I reread the introduction to Baruch Levine's Anchor Bible commentary on Numbers 1-20. Here are some points:
1. I've wondered why the Pentateuch has stories about the Israelites' rebellion against Moses. On page 64, Levine offers his answer:
"The JE tradition explains the forty-year delay in arriving at the Promised Land as God's punishment for Israel's lack of trust in him. The Israelites doubted his power and reliability, which is a traditional way of interpreting more realistic factors that would actually have caused the postponement of the conquest, or that might explain the refusal of the Israelite leadership to endorse a policy of conquest."
Does this imply that the Conquest is historical? It must, if an explanation was contrived to account for the Israelites' delay in carrying it out.
Levine also appears to take another passage in Numbers as historical, on page 93:
"There is a reason to assume from the narrative of Num 10:29-32 that Hobab and his tribe responded positively to Moses' persuasion and decided to join the Israelites in their journey to the Promised Land. This background would have laid the foundation for actual relations between the two groups in the settlement period and the early monarchy, when Kenites would have been helpful to the Israelites and at the same time protected by them, and exempted from the decreed fate of the other Canaanite peoples."
Levine seems to assume that, when a narrative can elucidate later situations in history (and he must take the Israelite relationship with the Kenites to be historical), then it is historically accurate.
2. But Levine doesn't believe that all of Numbers is historically accurate. Levine dates JE to the time between the tenth-seventh centuries B.C.E. During this time, Kadesh-Barnea "served as the hub of a network of fortifications in southern Judah", as the Kingdom of Judah sought to "control the southern border of Judah and the territories south of it" (page 90). Consequently, under the Judahite monarchy (and maybe even the United Monarchy), Kadesh was narrated to be a prominent Israelite site during the wilderness period. Numbers 20:14-21 presents Kadesh as bordering Edom, and that was not true in the Late Bronze Age (the literary setting of Numbers 20), but in the early eighth century B.C.E. Moreover, for Levine, JE's negative portrayal of various countries in Exodus and Numbers serves "to sanction or to justify later Israelite policies toward them" (page 97). Levine's argument is that JE largely reflects the time of the monarchy of first millennium B.C.E. Israel.
But why aren't the Arameans in JE, since they were a prominent nation in the first millennium B.C.E.? Levine isn't entirely satisfied with the solution that "some of the J and E traditions" go back to the time before Aram threatened Northern Israel and the Transjordan, for "the Edomite factor is included in the JE record (Num 20:14-21; 21:4), in a manner that recalls the eighth century"---which was after Aram rose as a threat (page 99). Levine's solution is that "Like Philistines, the Egyptians and Arameans were foreign elements in Canaan, and there need have been no sensitivity about doing battle with them or about seizing land from them" (page 99). So Levine's answer is that these peoples were unrelated to the Conquest, and so they went largely unmentioned in JE?
I want to note: conservative scholar Kenneth Kitchen argues against those who dispute the historical accuracy of the Pentateuch's statement that there were kings in Edom and Moab during the Late Bronze Period---on the grounds that there can be no kings without sedentarization and a degree of urbanization. Kitchen attempts to refute this by referring to kings who lived in tents (see pages 196-197 of here). But Numbers 20:17 refers to Edomite vineyards and fields. How sedentary (or, rather, non-sedentary does Edom have to be for Kitchen's argument to hold up?
3. Levine also makes points about P. He states that P "went to great lengths to keep the Israelites in Sinai rather than elsewhere for most of the wilderness period," perhaps "to retain geographic proximity between the formative sinaitic theophany and the revelation of the elaborate laws and rituals they had retrojected into the wilderness period" (pages 56-57). While P "reported the arrival of the Israelites only in the fortieth year (Num 20:1, 33:38)" (page 56), "the JE tradition scheduled the arrival of the Israelites in Kadesh near the beginning of the forty-year wilderness period" (page 68). So we have two different stories about Israel's journey.
While Hurvitz argues that P's language is pre-exilic, Levine refers to the term issar ("ban") in Numbers 30, which occurs in the fourth century Samaria Papyri from Wadi Daliye, and "occurs nowhere else in the Hebrew Bible but, significantly, it does occur in the Aramaic sections of Daniel---repeatedly in Daniel 6, in its final form a product of the second century" (page 108). Levine dates the term to the fifth-fourth centuries B.C.E., and, because the term is so integral to Numbers 30, he does not see it as a later interpolation; rather, he concludes that Numbers 30 was composed in the fifth-fourth centuries, the time of Israel's post-exilic period.
Levine believes that P still uses old terminology to create a flavor of antiquity (page 107), and that P "most probably took shape over a protracted period of time, beginning in the late preexilic period; it preserved some quite early material and continued to develop during the postexilic period" (page 104). For Levine, P itself has duplicates, as it "preserves more than one record of the erection and dedication of the Tabernacle" (page 66), one of them being in Numbers 9. And Levine claims that there is "internal development within priestly literature itself" (page 106), since Numbers P builds on "the priestly traditions of Exodus and Leviticus", by assigning the sanctuary an encampment of priests, Levites, and Israelites. So P was not monolithic, according to Levine.