ISIS site - J.J. Bimson's article from JACF2

Exodus and Conquest - Myth or Reality? Can Archaeology Provide the Answer?

By John J Bimson

The subject of this paper formed the basis of the second ISIS Fellowship Lecture, delivered after the AGM on 3rd September, 1988. Since the 1930s, the majority view has dated the Israelite Exodus and Conquest to the 13th century BC, at the end of the Late Bronze Age. A re-examination of the evidence suggests that the archaeology of this period is incompatible with the biblical narrative, and the campaign of conquest related in the Book of Joshua. Dr Bimson's own research concludes that a date for these events in the late 15th century would bring the narrative into accord with the archaeology of the Middle Bronze Age and the traditional biblical date for the Exodus of c. 1450 BC.

***

To begin by grasping the nettle offered by the second half of our title, it has to be said that archaeology cannot usually tell us whether biblical traditions are historical or mythological. Archaeology is not, strictly speaking, a science (although it employs scientific tools). One can rarely set up controlled experiments to test whether particular events (biblical or otherwise) actually happened. Rather, the archaeologist is at the mercy of the surviving evidence, and this imposes quite severe limits on what can be deduced with certainty. In the case of the cities of the Ancient Near East, limited time and resources mean that the archaeologist can only excavate a relatively small proportion of a tell (the Arabic term for a ruin-mound, in Hebrew spelt tel). For example, Yigael Yadin estimated that to excavate every level of the tell of Hazor (in northern Galilee) in its entirety would take eight hundred years! This emphasizes the small proportion which can be uncovered in a few seasons. Furthermore, only a limited amount of buried material survives the centuries for the archaeologist to discover it. Archaeology therefore has serious limitations when it comes to answering the kind of question posed in our title. One cannot guarantee that the appropriate evidence has survived, or (if it has) that the archaeologist will find it.

On the positive side, however, archaeology can significantly affect the balance of probabilities. I hope to show that it suggests the basic historicity of those biblical traditions which deal with the origins of Israel in Canaan.

Those traditions, contained in the books Exodus-Joshua (and referred to many times in the Prophets and the Psalms) relate that the Hebrews suffered slavery in Egypt and were led to freedom by Moses at a time of dramatic natural catastrophes; after forty years spent in the area south of Canaan, they migrated northwards through Transjordan, crossed the Jordan under the leadership of Joshua and conquered several key fortified cities.

Today most biblical scholars and archaeologists doubt the historicity of even this basic outline of events. The biblical traditions as we have them are seen as the result of a long and complex process of development, only taking their final shape during or after the Babylonian exile (6th century BC) and reflecting the political and theological concerns of that late period. Most scholars are therefore pessimistic about the possibility that these traditions preserve historical facts from a much earlier time. The majority view today is that the nation Israel arose within Canaan as an indigenous development. N. K. Gottwald is typical of many in affirming that the traditions concerning Israel's origins outside the land of Canaan 'are of questionable historical credibility' [1985: 35]; N. P. Lemche is confident that in its present form 'the account of Israel's pre-Palestinian existence ... can hardly be described as other than a fiction' [Lemche: 409]; G. W. Ahlstršm states that the story of the Exodus from Egypt 'is concerned with mythology rather than with a detailed reporting of historical facts' [Ahlstršm: 46].

The term 'mythology', when used in this context, is not intended to denigrate the biblical traditions, but simply to say that they embody religious convictions rather than true history. Nevertheless, in view of the way in which the traditions of Israel's origins pervade the Hebrew Bible, it is worth challenging such a view.

The scepticism of these scholars is based in part on the view that the traditions took shape at such a late period that they cannot possibly contain historical reminiscences from almost a thousand years before [Lemche:377-78, 384]. This view cannot be challenged here; suffice it to say that many scholars reject it, believing that at least some of the traditions concerning Israel's early history, especially those preserved in poetic form, do go back to the time before the monarchy [Cross; Freedman; Halpern]. However, another source of such scepticism is undoubtedly the perceived clash between the biblical traditions and archaeological evidence. Searching for evidence that Israel's conquest of Canaan occurred at the close of the Late Bronze Age (end of 13th century BC), scholars have failed to find any convincing correlations. Hence, Lemche concludes: '... It is no longer possible to offer even a reasonable defense of the Conquest narratives' [Lemche:413].

It is my contention that the failure to find appropriate evidence of Israel's conquest of Canaan is actually the result of looking for it in the wrong archaeological period. I have therefore tried in recent years to reopen the question of the date of the Exodus and Conquest. The first part of this paper is devoted to challenging the conventionally accepted date in the 13th century BC and defending an alternative date some two centuries earlier - a date suggested by the Bible itself.

PART ONE: EVIDENCE FOR A 13TH-CENTURY DATE EXAMINED

Between the 1930s and 1950s evidence accumulated in favour of dating the Exodus and Conquest to the 13th century BC. That date has remained the majority view. Even some of those scholars who reject the historicity of the Exodus and Conquest traditions still look to the decades around 1200 BC as the time when Israel emerged as a recognizable entity in Canaan. I will argue here that retention of the 13th-century date is an example of scholarly inertia, and that the evidence in its favour has long since been eroded away.
The evidence of Exodus 1:11

Figure 1: The biblical and archaeological sites of the Eastern Delta mentioned in the text.
Exodus 1:11 tells us that the enslaved Hebrews 'built for Pharaoh store-cities, Pithom and Raamses'. It has been recognized by the majority of scholars that the name Raamses is an appropriate rendering in Hebrew of the Egyptian Pi-Ramesse (= abode, or estate, of Ramesses), the name of the Delta-residence developed by and named after Ramesses II [Kitchen 1987]. The occurrence of this name in Exodus 1:11 has therefore been taken as an indicator that the enslaved Hebrews actually laboured for Ramesses II (1279-1213 BC). [1] If this reasoning is sound, the Exodus cannot have happened before the 13th century BC.

