Radical Exodus Redating Fatally Flawed

056

In the September/October BAR, John Bimson and David Livingston wrote an article entitled “Redating the Exodus,” BAR 13:05, in which they radically revise a number of generally accepted dates and conclude that the Exodus occurred in the latter half of the 15th century B.C. instead of in the 13th–12th centuries. Part of their argument maintains that the invading Israelites were responsible for the destruction of cities in Canaan at the end of the Middle Bronze Age. The generally accepted date for the end of the Middle Bronze Age, however, is about 1550 B.C. Bimson and Livingston therefore move this date down more than a hundred years in conformance with their proposed late 15th-century date for the Exodus.

Herewith is a response to Bimson and Livingston’s article.—Ed.

Nothing more gratifies a scholar’s ego than siring a revisionist theory. To revolutionize a field, to sweep away all the conformist clutter of a whole academic era is to join the ranks of Einstein, Freud, Pasteur.

Not all the members of this company, however, are so distinguished. For every Einstein, there is a Vladimir Lysenko, for every Freud, an Erich von Däniken. Revisionism is a boom-and bust occupation.

John Bimson and David Livingston (hereafter, B&L) are excellent in revisionism: The Biblical account of the Israelite conquest of Canaan, they say, can be salvaged, against the drift of all recent scholarship on the subject. True, they confess, there is no evidence of a unified invasion of Canaan in the late 13th or early 12th centuries B.C., when it is generally agreed the Israelites emerged in Canaan. In the late 13th and early 12th centuries B.C., many of the sites the Book of Joshua claims Israel conquered were uninhabited. But, say B&L, we can match the Biblical account of the conquest to the archaeology of earlier times: Move the conquest back 200 years, and postdate by 150 years an archaeological stratum with destroyed cities, and presto, they actually coincide! (See chart.)

Their first step—antedating the conquest—is easy for them. The Biblical account of the conquest was written late in the seventh century B.C. and fails to link the conquest to any event that external sources permit us to date. But the Biblical account does contain references to times that place the conquest in the 15th century (e.g., the Exodus occurred 480 years before Solomon started to build the Temple [1 Kings 6:1]). So, by taking elementary precautions against skepticism about the Biblical text, by pressing one’s eyelids down tightly on the cheekbone, one can pretend that the Book of Joshua is the unvarnished, untarnished truth, and that it all occurred in the 15th century B.C.: Israel conquered Canaan in a single decisive campaign.

B&L do not go so far as to claim—against George Gershwin’s immortal song—that “Methuselah lived 900 years” or that “Jonah, he lived in a whale.” Indeed, they do not even argue, as Joshua reports, that the Jordan River parted (Joshua 3:13), that Jericho’s walls “came a-tumblin’ down” (Joshua 6:20), or that the sun screeched to a celestial halt for the span of an entire day (Joshua 10:13). These kinds of miracles are not the stuff of workaday historical reportage. B&L do not stop to ask how such miracles found their way into an otherwise reliable record. Conversely, how did the other details of Joshua’s campaign avoid the taint of such exaggeration? If the texts themselves are late, as B&L concede when explaining away the use of the names Pithom and Raamses from a later period, what kinds of sources did their authors have about the early history of Israel? Were they parroting absolutely reliable recollections, handed down from generation to generation without a single fumble? Or is it more likely that the Biblical authors were doing their best to reconstruct the past from a jumble of invention, poetic or folkloristic glorification, and legitimate oral tradition?

What B&L have done is to accord unquestioning credulity to their own—highly idiosyncratic—reading of the Biblical conquest accounts. They have then brought the more malleable archaeological data into line with this reading. This is easier than wrestling with questions of historiography, the way in which history is written and constructed. But it simply 057doesn’t work!

B&L correlate the Israelite conquest with the end of Middle Bronze II. Using Biblical dates, they place the Exodus at around 1450 B.C., and the conquest, starting 40 years later, about 1420–1400 B.C. So B&L bring the end of MB II down from 1570 B.C., where most scholars place it, to 1420 B.C.

But this creates a major problem for them: They must compress the succeeding archaeological era, Late Bronze I, into just 20 years or so. This is because the period after Late Bronze I, namely Late Bronze II, cannot under any circumstances begin much after 1400 B.C. If LB II begins about 1400 B.C. and MB II ends (as B&L argue) at 1420–1400 B.C. (when the Israelites supposedly destroyed the MB II C cities in Canaan), this leaves, at a maximum, 20 years for LB I.

