Gezer’s Contribution to the Iron Age Chronological Debate

Lyndelle Webster, S. Wolff, and S. Ortiz’s contribution to this Forum is based on their detailed and well-presented earlier article (Webster et al. 2023). In the following, I will refer to both articles. The authors touch upon some of the most debated issues in the archaeology and history of the Iron Age in the southern Levant, known as the debate between high chronology vs. low chronology (Finkelstein 2005 vs. Mazar 2005, with earlier literature). In fact, the debate is over wider issues such as the dating and understanding of the Philistine immigration process and the historicity of the “united monarchy” of David and Solomon. When this debate started in the 1990s, radiocarbon dates were not yet widely used for historical periods in the Levant. However, since the early 2000s, the use of 14C for the dating of the Bronze and Iron Ages has become common, and it revolutionized our ability to date archaeological contexts. Early studies in this realm supposedly supported the low chronology (e.g., Sharon et al. 2007). However, later studies have pointed out the complexity of this subject, and several studies resulted in higher dates that fit my suggested modified chronology (summaries in Finkelstein and Piasetzky 2011; Mazar 2011). Based on dates from stratified contexts at Gezer, Webster et al.’s study clearly supports the high chronology for the entire period from the twelfth to the early ninth centuries BCE.The ability to resolve minute chronological questions of less than half a century by utilizing 14C dates was questioned in light of the inherited uncertainties and problems embedded in the method. Often, the calibrated dates from a single context provide wide ranges, as can be seen in many studies, including the present one. For example, at Gezer there are significant ranges of the uncalibrated dates (yrs BP) (i.e., in Stratum 12B: difference of up to 129 years among seven measured dates; in Stratum 11: 117 years among five measured dates; in Stratum 10B: 75 years among three measured dates). These differences and wiggles in the calibration curve provide wide ranges of calibrated dates in each stratum, even in the 68.3% hpd (“highest probability density,” following the terminology used by Webster et al.) and even more so in the 95.4% hpd (see Webster et al. 2023: table 3). The standard method to overcome this inherited problem is to construct a Bayesian statistical model, most often carried out with OxCal software. In such models, constraints can be inserted based on stratigraphic data (Bronk Ramsey 2005). However, the results depend on variables such as the quality of the archaeological samples and contexts, the number of samples from each context, the quality of the dating methods, and the laboratories involved. I am far from being qualified to judge the results of such Bayesian models, yet as a consumer of such studies (and partner to several publications of such models) I believe it is essential to recall that the models rule out substantial ranges of dates that do appear in the unmodeled dates (see my view in Mazar and Streit 2020: 487–91). This is particularly disturbing when there are only a few dates from a particular context and they provide a wide range of dates, as shown below in a few cases. Interestingly, in the case of Gezer, in many instances the range of unmodeled calibrated dates provides earlier options in addition to those suggested by the Bayesian model. This fact provides even more substantial support to the high dates suggested in the Gezer study. In the following, I will refer to the Bayesian model’s results as Webster et al. (2023) present them but will point out specific difficulties and options.As the researchers emphasized, Gezer is one of the few sites where a continuous urban tradition was detected throughout the Late Bronze II to Iron IIA with no detectable occupation gaps. There are very few sites with similar continuity; notable among them is Tel Reḥov in the Beth Shean Valley, where a series of radiocarbon dates starting in the late twelfth century until the ninth century BCE was published (Mazar and Streit 2020).1Philistine Bichrome (“Philistine 2”) pottery appears at Gezer for the first time in Stratum 11 in the Tandy expedition (corresponding to HUC Strata XIII–XII). The Bayesian model dates for this stratum (1162–1112 BCE, transition to Stratum 10 between 1131–1088 BCE) are in agreement with the high chronology of the Philistine culture (called “middle chronology” for the Philistine culture by Webster et al. 2023: table 2).2 According to this system, the Philistine Monochrome pottery (“Philistine 1”), which is not represented at Gezer, was produced in the heartland cities of Philistia since Ramesses III’s eighth year, and Philistine Bichrome pottery appeared a few decades later, during the second half of