David Livingston1
BSP 1:1 (Winter 1988) p. 25
[This article is a revision of our paper given at the Memphis Symposium on the Exodus and Conquest in April, 1987.]
One of the major reasons why many scholars today reject the truthfulness of the biblical account of the Conquest is the site of Ai. No one lived at et-Tell (the modern location for Ai) when the Israelites arrived. So, for this and other reasons, many scholars believe the story of the destruction of Ai (and the whole Conquest “saga”) is merely legend, embellished over hundreds of years and finally written down during the time of Israel’s kingdom. They ask, “How could it possibly be accurate history?”
Throughout the Old Testament, both Bethel and Ai are closely associated. To find one is to find the other. Conversely, to mislo-cate one is to mislocate the other. Furthermore, if one excavates a site which is not the biblical city he supposes it is, he will obviously come to some wrong conclusions.
ABR staff have been working on the Bethel-Ai problem for many years. We may not have all the answers yet, but we do have some suggestions.
Has Bethel Been Correctly Located?
The city of Bethel was on the border between Benjamin and Ephraim (Joshua 16:1–2, 18:13). It separated the northern and southern kingdoms of Israel and Judah and thus had every reason to be the most-mentioned city, next to Jerusalem, in the Old Testament. Its correct location is centrally important in determining geographical as well as chronological relationships. And its twin-city of Ai should yield important archaeological information to help in dating the Conquest.
Although almost all scholars agree that Bethel has been correctly located for over 150 years, there have been some serious problems with the identification of Ai. We will examine the procedure used to identify these two cities and the problems that have resulted.
In doing so, we will assume the biblical documents are accurate historical accounts. Contrary to the attitude of many scholars, we do not believe they are simply “folk legends.”
Problems with the Identification of Bethel and Ai:
Bethel and Ai were “twin cities. “ Genesis 12:8 is the first mention of this relationship. Then throughout the Old Testament they are associated, as late as post-exilic times (Ezra 2:28, Nehemiah 7:32).
Geographically, Ai must be east of Bethel. Topographically, there must be a mountain between them – not just a hill, but a mountain (Hebrew bar). To fulfill a twin-city relationship, they ought to be close together.
Other details added to these appear to make Finding both Bethel and Ai fairly easy. In fact, ff it were not for some doubts about the traditional location of Ai at et-Tell, the case might have been closed long ago.
The problems with Ai at et-Tell, however, are serious. So much so that they caused the excavator of et-Tell, Joseph Callaway, to say,
BSP 1:1 (Winter 1988) p. 26
“Ai is an embarrassment to anyone who takes the biblical account of the Conquest seriously.” He says this because et-Tell was not occupied at all at the traditional date of the Conquest in the late thirteenth century (1250 BC). But it wasn’t occupied at the earlier date (ca. 1400 BC), either. So it is an embarrassment no matter what one’s viewpoint.
The first settlement there after the Early Bronze occupation ceased (in about 2400 BC) was Iron Age (Israelite?), beginning about 1200 BC. This, of course, precludes a Canaanite occupation during the time of the Israelite Conquest. Some have suggested that the excavators may have just overlooked areas where the Middle or Late Bronze Canaanites had their city. But that cannot be. The excavations have been both thorough and carefully done.
Traditional Ai Cannot be Biblical Ai
The conclusion from this seems obvious to us: Traditional Ai cannot be Biblical Ai. But if it is not biblical Ai, is there another site east of traditional “Bethel” at Beit-in which might be Ai? No. Every potential slte within a radius of 5 miles east of Beitin has been thoroughly explored and excluded.
At this point we make what is a radical, but in light of all the above, logical suggestion. It is this: If Bethel and Ai were twin-cities and traditional Ai cannot be biblical Ai, then traditional “Bethel” (at Beitin) may not be biblical Bethel.
Is Modern Beitin Really Bethel?
Can it be possible that “Bethel” was actually mislocated? How was Beitin originally identified as Bethel?
Edward Robinson was the first to identify it – in the 1830’s. He equated the modern Arabic name of “Beitin” with “Bethel” (which is feasible, but not compelling). Actually, there was no village at the site at all in Robinson’s day. Apparently, it was an area name rather than a village name. In fact, for over 1400 years the very name “Bethel” had been completely forgotten in the area.
Besides the name, the only other evidence Robinson used in the identification was the distance of Bethel from “Aelia” (Jerusalem) mentioned by the early church fathers, Eusebius and Jerome. His measurement of the distance they gave was done on horseback, estimated by the length of time his horse traveled from Jerusalem to Beitin. This way of measuring distances may have some merit, but seems rather uncertain.
