Redating the Exodus

040

Among Biblical scholars and archaeologists it is almost axiomatic that the Israelites entered Canaan about 1230–1220 B.C. In terms of archaeological periods, this would be towards the end of the Late Bronze Age, for which the Generally Accepted Date (GAD) is 1550–1200 B.C.

Yet there are enormous problems with this dating. In recent decades an increasing number of scholars have recognized that if we accept the GAD of 1230–1220 B.C. for the Israelite entry into Canaan, we must reject the Biblical account of Israel’s conquest of Canaanite cities. This is because the Biblical account conflicts so strongly with the archaeological record. The Bible describes the Israelite conquest of Canaan at length and refers to a number of cities encountered by Joshua and his armies. In almost every case the archaeological evidence is inconsistent with the Biblical evidence—if we date the Israelite entry into Canaan to the GAD of 1230–1220 B.C.

Jericho was the first city encountered by Joshua and the Israelites when they crossed the Jordan (Joshua 2 and 6). According to the Bible, the Israelites conquered and destroyed Jericho. But according to the archaeologists—and the site has been very extensively excavated—there was no city at Jericho in 1230–1220 B.C. for the Israelites to destroy. Indeed, no trace of occupation at Jericho has been found between about 1300 B.C. and the 11th century B.C.,¹ the probable date of the earliest Iron Age remains.

Ai was the next city on the Israelite route. The Bible gives a detailed account of the battle of Ai that led to the city’s destruction (Joshua 7–8). Despite extensive excavations at the site commonly identified as Ai, the archaeologists have discovered no evidence of occupation between about 2400 and 1200 B.C. About 1200 B.C., a small unwalled village grew up on the site, lasting until about 1050 B.C.²

At Gibeon, with whose people Joshua made a treaty, according to the Bible (Joshua 9), no Late Bronze Age city has been found.³ James B. Pritchard, Gibeon’s excavator, commented that the anomalies encountered at Jericho, Ai and Gibeon “suggest that we have reached an impasse on the question of supporting the traditional view of the conquest with archaeological undergirding.”⁴

But this is just the beginning.

The Bible tells us that Joshua gave Hebron to Caleb at the time of the conquest (Joshua 14:13–15, 15:13–14; Judges 1:20). At Hebron, excavations in the 1960s produced only scanty remains from between the end of the Middle Bronze Age (c. 1550 B.C.) and a late phase of Iron Age I (11th century B.C.) and no evidence of occupation in the 13th century B.C.⁵

Hormah/Zephath figures as a flourishing town at the time of the conquest (Numbers 21:1–3; Judges 1:17), yet excavations at the site thought to be Hormah/Zephath (Tel Masos) have revealed a fortification from the Middle Bronze II period (c. 1900–1550 B.C.) and an Iron Age (12th century B.C.) settlement,⁶ but no intervening occupation.

On their march to Canaan, the 041Israelites were opposed by the king of Arad (Numbers 22:1, 33:40), yet Tel Arad was abandoned from the end of the Early Bronze Age (third millennium B.C.) until the Iron Age.⁷ No Late Bronze Age settlement was found. The late Professor Yohanan Aharoni’s suggestion that Canaanite Arad lay at Tel Malhata, about eight miles to the southwest of Tel Arad, does nothing to solve the problem, for there is no evidence of a settlement between the end of the Middle Bronze Age (c. 1550 B.C.) and Iron Age II (tenth century B.C.).

Thus at six sites that figure in the Biblical account of the conquest, we have little or no evidence of Late Bronze Age occupation. What we have is six cities that did not exist at the time to which Joshua is conventionally assigned.

There are further problems. Debir, one of the cities conquered by Israel (Joshua 10:38–39, 15:15–19; Judges 1:11–15), is now almost certainly to be identified with Khirbet Rabud, as Moshe Kochavi argues, rather than with Tell Beit Mirsim, as William F. Albright once urged.⁸ While there was certainly an important Late Bronze Age town on the site, there is no evidence for its conquest in the late 13th century B.C.

At Lachish, another town taken by the Israelites (Joshua 10:31–32), there was indeed a destruction at the end of the Late Bronze Age, which Albright dated about 1230–1220 B.C. and attributed to the Israelite invaders.⁹ However, recent excavations have led the excavator, David Ussishkin, to redate this destruction to about 1150 B.C. or even slightly later,¹⁰ so the end of Late Bronze Age Lachish can no longer be treated as evidence for a conquest by Israel in the late 13th century B.C.

Even this does not exhaust the list of cities mentioned in the conquest traditions, but the remainder cannot be discussed profitably here, either because identification is uncertain or because there has not yet been adequate excavation. But even if every one of the remaining sites produced a perfect match with the traditions, there would still be more problem-cities for a 13th-century conquest than good correlations with the Biblical narratives. Of the various cities said to have been conquered by the Israelites in Joshua and Judges 1, only Hazor and Bethel (Beitin) have destruction levels datable to the second half of the 13th century B.C.¹¹ Two destroyed cities hardly amount to evidence for a conquest, especially when there is no evidence that their attackers were the Israelites. These destructions may have resulted from any one of a number of other causes, for example, Egyptian campaigns or local intercity warfare.

This admittedly poor “fit” between Biblical tradition and archaeological evidence is universally recognized by scholars, the majority of whom nevertheless accept a date of 1230–1220 B.C. for Israel’s entry into Canaan.

How do they deal with the problem?