B&L do not dispute the fact that LB II begins about 1400 B.C. This is a measure of how firm that dating is. The basis for the dating is sketched out in an endnote.¹

If the beginning of LB II cannot be budged past 1400, and if, as most scholars recognize, the previous period (LB I) spanned roughly 150 years, then this necessarily places the end of MB II (and the destruction of the MB II C Canaanite cities) at about 1550 B.C. If this is correct, B&L have misdated the end of MB II by a century or more, and the Israelite armies missed their tryst with all those destruction layers by at least as much.

B&L have blithely turned their backs on the chronological chaos their theory wreaks on LB I. They have borrowed 100 to 150 years from LB I, and offer no prospect of paying off the debt.

The tells of Canaan all but compete with one another to clamor against B&L’s theory that LB I can be compressed into 20 years or less. Unfortunately, the evidence is complicated, and I shall provide here only a few of many possible examples.

Let’s look at Megiddo first. Megiddo stratum VIII represents Megiddo of the LB II A, the city of the Amarna era.^a It was therefore built no later than 1400 B.C. (and may well have been built considerably earlier). Beneath stratum VIII is stratum IX. This is the LB I city. So far there can be no argument. B&L concede that LB II started no later than 1400 B.C. Megiddo stratum VIII exhibits all the characteristics of an LB II A city. Stratum IX, below it, exhibits all the characteristics of an LB I city, in particular a plethora of so-called bichrome ware, the hallmark of LB I.

A word about bichrome ware.² Bichrome ware is a form of pottery with two-tone painted decoration and with geometric, and often animal, motifs, probably originally imported from Cyprus.³ In any significant quantity, it is diagnostic of the LB I period in Canaan. Although not the only indication of LB I, bichrome ware is certainly an important one. LB I is also marked off from MB II C by a number of changes in pottery forms. Several MB types go out of use in LB I. And, of course, Cypriot imports burgeon; bichrome ware either first arrives in LB I or enters an era of extremely broad distribution. And then bichrome ware vanishes by the end of LB I.

At Megiddo, stratum IX is full of bichrome ware, so it is absolutely clear that this is an LB I city.

This LB I city at Megiddo ended in total devastation. This destruction is universally attributed to Pharaoh Thutmosis III, who in a contemporary record recounts his campaign into Canaan and his destruction of Megiddo.⁴ This Egyptian campaign that destroyed Megiddo can be absolutely dated to 1468 B.C.

Already B&L are in considerable trouble. The LB I city of Megiddo was destroyed in 1468 B.C. Yet, according to B&L, LB I began in 1420–1400 B.C. If LB I did not begin until 1420 B.C., an LB I city could not be destroyed 50 years earlier, in 1468 B.C.

But this is not the worst of it. If the LB I city of Megiddo was destroyed in 1468 B.C., it must have been built much earlier; consequently LB I must have started much earlier, and certainly not in the period 1420–1400 B.C.

How much earlier? It turns out that the LB I city at Megiddo (stratum IX) contained two major architectural phases. If we allow 37 ½ years per phase (this is the number of years B&L suggest in their analysis of Shechem), then we are back to 1543 B.C. for the beginning of LB I, about the time archaeologists almost universally place it—about 1550 B.C., which marks the end of MB II (and the city destructions that B&L would like to attribute to the Israelites), as well as the beginning of LB I (which B&L would like to date to 1420–1400 B.C.).

In short, LB I must begin somewhere around 1550 B.C., and cannot begin as 058late as 1420 B.C.

This analysis of the Megiddo evidence can be repeated on numerous sites, including Gezer, Tel Mevorakh, Lachish, Alalakh and especially Tell el-Ajjul, to name a few. The evidence at Tell el-Ajjul is exceptionally complex, and does not lend itself to presentation in this context. But here, it is almost certain that a town with LB I material culture was destroyed already around 1565 B.C.⁵

Let us look briefly at only one more site—Alalakh. We know that Alalakh stratum IV must have been built before 1468 B.C. This is because Alalakh stratum IV contained the famous statue of King Idrimi; from the cuneiform inscription on this statue, we know that Idrimi mounted the throne at the dispensation of the king of Mitanni; but the king of Mitanni only had this power before 1468 B.C., when Pharaoh Thutmosis III overcame the coalition of Syrian kingdoms at Megiddo and grabbed off northern Syria for Egypt.

Below stratum IV at Alalakh are two LB I layers. If we allow 37 ½ years for each of these two LB I cities, we are again forced back to about 1550 B.C. for the beginning of LB I (and for the end of MB II).

