The Patriarchs in Their Ancient Setting

Summary

From the 1930s to the 1970s there was a scholarly consensus that Genesis 12–50 fitted a second-millennium BC world.  That consensus collapsed, and the collapse was conceded by conservatives as well as critics.  This article reports what was lost, what survives, and what the anachronism charges are actually worth.

• The Nuzi case is dead, and a conservative scholar wrote its obituary. Cuneiform tablets from Nuzi, an East Tigris site of the mid-second millennium, were held to explain patriarchal customs that "cannot otherwise be explained either from the local background or from any but Hurrian sources," in E. A. Speiser's words.¹ Martin Selman's 1974 Tyndale Biblical Archaeology Lecture — a conservative venue — examined the parallels one by one and concluded: "One important conclusion of our investigation is that no special relationship exists between the Nuzi tablets and the patriarchal narratives."¹ Thomas Thompson (1974) and John Van Seters (1975) reached the same verdict from the critical side. Anyone still citing Nuzi to date Abraham is forty years out of date.
• Both sides agree the first-millennium comparative record is barely studied, and that agreement settles the customs argument. Van Seters: "How can the claim be repeatedly made that the customs in the patriarchal stories reflect only the 'early or mid-second millennium' unless it is shown that they were no longer current at a late date? … nothing comparable exists for the first millennium."² Selman concedes the same fact.¹ A conservative cannot use the absence of first-millennium parallels as evidence while granting that nobody has looked — and Van Seters cannot use it the other way either. The comparative-customs argument cannot currently date these narratives in either direction.
• What Selman salvaged is much narrower and is still worth having. His positive point is not about Nuzi but about distribution over time. The patriarchal customs he examined "were quite at home in Mesopotamian culture of the second millennium, whereas some of them, such as the provision of a slavegirl by a barren wife for her husband, are unknown in Israel and Judah in the first millennium BC."¹ He argues that knowledge of that kind "can only be gained through sustained rather than occasional contact," and that no such opportunity existed in pre-exilic first-millennium Israel.¹ Note what this is: an argument about the tradition's origin, not a date for Abraham. Selman says so himself — "the chronological limits cannot be fixed by data based on customary law alone."¹
• The camel charge is the strongest anachronism argument and it is narrower than the headlines. Sapir-Hen and Ben-Yosef radiocarbon-dated camel remains from the Aravah copper-mining sites and concluded the first significant appearance of domestic camels there was not earlier than the last third of the tenth century BC.³ That is a real result about a real region. It is also a result about when camels became common as industrial pack animals in the southern Levant copper industry, which is not quite the claim Genesis makes — Genesis has a wealthy pastoralist owning some camels among sheep, cattle, and donkeys (Genesis 12:16).⁴
• The counter-evidence for second-millennium camels is stronger than critics allow, and its best items are not from apologists. Peter Magee, on two thirteenth-century Middle Assyrian texts from Ugarit and Nippur: "it might be concluded that domesticated dromedaries are known during the thirteenth and twelfth centuries BC."⁵ Martin Heide, whose 2010 survey is the fullest specialist treatment, holds that the Mesopotamian inscriptional evidence "requires [the dromedary's] domestication around the middle of the 2nd millennium or before."⁶ Old Babylonian already has distinct standard terms for the two species — am.si.kur.ra for the Bactrian, anše a.ab.ba, "donkey of the sealand," for the dromedary.⁷ Add an Alalakh ration tablet allocating fodder to a camel, Egyptian images from the eighteenth and nineteenth Dynasties, and Wadi Nasib petroglyphs.⁸⁹ Redford's "not until the ninth century" is refuted and should stop being quoted.
• But the residual problem is chronological and it does not go away. Heide's estimate is about 1500 BC. A patriarchal date near 2000 BC is still several centuries earlier than the best specialist figure, and Sapir-Hen and Ben-Yosef's southern-Levant result is untouched by Mesopotamian and Egyptian evidence. Genesis 12:16's handful of camels can now be accommodated; Genesis 24's caravan of ten fits less comfortably. A further caution about the conservative literature: it repeatedly runs Bactrian and dromedary evidence together, and third-millennium Bactrian domestication in Turkmenistan is irrelevant to Genesis.
• Millard's counter-argument is better than the camel argument itself, and it is rarely answered. He points out that the horse does not appear in the patriarchal narratives at all until Joseph's Egypt (Genesis 47:17).¹⁰ Horses were expensive and prestigious throughout the Iron Age, and "a picture of a wealthy man moving in Palestine drawn during the Late Bronze or Iron Ages might be expected to mention horses at some point, as a sign of his riches."¹¹ A first-millennium writer padding Abraham's wealth would reach for the horse. Genesis does not. That is an argument from what is absent, which is the kind that is hard to fake.
• The Philistine references are a genuine anachronism, but Kitchen's defence of them is better than the usual one and this article previously underrated it. Genesis 21 and 26 have Abraham and Isaac dealing with Abimelech "king of the Philistines" at Gerar;¹² the archaeological Philistines arrive around 1175 BC. The standard answer is later editorial updating, which is plausible and normally unfalsifiable. Kitchen adds two things. He shows the older terminology surviving inside the Hebrew Bible — Joshua 13:3's Philistine cities "reckoned as Canaanite" with the Avvim, Deuteronomy 2:23's Caphtorim, the prophets' archaic "Kerethites" and "Caphtor," and Caphtor as the well-attested early-second-millennium Kaptara.¹³ And he asks why, if this is a late insertion, the "Philistine" king sits at Gerar rather than Gath or Gaza, with no army, in stories that "would read just as well with Canaanite characters."¹³ A fabricator does not produce that. The anachronism stands; it no longer looks like a clean one.
• "Ur of the Chaldeans" is a smaller version of the same problem. Genesis 11:31 has Terah leave "Ur of the Chaldeans."¹⁴ The Chaldeans are attested as a people in southern Mesopotamia only from about the ninth century BC. The most economical reading is that "of the Chaldeans" is a later clarifying gloss telling readers which Ur was meant — a scribal helpfulness, not a claim about Abraham's neighbours. This is a reasonable explanation and it is also, again, unfalsifiable.
• Thompson's own book, read directly, damages the surviving conservative argument more than Selman's concession did. His sampling-bias point is the sharpest thing in the literature on either side: the second millennium is both better documented and more heavily searched for patriarchal-type names, so hits there are expected whenever the names were coined. "A fourth century occurrence of the name of Abraham in South Mesopotamia would hardly create much excitement!"¹⁵ The onomastic argument needs a rate, not a list, and nobody appears to have computed one. And where Selman says the slave-girl custom is unknown in first-millennium Israel and Judah, Thompson concludes the customs were not confined to a short period and were not lost in the first millennium at all.¹⁵
• The strongest positive argument is Genesis 14's coalition, and this article previously omitted it. Kitchen shows the four eastern kings' names fitting their regions — Chedorlaomer Elamite, Tidal the Hittite Tudkhalia with an Anatolian paramount-chief title, Arioch attested at Mari as Arriwuk — and then makes the better argument, that four-or-five-king power alliances were a political configuration of one window: "from circa 2000 to 1750 (1650 at the extreme), we have the one and only period during which extensive power alliances were common in Mesopotamia."¹³¹³ That is period-specific rather than merely period-compatible, which no other argument here is. Its weakness is Thompson's: the Mari archive is where those alliances are attested and where the name parallels come from, so the window may be an artefact of where the evidence survives. Kitchen notes Mari "covers only about fifty to seventy years." And no king in Genesis 14 has been identified.
• The strongest surviving positive evidence is onomastic and legal, and both are modest. Patriarchal names are of Amorite type — the Ya'qub-el and Abi-ram patterns are well attested in second-millennium sources and are not first-millennium Israelite naming fashion. And Genesis 23's purchase of the cave of Machpelah has Abraham weighing out 400 shekels "according to the standard measurement at the time," in a negotiation conducted publicly at the city gate before witnesses.¹⁶ Weighed silver rather than coined money is correct for any pre-Persian period and is an incidental detail a late writer could get wrong.
• Honest weaknesses. There is no external attestation of any patriarch. No inscription, no archive, no site. The patriarchal narratives are the only part of the Old Testament's national story for which this is true and remains true. The Sodom question, which is the one place archaeology could in principle intersect, is unresolved — see Sodom and Gomorrah. And the "second-millennium colouring" arguments that survive are compatible with an accurately transmitted tradition and with a well-informed later author, which means they cannot settle the date of composition.