One approach, which has gained considerable support in mainstream scholarship, is to explain Israel’s emergence in Canaan by processes other than conquest—that is, by thoroughly rejecting the Biblical account. Among these alternative views is the “peaceful infiltration” theory, long favored by German scholars. Another, newer proposal is the “peasant revolt” theory 042advocated by George E. Mendenhall and Norman K. Gottwald.¹² According to this, what the Bible describes in terms of an Israelite conquest was in fact a revolt of local peasants against the urban centers that previously dominated them.^a

We believe there is a better solution, one that does more justice to the Biblical traditions, which we would like to present to BAR readers.¹³ Our solution requires us to make two radical chronological adjustments, which we will discuss in turn.

The first is simple: Move the date of the conquest back about 200 years, to shortly before 1400 B.C. Although this conflicts with the GAD for Israel’s emergence in Canaan, it is in fact the date implied by the Bible itself.

In 1 Kings 6:1, we are told that Solomon began building the Temple in the fourth year of his reign and that this was 480 years after the Exodus. Solomon’s reign can be dated with considerable confidence to about 971–931 B.C., so the fourth year of his reign would be 967 B.C. According to the Biblical chronology, this would place the Exodus 480 years earlier—about 1447 B.C., or say 1450 B.C. for convenience. If we allow 40 years for the desert wanderings before the Israelite conquest of Canaan, we arrive at a date of about 1410–1400 B.C. for the Israelite entry into Canaan. This is almost 200 years earlier than the GAD of 1230–1220 B.C.

Another Biblical text—Judges 11:26—indicates that the Israelites had been settled in Transjordan for 300 years by the time of Jephthah, one of the Judges. Jephthah, by common agreement, can be dated to about 1100 B.C. This would place Israelite settlement east of the Jordan 300 years earlier—about 1400 B.C., again, almost 200 years before the GAD.

Admittedly, the figure of 480 years used in 1 Kings and the 300 years used in Judges sound like approximations, round figures rather than precise calculations. The figure 480 is perhaps especially suspect; it is 12 times 40, a frequently used figure in the Bible. Yet, we doubt whether these figures should be dismissed as meaningless, as they have been by most critical scholars. As we will show below, the reasons why scholars originally preferred the GAD to the Biblical chronology have all been undermined in recent years. In view of this, it is reasonable to take up the Biblical chronology once again as a working hypothesis, i.e., to test it against the non-Biblical evidence to see how it fares.

If the GAD for Israel’s conquest of Canaan provides such a poor fit between the Biblical account and the archaeological record, how did 1230–1220 B.C. become the GAD in the first place? The answer is tied up with the dating of the Exodus. Because the GAD of the conquest is linked with the GAD for the Exodus, we need to understand how this was determined.

Since the early decades of this century, three main arguments have been built up for dating the Exodus to the 13th century B.C. We will examine these in turn.

We are told in Exodus 1:11 that the Israelites enslaved in Egypt “built for Pharaoh store-cities, Pithom and Raamses.” This statement has been used to date the Exodus in two ways. First, the very name Raamses recalls the Egyptian name Ramesses, borne by one of Egypt’s most illustrious Pharaohs, Ramesses II (the Great), who reigned from 1290 to 1224 (or 1279–1213) B.C. This Pharaoh did indeed build a royal residence-city called Pi-Ramesse in the eastern Nile Delta. Exodus 1:11 has therefore been taken to indicate that the Israelites cannot have left Egypt before the reign of Ramesses II. However, the Bible does not use the name Raamses with chronological rigor, as is evident from its use in Genesis 47:11. In this verse the name (spelled Rameses) is used to describe the region in the Delta in which the patriarch Jacob and his sons settled. Since no scholar dates the original Hebrew settlement in Egypt as late as the 13th century B.C., it is acknowledged that here a name in common use for the area at a later time is being used retrospectively (just as a modern historian might write of Julius Caesar crossing “the English Channel”). The same explanation may apply to the use of Raamses in Exodus 1:11.

Is there evidence that building activity went on at the site of Raamses at an earlier time than the reign of Ramesses II? Yes there is, though this evidence does not prove that work was going on around 1450 B.C. (the Biblical date for the Exodus). The evidence of earlier occupation at Pi-Ramesse dates to the 19th–17th centuries B.C. This was when a major administrative center was first developed on the site. Archaeologists have been unearthing the long history of this important city 043since the 1930s.¹⁴ It lies in the Khata’na-Qantir district of Egypt’s Eastern Delta, in the region of Goshen, where the Israelites initially settled (Genesis 47:4–6). Indeed, the culture excavated from this early period has a strong Syro-Palestinian element, closely resembling that of contemporaneous Canaan (Middle Bronze II). We suggest this reflects the migration of Canaanite groups (including the Israelites) into Egypt—groups that were subsequently pressed into slavery to work on the sites of “Pithom and Raamses.” In other words, we associate the building of these two cities with the beginning of Israel’s enslavement, not with the eve of the Exodus several centuries later.

Let us now turn to the evidence regarding Pithom. There are two possible sites for Biblical Pithom: Tell el-Maskhuta and Tell er-Retabah. These two sites lie about eight miles apart in the Wadi Tumilat, west of Lake Timsah. We need not debate which of the two should be identified with Pithom. The important point is this: The same Syro-Palestinian (Middle Bronze II) culture which marks the early period at the site of Raamses has now been found at both these candidates for Pithom as well.¹⁵ At Tell el-Maskhuta (the site favored for Pithom by the majority of scholars), the early remains include probable grain-storage facilities, perhaps explaining the term “store-cities.”