All this evidence has two important implications. First, links between stratified ancient texts (such as the Amarna letters and the text on the statue of Idrimi) and excavated strata at numerous sites such as Megiddo and Alalakh all converge on the point that MB II cannot have ended much after 1550 B.C. or so. LB I obviously began immediately thereafter. Second, LB I, which ends about 1400 B.C., cannot be compressed into a period of 20 years—or even 50 years—as B&L would like to have it.

Against all this, B&L appeal to evidence from one site only—Shechem—in their effort to expand MB II C to a period of 150 years (1570–1420 B.C.). In one field at Shechem, they observe, there are four building phases; this required 150 years, they say, or 37.5 years per phase. If we apply this same logic to the numerous sites of which Megiddo and Alalakh are only examples, what must have occurred in the 20 years that B&L leave for LB I?

Answer: More than what B&L claim required 150 years at Shechem—at Megiddo, four major building phases, each with several resurfacings; at Alalakh, four major building phases in strata VI and V, and probably the construction of stratum IV. These examples could be multiplied if we had the space.

To this stratigraphic data may be added all the evidence from Egyptian epigraphy and from dated scarabs, which place Egyptian kings from Ahmose I through Thutmosis III in LB I; there is no evidence that permits us to put them in MB II C, or to put Amenophis IV in LB I. There is only one way to evict the end of MB II C from the 16th century: B&L must reshuffle all of Egyptian chronology, astronomical fixed-points included.⁶

But even looking solely at the Biblical account, B&L do not fare well. The Biblical text itself does not unequivocally support a conquest in the late 15th century B.C.

B&L make it seem as if the Biblical text is on their side by taking a smorgasbord approach to the Bible. They pick their Biblical texts arbitrarily and never justify their omissions. Their argument therefore must be regarded as either naive or disingenuous.

For example, B&L observe that Pithom and Raamses were inhabited in the 19th through 17th centuries B.C. They suggest that these cities were built at the start of Israel’s enslavement, and the fact that they were unoccupied in the late 15th century B.C. implies no inaccuracy in the text. They gloss over the following problem this plea of theirs creates: Exodus 1:11 says that Pithom and Raamses were built when a “new king” of Egypt arose, “who ‘knew’ not Joseph” (Exodus 1:8). This Pharaonic persecutor of the Hebrews dies in Exodus 2:23, and a new Pharaoh (the Pharaoh of the Exodus) ascends the throne. Thus, the text suggests that Israel built Pithom and Raamses about a generation before the Exodus, not, as B&L suggest, hundreds of years earlier. B&L ignore their own implied rejection of this Biblical datum. But this is not the worst of it. Let’s take the Biblical text as uncritically as B&L want us to. If there were only two Pharaohs between the time when, according to B&L, the Hebrews were forced to build Pithom and Raamses by the Pharaoh who knew not Joseph and the time when, according to B&L, the Hebrews left Egypt to escape the Pharaoh of the Exodus, then only two Pharaohs ruled between 1700 B.C. and 1450 B.C. This evades the historical data in the same measure as it defies logic.

B&L accept the claim that Israel built Pithom and Raamses, but (silently) reject the claim that this occurred shortly before 1450. How, except arbitrarily, do B&L figure out what to swallow and what to leave? B&L treat the text like a brunch buffet, picking what they like (i.e., what suits their theory) and leaving the rest.

Or consider the case of Ai, which B&L suggest is located at Khirbet Nisya instead of at et-Tell, although et-Tell better matches the description in Joshua 8:28. Joshua 7–8 makes it clear that Ai was heavily fortified: It had a “gate” (Joshua 7:5), and Joshua is able to breach its security only by decoying the soldiery from the town. Yet B&L have found no gate or fortifications at Khirbet Nisya. Where else in Israel have such imposing fortifications been removed, as B&L propose at Khirbet Nisya, along with every building from a town housing 12,000 people (Joshua 8:25)? And how did 12,000 people shoehorn into Khirbet Nisya—high-rise condominiums?

B&L propose that Ai—taking it to be Khirbet Nisya—was unwalled and small. In short, they reject the Biblical account here, but then rely on the tradition that Joshua needed 3,000 troops to take it, and of course on the tradition that Joshua actually took it. B&L assume that the figure of 12,000 inhabitants and the tradition of heavy fortification are unreliable. But they 059rely on “true” elements (the 3,000) to suit their theory. How do they divine which Biblical number was accurate and which was inaccurate?

These cases illustrate that B&L treat the Bibilical text arbitrarily; they also show that the Biblical testimony is itself rather ambiguous.