Verdict: the patriarchal narratives contain real second-millennium colouring and no second-millennium anchor.  The maximalist position of the mid-twentieth century — that Nuzi and Mari had proved Genesis 12–50 a second-millennium document — is dead, and conservatives helped kill it.  The minimalist position, that the narratives are wholly first-millennium fiction, has to explain a set of customs unknown in first-millennium Israel, a naming pattern that is not first-millennium Israelite, weighed silver, and a rich man with no horses.  What survives is a tradition that knows things a late inventor would be unlikely to know, embedded in a text that also contains at least one anachronism — though Kitchen's four questions about that anachronism (why Gerar, why no army, why Philistines at all) make it a much untidier one than the charge assumes.  The most defensible reading is that Genesis preserves genuinely old material that has been transmitted and updated, and that no external evidence bears on whether Abraham himself existed.

Part 1 — What is being tested

"Were the patriarchs historical?" is not one question.

Did Abraham, Isaac and Jacob exist as individuals?  Untestable by current means.  There is no inscription, no archive entry, no burial, no site.  This is not a suspicious silence — private individuals from the second millennium BC do not normally leave records unless they were kings or scribes — but it means no evidence bears directly on the question.

Do the narratives reflect a real second-millennium world?  Testable, and the subject of this article.  If Genesis 12–50 contains customs, names, prices, and legal procedures characteristic of the second millennium and not of the first, that is evidence about the age of the tradition.

Do they contain anachronisms that place their composition later?  Also testable, and the other half of this article.

These two tests can both come back positive.  A late writer using old sources produces exactly that result.  It should be said clearly at the outset that neither test can settle whether Abraham lived.