Archaeologists and Egyptologists have traditionally held a different view of this Syro-Palestinian culture in the Eastern Delta. They associate it with the forerunners of the Hyksos, a Semitic people from somewhere in the Syria-Palestine region who took over the throne of Egypt around 1650 B.C. and ruled the land for at least 100 years. There may be some truth in this. But it is also plausible to suggest that there were other Semites among the new settlers, as well as the ancestors of the later Hyksos. There is indeed evidence from Tell el-Maskhuta that some Semitic settlers were treated with brutality by the Hyksos. The MB II finds at Tell el-Maskhuta include the tomb of a woman and her dog, both killed by blows from a type of battle-axe used by the Hyksos. We suggest that these other Semitic settlers were (or at least included) the Israelites, whom the Hyksos treated as slaves—perhaps following an example already set by the Egyptians.¹⁶

In short, the reference to “Pithom and Raamses” in Exodus 1:11 cannot be used to date the Exodus to the 13th century B.C. Rather, the archaeological evidence makes best sense if Exodus 1:11 refers to the beginning of the Israelites’ enslavement (in about the 18th century B.C.), and not to the time of the Exodus.^b

Another traditional reason for dating the Exodus to the 13th century B.C. involves archaeological evidence from Transjordan that more recent scholarship has undermined.

Surface surveys of Transjordan, carried out by Nelson Glueck chiefly in the 1930s, led Glueck to conclude that much of the region was without a settled population during the Middle Bronze II and Late Bronze periods, that is, between the 19th and 13th centuries B.C. The traditions preserved in the Bible (Numbers 20–22) require the existence of strong kingdoms in Edom and Moab (as well as farther north; see Numbers 21:21–35) at the time when Israel was moving northward through this region on her final march to Canaan. This led many scholars to 044conclude that Israel’s encounter with Edom, Moab and the other kingdoms east of the Jordan could not be dated before the 13th century B.C. This has become a standard line of argument for dating the Exodus and conquest to that century and no earlier.

However, Glueck’s conclusion has been heavily modified during the last 30 years, although many authorities have been slow to recognize the fact. Further surveys and excavations have brought to light numerous Middle Bronze II and Late Bronze settlements that were missed by Glueck. For example, a survey in 1963–66 found 18 sites from the Middle Bronze II period, and almost as many from the Late Bronze I–II periods; a survey in 1975 discovered 14 Middle Bronze II and six Late Bronze sites; more sites from both periods were found during a further survey in 1978.¹⁷ J. Maxwell Miller reports that the findings of a recent survey of central Moab “seem to indicate at least a scattering of settlements even during the Middle Bronze Age which gradually increased in number during the Late Bronze and Iron Ages.”¹⁸ Miller goes on to comment: “Certainly there is nothing here that will provide us with a precise date for the emergence of the Moabite kingdom or for the Israelite Exodus from Egypt.” Significantly, at sites where Glueck’s own investigations were more intensive than usual, he often found pottery from the Middle Bronze II and Late Bronze periods, but when he initially published his conclusions he did not give these finds the emphasis they deserved. Glueck did, however, revise his views shortly before he died.¹⁹

The case of Dibon deserves a separate discussion. Excavations at Dhiban, the supposed site of the Moabite capital Dibon, have shown that no town existed there before Iron Age I (about 1200–1000 B.C.). Yet Dibon is mentioned in Numbers 21:30, 32:3, etc., as a town existing in the time of Moses. While some scholars have been inclined to solve the problem by treating the Biblical references as anachronisms, this cannot be correct because there is contemporaneous evidence for Dibon’s existence at an earlier time. Dibon is almost certainly mentioned in an Egyptian topographical list from the reign of Thutmosis III (1490–1436 B.C.), and it definitely occurs in a text of Ramesses II (1290–1224 B.C.) thus proving its existence at least as early as the Late Bronze Age.²⁰ The explanation for this discrepancy between the written evidence (including the Bible) and the archaeology of Dhiban is probably that Bronze Age Dibon was located at another site and is so far undiscovered.

The same may apply to Heshbon, another town that existed in the time of Moses according to Numbers 21:21–31, etc. Excavations at Tell Hesban have produced no remains from before about 1200 B.C.; it is possible, however, that the Heshbon of earlier periods should be sought at Tell Jalul, a large unexcavated mound farther to the south, with Middle and Late Bronze Age surface sherds.²¹

The important point, which has been reinforced again and again in recent decades, is that Glueck’s initial conclusions were definitely wrong, and it is disappointing to find scholars citing them as if they were still valid evidence against an early date for the Exodus.²² All too often the 13th-century date for the Exodus has been perpetuated by the baseless repetition of outmoded views.

The view that Israel arrived in Canaan at the end of the Late Bronze Age (about 1230–1220 B.C.) is now so entrenched that Biblical scholars and archaeologists regard the culture of the Iron Age as “Israelite” in contrast to the “Canaanite” culture of the Late Bronze Age. This prejudges the issue and encourages a circular argument: some have used the appearance of “Israelite” culture in the late 13th century B.C. to support their dating of the conquest/settlement!

