A Reply to Baruch Halpern’s “Radical Exodus Redating Fatally Flawed”

052

“Radical Exodus Redating Fatally Flawed,” BAR 13:06, Baruch Halpern’s swashbuckling response to our article, “Redating the Exodus,” BAR 13:05, was disappointing in its refusal to consider the issues more open-mindedly. In some instances Halpern’s polemic actually misrepresents our arguments instead of engaging them. We are grateful for space to answer bravado with fact.

We argued that the destruction of Canaanite cites at the end of the period known as Middle Bronze II (MB II) occurred around 1420 B.C. instead of the GAD (Generally Accepted Date) of around 1550 B.C. Halpern tries to show that this redating plays impossible havoc with the following period, Late Bronze I (LB I). He says that LB I cannot be compressed into a mere 20 years, and that it cannot end any later than 1400 B.C. (He says the latter is a fact conceded in our paper; actually we avoided raising the issue for the sake of simplicity, and the date can easily be contested, as shown below.) In the course of this argument Halpern makes several misleading statements.

Halpern tries to fix the end of LB I at 1400 B.C. by arguing that the following period, LB II A, cannot have begun any later than that date. Here he appeals to the Amarna letters, correspondence written in the 14th century B.C. between Canaanite city-chiefs and the Egyptian administration. The earliest letter in this collection is probably one dating from the 32nd year of the pharaoh Amenophis III. If this pharaoh reigned 1405–1367 B.C., then Halpern is correct to state that the Amarna letters are from roughly 1375–1345 B.C. Now there is (as Halpern notes) sound stratigraphical evidence to show that LB II A was at least partly contemporary with the Amarna period. But Halpern is going beyond his evidence when he uses it to claim that LB II A “begins no later than 1400” (see his endnote 1). He seems to assume that the towns of LB II A were very different from those of LB I. He speaks of “the great citadels of the Late Bronze II A period,” and of LB II A as the period “during which Canaan developed to the state reflected in the Amarna archive.” Actually, walled towns of any size were rare in both LB I and LB II A, and there was no dramatic increase in the size of most towns in the latter period.¹ In each case where a town mentioned in the Amarna letters has been firmly identified and adequately excavated, there is evidence of occupation in LB I as well as LB II A. In theory, therefore, there is no reason why LB I cannot extend to the very beginning of the Amarna period, or even overlap slightly with it. And there is more.

Halpern cites only one set of dates (the highest) for Amenophis III and his successor Akhenaten. Lower dates for Egypt’s 18th Dynasty are now common currency, and such notable Egyptologists as Kenneth Kitchen are favoring the lowest dates possible. Thus Amenophis III can be dated as late as 1390–1352 B.C., or even (if he shared a coregency with Akhenaten) 1382–1344 B.C.² These low dates would place the beginning of the period covered by the Amarna correspondence at around 1360 B.C. or even slightly later.

In short, there is much more flexibility in dating the end of LB I than Halpern seems to be aware of. It is possible to defend a date as low as about 1360 B.C. on the various grounds discussed above. If MB II ended around 1420 B.C., LB I could therefore have lasted up to 60 years, not the mere 20 assumed by Halpern. While this is much less than the time usually allotted to it (between 100 and 150 years), it is adequate for the LB I strata so far known from various sites.

Let us now turn to Megiddo, which Halpern uses to bolster his argument. Halpern says that the town of stratum IX belonged to LB I and was destroyed by Pharaoh Thutmosis III in 1468 B.C. Here he is taking at face value the dates assigned to the various strata by Megiddo’s excavators in the 1930s. But it is now generally acknowledged that the strata were often crudely defined; and that the correlation between strata in different areas of the tell was sometimes inaccurate. Dame Kathleen Kenyon’s fresh analysis of the Megiddo material led her to conclude that in the excavated area labeled BB, strata IX, VIII and VIIB all belonged before Thutmosis III’s campaign.³ Thus in area BB Kenyon’s revision compresses all three of these strata into about 80 years—whereas Halpern tries to argue that stratum IX alone must have lasted 75 years! Graham I. Davies has recently confirmed Kenyon’s conclusion regarding the Late Bronze Age strata in area BB.⁴ (However, in area AA Kenyon placed stratum VIII in the 14th century B.C., which Davies rejects in favor of a date in the 15th century.) Davies also assigns stratum IX to late MB II in area AA, and he is uncertain how far the stratum’s buildings may have survived through LB I.⁵

If this were not enough to undermine Halpern’s dogmatism concerning the date of stratum IX, Halpern adds to the evidence himself when he tries to pin down the stratum’s end. He asserts that “in a contemporary record” Thutmosis III “recounts … his destruction of Megiddo” in 1468 B.C. There are actually four accounts of Thutmosis III’s campaign against Megiddo, and not one of them mentions the destruction of the town. This point has been made by several scholars over the last 20 years.⁶ One of these scholars, William H. Shea, has shown that the destruction of Megiddo would actually have been contrary to Thutmosis III’s policy, and he concludes: “ … Any interpretation of Palestinian archaeology that utilizes such a chronological correlation should 053probably be revised.”⁷ Halpern is therefore quite wrong to claim that the destruction of LB I Megiddo “is universally attributed to Pharaoh Thutmosis III” and that it “can be absolutely dated to 1468 B.C.”