The first thing to note is that the Hebrew Bible does not use the name Raamses with chronological rigour. It uses it in Genesis 47:11 (actually in the form Rameses; the variation is not significant) to indicate the area where the ancestors of the Hebrew tribes first settled in the time of Jacob. By anyone's reckoning this must have been before any king called Ramesses ruled Egypt, [2] so the name is clearly being used retrospectively here (just as a modern historian might speak to Julius Caesar crossing the English Channel, or the Romans building York, neither name having been in use at the time referred to). We have a very clear biblical example of such retrospective usage in Genesis 14:14, where the city of Dan is mentioned in a narrative concerning Abraham; the city was actually called Laish in Abraham's day, and was not called Dan until much later, when the tribe of Dan conquered it and gave its own name to it, as narrated in Judges 18. Now, if the toponym Rameses/Raamses is being used retrospectively in Genesis 47:11, why not also in Exodus 1:11? In short, the name itself does not provide the date of the building activity in which the Hebrews were engaged, only the date when the narrative was last worked over by an editorial hand.

Against the use of Exodus 1:11 as dating evidence we must balance two other biblical references. 1 Kings 6:1 places the Exodus 480 years before the 4th year of Solomon, which points to a date (in round figures) of about 1450 BC. Judges 11:26 indicates a similar date, since it refers to Israelites settling in Transjordan 300 years before the time of Jephthah; as Jephthah seems to have been active around 1100 BC, this phase of Israelite settlement (at the end of their forty years of wandering in the wilderness) would have happened (again, in round numbers) roughly 1400 BC, which pushes the Exodus back to the mid-15th century BC. Both these verses have been either interpreted as symbolic or otherwise explained away on the strength of evidence favouring a later date [e.g. Wright:84; Kitchen 1966:72-75]. But as that evidence has now evaporated, the 15th-century date should be reconsidered. In connection with Exodus 1:11 we must ask whether an Exodus in the middle of the 15th century BC is compatible with archaeological evidence from Pithom and Raamses.

Taking Raamses first: is there evidence of building activity at the site as early as the 15th century BC? The site of Pi-Ramesse already had a long history of occupation before Ramesses II built the Delta-residence bearing his name. This history goes back to the 19th century BC, but is not unbroken. The site shows little evidence of occupation between the end of the Hyksos period (c. 1530 BC) and the late 18th Dynasty (c. 1310 BC) [Bietak 1986:236, 268].

This apparent gap in occupation would seem to seriously damage the case for a 15th-century Exodus. However, it would be unwise to assume the abandonment of the site on the basis of present evidence. We need to recall the limitations of archaeology, as outlined in our Introduction. In the present case those limitations are well summed up in the dictum that absence of evidence is not necessarily evidence of absence. It is a salutary fact that at another Eastern Delta site, Tell el-Maskhouta (the site of ancient Tjeku, = Succoth in Exodus 12:37), no trace has yet been found of a military base from the reign of Thutmose IV, nor of forts and other buildings from the 19th Dynasty, although the existence of such is attested in Egyptian texts. This is an important reminder that archaeological evidence can be extremely elusive at sites in the Eastern Delta. This is widely acknowledged, but is sometimes conveniently forgotten when the lack of evidence can be used to bolster a favourite theory.

The site of Pi-Ramesse, in today's Khata'na-Qantir district, covered an area of perhaps 4-5 square kilometres [Bietak 1986:269], and only a very small proportion of this has so far been explored. Furthermore, in many places ancient occupation-levels have been destroyed during the last hundred years through peasants digging for sebakh (soil used as fertilizer and for brick-making). The area has been greatly despoiled since it was explored and described by F. Lloyd Griffith and E. Naville in the 1880s [Bietak 1986:226, 228].

Most importantly, as W. H. Shea has pointed out, logic would suggest that some part of the site was occupied in the 15th century BC; Thutmose III, Amenophis II and Thutmose IV between them 'conducted well over twenty campaigns into Asia, and one would expect that they had a base of operations somewhere in this vicinity' [Shea:237]. The site lay at a strategic point on the eastern side of the Nile's easternmost arm, where there was an important route-junction (the name of the place in the Middle Kingdom was R3-w3ty, 'Mouth of the Two Roads'). It is therefore highly probable that an energetic pharaoh such as Thutmose III would have maintained a supply-base there for his many campaigns into Syria-Palestine. Indeed, the statement in Exodus 1:11 that the Israelites built 'store-cities' (Hebrew 'are miskenot, literally 'cities of storeplaces') for the pharaoh, could well refer to the building of such supply depots.


Figure 2: View of a section of Manfred Bietak's excavation works at Tell ed-Daba, now gradually becoming overgrown once more since the excavators have moved on to a different part of the site. These test squares revealed considerable 13th to 15th Dynasty occupation, but little 18th Dynasty material has come to light so far. In the background, the houses of Daba village can be seen. (Photo: D.Rohl)
It should also be noted that what evidence we already have is against a complete gap in occupation for most of the 18th Dynasty. M. Bietak, the excavator of Tell ed-Dab'a (in the south of the Pi-Ramesse area), has unearthed what he calls 'a massive filling wall' which he dates tentatively to the 'early 18th Dynasty' [Bietak 1986:236, 268]; and recently he has referred to evidence of occupation in the time of Amenophis III, which takes us back to the early 14th century BC [Bietak 1988:54]. So evidence of 15th-century activity may await discovery somewhere in the area if the occupation-levels have not been destroyed by sebakh-digging.

Turning to the site of Pithom, two candidates have traditionally been considered for this identification: Tell el-Maskhouta and Tell er-Retabah, sites about eight miles apart in the Wadi Tumilat. K. A. Kitchen, in the most recent and detailed study of this question [1987], argues convincingly for Tell er-Retabah. H. Goedicke has conducted excavations there and he reports finding remains of mud-brick buildings which he dates to the first half of the 18th Dynasty [Goedicke 1987]. Full publication is still awaited, so the details cannot yet be assessed, but in this case building activity in the right period seems fairly certain. [3]

One further requirement for a 15th-century Exodus is an explanation of how Moses was able to communicate so easily with the pharaoh. We have no evidence of a pharaonic residence-city in the Eastern Delta at this time, and this has long been seen as a stumbling-block for the early dating of the Exodus. However, in a forthcoming paper H. Goedicke will publish inscriptional evidence for the existence in the Eastern Delta, during the 18th Dynasty, of what he calls 'a royal domicile [used] during the recurrent tours of inspection the Egyptian king was supposed to do'. [4] This is all we would need to satisfy the requirements of Exodus 1-12, not an extensive residence-city on the scale of the later Per-Ramesse.

In short, archaeological evidence from the Eastern Delta, although not so clear-cut as we would like, does not rule out a 15th-century Exodus, as has so often been maintained.