One hundred years later, W.F. Albright accepted Robinson’s identification without checking the distance, either by horseback or automobile. We must remember that in the 1930’s et-Tell was not yet questioned as Ai. Everyone assumed it was. Anyone would have done as Albright did for it seemed obvious that Beitin must be Bethel.
On this basis then, Albright, and later Kelso, excavated Beitin for several seasons. Preliminary reports Finally culminated in The Excavation of Bethel published in 1968 as an Annual of the American Schools of Oriental Research (XXXIX). Because of our interest in the problem of Ai and its twin-city Bethel, we read the report looking for archaeological proof the writers might set forth that Beitin was truly Bethel. We could not find it in the report except to say that if one assumes Beitin is Bethel, the archaeological picture matches
BSP 1:1 (Winter 1988) p. 27
the criteria for biblical Bethel fairly well. So, we wrote Dr. Albright and asked to what proof he could point. (Perhaps we had overlooked something.) Albright’s first answer was that there was no archaeological proof (no inscriptions or anything specifically confirming that Beitin was really Bethel), but that the identification was maintained by the biblical and patristic (church fathers) evidence.
With that we restudied the biblical references and concluded that one could not locate Bethel precisely from them, either. So we wrote again asking about the biblical proofs, thinking he surely had something we had missed. His answer was that there was no biblical proof at all. The identification was made using the archaeological and patristic evidence. But, he had already eliminated the former himself.
So now we were left with only the patristic evidence. What was it, and how accurately could it be checked? We proceeded on all this very hesitatingly. How could Robinson and especially Albright and our own seminary professor, James Kelso, possibly be wrong?.
Roman Milestones Tell the Story
The patristic evidence Robinson and Albright used was that of Eusebius/Jerome in their Onomasticon. There they said that no matter which direction you were traveling, north or south, “Bethel” was near the 12th Roman milestone (RMS) on the road to Neapolis (modern Nablus). And since Bethel, according to their description, was right on the road, there was no mention of a turn-off. In other words, when one was at the 12th milestone, he was near Bethel.
Author standing beside typical Roman milestone.
Two resources were available for us to check this distance. They were the Roman milestones found and published in the last 100 years (incidentally, all of them since Robinson’s research in the 1830’s to 1850’s), and measurements by odometers on the road from Jerusalem to Beitin.
Odometer measurements done independently by several of us shows the mileage to be about 10.1 statute miles between the Damascus Gate and the center of modern el-Bireh, slightly over 12 statute miles to Beitin. This sounds good for Beitin, doesn’t it? Except that those distances in Roman miles are slightly over 11 for el-Bireh and a little over 13 for Beitin.
But, why measure from the Damascus Gate? Those working on this problem have assumed that
BSP 1:1 (Winter 1988) p. 28
the Roman column on the Medeba Map (giving the Arabic name of “Bab ‘al-Amoud” to the gate) was the zero mile marker in Roman/ Byzantine times. But other Roman cities had their zero milestone (miliareum aureum, “gold milestone”) at their center, in the forum (examples are Rome and London). Should Jerusalem be different? Why start counting at the northern gate, or any gate, for that matter? (Actually, the Roman column on the Map seems to have been commemorative and likely had nothing to do with mileage measurements.)
Using this column as a starting point has been shown to be wrong by comparing distances with fourth century milestones which have been discovered. They make measurements more precise. At least four between Jerusalem and Beitin have been found and published. Measuring backward from the first milestone clearly indicates that the zero milestone was at least as far south as the Church of the Holy Sepulcher and possibly as far south as the Nea Church (both on the Medeba Map) If so, it allows us to add several tenths of a mile to our measurements by odometer, putting Beitin almost 14 Roman miles north of Jerusalem and el-Bireh at least I 1 1/2.
What else can be learned from the milestones? In the late 19th and early 20th centuries several milestones still in place were pub-fished by explorers and scholars. Some had inscriptions. One already referred to was the First milestone. Although without an inscription, it was only one-half mile north of the Damascus Gate. The Third did not have an inscription, either. But the Fifth (over a mile south of er-Ram: biblical Rama?) was very clear, with the distance legible in both Greek and Latin. There was space for yet a Sixth milestone before turning off the main road to Eusebius’ “Ra-ma,” and he mentions this. He also mentions that one turned into “Rama” at the Seventh RMS when traveling from north to south. Obviously, his “Rama” was between the Sixth and Seventh RMS. This checks out today. It is
Medeba Map of Jerusalem, made of mosaic, found in a 4th century Byzantine church floor in Medeba, Jordan. Note the northernmost (Damascus) gate on the left and the column nearby. The colonnaded Cardo Maximus goes through the center with the Church of the Holy Sepulcher, lower middle, and the Nea Church at the end, upper right.