There is actually no unequivocal proof that the culture of Iron Age I (1200–1000 B.C.) should be associated with the arrival of the Israelites. Indeed, with the exception of the introduction of Philistine ware in the coastal region, there is sufficient continuity between the cultures of Late Bronze and Iron Age I to make it doubtful that the latter marks the arrival of any newcomers at all.²³ More important, as Patricia M. Bikai recently wrote:

“The shift from the richness of LBA [Late Bronze Age] to the apparent poverty of early Iron I is a phenomenon which extends into the whole of Canaanite (and eastern Mediterranean) culture and not just that section affected by the arrival of the Hebrews.”²⁴

So whatever the reason for the widespread cultural change at the start of the Iron Age (climatic change and socioeconomic decline have been suggested), it cannot be attributed to the 045Israelites.²⁵

We are not saying, however, that Iron Age I culture should not be associated with the Israelites at all. We are simply denying that it marks their arrival. It is our contention that the Israelites were already in Canaan during the Late Bronze Age. Along with other peoples of the eastern Mediterranean, they participated in the changes which marked the transition to the Iron Age, but they cannot be held responsible for them.

We have seen that there is no solid evidence for placing the Exodus and conquest (or settlement) in the 13th century B.C. We are therefore justified in experimenting with the neglected Biblical date of roughly 200 years earlier.

This is not the place to discuss the timing of the Exodus relative to Egyptian history, and we will simply refer the reader to other scholars who have recently dealt with this in terms of a 15th-century date. William H. Shea has offered an attractive scenario (though not without problems) for a date of 1450 B.C., at the end of the reign of Thutmosis III; Hans Goedicke’s theory, involving a setting in the co-regency of Thutmosis with Hatshepsut, about 1471 B.C., has already been well represented in the pages of BAR.²⁶ Both dates are subject to some variation, since more than one dating scheme is possible for Egypt’s 18th Dynasty, to which Hatshepsut and Thutmosis III belonged. In addition, other scenarios are conceivable.²⁷

But redating the Exodus is only one part of the problem. What about the conquest? If we redate the conquest to shortly before 1400 B.C., do we then find a good match between the Biblical account and the archaeological evidence? If we accept the dates currently given to archaeological periods, the answer still has to be “No.” The end of the 15th century B.C.—the date we propose for the Israelite conquest of Canaan—provides no more archaeological evidence for Israel’s conquest of Canaanite cities than does the GAD of 1230–1220 B.C. But we have a further proposal to make, which radically changes the picture.

Before outlining our second proposal, we need to give some explanation of archaeological periods. Scholars have divided ancient history into sequential periods, based on observed cultural discontinuities in the archaeological record. Thus, we have the Bronze Age followed by the Iron Age. The evidence supports the discontinuity between the two periods, without requiring us to provide absolute dates or to assign absolute lengths of time to each period. With accumulating evidence these archaeological periods have been subdivided—again based on observed discontinuities in the archaeological record—and absolute dates have been assigned to the beginning and end of each subperiod. This chronology has become quite nuanced and in general has attained wide scholarly acceptance, although there are still a number of areas of disagreement, some small and some major.

We would suggest a change in the date for the end of the period archaeologists designate Middle Bronze II (MB II, for short). We would move the end of MB II down by over a century, from around 1550 B.C. to around 1420 B.C. Since the Late Bronze Age began when the Middle Bronze Age ended, this means that the beginning of the Late Bronze Age is likewise redated from about 1550 B.C. to about 1420 B.C. (No change in the date for the end of the Late Bronze Age is required; it still ends at the GAD of around 1200 B.C., and Iron Age I begins at that time.)

The result of this adjustment is that events that occurred at the end of the Middle Bronze Age must have happened around 1420 B.C. instead of 1550 B.C. When we date the end of MB II to this later time, we find an almost perfect correlation between the archaeological evidence and the Biblical account of the conquest of Canaan.

The MB II period was characterized by strongly fortified cities. In this respect it fits exactly the Biblical description of Canaan at the time of the conquest. The complaint of Moses’ spies was that the cities of Canaan were too well fortified to be conquered—“fortified and very large” (Numbers 13:28); “great and fortified up to heaven” (Deuteronomy 1:28). The narratives of the conquest provide further confirmation that the Israelites were confronted by cities with walls and gates (see Joshua 2:15, 6:1, 5, 7:5, 8:29, 10:20, 14:12, etc.). Because of these fortifications Joshua avoided straightforward siege warfare, which offered little hope of success, and employed a variety of other tactics instead.^c

While the Biblical picture conforms perfectly to the situation in MB II, it does not ring true at all as a description of the Late Bronze Age. (Remember that the currently prevailing scholarly view is that Israel entered Canaan at the end of the Late Bronze Age, that is 1230–1220 B.C.) A recent study by Rivka Gonen has revealed that, contrary to the assumption of many scholars, most Late Bronze Age cities were unwalled settlements, much smaller than their MB II predecessors.²⁸

Moreover, many of the walled cities 046of Canaan were destroyed at the end of MB II.

Jericho was a large, fortified city during MB II, and was destroyed by fire at the end of that period (cf. Joshua 6:24). At Gibeon and Hebron there were cities during MB II that were followed in each case by a gap in occupation. If we accept Yohanan Aharoni’s suggestion that Canaanite Arad was located at Tel Malhata, there was an MB II city here as well, followed by a gap in occupation. Similarly with Hormah, if we identify it with Tel Masos. Indeed, in the case of Tel Malhata and Tel Masos, Aharoni himself recognized that the situation described in the Bible²⁹

“corresponds exactly to the situation during the Middle Bronze Age, when two tels, and two tels only, defended the eastern Negeb against desert marauders, and the evidence points towards the identification of these tels with the ancient cities of Arad and Hormah.”