Part 2 — The rise and fall of the Nuzi case

What was claimed

In the 1930s and after, roughly 4,000 cuneiform tablets from Nuzi — a Hurrian-influenced town east of the Tigris, mid-second millennium — were read as illuminating patriarchal customs.  A cluster of Genesis episodes appeared to have specific parallels:

• Sarai giving Hagar to Abram because she was barren (Genesis 16:1–4).¹⁷
• Adoption of a household servant as heir in default of a son (Genesis 15:2–3).
• The sale of a birthright (Genesis 25:29–34).
• Rachel taking her father's household gods (Genesis 31:19).

E. A. Speiser claimed some patriarchal customs "cannot otherwise be explained either from the local background or from any but Hurrian sources."  John Bright concluded that "the patriarchal customary law was at home specifically among the Hurrian population of approximately the same area at approximately the same time."¹  This was the standard position in a generation of textbooks.

Why it collapsed

Three problems, and Selman's lecture works through them in detail.

The sample was tiny and cherry-picked.  Of roughly 4,000 Nuzi tablets, "no more than a dozen are quoted with any regularity" in this connection.  Of the roughly 300 Nuzi texts on family law, "only four or five are regularly considered."¹  A parallel drawn from four texts out of three hundred is not a characterisation of Nuzi law.

The parallels were not exclusive.  The customs turn out to be attested across Mesopotamia over long periods, not distinctively Hurrian and not distinctively second-millennium.  The specific claim — that only Hurrian sources explain these practices — fails on its own terms.

Some of the readings were simply wrong.  Speiser's Hurrian interpretations of particular Nuzi texts have been rejected by Assyriologists on grounds internal to the tablets.

Thomas Thompson's The Historicity of the Patriarchal Narratives (1974) and John Van Seters's Abraham in History and Tradition (1975) pressed these points hard and are usually credited with ending the consensus.  What is less often noticed is that Selman's Tyndale lecture, delivered from a conservative platform in the same year, reached the same conclusion independently:

One important conclusion of our investigation is that no special relationship exists between the Nuzi tablets and the patriarchal narratives.  If this indicates that the results are largely negative, then it must be recognised that an approach of this kind is sometimes a necessary preparation for real progress.  To identify a false trail can be an essential task before the correct path is to be made clear.¹

This is worth pausing on.  A scholar giving a lecture named for a conservative research institute told his audience that the flagship argument of conservative Old Testament archaeology was a false trail.  That is how the thing is supposed to work, and it is a better advertisement for the field than any of the parallels were.

What Selman rebuilt

He does not leave the ground empty.  His positive argument abandons Nuzi specificity and rests on distribution over time:

the patriarchal customs discussed … were quite at home in Mesopotamian culture of the second millennium, whereas some of them, such as the provision of a slavegirl by a barren wife for her husband, are unknown in Israel and Judah in the first millennium BC.  Presumably the traditions in Genesis 12–50 containing references to these customs had a thorough acquaintance with a Mesopotamian way of life, since knowledge of this kind can only be gained through sustained rather than occasional contact.  No opportunity for a long term acquaintance of this nature existed in pre-exilic Israel in the first millennium BC, so that the second millennium appears to provide the only adequate alternative.¹

How strong is this?  Moderately.  Three cautions, one of which Selman states himself.

His own caution: "the chronological limits cannot be fixed by data based on customary law alone."¹  He is not dating Abraham.

A second: he notes that "much less is known about comparable customs from the first millennium BC in Mesopotamia and Syria" than about the second.¹  An argument from the absence of first-millennium parallels is weakened by an incomplete first-millennium record.

A third, which he does not address: Van Seters argued that Neo-Babylonian parallels are in fact closer to the patriarchal narratives than Mari or Nuzi.  If that is right, the exilic and post-exilic periods supply exactly the sustained Mesopotamian contact Selman says pre-exilic Israel lacked.  The Judean community at Āl-Yahūdu, doing business in Akkadian legal forms in Babylonia for a century, is documented — see Exile, Babylon, and Persia.  Selman's argument rules out a pre-exilic first-millennium origin.  It does not rule out an exilic one.

Van Seters, and the asymmetry both sides admit

Van Seters's 1975 book is the other half of the pair that ended the consensus, and read directly it makes one argument that deserves to sit at the centre of this whole question:

How can the claim be repeatedly made that the customs in the patriarchal stories reflect only the "early or mid-second millennium" unless it is shown that they were no longer current at a late date?  While there has been a great deal of interest in the social and legal history of the second millennium which has produced numerous studies on many different facets of society, nothing comparable exists for the first millennium.  The obvious prejudice for the second millennium, created largely by the mood in biblical studies, has resulted in very meager treatment of the first millennium materials.²

He adds that the resulting one-sidedness "does not inspire confidence in existing studies and more closely resembles an apologetic than a scholarly investigation."²

The striking thing is that Selman concedes the same fact.  In the Tyndale lecture whose positive argument this article reports above, he grants that "much less is known about comparable customs from the first millennium BC in Mesopotamia and Syria."¹

So both sides agree on the asymmetry, and that agreement is decisive.  Selman's surviving argument is that certain patriarchal customs are unattested in the first millennium.  Van Seters's reply is that the first millennium is barely studied, so "unattested" there carries almost no information.  A conservative cannot use the absence of first-millennium parallels as evidence while also granting that nobody has looked.  This is the same defect this article identifies in the onomastic argument, and in both cases the fix is the same: the argument needs a comparison of rates in comparably-searched corpora, and nobody has assembled one.