BSP 1:1 (Winter 1988) p. 29
about 5.5 statute miles from the Damascus Gate to a point just opposite er-Ram on the main road (6-6.5 Roman miles from the center of Jerusalem).
The only other milestone we are aware of was mentioned by Michael Avi-Yonah is his Map of Roman Palestine (1940). It was discovered about 1 mile south of el-Birch. He calls it the “Tenth” RMS from Jerusalem. However, in correspondence with him, he said there was no inscription on it. It got Its number from the location. If it truly was the Tenth, then the Eleventh was near el-Birch and the Thirteenth was near Beitin, now too far away. But, according to our other measurements, Avi-Yonah’s milestone should be the Eleventh, making el-Bireh near the Twelfth and putting Beitin near the Fourteenth.
Biblical Bethel at el-Bireh?
We might conclude, then, that modem el-Bireh covers ancient Bethel. And Beitin, instead of being “Bethel,” might conceal some other ancient town such as Zem-araim, Ophrah, or even Beth-Aven (as suggested by Captain Conder in the 1880’s and others since). Certainly the latter is not impossible. Phonologically it seems as feasible as equating “Beitin” with “Bethel.” Still another scholar has suggested that Beitin might even be biblical Ai, if it is not Bethel. Although the topography and location do not seem to fit the biblical requirements for Ai, it does have most of the required archaeological periods represented as well as the necessary gates and fortifications. Furthermore, and this may be very significant, if Beitin were Ai, it is located on the road between Jericho and Shechem. To grasp the importance of this, we recall that the Conquest sequence was Jericho, then Ai, then a renewal of the covenant at Shechem (Joshua 6–8).
How Verify New Bethel and Ai?
Even if our radical conclusion about relocating Bethel is reasonable, we have not been able to find a way to verify that el-Bireh might be Bethel. It is a heavily-populated modem city. One section of the city, however, has a high point called “Ras et-Tahu-neh.” It was surveyed by the Israel Department of Antiquities in 1969. Surface finds indicate that It was occupied in almost every period of ancient times. If Bethel is at el-Bireh, this high point is not necessarily Bethel, but it needs to be excavated because of its strategic location, whatever it is. However, Ramallah/el-Bireh is very tense politically. So excavation is not advisable for now. How can we test our proposal that Bethel and At must be relocated?
An interesting and possibly important item helping to identify Bethel is that roads in and out of Ramallah/el-Bireh converge at the narrow ridge on which the cities sit. It forms an hourglass-like configuration, in which roads north and south of the towns go in every direction. But they all come together while actually going through the towns. This is necessitated by the extremely deep and rugged wadis extending east and west of the towns. Taking all this into consideration, the high point on which the towns sit would be very strategic in controlling travel going north or south.
This is not true of Beitin. It lies in a relatively level area and does not seem strategic for controlling travel in the area, although a road to Jericho and one of the roads going to Nablus go past it. We think Jeroboam, in establishing a
BSP 1:1 (Winter 1988) p. 30
worship center with a golden calf at Bethel, would more likely have chosen the constricted area at Ramallah/el-Bireh. This way he could control the travel of pilgrims from the northern kingdom as they tried to go to Jerusalem (I Kings 12:25–33).
Returning to our efforts to locate Bethel, one way to try to solve this problem was to look for a site east of el-Bireh, on the other side of the mountain. We recall that in Genesis 12:8 Abraham camped on a mountain with Bethel on the west and Ai on the east. (Incidentally, there is no mountain at all east of Beitin.) Jebel et-Tawil (“the tall one”) is the Arabic name of the mountain on the east edge of el-Birch. An Israeli settlement recently built there retains this meaning in Its own name of “Psa-got,” “(The) Heights.”
We went to the other side of et-Tawil and found an ancient site named Khirbet (“ruins off) Nisya. Local Arabs told us the name “Nisya” means “forgotten.” No one knows what the ruins are. Although it sits on a natural rise, Nisya is not a “tell” with ancient remains in much depth. Bedrock can be seen at the surface all over the site. However, in spite of its having been quite denuded, the site was occupied almost continuously from early times, as we shall see. To prepare for its exploration, we requested an excavation permit from the Israel Department of Antiquities and received it in 1978. Since then we have dug for six short seasons, beginning in 1979.