Evidence from Transjordan
According to the biblical traditions in Numbers 20-25, after spending forty years in the area south of Canaan, the Hebrews moved north through Transjordan in order to enter Canaan from the east. Those traditions relate that the migrating Hebrews encountered various peoples during their northward trek: Edomites, Moabites, Amorites and the inhabitants of Bashan. With the latter two groups they even fought battles in which they conquered certain cities.

Surface surveys (i.e. studies of surface pottery finds, rather than excavations) of Transjordan, carried out by N. Glueck from the 1930s onwards, led Glueck to the conclusion that most of the region was without a settled population between the 19th and 13th centuries BC [Glueck 1940:125-140]. Pottery from the middle and Late Bronze Ages appeared to be absent or very scarce over much of the region. Glueck was followed by many other scholars in concluding that Israel's clashes with kingdoms east of the Jordan could not have happened before the 13th century BC [e.g. Wright:73; Kitchen 1966:61-62].

However, as a result of further surveys and full-scale excavations conducted during the last thirty years, Glueck's theory of an occupational gap has died the death of a thousand qualifications. A great many Middle and Late Bronze Age sites have come to light, requiring Glueck's theory to be modified beyond recognition [Mattingly; Bimson & Livingston:44; Boling:11-35]. There appears to have been some reduction in the population during the periods in question, but certainly not an absence of settlement. In fact Glueck himself revised his views shortly before he died [1970:141]. Unfortunately some scholars have lagged so far behind that as recently as 1985 the imaginary gap in occupation was cited against the 15th-century date for the Exodus [Stiebing:66]. The truth is that the evidence from Transjordan is quite neutral as far as dating the Exodus is concerned; it cannot prove a 15th-century date but it no longer constitutes evidence against it.

The argument from 13th-century destructions in Canaan
Between 1930 and 1960 excavations in Palestine uncovered evidence that a number of cities were destroyed at or near the end of the LBA (Late Bronze Age), i.e. in the decades around 1200 BC. These included cities which the Bible says were taken by the incoming Israelites: Debir (if identified with Tell Beit Mirsim), Lachish, Bethel (conveniently identified with Beitin) and Hazor. The fall of all these cities was dated to around 1220 BC, and seemed to provide evidence for a wave of destruction at that time. Therefore there seemed to be good grounds for viewing these destructions as the work of the Israelites under Joshua. Furthermore, with Conquest dated to c. 1220 BC, this implied an Exodus some forty years earlier, i.e. c. 1260 BC, in the reign of Ramesses II, which fitted nicely with the conventional understanding of Exodus 1:11 [e.g. Wright:60, 69-85].


Figure 3: Pink granite statue of Ramesses II, the popular choice for the pharaoh of the Oppression, depicting him as an old man with much of his 67-year reign behind him. This interesting statue was found at Memphis and is now to be seen in the open-air museum at Mit-Rahina. (Photo: D.Rohl)
This neat scenario has now been eroded utterly. The LBA destructions can no longer all be dated to the same time. Indeed, a recent study by B. G. Wood [1985; 1987a], analysing the pottery from a great many sites, shows that there were three waves of destruction spanning roughly a century.

The first wave occurred at the end of the subdivision of the LBA known as Late Bronze IIB1, and should be dated c. 1210 BC. Of the places mentioned in the Bible as taken by Israel, it included only one: Hazor.

The second wave occurred c. 1170-1160 BC, at the end of Late Bronze IIB2. This included Tell Beit Mirsim (once identified as Debir) and Beitin (generally accepted as the site of Bethel). However, it is now almost universally agreed that the true site of Debir is Khirbet Rabud, which was not destroyed in any of these three waves of destruction. The number biblical sites involved in this second wave is therefore no more than one (Bethel), and even this should probably be excluded; as we will see below, the location of Bethel at Beitin has recently been strongly challenged.

The third wave of destruction actually fell within the early Iron Age, at the end of Iron IA1, c. 1125 BC. Of the places Israel is said to have taken, this also included only one: Lachish.


Figure 4: Sketch map of Palestine showing the major sites destroyed during the LB/IA transition. (After B.Wood, 1985)
The whole range of sites included in Wood's study can be seen on the accompanying map. It is clear that either Israel's conquest of Canaan was a long, drawn-out affair, spanning about a century [Ussishkin:38-39], or the destructions of Hazor, Tell Beit Mirsim, Beitin and Lachish have nothing to do with Israel's arrival and we should find alternative explanations for them. Wood argues the latter view forcefully in a paragraph which is worth quoting at length, because it puts all three waves of destruction in a broader context:

In summary, it is apparent that the archaeological data do not support a conquest of Palestine by the Israelites at the end of the 13th century. The destructions that occurred in the Late Bronze/Iron Age transitional period can now be seen as part of a larger process that was taking place all around the Eastern Mediterranean basin. The underlying causes are not yet understood, but the end results are clear. The city-states ... became progressively weaker until they reached a stage where they could no longer maintain themselves. Since Egypt depended upon the city-states to support her forces in Palestine, as the city-states became weaker, so did Egypt's hold on her northern province. One-by-one the city-states fell; some were destroyed, others were not. In a weakened condition, they may have succumbed to attack by outsiders, revolts from within, or simply been abandoned when the citizens could no longer eke out a living. [1987a]

When the various Late Bronze/Iron Age destructions are seen from the perspective of the widespread economic and political collapse which affected the Eastern Mediterranean at that time, there is simply no reason to introduce invading Israelites in order to explain them.

Returning to the biblical account of the Conquest, it is also worth stressing that some cities which Israel is said to have conquered were definitely not destroyed in the decades around 1200 BC; indeed, some did not even exist at that time. Jericho (Joschua 6) was abandoned from c. 1275 BC until the early Iron Age; Gibeon (Joshua 9) was either abandoned or only sparsely settled in the LBA; Hebron (Joshua 10:37) shows no trace of LB occupation; Zephath (Judges 1:17) and Arad (Numbers 21:1-3) have similarly troublesome gaps (and this is true of Arad whether it is located at Tel Arad or Tel Malhata [5]. The city of Ai (Joshua 7-8) also comes into this category if its location at Et-Tell is maintained, but this will be discussed below.


