Conquest, infiltration, or revolt? Three models of Israel's origin

What the texts and the dirt actually give you

Joshua narrates a rapid coordinated campaign. The walls of Jericho fall (Joshua 6). Ai is burned (Joshua 8). A southern coalition is broken at Gibeon (Joshua 10). A northern coalition is broken at the waters of Merom and Hazor is burned (Joshua 11). The closing summary at Joshua 11:23 says Joshua took the whole land. Then Judges 1 opens with a long list of cities Israel did not drive out, tribe by tribe. Manasseh did not drive out Beth-shean, Taanach, Dor, Ibleam, Megiddo. Ephraim did not drive out Gezer. Zebulun did not drive out Kitron and Nahalol. Asher did not drive out Acco, Sidon, Achzib. The same canon that gives the totalizing summary also gives the catalog of where the conquest stopped.

The external evidence has two anchors. The Merneptah Stele, dated to about 1208 BCE, names a people called Israel in Canaan. The hieroglyphic determinative attached to the name marks Israel as a people rather than a city or land. That is the earliest extrabiblical mention of Israel, and it is the hard ceiling on any model of Israelite origins. Whatever happened, Israel existed as a recognizable group in Canaan by the late thirteenth century BCE. The second anchor is the highland survey work conducted from the late 1960s through the 1980s. About 250 new small unwalled villages appeared in the central hill country around 1200 BCE, in Ephraim, Manasseh, and Benjamin, where there had been only a handful in the Late Bronze Age. The villages share a building style (the four-room or pillared house), a storage style (the collared-rim jar), and a notable absence of pig bones compared with the coastal Philistine sites of the same period.

Every model has to account for both the literary picture (Joshua plus Judges plus the surrounding patriarchal and exodus traditions) and the empirical picture (the stele, the survey villages, the destruction layers or the lack of them at the named cities). The four positions below differ on which they prioritize and how they read the other.

How the debate developed

The four pieces of evidence that decide it

Four data points sit at the center of the debate. Each model reads each one differently. The table below sets the four positions against the four pieces of evidence so a reader can see, in one view, what each model has to say about each data point.

The Iron Age I village data, in detail

Finkelstein's central-highland survey is the empirical pivot the entire debate has turned around since 1988. The survey covered the territories of Ephraim, Manasseh, and Benjamin systematically, mapping every visible site and dating its surface pottery. The result was a count of around 250 new small unwalled villages appearing in Iron Age I (roughly 1200-1000 BCE), where the Late Bronze Age had only a handful. The villages are small (typically 0.5 to 2 hectares), use the four-room or pillared house plan, store food in distinctive collared-rim jars, and show a notably low frequency of pig bones in their faunal remains compared with coastal Philistine sites of the same period.

The data has two features the models read differently. The first is the timing. The expansion is gradual and runs through the twelfth and eleventh centuries BCE. It does not look like the sudden settlement after a single campaign. The second is the material continuity. The pottery, the building techniques, and the small finds are continuous with Late Bronze Canaanite material culture. There is no toolkit break. The conquest model has to explain why an incoming population did not bring different things. The infiltration and revolt and emergence models, in different ways, treat the continuity as the central piece of evidence.

Joshua 24 in light of the four models

Joshua 24 is the chapter where the choice of god, and so the choice of community, is publicly made. Joshua gathers the tribes at Shechem, recites the LORD's acts from Abraham through the conquest, and asks the people to choose. The chapter assumes a community already gathered, with a shared memory of how it got there. Each of the four models reads that community's prehistory differently.

On the conquest model, Joshua 24 is the covenant renewal of a population that arrived together from outside, sealing in a public ceremony what the campaign in the preceding chapters has secured. The Shechem location is loaded because Abraham and Jacob had been there. The recital from Abraham to the conquest is the community's actual history. On the infiltration model, Joshua 24 is the gathering of pastoralist groups that came in slowly over generations, now publicly aligning themselves around a shared Yahwistic identity. The recital is theological compression, not military history. On the revolt model, the chapter is the constituting moment of the new community: Canaanite peasants choosing the god of the marginal over the gods of the city-state aristocracy they have walked away from. The choice between 'the gods your fathers served beyond the Euphrates' and 'the gods of the Amorites in whose land you dwell' (24:15) is the social break formalized. On the emergence model, the chapter is the later literary projection back of how a Canaanite group came to self-identify as Israel, written when the canonical memory was being shaped in a much later period.

All four readings keep the chapter's plain sense (a public choice of whom to serve) and let the recital do its theological work. Where they differ is what historical event the recital is compressing.

What each model has to account for

Each position has a piece of the evidence it handles well and a piece it has to explain. The conquest model handles the Hazor destruction and the Joshua narrative directly and has to account for the absence of Late Bronze cities at Jericho and Ai and the material continuity of the highland villages. The infiltration model handles the gradual settlement pattern and the pastoralist background and has to account for the destruction layers and the literary specificity of Joshua's campaign. The revolt model handles the material continuity and the social-religious break and has to account for the absence of a documented uprising and the centrality of the exodus tradition. The emergence model handles the village data and the Late Bronze collapse context and has to account for the exodus tradition, the Hazor destruction, and the canonical reception of Joshua as historical campaign narrative.

There is no single model that handles all four pieces of evidence without remainder. That is part of why the debate has run for a century and why the better recent work tends to combine features. Killebrew and Faust both work inside the emergence model but argue for real ethnic markers that distinguish early Israel from surrounding Canaanites. Hess and Hoffmeier work inside the conquest model but acknowledge the highland villages as the resulting settlement pattern rather than insisting on city-by-city destruction. The clean four-position taxonomy is heuristic. Most working scholars hold positions that blend two or three of the models' features.

Reading Joshua with the question open

Joshua 24 is doing something the four models all agree on. It is making the case that this community's identity is bound to a specific story about a specific god. The recital from Abraham to the conquest is the historical prologue of a covenant ceremony, and the ceremony is asking for a public choice. Whether the recital compresses an external arrival, a slow infiltration, an internal revolt, or an emergence inside the same Canaanite world, the chapter's claim is that the community is constituted by remembering that story together. The historical question is real and worth asking. It is not the only question the chapter is asking the reader.