the twelfth century. This is in contrast to the low chronology, which dates the appearance of Monochrome Philistine pottery after the end of the Egyptian rule in Canaan ca. 1140/1130 BCE (Finkelstein 1995, 2018). At Gezer, Stratum 12A of the Tandy expedition, dated by the Bayesian model to 1203–1155 (transition to Stratum 11: 1183–1136), probably corresponds with the Philistine Monochrome phase in Philistia. As Webster et al. note in their 2023 article, similar dates for the appearance of Philistine Monochrome and Bichrome pottery come from Ashkelon, where the transition from Stratum 21 (pre-Philistine phase) to 20 (Philistine Monochrome pottery) was dated by Bayesian model to the range 1190–1150 BCE, and the transition from Phase 20 to 19 (appearance of Philistine Bichrome) was dated to 1170–1130 (Asscher et al. 2021). This evidence negates both the low chronology as well as the ultra-high chronology suggested for Tell es-Safi (Asscher et al. 2015). I fully support this conclusion. It is supported also by the precise resemblances between Philistine Monochrome pottery and Mycenaean IIIC pottery in Cyprus, dated to the Late Cypriote IIIA period, that is, the first half of the twelfth century BCE. It also fits the appearance of a Monochrome Philistine deep bowl at the site of Qubur el-Walaidah in the northern Negev in a context dated to the Egyptian Twentieth Dynasty (further discussion and references in Mazar 2023: 203–4).The Philistine Bichrome pottery appears at Gezer along three or four strata (HUC XIII–IX = Tandy Strata 11, 10B, 10A), recalling other Philistine sites like Ashdod, Ashkelon, Tel Miqne, and Tell Qasile where such pottery appears along several stratigraphic phases, indicating its longevity.Strata 9 of the Tandy expedition to Gezer (a poor phase, unparalleled in the HUC excavations), 8B–A (=HUC VIII), and 7 (=HUC VII) provide a stratigraphic and chronological framework for the large public building west of the monumental six-chambered gate bounded on the south by the casemate wall, and by analogy to the gate itself. The Bayesian-model dates for Stratum 8 are 998–957 BCE, and the transition between Strata 8 and 7 is in the range 969–940 BCE. The authors conclude that these buildings reflect Judah’s expansion to the west during the tenth century BCE. Though they avoid mentioning specifically Solomon as the builder, they believe that the latest date of Stratum 8 may fit the time of the Sheshonq I (biblical Shishak) raid. Being aware that the various dates suggested for this raid by most Egyptologists are somewhat later, the authors utilize high dates for this pharaoh (988–945 BCE), based on a few 14C dates (Webster et al. 2023: 2). However, dating Sheshonq by 14C is highly problematic due to the very minimal data. Such a high date for him would make the biblical correlation between his raid and the fifth year of Rehoboam (1 Kgs 14:25) impossible since, according to the inner-biblical chronology, this year would be no earlier than ca. 925 BCE. To resolve this, we need to look at the data: Three 14C dates from Stratum 8 were published, and two others were defined as outliers (Webster et al. 2023:15, table 3). Their unmodeled calibrated dates cover most of the tenth century BCE (982–901, 1000–930, 998–912 BCE in 68.3% hpd), so that all three may fit (in their lower range) the traditionally suggested dates for Sheshonq’s raid invasion, namely, ca. 925–920 BCE. The somewhat earlier dates indicated by the Bayesian model resulted from constraints created by a single date from Stratum 9 and nine dates from the succeeding Stratum 7. However, one may ask whether the unmodeled calibrated dates cited above should not be considered valid.The HUC gate area excavations (Field III) and the Tandy excavations are known only from preliminary reports, and the material culture (pottery assemblages and other artifacts) still needs to be published. Critiques may argue that we need better documentation of the stratigraphic data concerning the relationship between the gate structure and the large building to its west, accompanied by precise publication of pottery and artifacts from these areas. As for now, we have to rely on the excavators’ conclusions.Another issue is the date of Tandy Stratum 7. The excavators previously thought that it was destroyed by Hazael ca. 830. The Bayesian model dated this stratum to 957–913 and the end of Stratum 7 to 927–885, leading the authors to give up their former suggestion and conclude that Gezer was destroyed before 885 BCE and perhaps lost its power or was even abandoned during the Omride period in the ninth century. Such a conclusion has essential historical implications since it diminishes the power of the northern kingdom of Israel in the ninth century BCE in this region. To evaluate this conclusion, I suggest looking at the unmodeled calibrated dates. Nine dates from