Figure 5: Sketch-map of sites mentioned in the discussion of the Conquest narratives. (Italics indicate modern Hebrew or Arabic site-names.) An enlargement of the central square is provided on page 37 where the archaeological sites of this region are discussed in detail.
If it were not obvious already from Wood's analysis, this negative evidence shows clearly that there can be no neat 'fit' between the biblical accounts of the Conquest and the archaeology of the Late Bronze/Iron Age transition. The negative evidence is often paraded as proof that the Conquest narratives are unhistorical [Weippert 1971:46-55; Miller; Lemche:386-406, 413], but this is largely the consequence of a tunnel-vision which prevents the consideration of other periods as alternative settings for Israel's arrival in Canaan.
The new settlements of Iron Age I
At the beginning of the Iron Age a great many new settlements appeared in the hill-country of Palestine. Almost a hundred new settlements have been traced in the centre of the country alone, with others in Upper Galilee in the north and on the edge of the Negev in the south. These are mostly small, open, agricultural villages, though a few have protecting walls.

The rise of such settlements in the hills has been linked in a variety of ways with Israel's emergence. With their initial spread dated to roughly 1200 BC, and Israel's arrival dated only a couple of decades earlier on the strength of the destructions at the end of the LBA, it once seemed logical to view the settlements as the archaeological evidence for the Israelites beginning to settle down in their Promised Land. However, recent studies have shown that any connection which these settlements may have with the arrival of the Israelites is more complex than was previously envisaged.

It has been pointed out by a number of scholars that the agricultural villages show considerable cultural continuity (i.e. in terms of pottery styles etc.) with the preceding LBA. There is therefore no reason whatever to view them as evidence for the arrival of a new group from outside. While it is tempting to take them as an indicator of population increase, and hence to see them as indirectly attesting an influx of newcomers, there is still no reason to connect this with newly-arrived Israelites. Wood's aforementioned study redates the beginnings of highland village life to around 1160 BC, the time of the Philistine invasion of the coastal plain. This lends plausibility to a suggestion that the Philistine invasion displaced the populations of the coastal cities into the interior, and thus provided the impetus for colonisation of the hill-country [Callaway]. However, while this is possible explanation for the rise of Iron Age villages in the hills, we actually have no way of knowing whether or not the Philistine incursion significantly increased or displaced the local population.

Some scholars have suggested that the hill-country settlements are evidence for the withdrawal of a disgruntled peasant population from the city-states - a withdrawal which contributed to the collapse of the city-state system [Gottwald 1978:50; Chaney:60]. Another explanation for the new settlements reverses this cause and effect connection: a drift of part of the population into the hills occurred in response to the economic collapse of the city-states, as people sought new socio-economic structures in which to survive [e.g. Coote & Whitelam:117-138]. Neither of these explanations requires any link between the new settlements and the arrival of the Israelites, though proponents of both have suggested that the settlements mark the emergence of Israel as an indigenous development within Canaan. Such a view of Israel's origins naturally ignores the main thrust of the biblical traditions, which state that Israel was not autochthonous.

Another view is that the hill-country settlements are the work of semi-nomadic groups settling down [e.g. Finkelstein 1985:81-82; 1988]. However, as we noted above, the continuity which the settlements display with the preceding LBA culture rules out the possibility that these groups were newly-arrived in the land at the time of their sedentarization. V. Fritz concludes: '... This continuity is best explained by intensive, prolonged contact with Canaanite culture. This contact must have already occurred in the Late Bronze Age before the beginnings of sedentary life' [1987:97]. As a consequence of this conclusion Fritz has argued that the settlements mark the sedentarization of semi-nomads who had entered the land long before 1200 BC: 'Their "migration" into the land must therefore have occurred in the 14th century or already in the 15th' [1981:71].

In short, the new settlements which appear in the highlands of Canaan at the beginning of the Iron Age cannot be linked with the Israelites unless it is assumed that Israelite beginnings in Canaan go back a long way before 1200 BC. In other words, they do not provide evidence for an Israelite entry into Canaan in the late 13th century BC. On the other hand, it is noteworthy that Fritz's theory is compatible with an Israelite entry in the late 15th century BC, in line with the biblical chronology outlined earlier.

Merenptah's reference to Israel
The pharaoh Merenptah, successor to Ramesses II, left a victory-hymn celebrating a defeat of the Libyans in his fifth year (1208 BC). In the final strophe of the hymn, Merenptah mentions various entities in Palestine which he also claims to have subdued. Among these is Israel, written with the determinative for a people; Merenptah's other erstwhile foes are characterised by the determinative for a city or land. Some scholars have viewed the distinctive determinative as evidence that Israel had not yet become a well-settled group at the time of Merenptah, and therefore as evidence that the Israelites had only just entered Canaan in the late 13th century BC [e.g. Garner: 32-33].

The argument is weak for two reasons. Firstly it overlooks the fact that in the biblical period the name 'Israel' was first and foremost the name of a people and not of a state or territory. Hence an Egyptian scribe would have used the 'people' determinative even for a sedentary Israel. Secondly, it is completely illogical to argue that if the Israelites were semi-nomadic in the time of Merenptah they must have been newly-arrived. Having adapted to a semi-nomadic lifestyle during their wilderness wanderings, there is no obvious reason why they should have reverted to a sedentary existence on entering Canaan. They may well have retained a semi-nomadic lifestyle until external factors (such as the socio-economic changes which took place at the end of the LBA) forced change upon them.

Furthermore, some recent studies of the final strophe of Merenptah's inscription actually point to the conclusion that Israel (whether semi-nomadic or settled) was a well-established force in Canaan by Merenptah's reign, and had therefore been in the land for a considerable length of time.

The final strophe reads:
The princes are prostrate, saying 'Peace!'
Not one raises his head among the Nine Bows
Desolation is for Tehenu,
Hatti is pacified,
Plundered is Canaan with every evil.
Carried off is Ashkelon,
Seized upon is Gezer,
Yanoam is made non-existent.
Israel is laid waste,
His seed is no more,
Hurru is become a widow because of Egypt.
All lands together are pacified,
Everyone who was restless has been bound. A1


Figure 6: The 'Israel Stela' of Merenptah

Earlier studies regarded the four names in the middle of this strophe as a list of minor entities arranged in order from north to south: Ashkelon, Gezer, Yanoam, Israel [e.g. Yeivin 1971: 30]. This implied that Israel was a fairly small group which Merenptah had clashed with somewhere in the Galilee area. Recent analyses of the structure of the coda lead to a different conclusion [Ahlstr"m & Edelman; Stager; Wood 1987a]. The results of these analyses are reflected in the way the lines are arranged in the rendering given above. [6] The coda has a chiastic or envelope structure which hinges on the section marked C. Thus A1 mirrors A in referring to Egypt's traditional enemies in very general terms; B1 mirrors B in referring to specific major entities; C focuses on specific minor entities. Thus Israel features among the major entities, keeping company with Tehenu (Libya), Hatti (Syria-Palestine), Canaan (Western Palestine) and Hurru (another general term for Syria-Palestine or its inhabitants). This is confirmed by the parallelism within section B1; Israel is depicted as a bereaved father, in parallel with Hurru, a bereaved wife [Stager: note 30].