Halpern next refers to Tell el-Ajjul, describing it as a site where “a town with LB I material culture was destroyed already around 1565 B.C.” Here he is following Aharon Kempinski’s dating of “Palace I” at Tell el-Ajjul, but he does not mention a corollary of adopting this date that Kempinski himself acknowledges; that bichrome ware and associated pottery groups must be dated to start as early as 1600 B.C.⁸ Earlier in his article Halpern treats bichrome ware as “the hallmark of LB I,” which he asserts must begin around 1550 B.C. If Halpern is convinced by early dates for Tell el-Ajjul’s “Palace I” (which some archaeologists reject), he must either redate LB I to begin around 1600 B.C., or he must admit that bichrome ware (and associated pottery types) flourished already in MB II.

Halpern’s argument concerning Alalakh depends on the assumption that stratum IV was the level of King Idrimi and therefore “must have been built before 1468 B.C.” However, the discovery of a statue of Idrimi stratum IV does not prove (as Halpern seems to think) that the stratum was contemporary with his reign, but merely that it cannot have been earlier than his reign. Some archaeologists, placing more emphasis on the pottery evidence, have dated the beginning of stratum IV to roughly the middle of the 15th century B.C.⁹ As stated in endnote 56 to our original article, the dating of Alalakh’s strata will be discussed in some detail in the third edition of Redating the Exodus and Conquest (Sheffield, England: JSOT/Almond Press).

In his comments on Shechem, Halpern overlooks the fact that it was not Bimson and Livingston who found the four building phases of MB II C a tight fit within the century usually allotted to that period. It was William G. Dever (one of the excavators) and Dame Kathleen Kenyon. Even more significant, Halpern completely sidesteps the challange that Manfred Bietak’s excavations offer to the GAD for the end of MB II in Palestine. So far there has been no convincing reply to Bietak’s evidence.¹⁰ If Halpern has one he would do archaeology a service by stating it. Halpern also fails to mention our footnote that draws attention to another important challenge to the GAD.

Halpern next argues that, according to Exodus 1:8 and 2:23, “there were only two Pharaohs” between the building of Pithom and Raamses (Exodus 1:11) and the Exodus event. He says he “rejects” the Bible’s “claim” that the building of Pithom and Raamses occurred shortly before 1450 B.C. This is simply untrue; in a footnote we refer to the possibility “that Exodus 1:11 does refer to activity in the 15th century B.C.” However, we cannot agree with Halpern that the Bible claims this as fact. It is surely obvious that when the Bible narrates historical events it does so selectively and economically, with a motivation that is essentially theological (for example, we noted that the Biblical account of the conquest is not comprehensive). Exodus 1:8 and 2:23 record only those changes of pharaoh that are necessary to explain the turn of events (the Israelites’ changed status is the first instance; Moses’ freedom to return to Egypt is the second). There may have been many more, but it was not necessary for the writer to mention them. When we argue (in the main body of our article) that Exodus 1:11 refers to events centuries before the Exodus, we are suggesting one way in which the Biblical text may be read; we are certainly not rejecting a Biblical “claim.”

Halpern misrepresents us again when he says that Bimson and Livingston “propose that Ai … was unwalled and small.” We do not say it was unwalled, only that we have not found any wall remains at Khirbet Nisya. We mentioned the possibility that these have been removed by agricultural activity in later periods. We do assume Ai was small, because (a) Joshua’s spies reckoned that two or three thousand men were enough to take it, since its people “are but few” (Joshua 7:3), and (b) Ai is said to be smaller than Gibeon (Joshua 10:2); ancient Gibeon had an area of about 15 acres, so Ai should be smaller than that. Khirbet Nisya could have had an occupied area of six or seven acres, which would be about right for the 1,200 inhabitants that we suggested (as an amendment to the 12,000 referred to in Joshua 8:25; see endnote 45 to our article). We leave it to the reader to decide whether our reasoning is as arbitrary as Halpern claims.

When Halpern turns his attention to Transjordan he tries to force more misleading arguments on his readers. On the basis of Deuteronomy 2:10–12, 20–21 he states that the peoples of Ammon, Moab and Edom totally eradicated the Amorites, suggesting that 054this should be marked archaeologically by a change in material culture. In fact the Bible refers to Amorite peoples continuing to live in Transjordan alongside the Ammonites, Moabites and Edornites. Thus Sihon is described as “king of the Amorites” (Numbers 21:21–32), and Og of Bashan belonged to “the remnant of the Rephaim” (Deuteronomy 3:11), a people Halpern assumes were Amorites. If Amorite culture was distinct from that of the Ammonites, Moabites and Edomites, then we should find obvious differences in material culture between their respective territories. This is not the case.