Criteria for Biblical Ai
As we see it, for a site to be biblical Ai, it must meet at least four criteria: 1) The topography (hills and valleys) must fit the biblical description; 2) It must be properly located in relationship to other towns mentioned in Scripture; 3) It must be occupied or deserted at the same periods as the Bible has; 4) It must have been fortified with walls and a gate just prior to the Israelite invasion. Without going into the topographical and geographical details (presented in our 1970–1971 articles in The Westminster Theological Journal), let us look at the archaeological evidence. But before doing so, it can be noted that the topographical and geographical details are well-suited for Khirbet Nisya to be biblical Ai. The battle could be reenacted there today with every detail matching. This does not prove it is Ai. But, it meets the requirements of the first criterion.
Results of Six Seasons Excavations
Since we have only published the results of our excavation “in house” and with seasonal reports to the Department of Antiquities of the State of Israel, most scholars consider Roy Blizzard’s surface exploration in 1973 as the last word. The earliest occupation tie found was Iron Age I which, if it were all, would automatically rule out Khirbet Nisya as Ai. Actually, most scholars have done just that – ruled it out, based on the lack of archaeological evidence. We cannot blame them for we would do the same.
However, we are able to report that we have found considerable material earlier than Iron Age (IA) I, indicating that Khirbet Nisya was occupied throughout Middle Bronze (MB) II and into Late Bronze (LB) I (2000–1400 BC). (A Department of Antiquities survey in 1978 seemed to indicate occupation in Early Bronze also.)
The biblical/archaeological periods when finds should be rep-
BSP 1:1 (Winter 1988) p. 31
1. Oldest coin found at Kh. Nisya -Persian period, ca. 370 BC. Reverse shows war galley and waves. (Silver}
2. Obverse shows King (L) shying lion.
3. Royal seal found at Kh. Nisya from Hezekiah’s time, a/e melech seal. Bird is clear, writing is not.
4. Seal on jar handle from Hyksos time (ca. 1600 BC), found at Kh. Nisya. This is a coiled rope design, probably from a scarab seal.
5. Iron Age II seal with distinct lettering found at Kh. Nisya. 2”he two clear Hebrew letters are ayin and yod. It is probably someone’s name.
6. The most recent coins found at Kh. Nisya were on this torque, or headdress. These 28 coins are all silver (as well as the coil) and range from 1480–1550 AD.
BSP 1:1 (Winter 1988) p. 32
resented at Ai are: the Patriarchal (MB, possibly even Early Bronze), the Conquest (LIB), late Israelite (IA If), and the Return (Persian). Other periods might be present. But they do not mean much one way or the other biblically.
We have purposely not been very specific on the periods mentioned above, especially the first two, since there is controversy over the biblical dates and some diversity of opinion as to the naming and dates of the MB phases. Since we hold to the early biblical dates for the Patriarchs and the Exodus/-Conquest, we had hoped to find very early MB and LB I material, with no LB II. This should be the situation, since the site would be unoccupied by Canaanites for a time after a 1400 BC Conquest.
Fortunately, we do have finds for all the required periods, or we would not suggest the radical changes above. There are ceramics for what we will call MB II A,B and C; LB I local ware {no imports); both IA I, and particularly IA II to the end (586 BC); and Persian ceramics, as well as the two earliest coins (Persian, from 380 BC). We also have Hellenistic pottery and coins, with almost 30 coins from Alexander Janneus alone. Following this is Early Roman, but almost no Late Roman. Then, we have Early and Late Byzantine represented by ceramics, many coins and most of the remaining architecture on the site. And, finally, Early Arabic, after which the site seems to have been abandoned, for the most part. Even though quite a denuded site, it seems to have had fairly continuous occupation for long periods of time.
One other occupational period should be considered. Eusebius mentions that Ai was “deserted” in his day, about 330 AD. This is of interest because a gap exists in both coins and ceramics from the late first century until mid-fourth century. There are two coins from this “gap.” But they are widely separated in time and seem to be strays. (We have found about 160 coins to date, which in themselves give us a fairly clear picture of occupation and abandonment from 380 BC until the 8th century AD.)