In short, by Merenptah's day Israel was a well-established and significant political force in the area, and cannot have been there for only a short time. The inscription is therefore more in keeping with a 15th-century date for the Exodus and Conquest than with a date in the 13th century.

This brings us to the end of our investigation of the usual arguments for dating Israel's origins in Canaan to the 13th century BC. To sum up: some of the old arguments for the 13th-century date have been eroded by more recent evidence, while some were never very secure anyway; some evidence commonly employed in favour of the 13th-century date (the Iron Age settlements in the highlands and Merenptah's reference to Israel) are actually more readily compatible with the 15th-century date. This, of course, raises an important question: if Israel was in Canaan for two centuries before Merenptah's time, why do we have no evidence for its existence during that period?

This is really two questions in one: why do we have no archaeological evidence for Israel's existence in the land, and why do we have no inscriptional references to Israel until the one left by Merenptah? Both are readily answered. If the Israelites were semi-nomadic for the first two centuries of their existence in Canaan, we would not necessarily expect their presence to be attested archaeologically. In Palestine under the British Mandate (i.e. during the first half of the present century) between 55,000 and 65,000 Bedouin lived in the Negev. I. Finkelstein comments: 'This population left almost no material remains, however; without contemporary, documentary evidence, we would not know of its existence' [1986:51]. We should not expect semi-nomadic Israelites to have been any different in this respect. As for inscriptional references, the absence of such before Merenptah's reign needs to be put in context. After Merenptah's inscription of 1208 BC we do not encounter the name Israel again outside the Bible until 853 BC, when the Assyrian king Shalmaneser III refers to 'Ahab the Israelite' [Pritchard 1969:278-79]. Israel certainly continued to exist during the intervening three and a half centuries, and yet its name is absent from the extant texts. Returning to the centuries before Merenptah, it is possible that during that period the Israelite tribes were classed with such wider non-sedentary groups as the shasu/sutu and 'apiru/habiru [Weippert 1979:33-34; Coote & Whitelam:106-109; Lemche:152-163]. In any case the absence of specific references to an entity called Israel in that period cannot be taken as proof that no such entity existed, as the later silence of three and a half centuries makes clear.

PART TWO: THE SEARCH FOR A 15TH-CENTURY CONQUEST

When we look at the archaeological history of Palestine as conventionally understood, we find no evidence of a wave of destruction at the end of the 15th century BC which could be interpreted as Israel's conquest of Canaan. However, about a century earlier in the archaeological record we do find such a wave of destruction. This falls at the transition from the MBA (Middle Bronze Age) to the LBA - more precisely, at the transition from MBIIC to LBI. At that time a great many of the fortress-cities of Canaan were violently destroyed. Insofar as biblical cities have been confidently identified and adequately excavated, almost all those which the Bible says were taken by Israel were included in this wave of destruction. (The outstanding exception is Ai, which will be discussed separately below.) I have argued in detail elsewhere that these destructions are the missing evidence for Israel's arrival in Canaan, and that they should be redated accordingly [Bimson 1981:119-223; Bimson & Livingston:51-52]. Some of the evidence for that redating will be summarised briefly below. First we will see how well the destruction of one particular MBIIC city correlates with the biblical tradition.
A test case: the destruction of MBIIC Jericho.

Sketch plan showing the massive glacis walls of the MB II city of Jericho (in black) inside which the smaller earlier city is delineated by the EB city wall (in grey). [Illustration: D.Rohl]
The biblical account of the destruction of Jericho is particularly rich in detail, while the site of Old Testament Jericho has been confidently identified and extensively excavated. We therefore have an opportunity to compare the biblical account with archaeological discoveries in a way which is not often possible. We find no less than five points of correspondence:
  1. Like most cities constructed in the MBII period, Jericho was very strongly fortified [Kenyon 1957:218-220]. It therefore satisfies the biblical picture of a secure, walled city (Joshua 2:5, 15; 6:1).
  2. It was nevertheless destroyed, and its destruction involved a violent conflagration [Kenyon 1957:259-60]. This corresponds to the fate of Jericho in Joshua 6:24.
  3. Some of its buildings collapsed just before they were burned [Kenyon 1981:370]. This suggests earthquake activity [Wood 1987b], as does the collapse of the walls in Joshua 6:20. [7]

    Cross-section of the walls and glacis of MB II Jericho which Kenyon suggested were designed to prevent the use of battering rams. The total height of the defences was around 14m with a width of 20m at its base. [Illustration: D.Rohl]
  4. Storage-jars well-stocked with grain were found in the excavated buildings [Kenyon 1957:230], showing that the harvest was either underway or recently completed. The Israelites took Jericho at the time of harvest according to Joshua 3:15. Incidentally, Egyptian tactics would typically have been very different from those of Joshua, namely to lay siege to a city shortly before harvest, when the previous year's supplies were depleted (thus forcing an early surrender of the city), and when the standing grain could be used to feed the Egyptian troops if the siege was protracted. The timing of Jericho's destruction therefore goes against the usual view that it fell in an Egyptian campaign [Wood 1987b].
  5. The latest tombs of the MBA city contained multiple burials indicating that some catastrophe had caused a high death-toll shortly before the city was destroyed. Kathleen Kenyon, the excavator of these tombs, ruled out warfare (because there were no signs of wounds on the skeletons) and famine (because of various signs that the city was well-supplied with food, e.g. the jars of grain mentioned above) and concluded that some kind of plague had affected the city's population shortly before it fell to enemy (in her view Egyptian) attack [Kenyon 1957:254-55]. It is striking that the Israelites were affected by a plague shortly before launching their attack on Jericho, while they were encamped on the opposite side of the Jordan at Shittim (Numbers 25:9). The Israelite spies who penetrated the city in preparation for the attack (Joshua 2:1) may have carried the infection from Shittim to Jericho (or vice versa, since the order of events is not explicit in the biblical account). In any case, two cases of plague in a limited area is of significance for a theory which would make them synchronous events [Bimson 1981:121-22].