Aharoni described this as “a most startling conclusion,” yet he did not take the logical step of placing the conquest at the end of MB II; instead he simply wrote that:³⁰

“the Biblical tradition preserves a faithful description of the geographical-historical situation as it was some three hundred years or more prior to the Israelite conquest.”

Returning to our list of cities, Lachish and Hazor were both large cities in MB II, and both fell to attackers at the close of the period, as did Bethel if we accept its identification with Beitin. The very limited excavations at Debir (Khirbet Rabud) did not uncover Middle Bronze Age remains, but a few MB II sherds were found on the surface.³¹ These may well indicate the existence of a settlement in that period, and we hope more thorough excavations will be undertaken at this important site, including investigation of the lower slopes, to clarify the picture.

At Taanach, Megiddo, Gezer, Beth Shean and other cities that Israel failed to take from the Canaanites (Judges 1:27ff), the transition from the Middle to the Late Bronze Age occurred without the breaks in occupation attested elsewhere. Indeed, generally speaking, the area in which destructions occurred at the end of MB II corresponds with the area of Israelite settlement, while cities that survived lay outside that area.³²

Thus we find a remarkable correlation between Biblical tradition and archaeological evidence when we place the conquest at the end of MB II.

The only exception seems to be Ai. Ai lacks both Middle Bronze and Late Bronze remains—so it is a problem for both the conventional date of 1230–1220 B.C. (the end of the Late Bronze Age) and our proposed date of about 1420 B.C. (the end of the Middle Bronze Age). But we believe we have a solution to this anomaly. This brings us to our final major proposal. In our view, the site of Ai has been misidentified.

Ai is almost universally identified as Khirbet et-Tell (hereafter simply et-Tell). But we believe the identification is erroneous. It rests principally on two lines of argument. The first is the supposed correspondence between the Biblical name Ai and the modern name of the ruin, et-Tell. Et-Tell means “the tell,” that is, a ruin mound in modern Arabic. In the Bible, the site Joshua conquered is always called ha-Ai, “the Ai.” Ai has long been taken to mean “ruin,” so the site even in Biblical times was supposedly known as “The Ruin.” Many Biblical scholars have assumed that the modern Arabic name of et-Tell is effectively a translation of the name Ai (taken to mean “ruin”). However, Ziony Zevit has recently demonstrated a serious objection to this equation and has concluded that “any connection between the Hebrew name and the Arabic name for the site is to be rejected.”³³ Ai, he says, cannot be related to the word for ruin.^d

The second basis for identifying Ai with et-Tell is its proximity to Bethel. It is clear from a number of Biblical references that Ai lay close to Bethel, to the east of it (Genesis 12:8; Joshua 7:2, 8:9, 12, 12:9). Bethel has been traditionally located at Beitin. In the area east of Beitin, there are only three sites other than et-Tell that might qualify as Ai (Khirbet Haiyan, Khirbet Khudriya and Khirbet Haiy), but all were ruled out by soundings carried out in the 1960s during Joseph Callaway’s excavations at et-Tell; they produced 047only late remains. Thus, Callaway and many others have concluded that et-Tell is the only possible site for Ai.

But there is a fallacy in this reasoning—Bethel itself has been incorrectly located. Beitin was first identified as Bethel in 1838 by the American topographer and Biblical scholar Edward Robinson. He published his conclusion in 1841 and it has generally been accepted ever since. Robinson put forth two reasons for his identification. One was the name itself, Beitin, which he saw as an example of a well-attested phenomenon—the shift from Hebrew lamed (l) to Arabic nun (n) in the termination of names. Thus he recognized Beitin as the expected Arabic equivalent of Bethel.³⁴ The name alone, however, cannot be taken as proof because it is well known, as William F. Albright stated long ago, that “names of towns and villages are frequently displaced over a considerable local area.”³⁵ There are long gaps in the tradition of Bethel’s location, notably a gap of 1,400 years between about 400 A.D. and the first attested use of the name Beitin in the early 1800s. On this basis alone, it is quite possible that the name underwent a geographical shift.³⁶

Robinson’s other reason for the identification of Beitin as Biblical Bethel concerns the evidence of the patristic authors Eusebius (269–339 A.D.) and Jerome (345–419 A.D.). In the Onomasticon of Eusebius, which Jerome revised and amplified, it is stated that Bethel lay “at [or near] the twelfth Roman milestone from Aelia [Jerusalem]” on the east side of the road to Neapolis (Shechem-Nablus). In 1838 Robinson estimated the distance between Beitin and Jerusalem by the time it took him to make the journey on horseback³⁷ and concluded that Beitin lay the correct distance north of Jerusalem. We should not scoff at Robinson’s method of measuring distance; it was the only means available to him at the time. But it certainly left much to be desired! Measuring the distance with a modern odometer, following the Roman road almost the whole way and taking into account Roman milestones discovered since Robinson’s researches, modern measurements place Beitin at approximately the 14th milestone, not the 12th. (A Roman mile equals 1,618 yards, compared to an English mile of 1,760 yards.) In short, the patristic evidence which Robinson used to support his identification is actually against it. Beitin lies too far north of Jerusalem.