It should be said that this cuts against Van Seters too.  If the first-millennium record is too thin to support "these customs persisted," it is also too thin to support "these customs are late."  The honest conclusion is that the comparative-customs argument cannot currently date the patriarchal narratives in either direction, which is very close to what Selman said in his own last sentence on the subject.¹

Thompson's own argument, read directly

This article previously represented Thompson entirely through Selman, which was unsatisfactory for the book that ended the consensus.  Read directly, two of his arguments matter and one of them cuts at Selman's surviving position.

The sampling-bias argument, which is the sharpest thing in the book.  The whole second-millennium case rests on parallels — names, customs, legal forms — found in second-millennium archives.  Thompson asks why we should expect otherwise:

Do we have more evidence from the earlier period because more written texts have been discovered from that period (Mari, the Cappadocian texts, Nuzi, Alalakh, Ta'annek, Ugarit, Amarna, etc.), and also because more scholars have been looking for "patriarchal-type" names in these early texts than in later texts?  A fourth century occurrence of the name of Abraham in South Mesopotamia would hardly create much excitement!  How much has our evidence been determined by the understandable desire of the researchers to support their theses?¹⁵

This is a real problem and it bears directly on the onomastic claim above.  If the second millennium is both better documented and more heavily searched for patriarchal-type names, then finding them there is what one would expect whenever the names were coined.  The argument needs a rate — how often such names occur per thousand attested names in each period — not a list of hits.  This project has not found anyone who has done that calculation, and until someone does, the onomastic argument should be held loosely.  It is the same defect this project identifies elsewhere as reasoning from a trace rather than from the thing, run in the positive direction.

And his conclusion on the customs contradicts Selman's.  Having worked through the later material, Thompson writes that

there is no single element nor combination of elements which these later texts show, that is not found also in earlier material.  It, on the other hand, becomes clear, on the basis of this later material, that we are not dealing in Genesis with customs which existed in any single limited area for a short period of time, which then were no longer understood in the First Millenium when the stories of Genesis were written.¹⁵

Selman says the slave-girl surrogacy custom is "unknown in Israel and Judah in the first millennium BC."  Thompson says the customs are not confined to a short period and were not lost in the first millennium.

These are not quite contradictory, and the difference matters.  Selman's claim is about Israel and Judah;  Thompson's is about the ancient Near East at large.  Both can hold: a practice can persist in Mesopotamia and Egypt while dropping out of Israelite usage.  But that is precisely what weakens Selman's inference.  His argument was that a first-millennium Israelite could not have known these customs because no sustained Mesopotamian contact was available.  If the customs were still current in Mesopotamia, then a Judean in Babylonia could meet them — and Judeans in Babylonia are documented, doing business in Akkadian legal forms, in the Āl-Yahūdu archive.

Thompson also supplies a better parallel than Nuzi, which is worth noting because it is not the one either side usually cites: an Egyptian text from the reign of Ramesses XI concerning a childless marriage, a slave concubine, children who become heirs, and a wife with influence over the inheritance — with the children described as "no longer with him as servants, but … as brothers and children, being freedmen of the land."¹⁵  Compare Genesis 16 and 21, where Sarah likewise has influence and Hagar's status as slave is at issue.  Thompson thinks Van Seters overrates it, but it is a late-second-millennium Egyptian parallel to the episode Selman's case rests on.

And his thesis is more radical than "the parallels fail."

Not only has the historicity of Abraham not been proven, but it does not seem to be implied in the biblical narratives themselves.¹⁸

That second clause is a literary claim, not an archaeological one, and this article does not accept it.  Genesis 23's property transaction before witnesses at a city gate, with a stated price in a stated weight standard, is not written in the register of a text indifferent to whether it happened.  But it should be recorded that Thompson's argument runs deeper than the parallels, and that answering the parallels does not answer him.

Part 3 — The anachronisms

Camels

The charge.  Camels appear as domesticated pack and riding animals throughout Genesis — twenty-two times — and in Exodus, Judges and Job.  Extra-biblical evidence for domestic camels in the region is much later.  Donald Redford: "anachronisms do indeed abound … camels do not appear in the Near East as domesticated beasts of burden until the ninth century B.C."⁸

The 2013 study.  Lidar Sapir-Hen and Erez Ben-Yosef radiocarbon-dated camel remains from copper-mining sites in the Aravah valley and concluded that the first significant appearance of domestic camels there was not earlier than the last third of the tenth century BC.³  This is careful, dated, published work and it moved the standard estimate down from the twelfth century.