Halpern’s own date for the emergence of new populations in Transjordan is also problematic if he insists that the event must be marked by a change of culture. He says archaeology points to “the settlement of Transjordan by Moab, Ammon and Edom toward the end of the late Bronze Age.” He presumably has the conclusions of Nelson Glueck in mind. But recent studies have modified Glueck’s conclusions in several respects. There was no sudden crystallization of settlement along the King’s Highway in the 13th–12th centuries B.C. In Moab, for instance, there is only a gradual increase in the number of settlements during the Late Bronze and Iron Ages; as J. Maxwell Miller comments: “Certainly there is nothing here that will provide us with a precise date for the emergence of the Moabite kingdom. … ”¹¹ If one insists on archaeological evidence for the arrival of the Moabites, the 13th century B.C. is really no better than the 15th. Identifying Iron Age I culture in Transjordan with the arrival of these peoples is really no help to Halpern’s case.¹² There are Egyptian references to Edom and Moab from the 13th century B.C., whereas Iron Age I begins no earlier than 1200 B.C. and probably half a century later.¹³ There is also considerable continuity between the Late Bronze and Iron Age cultures of Transjordan,¹⁴ making it difficult to affirm that the latter signals the arrival of new peoples.

Before leaving the problems of Transjordan, we must ask how Halpern explains references to Dibon in Egyptian texts from the 15th and 13th centuries B.C. if he supports its location at Dhiban (where the earliest remains are from the second half of the 12th century B.C.).

Now we come to Halpern’s misunderstanding of our view of Israelite 055settlement. He says that we “hypothesize an Israelite presence in the lowlands” and “discard the history of Israel’s settlement given in Joshua and Judges (with the Israelites settling first in the highlands). … ” In fact we refer to “cave burials away from cities” as a strong indication of a seminomadic population, suggesting that such may be indirect evidence of Israel’s presence. We did not locate this population in the lowlands, and the evidence does not require us to do so; the Late Bronze Age cave burials occur chiefly in the highlands. We do believe, however, that there was contact between the seminomadic highlanders and the lowland cities, and our evidence for this is the continuity between LBA culture and that of the Iron Age settlements that flowered in the highlands after 1200 B.C. As Volkmar Fritz remarks: “The results of archaeological research indicate that early Iron Age Culture was highly dependent upon Late Bronze Age Culture. … ”¹⁵

In case we did not make our position clear in our article, it is that the breakdown of the Late Bronze Age city-states led to a redistribution of settlement and to the Israelites abandoning a seminomadic lifestyle.¹⁶ Our view is actually very close to that recently argued by Fritz.¹⁷ In proposing what he calls the “symbiosis hypothesis” of Israel’s origins, Fritz provides an answer to another of Halpern’s criticisms. Halpern says that our theory requires “invisible Israelites” in Canaan during the Late Bronze Age, since he sees no evidence for their presence until the start of the Iron Age. Fritz, on the other hand, notes that the character of Iron Age settlements points to “intensive, prolonged contact” between seminomadic Israelites and Canaanite culture before the start of the Iron Age; “This contact must have already occurred in the Late Bronze Age before the beginnings of sedentary life.”¹⁸ In other words, Fritz sees the evidence demanding an Israelite presence in Canaan during the Late Bronze Age.

Finally, Halpern tries to ridicule our reconstruction by contrasting a picture of Israel as a “mighty ethnic group,” victorious over Canaan, with our view of Israel borrowing from and assimilating to Canaanite culture. But surely the Book of Judges (and even parts of Joshua) shows that Israel did not totally vanquish the Canaanites, and that Israelites did assimilate to Canaanite culture. When Halpern asks: “Where are Israel’s memories of this long period (200 years or more) of servile dormancy?” the answer is: “In the Book of Judges.” There Israel is anything but a “mighty ethnic group,” but is frequently weak, idolatrous, servile and internally divided.

Much more could be said in reply to Halpern’s article, but I hope I have dealt with enough of his objections to show that our hypothesis is a great deal stronger than he is prepared to admit.

“Radical Exodus Redating Fatally Flawed,” BAR 13:06, Baruch Halpern’s swashbuckling response to our article, “Redating the Exodus,” BAR 13:05, was disappointing in its refusal to consider the issues more open-mindedly. In some instances Halpern’s polemic actually misrepresents our arguments instead of engaging them. We are grateful for space to answer bravado with fact. We argued that the destruction of Canaanite cites at the end of the period known as Middle Bronze II (MB II) occurred around 1420 B.C. instead of the GAD (Generally Accepted Date) of around 1550 B.C. Halpern tries to show that this redating plays impossible havoc […]