Town sites of the Middle Bronze Age destroyed at the MB/LB transition. The clusters of sites match well with the areas attacked by the Israelites during the Conquest.
In the case of Jericho there are therefore very good grounds for identifying the destruction of the MBA city with the Israelite capture of Jericho recorded in the book of Joshua. However, standing in the way of the identification is the conventional dating of the fall of MBA Jericho at least a century before the biblical date for the Conquest. We therefore need to ask: How secure is the dating of the destructions which occurred at Jericho and many other sites at the MB/LB transition?
The case for redating the MB/LB transition.
The currently accepted dating of the MB/LB transition depends largely on the assumption that the destructions which occurred at that time mark the expulsion of the Hyksos from Egypt and a subsequent Egyptian campaign of retaliation throughout Palestine [e.g. Dever:174-175]. This scenario has resulted in the destructions being given a round date of 1550 BC, though the new and lower dates for the 18th Dynasty now preferred by many Egyptologists would place them nearer to 1500 BC [Bietak 1988:54]. In fact there is no evidence at all to link the destructions with Egyptian campaigns against the Hyksos [Bimson 1981:132-40; 1987], and so a major plank for dating the MB/LB transition is actually a piece of fiction.


Chart showing the dates for Palestinian archaeological periods and early biblical history in the conventional and revised schemes; also the major periods of Egyptian history. All dates shown are approximate.

I will not repeat in detail here the arguments in favour of redating the MB/LB transition to shortly before 1400 BC. Briefly, two lines of recent research converge in support of such a revision. One is the chronological research of M. Bietak, the excavator of Tell ed-Dab'a in Egypt's Eastern Delta. At this site a Middle Bronze culture closely related to that of Palestine is represented in archaeological contexts datable by Egyptian finds. On the basis of his excavations Bietak would lower the dates for the period known as MBIIB by roughly a century [Bietak 1984]. In Palestine MBIIB is followed by MBIIC, the final phase of the MBA. In Egypt the equivalent of MBIIC is a very short period, ending with the expulsion of the Hyksos (now to be dated between 1530 and 1515 BC [Bietak 1988:54], but in Palestine, as is well-attested at sites such as Shechem, it must have lasted at least a century and probably more. As far as Palestine is concerned, Bietak's radically low dates for MBIIB therefore push down the end of MBIIC as well.

Bietak himself would lower the MB/LB transition in central Palestine to 1459 BC. [8] This is because he attributes the destructions which marked the transition to Thutmose III, whose campaigns began in that year according to the low chronology. However, Egyptologist J. Hoffmeier has shown that, contrary to popular opinion, the campaigns of Thutmose III did not cause widespread destruction in Canaan [Hoffmeier, forthcoming]. Other destroyers of the MBIIC cities must therefore be found, and a date a few decades later than Bietak's would allow us to identify their destroyers as the incoming Israelites. [9]

A later date than Bietak's becomes increasingly likely in the light of the second piece of research to be mentioned here. In a recent re-examination of the pottery from the MBA city at Jericho, B. G. Wood has shown that the city actually continued to thrive somewhat into the LBI period before it was destroyed [Wood 1987b]. This conclusion is radical enough by itself, but it opens up an even more radical possibility. Wood's conclusion is based on a careful study of local Palestinian pottery from the site, whereas previous work on Palestine's ceramic chronology has given more weight to imported wares. It may be that other cities supposedly destroyed at the end of MBIIC should also have their lives extended into LBI. This possibility needs to be tested by means of a detailed comparative study of pottery from a whole range of sites, applying Wood's dating criteria.

The radical conclusions of Bietak and Wood put the dating of the MB/LB destructions back into the melting-pot. Both studies imply a later date for those destructions than has conventionally been entertained. Bietak's work places the MB/LB transition later than has previously been suspected, while Wood's findings may require us to place the major wave of destructions some way into LBI instead of at the MBIIC/LBI transition. While it is too early to be dogmatic, it does seem likely that either Bietak's evidence, or Wood's evidence, or some combination of the two, will allow (or even require) us to date those destructions late in the 15th century BC.

With the Israelite Conquest assigned to shortly before 1400 BC, and with the wave of MB/LB (or in Wood's view LBI) destructions redated to correlate with it, the biblical tradition is archaeologically attested at every site where a city said to have been destroyed by the Israelites has been confidently identified and adequately excavated. This statement would not win universal assent, however; many biblical scholars and archaeologists would object that it is not true of the city of Ai, the city which Israel took immediately after Jericho according to Joshua 7:2-8:29. This city deserves a separate discussion.

The identification of Bethel and Ai.
Ai lay in the central highlands, not far from Bethel and roughly to the east of it (Genesis 12:8). With Bethel confidently identified with the site of Beitin, Ai has been identified with Khirbet et-Tell, the only site east of Beitin which has clear remains from the Old Testament period. However, these remains do not indicate occupation at the time of the Conquest. There is no evidence of any occupation at Khirbet et-Tell between the end of the Early Bronze Age (around 2300 BC) and the beginning of the Iron Age (c. 1160 BC) when a small, unwalled village, typical of the period, was established on the ancient mound. Israel's capture and destruction of Ai has therefore been a longstanding problem for scholars who have tried to correlate the biblical traditions with archaeological evidence. Furthermore, it remains a problem within the revised framework proposed here, since the gap in occupation at Khirbet et-Tell includes the whole of the MBA.


Sketch map showing the sites of et-Tell and Khirbet Nisya in relation to Jericho and Jerusalem. A scale in Roman miles is provided in order to compare the distances of the two candidates for Bethel (Birch & Beitin) from Old Jerusalem.
In recent years, however, a possible solution has emerged. This stems from D. Livingston's bold proposal [1970; 1987] that Beitin is not the site of biblical Bethel. Livingston points out that Beitin's location does not fit the biblical requirements for Bethel very well. There is no mountain between Beitin and Khirbet et-Tell, as there should be between Bethel and Ai (Genesis 12:8), and Beitin is rather too far north to fit neatly into the line of border towns between Benjamin and Ephraim listed in Joshua 16:1-3 and 18:12-14. Furthermore, Beitin does not fit the location of Bethel described by the early Christian authors Eusebius and Jerome.