What it does and does not show.  It shows when camels became a significant presence in one industrial region of the southern Levant.  It does not, and does not claim to, show that no domestic camel existed anywhere in the Near East earlier.  Press coverage at the time flattened this distinction badly.

The counter-evidence.  Advocates of an earlier date cite three second-millennium texts: an Old Babylonian Sumerian text from Nippur alluding to camel's milk;  a Sumerian lexical text from Ugarit listing a camel among domesticated animals;  and a ration tablet from Alalakh Level VII (269:59) allocating fodder to a camel.⁸  They also cite third-millennium camel material from Shahr-i Sokhta in Iran, terracotta camel-drawn cart models from Altyn-depe in Turkmenistan, and camel bones from Arad and Jericho.⁸

Assessment.  The Alalakh ration text is the strongest, because a fodder allocation in an administrative list implies an animal being kept and fed.  Against that, the cuneiform signs in these texts are read differently by different Assyriologists, and Levantine bones in third-millennium strata may be wild camel or intrusive.  The Bactrian evidence from Central Asia is stronger than the dromedary evidence from the Levant, and it is the dromedary that matters for Genesis.

The stronger version of the counter-case, which this article previously understated.  Two of the items are from specialists with no stake in Genesis, and they matter more than the rest.

Peter Magee discusses two Middle Assyrian texts of the thirteenth century BC containing the term anše.a.ab.ba, one from Ugarit and one from Nippur, and concludes: "On the basis of this evidence, it might be concluded that domesticated dromedaries are known during the thirteenth and twelfth centuries BC."⁵

Martin Heide, whose 2010 survey in Ugarit-Forschungen is the fullest specialist treatment, summarises the Mesopotamian inscriptional evidence as requiring the dromedary's "domestication around the middle of the 2nd millennium or before."⁶

And the lexical point is the best single item in the file.  Old Babylonian texts (c. 1894–1595 BC) already distinguish two animals: am.si.kur.ra and am.si.ḫar.ra.an for the Bactrian, and anše a.ab.ba, literally "donkey of the sealand," for the dromedary.⁷  A language does not develop separate standard terms for two camel species it does not handle.

Other items, rated.  A bowl fragment from Piramesse with an incised dromedary sketch, late eighteenth or early nineteenth Dynasty;  Petrie's statuette from Rifeh of a dromedary carrying two water jars, nineteenth Dynasty, which scholars have called "sufficient proof for the presence of camels in Ramesside Egypt";  petroglyphs in the Wadi Nasib dated to the sixteenth century BC showing a dromedary led by a man, beside proto-Sinaitic inscriptions near the Serabit el-Khadim mines;  an eighteenth-century Syrian seal showing figures seated on a Bactrian camel's humps;  and a camel skull from Tell Ta'anakh dated 2200–1550 BC.⁹  These are suggestive rather than decisive individually — a skull may be wild, a petroglyph is hard to date, an image may be exotic — but they are numerous and they point one way.

One thing the conservative literature does badly and should stop.  It repeatedly runs Bactrian and dromedary evidence together.  Bactrian domestication in Central Asia in the third millennium is well established and is irrelevant to Genesis, which needs dromedaries in Mesopotamia and the Levant.  The Old Babylonian anše a.ab.ba term, Magee's thirteenth-century texts and Heide's summary are the relevant items;  the Turkmenistan cart models are not.

A better argument from scale.  Shalmaneser III's record of the battle of Qarqar in 853 BC has him capture a thousand dromedaries from Gindibu the Arab.¹⁹  A thousand camels fielded in a coalition army is not the beginning of a domestication process;  it is a mature one, and it sits only a century or so after Sapir-Hen and Ben-Yosef's first significant Aravah appearance.  How long the build-up took is a matter of judgement, but the direction of the inference is sound.

Kitchen's version, which is the responsible one

Kenneth Kitchen's treatment in On the Reliability of the Old Testament is more careful than the apologetic literature that cites him, and it is worth setting against ABR's "camels were domesticated thousands of years before Abraham."

He starts from the text, and notices something the anachronism charge overlooks.  Between about 2000 and 1200 BC, camels barely appear in the Bible.  They are "last and least of Abraham's possessions (Gen. 12:16)," used "solely for the long-distance, desert-edge trip to Harran and back" for Isaac's bride, then for the trip from Harran to Canaan, then by the desert-travelling Midianites.  "This is remarkably little."  After the Exodus they occur "once among Pharaoh's transport animals (Exod. 9:3) and twice in lists of creatures not to be eaten."  His summary: "Not much of a presence at all!"¹³

That is an argument of the same shape as Millard's missing horse.  A later writer retrojecting a familiar animal into the patriarchal age would use it freely.  Genesis uses it grudgingly and for exactly the journeys a marginal desert pack animal would be used for.