There are other objections to the equation between Beitin and Bethel. According to Genesis 12:8, a mountain lay to the east of Bethel, between Bethel and Ai.^e But there is a valley east of Beitin instead of a mountain. Robinson himself admitted that in the vicinity of Beitin “there is no major summit or hill.”³⁸ This does not conform at all to the situation described in Genesis 12:8.³⁹

According to the Bible, Bethel lay on the border between Benjamin and Ephraim (Joshua 16:1–3, 18:11–13). However, the natural geographical border, to which the Bible lists otherwise conform, lies south of Beitin. If Beitin is Bethel, the border requires an abnormal bulge northward at that point. Modern historical geographers have acknowledged the anomaly: “The boundary lines east and west of the central section are definitely not so far north as the town of Beth-El [Beitin],” writes a prominent modern Israeli scholar.⁴⁰

The location of Bethel must be sought farther south, nearer Jerusalem. We suggest that the modern town of Bireh, east of Ramallah, stands on the ancient site of Bethel.⁴¹ In favor of this, we note the following:

1. Modern odometer measurements place the 12th Roman milestone about 550 yards north of Bireh, considering that the “0” milestone was located well inside the Damascus Gate, perhaps as much as a full half-mile. Thus Bireh fits the patristic evidence that Bethel was at or near the 12th milestone.

2. There is a prominent mountain east of Bireh. Its name, Jebel et-Tawil, means “the tall one” and it can be seen clearly at a great distance from any direction. Thus the topography suits Genesis 12:8, where Abraham is said to have built his altar to the Lord on this mountain.

3. Bireh lies exactly on the natural east-west geographical border, the border between Benjamin and Ephraim, satisfying that requirement also.⁴²

No excavation has been done at Bireh, and modern occupation seems to exclude it for the foreseeable future. But a surface survey of Ras et-Tahuneh, the highest point in the 048town, has produced pottery from the Chalcolithic, Early Bronze, Middle Bronze II, Iron I, Iron II, Persian and Arabic periods.⁴³ Therefore, there should be no doubt that this area was the site of a town in centuries past, just as it is today—including those centuries when Bethel flourished, according to the Biblical references.

A new location for Bethel also presents a new possibility for the location of Ai. About one mile southeast of Bireh lies the small site of Khirbet Nisya,⁴⁴ which we propose as the site of Biblical Ai. The topography suits all the requirements. The mountain Jebel et-Tawil, mentioned previously, lies between Bireh and Khirbet Nisya (Genesis 12:8); a valley lies to the north of Khirbet Nisya, with a hill on the far side (Joshua 8:10–13). The valley descends into Wadi Suweinit, which in turn goes on down to Jericho (cf. Joshua 7:2, etc.), and is steep enough to qualify as the “descent” (Hebrew, morad) mentioned in Joshua 7:5. About a mile below, the gorge narrows to little more than a split in the rocks, with sheer cliffs on either side; this feature may be the “Shebarim” (breaks) referred to in Joshua 7:5 (though other interpretations could be suggested with the environs of Khirbet Nisya in mind). To the west of the site there is a ridge where a force could wait in ambush (Joshua 8:9, 12–13), without being seen from either Khirbet Nisya or Bireh. Furthermore, the site is a small one, and is therefore appropriate for a town that Joshua’s spies said could be conquered by only 2,000 or 3,000 men, its people were so few (Joshua 7:3).⁴⁵

In 1970, when the proposal to locate Ai at Khirbet Nisya was first published, no excavations had been done at the site. In 1973 Roy B. Blizzard did a thorough pottery survey over about half the surface and concluded that there had been no occupation there earlier than the Iron Age. This conclusion has been cited by various critics as definitely ruling out Khirbet Nisya as the site of Ai.⁴⁶ But Blizzard’s results have been invalidated by subsequent surveys and excavations.

In 1978, when an excavation permit was requested for Khirbet Nisya, the Israel Department of Antiquities surveyed the site (of which they had no previous record) and reported Chalcolithic (fourth millennium B.C.) and Early Bronze I (third millennium B.C.) sherds among those found on the surface, along with those of later periods. Since then, there have been six short seasons of excavations (1979, 1981, 1982, 1984, 1985 and 1986) and the results leave no doubt that there was a settlement there in the Middle Bronze II period as well. This was already clear by the end of the 1981 season, which produced Middle Bronze II sherds from a wide range of vessels and also a typical MB II five-ribbed dagger. The 1985 season added considerably to the repertoire of diagnostic MB II pottery, some of which is illustrated here.⁴⁷

So far no building remains have been found that can be attributed to the Middle Bronze II period. The search for such remains is difficult because of the nature of the site. Khirbet Nisya, like many other hill-country sites, has suffered a great deal from the ravages of time and human activity. At the top of the hill bedrock is frequently exposed and is never far from the surface, while lower down the slopes, large areas have been converted into agricultural terraces. An abundance of pottery from many periods (Iron Age I and II, Persian, Hellenistic, Herodian/Early Roman, Byzantine and Early Arabic, as well as MB II) occurs in these terraces, and it is clear that the people who created the terraces (in the Byzantine period and later) removed building stones and ancient occupation levels from farther up the hill to build their terrace retaining walls and to provide 051the fill behind them.

However, the pottery is so plentiful that it provides a clear picture of the site’s history. Khirbet Nisya was occupied throughout MB II. And occupation ceased at the time of the MB II/LB transition.