Eusebius (AD 269-339) wrote a work known as the Onomasticon which was subsequently revised and amplified by Jerome (AD 345-419). This gives the location of various biblical sites in relation to contemporary landmarks, including Roman milestones. According to the Onomasticon, Bethel lay 'at [or near] the twelfth Roman milestone from Aelia [Jerusalem, renamed Aelia Capitolina by the emperor Hadrian]', on the east side of the road leading north to Neapolis (i.e. Old Testament Shechem, modern Nablus). In the last hundred years a number of the Roman milestones along this road have been discovered. Their locations make it quite clear that the Beitin lay near the fourteenth milestone, not the twelfth. In other words, this evidence agrees with that of the biblical boundary lists in showing Beitin to be too far north for identification with Bethel.

Ironically, it was a rather loose application of the Onomasticon which led to the identification of Beitin with Bethel in the first place. In 1838 the American biblical scholar and explorer Edward Robinson estimated the distance between Beitin and Jerusalem by the time it took him to make the journey on horseback, concluding that it lay the correct distance north of Jerusalem to be biblical Bethel [Robinson 1856:449-50]. Modern measurements with odometer, and the discovery of some of the Roman milestones, show that he simply underestimated the distance. Beitin is too far from Jerusalem to be Bethel if Eusebius's information is correct.

If Beitin is not Bethel, what is it? It is certainly a significant site, with archaeological remains from virtually all of the Old Testament period. It may be the site of biblical Bethaven. Its name is a possible reflex of Bethaven (spelt Bethaun in the Onomasticon), and there is no evidence to stand in the way of the identification. But if Beitin is Bethaven rather than Bethel, where is Bethel?

A site which fits Eusebius's location of Bethel (i.e. near the twelfth Roman milestone north of Jerusalem) is present-day el-Bireh. The twelfth Roman milestone itself has never been found, but the 3rd, 4th and 5th have, along with another from Khirbet esh-She which unfortunately lacks an inscription. The locations of the 3rd, 4th and 5th indicate that the one found at Khirbet esh-She must have been the 11th. This place lies south of el-Bireh, putting el-Bireh near the twelfth milestone [Livingston 1987].

El-Bireh has never been excavated and the existence of a thriving modern town makes excavation unlikely. However, a surface-survey of the highest point in the town produced pottery from most of the major archaeological periods, suggesting the site was an important one in Old Testament times.The early Christian pilgrim Egeria, who visited Palestine in the fourth century, has left an account which confirms the location of Bethel at el-Bireh rather than Beitin. She says that twenty-eight miles south of Neapolis lay a village called Bethar, and a mile south of that 'the place where Jacob slept on his way from Mesopotamia' - i.e. Bethel (Genesis 35:1-15); twelve miles further south lay Jerusalem [Wilkinson 1971:155]. This makes sense if Bethel stood at present-day el-Bireh, for the village she calls Bethar would then be Eusebius's Bethaun and biblical Bethaven; if Bethel is located at Beitin, there are no ruins north of it to equate with Egeria's Bethar. [10]


The hill of Khirbet Nisya has been sculpted into a series of agricultural terraces, creating the horizontal striation effect which can be clearly seen on its slopes. The excavators believe that the ancient site was denuded to provide materials for the terracing, leaving little trace of the original Bronze Age structures. What does remain is the pottery which shows that the site was occupied in the MBA. The summit of this barren site is now sparsely covered with trees. The modern settlement at the top right of the picture lies beyond the hill. (Photo: D.Livingston)
If we accept Livingston's arguments and locate Bethel at el-Bireh (for which the evidence seems overwhelming), does this help us find an alternative location for Ai? Livingston himself has combed the area east of el-Bireh very thoroughly and has suggested identifying Ai with a small site known as Khirbet Nisya. This fits the requirements of Genesis 12:8 in that a significant mountain lies between it and el-Bireh. The terrain also makes detailed sense of the accounts of the Israelite attacks on Ai in Joshua 7 and 8.

Livingston has conducted a number of short excavation campaigns at Khirbet Nisya since 1979 [Bimson & Livingston 48-51; Livingston 1987]. Detailed publication of the finds is still forthcoming, but two major facts have emerged, one favouring the site's identification with Ai and one weighing against it. In favour is the pottery record from Khirbet Nisya. Pottery has been found from the Chalcolithic, Early Bronze I, MBII, LBI, Iron Age I and II, Persian, Hellenistic, Roman, Byzantine and Early Arabic periods. These finds cover all the periods when Ai is known to have been inhabited according to the Bible and Eusebius. It is particularly significant that there was a settlement there in MBII, and that at the transition to LBI (or shortly thereafter) the site was abandoned. This correlates well with a destruction and abandonment of Ai at the MBII/LBI transitional period.

On the negative side, no building remains have yet been found from that period, even though Ai appears from the biblical account of the Conquest to have been a fortified town (Joshua 7:5; 8:29). Nor has any trace of a destruction level been discovered. A possible explanation for this may lie in the activities of the Byzantine and later inhabitants, who converted the entire hill into farming terraces, re-using building remains to construct the terrace walls and removing ancient occupation levels to provide the fill behind them. (Indeed, it was in the fill of one of the terraces that much of the MB II pottery was discovered in the 1985 season). From the point of view of the ancient farmers this vastly improved the site's agricultural potential, but from the archaeologist's point of view it may have been a gross act of vandalism, removing all evidence that the MB II settlement was a walled town. In more recent centuries wind and rain have contributed further to the process of denudation. If this is not the explanation for the lack of building remains and traces of burning, we must conclude that Ai has not yet been found. On the other hand the possibility always remains that some traces of buildings and fortifications still await discovery at Khirbet Nisya during a future season of excavation.

The small size of Khirbet Nisya may also seem to stand against its identification with Ai, and so a word needs to be said about this.