Three Problems with Khirbet Nisya Being Ai
There are problems, however, in equating Khirbet Nisya with Ai. The first is that to this date we have not found any architecture earlier than Hellenistic (except for, possibly, one wall of a Persian building and one or two walls from the Iron Age). It is very clear that each succeeding occupation has gone down to bedrock for their building activity, often carving installations into it. A very large mikveh (Jewish immersion pool) with adjoining cistern, for instance, was almost certainly cut out in the Hasmonean period. Pits cut into bedrock all over the site are similar to those found by Prit-chard at Gibeon for the Hellenistic period. Byzantine foundations are laid everywhere on bedrock. We could give more illustrations. But added to this is the lack of soil accumulation. It is only one or two meters above bedrock at its deepest. Considering all this, should we expect architecture from earlier periods, even though there may have been substantial settlements from time to time?
The second problem is the lack of walls and gate. To fit the biblical description, which speaks of a gate (and you don’t usually have a gate without walls], any candidate for Ai should have both. We have neither. While some may eventually be found, none have so far,
BSP 1:1 (Winter 1988) p. 33
even though we have looked for them. We have also looked for some evidence of an emergency (siege) water supply, but have not found any yet.
The third problem is that we do not seem to have any occupation for the time of Abraham (no EB III or IV, nor MB I materials).
Can the Problems be Reconciled?
As we consider the first problem mentioned, we need to recall that neither is there early architecture (EB, MB, LB) at several nearby sites [Tell en-Nasbeh (Mispeh?), Tell el-Jib (Gibeon), and Tell el-Ful (Gibeah?)]. This is the case even though eramics were found for most of those periods at them.
As for the second problem – lack Of walls and gates at Khirbet Nisya – it is important to note how city walls were built in late MB II and probably into LB I (where MB walls seem to continue in use in some cases). Almost without exception in the MB sites studied, city walls were butt on a field stone base about one meter high (for drainage?] which were topped with mudbricks to the desired height. Examples of this kind of wall can be found at Beth-Shem-esh, Shechem, Beitin, Jericho, Dan (notably sot), Hazor, Gezer, Aphek, Tel Zeror, Khirbet Zuk-eriya (only I acre in size) and at many other sites, as well.
If this kind of wall was covered with debris, it has been preserved. If not, it has “melted.” It may be seen how fast this could happen by examining the recently discovered mudbrick walls and gate towers at Tel Dan. They are deteriorating so rapidly that there is real consternation as to how to preserve them at all. We do not seem to have discovered how the Canaanites were able to build walls and gates of mudbrick and have them even be preserved long enough to be useful for defense!
If there ever was a wall around Khirbet Nisya, it may only have been during the Canaanite occupation. If so, it should have been of mudbrick with a fieldstone base like those mentioned. But should we really expect to find mudbrick walls and gate at a site as denuded as Khirbet Nisya? Further, a fieldstone base could easily have been completely dismantled or incorporated into a terrace wall and therefore not be discovered.
What we have tried to do in the above is to suggest how there could be an MB-LB presence at Khirbet Nisya (which there obviously was) without there being any wall or gate remaining, even though there might have been one there at that time. On the other hand, if this turns out to be the case, it would be the first walled site excavated which has left no trace of a wall, gate, or water system. This may be a real minus in the identification of Khirbet Nisya with Ai.
One other important observation is that the three nearby sites mentioned at the beginning of this section all have Iron Age city walls with foundations laid on bedrock, thus obliterating any evidence of earlier walls. Nisya has nothing like this to account for the absence of city walls.
As for the third problem, we may yet discover remains from the patriarchal period. However, as excavation has proceeded without finds from this period, prospects of ever finding them become less likely.
Conclusion
To keep a proper perspective on the problems involved in identifying Bethel and Ai, there are at least four things to consider: 1) That topography alone cannot be
BSP 1:1 (Winter 1988) p. 34
used to identify a site. But, in this case, there are few other places in the area that fit the detailed description given in the Bible. 2) Geographical relationships should meet all the biblical specifications. Even though they do, they will not alone prove the identification, either. Butifboth topography and geography meet the requirements, we may be closer to a positive identification. 3) Even though we have not been able to determine that it was a fortified city or a substantial settlement in earlier periods (serious detractors), Khirbet Nisya nonetheless does have almost all of the necessary archaeological periods represented. This archaeological evidence, coupled with its meeting biblical specifications for both topography and geography, should make it a candidate for biblical AI. 4) As more evidence becomes available, it will either eliminate Khirbet Nisya as a candidate or it will become heavier in its favor.
Bible and Spade 1:2 (Spring 1988)