Moreover, it appears now there was no occupation at Khirbet Nisya after MB II until the Iron Age.⁴⁸ If Joshua conquered the site at the end of MB II, this termination of settlement at the very end of MB II is exactly what we would expect. The later history of the site is also what we would expect of Biblical Ai. There must have been a town at Ai by the end of the Iron Age, because “men of Bethel and Ai” were among the returnees after the Exile (Ezra 2:28; Nehemiah 7:32); Ai must have existed in 587 B.C. (when the Babylonians conquered Judah) for these people to have been deported from it! In the days of Nehemiah (445 B.C., i.e., in the Persian period), Ai was resettled along with Bethel (Nehemiah 11:31). (These Biblical references, incidentally, conflict with locating Ai at et-Tell; et-Tell was never resettled after 1050 B.C.⁴⁹ Pottery evidence indicates Khirbet Nisya was occupied during the Iron Age. It was again occupied from the Persian period until the first century A.D. (as we know from both pottery and coins), after which it was apparently abandoned again until the Byzantine period (c. 350 A.D.). This gap fits with the testimony of Eusebius that Ai was deserted in his day (c. 330 A.D.). The latest significant occupation seems to have been in the Arabic period (c. 640–800 A.D.).⁵⁰

Further excavations are planned at Khirbet Nisya, which may well throw more light on all these periods. But it should already be apparent that Khirbet Nisya is a far better candidate for Ai than et-Tell.

Let us sum up our argument so far. We have proposed: (1) a return to the Biblical date for the conquest of Canaan (i.e., shortly before 1400 B.C.), and (2) a lowering of the date for the end of the Middle Bronze Age, from 1550 B.C. to shortly before 1400 B.C. The result is that two events previously separated by centuries are brought together: the fall of Canaan’s MB II cities becomes the archaeological evidence for the conquest. These twin proposals create an almost perfect match between the archaeological evidence and the Biblical account. The only outstanding anomaly is Ai, because of the lack of MB II remains at et-Tell. But even this problem is removed by our further proposal that Ai should be relocated.

We still need to show, however, that the end of the Middle Bronze Age should be redated by over a century. There are two questions here: (1) Why has the end of MB II been dated to about 1550 B.C.? (2) Is there any independent evidence for our drastic reduction of this date?

The answer to the first question is that the end of MB II has been fixed at about 1550 B.C. by its association with an event in Egyptian history. As we have seen, the end of MB II is marked by the widespread destruction and abandonment of major Canaanite cities. The destroyers of these cities have conventionally been identified as the Egyptians, pursuing their Hyksos overlords after driving them out of Egypt. The date for the expulsion of the Hyksos and the beginning of a new dynasty of native Egyptian kings (the 18th Dynasty) is fixed by Egyptologists at around 1550 B.C. Hence this has become the date for the fall of Canaan’s MB II cities, and therefore the date when the Middle Bronze Age ended and the Late Bronze Age began.

But recently scholars have become increasingly critical of attributing the destruction of all these sites to the Egyptians. We are now beginning to recognize that the evidence for an Egyptian destruction is very scanty—a few scraps of inscriptional evidence which are suggestive at best. Furthermore, the prominent Egyptologist Donald Redford has recently pointed out that at the start of the 18th Dynasty—the first Egyptian dynasty after the expulsion of the Hyksos—the Egyptians were simply not capable of besieging fortified cities throughout Canaan.⁵¹ Noting the lack of evidence for such Egyptian campaigns in Canaan, William H. Shea has also written of the need to find alternative destroyers for the Middle Bronze II cities.⁵²

If the Egyptians were not the destroyers of these cities at the end of the 052Middle Bronze Age, their destruction may well have occurred later than 1550 B.C. And indeed there is now strong evidence for placing the end of MB II much later than the expulsion of the Hyksos from Egypt.

The evidence comes from recent excavations in Egypt. Manfred Bietak of the Institute of Archaeology of the University of Vienna and the Austrian Archaeological Institute has been conducting excavations at a site called Tell el-Dab’a since 1966. This is a mound in the Khata’na-Qantir district, that is, the region of Biblical Raamses. Here Bietak found Middle Bronze Syro-Palestinian pottery associated with Egyptian finds of the Hyksos era. After analyzing this material, Bietak has suggested that the period known as Middle Bronze II B—one of three subdivisions of MB II—be lowered by roughly a century; instead of MB II B’s conventional dates of 1750–1650 B.C., he proposes dates of 1650–1570 B.C. This, in turn, displaces the final phase of Middle Bronze II, known as MB II C, which must now start about 1570 B.C. instead of about 1650 B.C. as was previously held.⁵³ In light of this, the conventional end-date for MB II C, namely 1550 B.C., cannot be retained. Even the later end-date of about 1500 B.C., now favored by some archaeologists, is impossibly early. In light of the evidence from Tell el-Dab’a, we must lower the date for the end of MB II C (the last phase of MB II)—and the period of the destruction of the Canaanite cities—well into the 15th century B.C. But can we be more precise?

Helpful evidence for dating the end of MB II comes from Shechem, in Canaan itself. At Shechem, four major building phases have been dated to MB II C. As Dame Kathleen Kenyon pointed out, it seems impossible to compress all these four building phases into the time normally allowed for the period (1650–1550 B.C.). Her solution was to assign the last building phase to the Late Bronze Age instead.⁵⁴ But the attribution of all four phases to MB II C is now proven beyond doubt, and the phasing is therefore described by William G. Dever as “tight.”⁵⁵ So, logically, rather than compress MB II C into less than the century generally allotted to it, we should allow it considerably longer than a century. Indeed, some archaeologists are now inclined to allow about 150 years for the period. Combining this length of time with Bietak’s new date of about 1570 B.C. for the start of MB II C, we arrive at a date of about 1420 B.C. for its end. These are the dates we provisionally propose for MB II C here.⁵⁶

Thus, about 1420 B.C. becomes the suggested date for the Israelite conquest of Canaan. The Exodus should be placed 40 years earlier, around 1460 B.C. This is in harmony with the Biblical data.

