Did the walls of Jericho fall?

What everyone agrees on

Tell es-Sultan is the agreed identification of biblical Jericho. The mound sits beside the spring ʿAin es-Sultan, two and a half kilometers north of the Roman and modern town. The site has been trenched or excavated by Charles Warren in 1868, by Ernst Sellin and Carl Watzinger from 1907 to 1909, by John Garstang from 1930 to 1936, by Kathleen Kenyon from 1952 to 1958, and by an Italian-Palestinian team led by Lorenzo Nigro and Hamdan Taha since 1997. Across all five campaigns, one stratigraphic fact is fixed. There was a massive destruction at the end of the Middle Bronze IIC city, with collapsed mudbrick fortifications, burned houses, and storage jars full of carbonized grain left behind. That destruction is not in dispute. What is in dispute is when it happened, and what was on the site afterward.

The Late Bronze Age (LBA) runs roughly 1550 to 1200 BCE, subdivided as LB IA (1550-1450), LB IB (1450-1400), LB IIA (1400-1300), and LB IIB (1300-1200). The conquest on the early-date reading lands in LB IB or early LB IIA. On the late-date reading it lands in LB IIB. So the question 'was Jericho fortified and inhabited at any point in the LBA?' is what decides whether Joshua 6 has an archaeological footprint.

The three lines of evidence

Three classes of evidence have done most of the work in the debate. The pottery from City IV is the centerpiece, because Cypriot bichrome ware would settle the LB I identification if it could be secured. The Egyptian royal-name scarabs from the cemetery are the secondary witness. The radiocarbon series on charred grain from the destruction debris is the most recent entry and the one most often cited in the 2000s and 2010s. The columns below lay out how each position handles each line.

Cypriot bichrome ware as the type fossil

The bichrome ware question deserves a closer look because it is what the pottery argument turns on. Cypriot bichrome ware is a tradition of two-color painted pottery (typically red and black) produced on Cyprus in the late seventeenth through fifteenth centuries BCE and traded across the Eastern Mediterranean. The Levantine type series was worked out by Claire Epstein in 1966 and refined by Robert Merrillees and Jacqueline Karageorghis. The diagnostic forms (jugs, bowls with concentric circle motifs, the 'bird and fish' motifs) appear at Tell el-ʿAjjul, at Megiddo, at Tell Abu Hawam, and across the Cypriot exports. Their presence in a sealed context places that context in LB I.

Wood's 1990 BAR article (and his subsequent publications) argued that Garstang's photographs of City IV pottery show LB I features, and that several sherds in Kenyon's published plates have LB I or transitional MB-LB characteristics that Kenyon misclassified. Bienkowski's response (BAR 16/5, 1990) replied that Wood's identifications confuse local Canaanite painted ware with the diagnostic Cypriot type fossils, and that no diagnostic Cypriot bichrome import has been recovered. Both authors are working from the same published material; the disagreement is over typological identification of the ware, not over its physical existence on the site.

The Amenhotep III scarabs

Garstang excavated several rock-cut tombs north of the tell between 1930 and 1936 and published the funerary material in successive Liverpool Annals of Archaeology and Anthropology volumes. The scarab series from those tombs includes both private-name and royal-name examples, with the royal series running Hyksos, Eighteenth Dynasty (Hatshepsut, Thutmose III, Amenhotep II, Amenhotep III), and beyond. The Amenhotep III scarabs are the latest royal-name examples, and they place at least some tomb use in the second half of the fourteenth century BCE.

What the scarabs do not establish, by themselves, is that the mound above the necropolis was inhabited at the time of the latest scarabs' deposition. Cemeteries are often reused across centuries by populations living elsewhere; the necropolis at Megiddo, for example, includes burials from periods when the tell shows occupation gaps. The early-date position reads continuous tomb use as evidence of continuous settlement; the MB IIC and unoccupied-LB positions read tomb reuse as compatible with an abandoned or much-reduced mound. The scarabs cannot adjudicate the question on their own.

The Bruins and van der Plicht carbon-14 series

In 1995, Hendrik Bruins and Johannes van der Plicht published a radiocarbon analysis of six short-lived cereal samples and twelve charcoal samples from Tell es-Sultan in Radiocarbon 37. The samples came from sealed contexts in the destruction layer of City IV, recovered originally by Kenyon and stored in Jerusalem. The short-lived cereals are the most diagnostic, since charcoal from long-lived trees can give 'old wood' dates that pre-date the destruction by decades or centuries. The six cereal samples averaged 1562 ± 38 BCE calibrated (1σ), which centers on the MB-LB transition and substantially predates either standard exodus date.

Wood's 2008 response (in Critical Issues in Early Israelite History) argued that the samples may have been contaminated during forty years of storage, that calibration uncertainty was understated, and that the carbon-14 result by itself should not override the pottery and scarab evidence. Bruins and van der Plicht have stood by the series in subsequent publications. The carbon-14 question is in stasis pending new sampling.

What each side has to account for

The LB IIA position has to explain the controlled stratigraphy, the absence of Cypriot bichrome imports from sealed contexts, and the 1562 ± 38 BCE carbon series. It has to read the Cypriot-style sherds Wood points to as the diagnostic type fossil rather than as local painted ware. It has to read the cemetery scarabs as evidence of mound occupation rather than as funerary reuse of an older necropolis. The position is not without defenders, and Wood's argument has been worked over in detail at Associates for Biblical Research and in evangelical academic publications. It remains a minority view in the broader field.

The MB IIC position has to explain the cemetery scarabs that run through Amenhotep III, the limited LBA architectural evidence (the 'Middle Building'), and the residual question of whether the eroded upper mound once held an LBA fortification that no excavator has recovered. It has to leave Joshua 6 without a direct archaeological correlate. The position holds that the destruction is the end-of-MB regional event, that the LBA occupation was minimal, and that the Joshua narrative either reflects theologized memory of the visible MB ruin or belongs to a literary tradition older than the archaeology of the tell.

The 'unoccupied in LB' position has to explain why Joshua 6 names Jericho at all if there was nothing there to take. The standard answers are that ha-yericho means 'the moon-city' or 'the fragrant place' and was attached to the spring and the visible ruin regardless of occupation, that the Joshua narrative reflects an older memory of a destruction the audience could still see, or that the conquest narrative is foundation literature shaped by editorial concerns later than any reconstructed settlement event. The position takes the absence of LBA fortifications as decisive, treats the carbon series as confirmation, and accepts that the conquest of Jericho on this reading is a story without a site.

Reading Joshua 6 with the question open

Joshua 6 describes a city Israel marches around for seven days, with priests carrying the ark and trumpets, and walls that fall when the people shout. The narrative does not depend on the archaeological identification of the destruction layer for its theological work. If the LB IIA position is correct, the chapter has a direct archaeological correlate. If the MB IIC or unoccupied-LB positions are correct, the chapter is a memory or a literary construction attached to the visible ruin beside the spring. The pottery, the scarabs, and the carbon series will keep being weighed against each other. The decision on which way the evidence tips has not closed in eighty years of excavation, and is unlikely to close before the next set of samples is run.