Then he catalogues the external evidence, and it is a shorter and better-disciplined list than ABR's: a camel skull from the Fayum, Egypt, c. 2000–1400;  a kneeling-camel figurine from Byblos, nineteenth or eighteenth century;  a camel jaw from a Middle Bronze tomb at Tell el-Far'ah North, c. 1900/1550;  the eighteenth-century north Syrian cylinder seal in the Walters;  and the camel in the Sumerian lexical work HAR-ra-hubullu, "going back in origin to the early second millennium."  For the late second millennium: a kneeling camel loaded with two jars from a later-thirteenth-century tomb south of Memphis;  a camel on Qurraya painted ware of the thirteenth or early twelfth century;  and a camel on an early-thirteenth-century sherd from Pi-Ramesse.¹³

And his conclusion claims far less than his popularisers do:

the camel was for long a marginal beast in most of the historic ancient Near East (including Egypt), but it was not wholly unknown or anachronistic before or during 2000–1100.  And there the matter should, on the tangible evidence, rest.¹³

"Not wholly unknown" and "marginal beast" is the defensible claim.  This article's own assessment above was reached independently and lands in the same place.

And Kitchen's Philistine argument, which is better than the one usually given

The standard conservative answer to the Philistine anachronism is editorial updating, which this article rated as plausible but unfalsifiable.  Kitchen makes that argument — with a good analogy, that we say "the Dutch founded New York" although they founded New Amsterdam — but he also does two things the usual version does not.¹³

He shows the substitution leaving traces in the text.  Joshua 13:3 lists the five Philistine cities as "reckoned as Canaanite" and brings in the obscure Avvim;  Deuteronomy 2:23 has "the Caphtorim that came from Caphtor" replacing the Avvim;  the prophets use "Kerethites" and "Caphtor" as archaisms for Philistia (Zephaniah 2:5, Ezekiel 25:16, Amos 9:7, Jeremiah 47:4).  And Caphtor is ancient Kaptara, well attested from the early second millennium — the Mari archives record a king of Hazor sending gifts to Kaptara.¹³  So the older name for the region's population is recoverable from inside the Hebrew Bible, which is what an updating hypothesis predicts and what makes it something other than a bare escape.

And he asks four questions that nobody asks.

Why does the biblical narrator locate his "Philistine" king at Gerar and not in the pentapolis cities, e.g., in Gath or Gaza?  Why is there no Philistine army force to deal with the unwanted Hebrew pastoralists?  Why introduce a "Philistine" element at all?  (The stories would read just as well with Canaanite characters, as at Shechem!)¹³

This is the strongest form of the argument and this project had not encountered it.  A writer inserting Philistines anachronistically would put them in Philistia, with the military power Philistines actually had.  Genesis puts a single treaty-making king at Gerar, outside the pentapolis, with no army.  That is not what a fabricator produces.

How much it is worth.  It converts the Philistine problem from a clean anachronism into an untidy one.  It does not remove it: "the land of the Philistines" is still a first-millennium designation in a second-millennium narrative, and Kitchen's explanation of why it is there remains an inference.  But this article's earlier verdict — that the editorial-updating defence is "reasonable and unfalsifiable" — was too flat.  It has textual traces and it makes predictions about how the anachronism should look, and those predictions hold.

Where this leaves the charge.  Substantially weakened in its strong form, and surviving in a weaker one.

The strong form is refuted.  Redford's "camels do not appear in the Near East as domesticated beasts of burden until the ninth century" cannot be sustained against Magee's thirteenth-century texts, Heide's summary, the Old Babylonian lexical distinction and the Egyptian images.⁸  That claim should stop being made.

The weaker form stands.  Sapir-Hen and Ben-Yosef's result — that domestic camels were not a significant presence in the southern Levant before the late tenth century — is about a different question and is not touched by Mesopotamian and Egyptian evidence.

And the residual problem for Genesis is chronological, not zoological.  Heide's "middle of the 2nd millennium or before" is roughly 1500 BC.  A patriarchal date around 2000 BC is still several centuries earlier than the best specialist estimate of dromedary domestication, and "or before" is doing real work in that sentence.  Genesis has camels in small numbers among a pastoralist's holdings (Genesis 12:16), which the second-millennium evidence can now accommodate.  It also has Rebekah's family running a caravan of ten camels (Genesis 24), which fits less comfortably.

A note on method.  Millard makes a point about sampling that applies generally: "the camel is not an urban creature;  it is kept outside settlements and used primarily in the steppe."  So "the likelihood of camel bones being found is small, as is the likelihood of city scribes frequently naming it."¹¹  Excavations dig settlements.  This is a real bias and it should be stated — while noticing that it is also exactly the kind of argument that can excuse any absence, and that the Aravah study is a case where the bones were found and dated.

Philistines

Genesis 21:32–34 has Abimelech and his commander Phicol return "to the land of the Philistines," and Abraham stay there.  Genesis 26:1 has Isaac go to "Abimelech king of the Philistines at Gerar."¹²

The Philistines of the archaeological record — the Aegean-derived population with distinctive pottery, architecture and diet — appear in Canaan around 1175 BC, in the wake of the Sea Peoples' movements.  On any patriarchal chronology this is centuries too late.

The defences, honestly rated.