Finally, we would like to forestall two objections which may be leveled at our reconstruction.

We have suggested that the destruction of Canaanite cities at the end of the Middle Bronze Age represents the Israelite conquest. But a number of sites were destroyed at this time that the Bible doesn’t even mention. In the Bible, only 14 cities are said to have been conquered by Israel (though 31 defeated kings are listed in Joshua 12), whereas almost 40 cities were destroyed or abandoned at the Middle Bronze/Late Bronze transition.

It is not difficult, however, to account for these “surplus” destructions within the framework we are proposing. There are three possibilities, which should be seen as complementary rather than mutually exclusive.

First, cities that fell at that time but that are not said to have been taken by the Israelites could have been destroyed by Egyptian campaigns. Thutmosis III (1490–1436 B.C.) and Thutmosis IV (1412–1402 B.C.) both campaigned in Palestine. Some of the “surplus” destructions may perhaps be attributed to them.

Second, cities not mentioned in the conquest traditions may nevertheless have been destroyed by the Israelites, since the Biblical account is probably not a comprehensive one; for example, Joshua 12:17 and 24 mention the defeat of the kings of Tappuah and Tirzah, both in the central hill country, though no battles in that region are described in either Joshua or in the first chapter of Judges.

Third, some cities may have been overthrown by Canaanite elements making common cause with the Israelites, or at least exploiting the upheaval that the invaders brought, in order to pursue their own ends. The conquest traditions make it plain that some elements in Canaan did align themselves with the Israelites (Joshua 2:1–21, 9:3–10:7; Judges 1:22–25), and, while the “peasant revolt” theory of Israel’s origins has overstressed this aspect, it cannot be left out of account altogether.^f

Another problem that may be posed to our revised chronology is that the enormous increase in hill-country settlements in Canaan, which has conventionally been identified with the coming of the Israelites, did not occur until about 1200 B.C. Where, our critics may ask, is the evidence for Israel’s presence in the land after the MB II destructions of about 1420 B.C.—and before 1200 B.C.?

The Israelites had lived in Egypt as enslaved pastoralists, and then spent 40 years as seminomads before entering Canaan; this makes it unlikely that they brought a distinctive material culture into Canaan. They would have 053bought their pottery at the local Canaanite markets, so there would be no distinctively “Israelite” pottery. They arrived as tent-dwellers (Joshua 3:14, 7:21–24, 22:4–8, etc.) and that lifestyle was probably maintained for a considerable time by those who were not assimilated into Canaanite society (as some were, e.g., Judges 1:29, 32). Recent work by Rivka Gonen shows that the MB II destructions were followed by a period of reduced urban population, while cave burials away from cities seem to indicate the presence of a significant seminomadic population.⁵⁷ In indirect evidence such as this we can expect to find Israel’s arrival in Canaan reflected. It is a mistake to expect the sudden appearance of a distinctive “foreign” material culture to mark the event.⁵⁸

At the end of the Late Bronze Age and the start of the Iron Age, around 1200 B.C., a major change occurred in settlement patterns, with new settlements appearing in hill-country areas. Contrary to the views of some earlier scholars, these are not evidence of nomadic groups beginning to settle down. In the opinion of most scholars today, the new settlements mark a shift of the existing population away from the old lowland cities of the Bronze Age. This shift was made possible by new technological developments, including iron tools for clearing forests, and slaked-lime cisterns for storing water. While we do not believe the new settlements mark the arrival of the Israelites, we are still happy to call them “Israelite” settlements. This is because, in our view, the Israelites had been in the land for some two centuries by 1200 B.C. and were therefore involved in the changes that took place at that time.⁵⁹

We would not claim that our proposed 15th-century chronology for the Exodus and conquest is entirely without problems. What we do say is that our proposal presents far fewer problems than the conventional 13th-century chronology. As we noted earlier, James B. Pritchard wrote in the 1960s that scholarship had “reached an impasse on the question of supporting the traditional view of the conquest with archaeological undergirding.” The response to the impasse has been to 066reconstruct Israel’s origins in ways that do not involve the Biblical picture of a conquest at all. But this will not do. As Abraham Malamat observed:⁶⁰

“A basic element of Israelite consciousness is that Canaan was ‘inherited’ by force. This tenet is like a leitmotif that runs through the Biblical sources.”

Consequently Malamat states that, even though some embellishment of the account may have occurred, “at the core, a military conquest remains.” We believe that the revised chronology we have proposed does justice to both the Biblical picture of the conquest and the archaeological record. We offer it as an alternative to be considered and explored.

A response to this article by Baruch Halpern of York University, Toronto, Canada, will appear in a subsequent issue [“Radical Exodus Redating Fatally Flawed,” BAR 13:06].

Among Biblical scholars and archaeologists it is almost axiomatic that the Israelites entered Canaan about 1230–1220 B.C. In terms of archaeological periods, this would be towards the end of the Late Bronze Age, for which the Generally Accepted Date (GAD) is 1550–1200 B.C. Yet there are enormous problems with this dating. In recent decades an increasing number of scholars have recognized that if we accept the GAD of 1230–1220 B.C. for the Israelite entry into Canaan, we must reject the Biblical account of Israel’s conquest of Canaanite cities. This is because the Biblical account conflicts so strongly with the […]