Stratum 7 were published: Six provided dates in the tenth century BCE (and even late eleventh century BCE), similar or even earlier to the three dates from Stratum 8. However, three others are later: they cover the last quarter of the tenth century and much of the ninth century, and their lowest dates are in the years 844, 840, and 831 BCE. Could these three latest dates still support the older hypothesis, or is the Bayesian model conclusive enough to negate this hypothesis?The evidence from Gezer Stratum 8 significantly impacts the debate concerning the united monarchy and the rise of the early state in Judah and Israel. Yigael Yadin’s attribution of the six-chambered gates at Hazor, Megiddo, and Gezer to Solomon, based on 1 Kgs 9:15 (and 1 Kgs 9:17 concerning Gezer) (Yadin 1972: 147–64) was severely criticized by later scholars (e.g., Finkelstein 1996 and many additional publications by him and others). A new paradigm was constructed, based on obliterating the united monarchy from historical discourse and emphasizing the more robust status of Israel over Judah until the ninth century BCE (e.g., Finkelstein 2013: 37–118; Frevel 2023: 117–258, and others). In this debate the monumental architecture, including the six-chambered gates at Hazor, Megiddo, and Gezer, is of special importance. The gate at Megiddo was first attributed to Stratum V and dated to the eleventh century BCE by its excavators, then to Solomon (by Yadin, who attributed it to the combined Stratum VA–IVB) and to David (by Y. Aharoni); later, I. Finkelstein and D. Ussishkin attributed this gate to Stratum IVA of the eighth century BCE, and more recently Finkelstein and his team moved the gate back to Stratum VA–IVB and dated it to the Omride period in the ninth century (Finkelstein et al. 2019, with references to previous literature). The recent years of excavations at Megiddo have shown that Stratum VA–IVB had at least two major phases; the lower one (Phase Q-5 in Area Q) was dated by 14C to the late tenth century or early ninth century BCE, prior to the Omride period (Kleiman et al. 2019). This date is also supported at Tel Reḥov, where Stratum V, with similar material culture to that of Megiddo VA–IVB, was dated by 14C to the late tenth and early ninth centuries BCE (Mazar 2020b: 118–22), and at Beth Shean, where Stratum S-1a, with similar material culture, was proved to be contemporary with Tel Rehov V, based on paleomagnetic study (Vaknin et al. 2023). The Megiddo excavators assume that the six-chambered gate was constructed in the later phase of Stratum VA–IVB (corresponding with local Stratum Q-4) of the Omride period. Yet, this claim lacks objective stratigraphic or ceramic evidence since, in the gate area, no subphases of Stratum VA–IVA were observed. Could the Megiddo gate be earlier than the Omride period?At Hazor, the excavators Y. Yadin and A. Ben-Tor attributed the six-chambered gate and casemate wall to Stratum X, which they dated to tenth century and identified as Solomonic, while other scholars suggested a later date in the Omride period (Finkelstein 1999; Shochat and Gilboa 2019; Sergi and Kleiman 2018). Recently, two 14C dates were measured from samples found on a floor of Stratum XB and provided dates in the tenth century, while dates from the following phase Stratum XA are in the ninth century (announced in 2024 by I. Kreimerman, the current excavator of Hazor, in public lectures). The dated floors are located in a building close to the gate structure and associated with the gate by the excavators. If they were correct, it would mean that the gate was founded in the tenth century BCE, recalling the results at Gezer. This new data from Gezer, Hazor, and, to a lesser extent, Megiddo must lead to a renewed evaluation of Yadin’s thesis (Mazar 2007: 155–57; 2010: 33, where I suggested that these gates might have been constructed in the tenth century).The results from Gezer may imply the debate concerning the dates of monumental architecture in Jerusalem. Studies suggested that the Stepped Stone Structure complex in the upper part of the City of David and some of the buildings uncovered in the Ophel were founded during the tenth century BCE (Mazar 2010, 2020a; E. Mazar 2018: 315–93; Winderbaum 2022). Though these dates are subject to debate, the early dates may illuminate Jerusalem’s role during this time. This subject and others, like developments in the Shephelah and the role of the Arabah copper industry in the tenth century, are beyond the scope of the present response but should be mentioned as they are crucial for understanding state formation in Judah and the evaluation of the biblical narrative concerning the reigns of David and Solomon (Mazar 2007; 2010: 30–34; for recent treatment see Sergi 2023: 101–40, 249–340). The evidence from Gezer and other sites mentioned above calls for a reevaluation of this subject.