, or PFM9K (Nilsson et al. 2014; Korte et al. 2011; Korte and Constable 2005) have been built predomi- nantly on well-stratiied records such as lake varves or volcanic sediments. Herein lies one of the main issues of the paper. he authors argue that while there is broad agreement between data col- lected in the southern Levant and more global records such as ARCH3k.1e, CALS10k.1, or PFM9K, there seem to be very lo- cal phenomena, such as the increase in magnetic ield strength in the tenth century B.c.e. Consequently, they aim to “provide new intensity data to improve the reference curve for the Levant Archaeomagnetism, Radiocarbon Dating, and the Problem of Circular Reasoning in Chronological Debates: A Reply to Stillinger et al. 2016 This journal was published by the American Schools of Oriental Research and is available on JSTOR at http://www.jstor.org/journal/neareastarch. You may receive the journal through an ASOR membership or subscription. See http://www.asor.org/membership/individual.html for more information.