The situation in Transjordan also presents B&L with insurmountable difficulties. The Bible is clear: En route from Egypt to Canaan, Israel encountered a strong Edomite kingdom (Numbers 20:14–21), an Ammonite nation (Numbers 21:24), and a king of Moab (Numbers 22–24)—all located in Transjordan.

B&L claim that Israel encountered Ammon, Moab and Edom in the MB II settlements of Transjordan. But no text mentions Ammon, Moab or Edom before the 13th century B.C. On what basis do B&L conclude that the MB II settlements in Transjordan represent those nations? According to Biblical tradition, these Transjordanian nations were, like Israel, formed elsewhere, and they later displaced peoples in Transjordan—they “eradicated” the Amorites (Deuteronomy 2:10–12, 20–21).^b Israel too supplanted the Amorites.

B&L reject the obvious conclusion, surely implied in the Biblical tradition, that these Transjordanian nations crystallized along the King’s Highway in Transjordan at the same time as Israel in Cisjordan—when the Egyptian empire began to recede in the 13th–12th centuries B.C.

B&L assume that Ammon, Moab and Edom eradicated the Amorites by the end of MB II, so that these nations were there to resist the Israelites’ passage through their land. That these Transjordanian nations were there in the 15th century B.C. is contradicted by archaeology as well as by the Biblical text. The material culture of MB II–LB I Transjordan is absolutely continuous.⁷ Did these Transjordanian nations, like Israel in Cisjordan (according to B&L), conquer an area only to be immediately enslaved by the Amorites? That seems to be B&L’s naive suggestion. They are explicit as to the Israelites: They must account for the fact that although the Israelites, on their own hypothesis, conquered Canaan in about 1400 B.C., Israel leaves no archaeological record, according to B&L’s own chronological reckoning, until 200 years later. Did the Edomites, Ammonites and Moabites, too, remain archaeologically invisible for 200 years?

We can agree that there were settlements in MB Transjordan. But what suggests that the Edomite, Ammonite and Moabite kingdoms were up and running at that time? B&L produce no evidence—only hollow assertions based on their arbitrary excerpts from the Biblical text. All the archaeological evidence—poorer Iron I culture, the lack of settlement at Dibon and Heshbon (important cities, according to the Bible, in Moab and Ammon) in the Late Bronze Age—converges on a settlement of Transjordan by Moab, Ammon and Edom toward the end of the Late Bronze Age. Against B&L, the ruin of Dhiban is probably Dibon, and Hesban, Heshbon; there is no need to imagine that these names, or those of Arad or Bethel, “wandered,” a phenomenon so liberally invoked by B&L that one wonders that any site in Canaan actually retained its own name.

There is simply no evidence of Moabite, Ammonite or Edomite kingdoms in MB II or LB I or LB II A. This is admittedly an argument from silence. But it is a silence of about 300 years duration. Is it preferable to accept this argument from silence, or is it better to imagine, as B&L do, invisible kingdoms of Ammon, Edom and Moab skulking around pretending to be Canaanite during MB II?

B&L find evidence to make their case, but do no justice to the many points that contradict it. They arbitrarily accept congenial elements of the Biblical tradition and dismiss those that are intractable. They ignore the complexities of the literature, and never stop to determine what the character of any particular narrative is—after all, there are gradations of historical literature: from the apocryphal Davy Crockett, who killed a ‘bar’ at the age of three; to Walt Disney’s Davy Crockett, that giant of the wilds; to the David Crockett known through documents to American historians. B&L treat the Bible like the first when it is troublesome, and like the last when it is not. Their sole criterion for the choice seems to be expediency.

The idea that the Ammonites, Moabites and Edomites hid for hundreds of years behind a facade of Canaanite material culture (throughout the Late Bronze Age) is sufficiently improbable in itself to undermine B&L’s radical reconstruction of history. But there is more to come.

On the Israelite side of the Jordan, the evidence is just as devastating to B&L’s theory, which attempts to place the Israelites there in the late 15th–early 14th centuries B.C.

In Iron Age I, there is an explosion of population in the hill country of Canaan. This begins with the regions best suited to a pastorally based economy, and expands into hill districts that invite a mixed and then a cash-crop economy.⁸ These Iron I hill settlements are distinctive—not that the pottery or the housing stock is unique, but the mix of them is distinctive and stands in stark contrast to Canaanite plains sites in the Late Bronze Age or in Iron I.

These hill settlements can be identified by virtue of their continuity into Iron II, when texts and epigraphs indicate that Israel inhabited them. From the start of Iron I, then, the Israelites can be located on the ground. Their counterparts in Ammon, Moab and Edom exhibit the same material culture at the same time.⁹

The logical inference is that Israel and these other states crystallized in Transjordan and the Canaanite interior as migrants from Syria, and, in the face of Assyrian advances (and taxation), homesteaded the land along the trade routes into the Canaanite interior, in a gradual process culminating at the end of the 12th century B.C.