Editorial updating.  A later scribe replaced an obsolete ethnic name with the current one for the same coastal region, exactly as Exodus 1:11's "Rameses" appears to update an older Delta toponym.  This is entirely plausible as a scribal practice and there is a documented parallel in the same corpus.  It concedes that the text as it stands is not the text as it was.  In Kitchen's hands it is also more than a bare escape — he shows the older terminology surviving inside the Hebrew Bible and asks why the supposed anachronism does not look like one.  See below.

Earlier Aegean presence.  Some argue for pre-1200 Aegean populations on the Canaanite coast.  The evidence is thin and does not amount to "Philistines" in the sense Genesis uses.

This project's assessment: the Philistine references are an anachronism, and an untidy one.  The editorial-updating explanation is reasonable and, in Kitchen's version, does more work than an explanation usually does — it predicts that the older terminology should be recoverable from the text, and it is.  A reader who accepts the move for "Rameses" should accept it here;  a reader who rejects it here should stop using it there.

Ur of the Chaldeans

Genesis 11:31: Terah "set out from Ur of the Chaldeans."¹⁴  The Chaldeans (Kaldu) appear in Assyrian records as a people of southern Mesopotamia from around the ninth century BC.

The natural explanation is a gloss: a later editor specifying which Ur, for readers who knew southern Mesopotamia as Chaldea.  This is a smaller and more obviously editorial matter than the Philistines, and the same assessment applies.

The Hittites at Hebron

Genesis 23 has Abraham buy the cave of Machpelah from "Ephron the Hittite" among "the sons of Heth."¹⁶  The Hittite empire of Anatolia is second-millennium;  the neo-Hittite states of northern Syria are first-millennium;  neither obviously puts Hittites in Hebron.

The likeliest reading is that bene-Heth is a local Canaanite group whose name resembles the Anatolian one, rather than Anatolian Hittites resident in the southern hill country.  This is a genuine philological puzzle rather than a decisive anachronism, and it is not much use to either side.

Part 4 — What survives on the other side

Four items, listed in descending order of strength.

1.  The missing horse.  Genesis mentions no horses among the patriarchs.  The horse first appears at Genesis 47:17, when Egyptians trade horses to Joseph for food.¹⁰  Millard's argument:

A picture of a wealthy man moving in Palestine drawn during the Late Bronze or Iron Ages might be expected to mention horses at some point, as a sign of his riches or his eminence.¹¹

He adds the related point that a later writer inflating patriarchal wealth would substitute "horse," not "camel," because horses were the expensive prestige animal of the Iron Age.⁸  This is a genuinely good argument and this project has not found a published answer to it.

2.  Customs absent from first-millennium Israel.  Selman's surviving point: the barren wife providing a slave girl, and related practices, are at home in second-millennium Mesopotamia and are unknown in Israel and Judah in the first millennium.¹  Weakened by the Van Seters objection about Neo-Babylonian parallels and by the thinness of the first-millennium comparative record, but not eliminated.

3.  Weighed silver.  Genesis 23:16 has Abraham weigh out 400 shekels "according to the standard measurement at the time."¹⁶  Coined money reaches the Levant in the Persian period.  A writer after that could easily slip.  This is a small point — weighed silver persisted alongside coinage — but it is the right kind of incidental detail.

4.  Amorite-type personal names.  The patriarchal names follow patterns attested in second-millennium Amorite onomastics rather than first-millennium Israelite naming fashion.  This article has not verified the onomastic claim against a primary corpus and states it as commonly reported;  it should be checked before publication.

Part 4b — Genesis 14, which this article previously omitted

Genesis 14 has four eastern kings campaign against five kings of the Dead Sea plain, carry off Lot, and be pursued and defeated by Abraham.  It is the most historiographical chapter in the patriarchal cycle and this article did not treat it.  Kitchen's discussion is the fullest conservative case and it is stronger than his camel argument.

The names fit their regions.  Chedorlaomer is an Elamite name — Kutir plus a deity — and he is titled king of Elam.  Tidal, king of Goyim ("peoples"), bears the early Hittite name Tudkhalia, and his title corresponds to the rubaʾum rabium, the "paramount chiefs" attested in Anatolia in the twentieth and nineteenth centuries.  Amraphel's name is Semitic and his kingdom Shinar stands for Babylonia in Hittite, Syrian and Egyptian sources.  Arioch is well attested in the Mari archive as Arriwuk or Arriyuk in the early second millennium, and as Ariukki at Nuzi.¹³

Kitchen's own summary is properly limited: "the personal names fit the regions they ruled and correspond with real names and known name types, even if the individuals are not yet identified in external sources."¹³  No king in Genesis 14 has been found.