Rhyton-form urn found in Tomb 9 of MB II Jericho.
Livingston has estimated that the area containing occupational debris (chiefly pottery) is about seven acres [Livingston 1980: 24]. Even if we assume an occupation density of 200 people per acre (which is rather high for an ancient settlement of the type characteristic of biblical Palestine), this gives us a population of only 1,400 people. The Old Testament, on the other hand, speaks of Ai as having 12,000 inhabitants (Joshua 8:25), implying a truly vast city. However, before Livingston suggested locating Ai at Khirbet Nisya, J. W. Wenham [1967: 21, 26, 41] had argued that the original population figure for Ai must have been 1,200, which had been distorted by a factor of ten through textual corruption. Wenham points to Joshua 7:3 as support for the smaller figure: the Israelites sent to reconnoitre the town suggest to Joshua that he should send only two or three thousand men to capture it, because its inhabitants 'are but few'. This certainly does not sound like a description of a town of 12,000, which would have been relatively huge in the Old Testament period. (For comparison, Jericho, with an area of no more than 10 acres, would have had a population of between 1,600 and 2,000 if present estimates of population density are dependable.) Furthermore, Ai is said to have been smaller than Gibeon (Joshua 10:2); the tell of Gibeon has an area of 15 acres, so Ai must have been smaller than that. In these respects the site of Khirbet Nisya is actually a good candidate for Ai.

In conclusion, Khirbet Nisya is undoubtedly a better candidate for identification with Ai than is Khirbet et-Tell. It has the correct topographical relationship to the true site of Bethel, is the right size and was occupied at the right periods. But whether or not Khirbet Nisya is the true site of Ai, it is clear that we are no longer compelled to look for Ai at Khirbet et-Tell. It follows that the gap in occupation at Khirbet et-Tell is not evidence against the historicity of the Conquest, nor does it weigh against our theory for placing the Conquest at the MBII/LBI transition.

Conclusion

Writing of the difficulty of establishing, with any degree of confidence, that any given archaeological evidence pertains to Israel's entry into Canaan, H. D. Lance remarks: 'If the biblical list of cities destroyed by Joshua could be correlated site by site with massive destructions at the end of the Late Bronze Age, one could begin to find the probabilities persuasive. But no such correlation exists' [Lance:64]. His final comment is true of the situation at the end of the LBA, as we have seen; but it is not true of the situation at the end of MBA, when enough correlations exist to make the probabilities very persuasive indeed. It is undoubtedly true that the failure to find such correlations in the past has contributed to a radical scepticism concerning the traditions of Israel's origins in Canaan. Perhaps the future recognition of such correlations will lead eventually to a rehabilitation of those traditions.

Notes

  1. In this paper I consistently adopt the 'ultra-low' chronology for New Kingdom Egypt, currently gaining in popularity among Egyptologists. [See Kitchen 1977/78; various papers in Astršm 1987].
  2. The short (between one and two years) reign of Ramesses I fell only fifteen years before the reign of Ramesses II. The theory of some scholars [Courville:118-122; Merrill:107; Dyer:226-27] that other pharaohs with the name Ramesses ruled in a much earlier period are not supported by inscriptional evidence and are totally unnecessary.
  3. If Tell el-Maskhouta is preferred as the site of Pithom, lack of 18th Dynasty remains at that place should not be seen as evidence against a 15th-century Exodus. As noted above, no remains have yet been found of 18th and 19th Dynasty installations referred to in Egyptian texts and inscriptions. Lack of remains from the 15th-13th centuries therefore has more to do with conditions at the site than with the site's occupational history.
  4. Goedicke in personal correspondence dated 9th September 1987.
  5. For details, with references, see Bimson 1981:188-196 and (brief, but more up to date) Bimson and Livingston:40-41.
  6. My arrangement of the lines is also indebted to a forthcoming study by W. H. Shea.
  7. For further possible evidence of earthquake activity at the end of the MBII city, see Bimson 1981:122-124.
  8. Bietak in personal correspondence dated 17th March 1987.
  9. Bietak has recently resisted my suggestion that the date of the fall of the MBIIC cities could be dated to the late 15th century BC [Bietak 1988]. My reply to Bietak will appear in Biblical Archaeology Review 14/6 (Nov-Dec 1988).
  10. I am indebted to David Livingston for the information contained in this paragraph.

PICTURE CAPTIONS

Part One
  1. The biblical and archaeological sites of the Eastern Delta mentioned in the text.
  2. View of a section of Manfred Bietak's excavation works at Tell ed-Daba, now gradually becoming overgrown once more since the excavators have moved on to a different part of the site. These test squares revealed considerable 13th to 15th Dynasty occupation, but little 18th Dynasty material has come to light so far. In the background, the houses of Daba village can be seen. (Photo: D.Rohl)
  3. Pink granite statue of Ramesses II, the popular choice for the pharaoh of the Oppression, depicting him as an old man with much of his 67-year reign behind him. This interesting statue was found at Memphis and is now to be seen in the open-air museum at Mit-Rahina. (Photo: D.Rohl)
  4. Sketch map of Palestine showing the major sites destroyed during the LB/IA transition. (After B.Wood, 1985)
  5. Sketch-map of sites mentioned in the discussion of the Conquest narratives. (Italics indicate modern Hebrew or Arabic site-names.) An enlargement of the central square is provided on page 37 where the archaeological sites of this region are discussed in detail.
  6. The 'Israel Stela' of Merenptah
    Part Two
  7. Sketch plan showing the massive glacis walls of the MB II city of Jericho (in black) inside which the smaller earlier city is delineated by the EB city wall (in grey). [Illustration: D.Rohl]
  8. Cross-section of the walls and glacis of MB II Jericho which Kenyon suggested were designed to prevent the use of battering rams. The total height of the defences was around 14m with a width of 20m at its base. [Illustration: D.Rohl]
  9. Town sites of the Middle Bronze Age destroyed at the MB/LB transition. The clusters of sites match well with the areas attacked by the Israelites during the Conquest.
  10. Chart showing the dates for Palestinian archaeological periods and early biblical history in the conventional and revised schemes; also the major periods of Egyptian history. All dates shown are approximate.
  11. Sketch map showing the sites of et-Tell and Khirbet Nisya in relation to Jericho and Jerusalem. A scale in Roman miles is provided in order to compare the distances of the two candidates for Bethel (Birch & Beitin) from Old Jerusalem.
  12. The hill of Khirbet Nisya has been sculpted into a series of agricultural terraces, creating the horizontal striation effect which can be clearly seen on its slopes. The excavators believe that the ancient site was denuded to provide materials for the terracing, leaving little trace of the original Bronze Age structures. What does remain is the pottery which shows that the site was occupied in the MBA. The summit of this barren site is now sparsely covered with trees. The modern settlement at the top right of the picture lies beyond the hill. (Photo: D.Livingston)
  13. Rhyton-form urn found in Tomb 9 of MB II Jericho.

BIBLIOGRAPHY


borrowed from http://www.nunki.net/isis/downloads/jacf2article1.zip