Instead, B&L suggest that the Israelites conquered Cisjordan in the 15th century B.C. and lived in symbiosis with the Canaanite fortified cities of the plains. Two internal difficulties beset this hypothesis.

First, the Biblical accounts claim that Israel supplanted the denizens of the hills (Joshua 10–11). Joshua 17:14–18 states that the Israelites first settled the hilly regions and could not move into the lowlands. But hill settlements are few 060until Iron I. Faced with this difficulty, B&L hypothesize an Israelite presence in the lowlands (where most of the LB cemeteries unconnected with settlements are found).^c But this conflicts with the Biblical tradition that Israel first settled the highlands. Again B&L’s treatment of the Biblical text is arbitrary: Did the Israelite conquerors remember a conquest, but simultaneously forget that they settled originally in the plains?

According to B&L, the chronology of 1 Kings 6:1^d and the conquest story of Joshua 10–12, both of which conflict with the archaeological and historical data, are correct; but they discard the history of Israel’s settlement given in Joshua and Judges (with the Israelites settling first in the highlands), which matches the archaeological evidence precisely. This is a perverse approach to reconstructing the past.

The second internal problem that B&L evade is this: Biblical accounts preserve no record of Egyptian domination in Canaan, a domination that lasted from LB I to Iron I, as excavations and texts such as the Amarna letters make clear. How is it that the Bible can place the conquest in 1410 B.C., but suffer from amnesia about Egyptian domination in Canaan from 1410 to about 1140 B.C.?¹⁰ B&L’s smorgasbord approach is again evident.

The Transjordanian nations of Ammon, Moab and Edom had distinctive onomastica (repertoires of personal names). In the onomastic evidence from the 16th to the 12th centuries, there is not a single indication that Ammonite, Moabite or Edomite elements are present in the names. And there are no names compounded with theophoric elements mentioning the Transjordanian national gods,^e such as Qaus or Chemosh, as we find in later periods.

The onomasticon in the area of Canaan occupied by the Israelites also gives B&L trouble. The Amarna archive (Late Bronze Age) contains hundreds of personal names, but not one includes as a theophoric element the name of the Hebrew God, YHWH. Even in exile in Babylon, a high percentage of Jews—probably half—had names with Yahwistic theophorics. Why are there no Yahwistic names in any of the many 13th-century B.C. sources? The indication is that Israel was probably absent from Canaan.

One last point deserves mention. B&L suppose that Israel conquered MB II Canaan, but occupied the lowlands for 200 years, without enjoying any period of dominion or superiority. But the Late Bronze sites are generally smaller than those of MB II, and they 061suggest nothing like an influx of population (unlike the situation in Iron I—beginning about 1200 B.C.—with the founding of scores of new hill-country sites). So here is Israel, according to B&L, buying pots from, and assimilating to, local Canaanite culture—a culture they have vanquished and should be ruling. Their theory simply makes no sense. Where is this mighty ethnic group (the Israelites) in any text from the pre-Iron I period? Where are these conquerors archaeologically, who, in a fit of apparent masochism, made obeisance to the population they had just decimated? Where are Israel’s memories of this long period (200 years or more) of servile dormancy? B&L’s invisible Israelites, those stalwarts who so jammed the historical radar as to efface every trace of their presence for 200 years, are a vapor. B&L’s theory is as insubstantial as its subjects.

B&L’s smorgasbord approach is attractive because it masquerades as a defence of the Bible. But it is not. B&L dismiss as much Biblical evidence, in the end, as they embrace—as they do the archaeological evidence, picking and choosing. Their textually arbitrary, historically unconvincing, archaeologically improbable hypothesis hides its warts behind a veneer of benevolent piety. Piety has its place, no doubt, but it also has its price. And the going price for B&L’s piety is about 200 years of Israelite history.

In the September/October BAR, John Bimson and David Livingston wrote an article entitled “Redating the Exodus,” BAR 13:05, in which they radically revise a number of generally accepted dates and conclude that the Exodus occurred in the latter half of the 15th century B.C. instead of in the 13th–12th centuries. Part of their argument maintains that the invading Israelites were responsible for the destruction of cities in Canaan at the end of the Middle Bronze Age. The generally accepted date for the end of the Middle Bronze Age, however, is about 1550 B.C. Bimson and Livingston therefore move this […]