The stronger argument is about the shape of the coalition, not the names.  Kitchen walks through Mesopotamian political history period by period.  Before about 2300, Sumerian city-states;  then Akkad;  then the Third Dynasty of Ur.  After Ur III fell around 2000, Mesopotamia was divided among Isin, Larsa, Eshnunna, Assyria and Mari — until Hammurabi eliminated his rivals in the eighteenth century.  From about 1600 onward Assyria and Babylon dominated.  His conclusion:

Thus, from circa 2000 to 1750 (1650 at the extreme), we have the one and only period during which extensive power alliances were common in Mesopotamia and with its neighbors.  Alliances of four or five kings were commonplace and modest then.¹³

He cites a Mari letter for the scale: "There is no king strongest by himself — 10 or 5 kings follow Hammurabi of Babylon," and similarly for Rim-Sin.¹³

Why this is a better argument than the others in this article.  It is period-specific rather than merely period-compatible.  Camels, weighed silver and Amorite-type names are all consistent with a second-millennium setting and also with later ones.  A four-king eastern coalition raiding the Levant is, on Kitchen's reconstruction, a political configuration that existed in one window and not before or after.  That is the difference between consistent and diagnostic, and it is the distinction this project applies everywhere else.

Three caveats, and the first is serious.

Thompson's sampling-bias objection applies with full force here.  The Mari archive is the single best-documented body of evidence for exactly the window Kitchen identifies, and it is where the name parallels come from.  Whether multi-king alliances were genuinely confined to 2000–1750, or are simply best attested there because that is where the archives are, is precisely the question Kitchen's argument assumes an answer to.  Kitchen himself notes that "even the great Mari archive covers only about fifty to seventy years."¹³

No individual is identified.  Four kings, no matches.

And the reconstruction is contested.  Kitchen's survey of Mesopotamian political history is his own, and the claim that extensive alliances were impossible after 1600 is an argument from the character of Assyrian and Babylonian dominance rather than from an absence anyone has demonstrated.

Assessment.  This is the strongest positive argument for a second-millennium patriarchal setting that this article has found, and it is not decisive.  It would become decisive if someone showed that four-power coalitions of this type are absent from the first-millennium record in a corpus searched as hard as Mari has been.  Nobody has done that, and the same defect runs through every argument in this article's Part 4.

Part 5 — The binding of Isaac

Genesis 22 has God command Abraham to offer Isaac "as a burnt offering" and then stop him.²⁰

The historical question here is different from the others.  Nothing about the episode is externally testable.  What can be said is contextual, and it cuts against the usual modern reading in one respect: child sacrifice was a real practice in the ancient Levant and Phoenician world, attested archaeologically at Carthage and textually in the Hebrew Bible's own denunciations of the practice at the Tophet.  A story in which a god demands a child and then forbids it is intelligible as a text addressed to a world where the demand was not unthinkable.

That observation is not evidence for the episode's historicity and is not offered as such.  It is evidence that the episode is at home in an ancient Near Eastern setting rather than an invented moral puzzle of a later, more squeamish period — which is a claim about the tradition's age, of the same weak kind as the rest of this article's positive evidence.

Common objections

"Nuzi proves Abraham lived in the second millennium."  It does not, and conservative scholarship conceded this in 1974.  See Selman.¹

"The camel bones prove Genesis is late."  They show camels were not common in the Aravah before the late tenth century.  That is a narrower claim than the one usually built on it, and the second-millennium textual evidence, though contested, is not nothing.

"The Philistines are a clear anachronism, so Genesis is fiction."  The anachronism is real.  It shows the text was edited later than the events it describes, which almost nobody disputes about any part of the Pentateuch.  It does not show the underlying tradition is fictional, any more than a modern edition's spelling modernisation shows a medieval chronicle is fictional.

"Absence of evidence for Abraham is evidence of absence."  Not for a private individual in the second millennium BC.  We have no external evidence for any non-royal person from that period in Canaan.  The correct statement is that there is no evidence either way, which is less satisfying than either side wants.

"The conservative arguments are all unfalsifiable."  Some are — editorial updating and the camel-sampling bias both explain away absences and cannot be tested.  But the horse argument is not: it makes a prediction about what a late writer would have written, and the text either has horses or it does not.  It does not.

"So the whole thing is a wash."  Not quite.  The best current summary is that the patriarchal narratives carry second-millennium features that a first-millennium Israelite writer would have been unlikely to invent, and at least one first-millennium feature that an eyewitness could not have written.  Both facts are real.  A tradition transmitted and updated over a long period explains both;  so does a well-read late author;  and nothing currently distinguishes those two.

What would settle it

• A second-millennium archive from the Negev or the southern hill country. The patriarchal narratives are set in a region that has produced almost no second-millennium documents. This is the largest single gap.
• A securely identified Sodom. The one episode in the patriarchal cycle with a potential archaeological signature. Currently unresolved — see Sodom and Gomorrah.
• A published onomastic study comparing the patriarchal names against dated Amorite and first-millennium Israelite corpora with statistics rather than examples. The claim is widely made and, as far as this project can tell, not rigorously demonstrated.
• A survey of first-millennium multi-power coalitions searched as hard as the Mari archive has been. This would make or break Kitchen's Genesis 14 argument, which is currently the strongest positive case in the file and rests on an absence nobody has demonstrated.
• Better first-millennium Mesopotamian family-law evidence, which would test Selman's surviving argument directly.
• A dated second-millennium dromedary in the southern Levant. One securely stratified, radiocarbon-dated domestic camel from a Middle Bronze context would remove the strongest anachronism charge.

Sources and further reading

The literature reviewed for this article, including sources considered and not used, is in literature-review.md.