Biblical Minimalism and How to Test an Ancient Text

Summary

Every other article in this section is an argument about a mound, an inscription, or a date.  This one is about the rules under which those arguments are conducted, because most of the disagreement is really about the rules.

• The debate has a real history, and the founding book is better than its reputation. In 1992 Philip Davies published In Search of 'Ancient Israel'. Its organising move is a distinction: "I shall be dealing with three Israels: one is literary (the biblical), one is historical (the inhabitants of the northern Palestinian highlands during part of the Iron Age) and the third, 'ancient Israel', is what scholars have constructed out of an amalgamation of the two others."¹ Niels Peter Lemche and Thomas Thompson at Copenhagen, and Keith Whitelam, pressed the same line. The label "minimalism" is theirs by adoption rather than by choice.
• Davies's methodological argument is one this project accepts, and it is not the argument he is usually attacked for. On the limits of setting as evidence: "It is well known that authentic geographical settings and genuine chronological settings do not of themselves guarantee the historicity of anything described as happening there (e.g. A Tale of Two Cities). The existence of common names, places and events between an historical construct and a literary one does not necessarily make the literary construct historical."¹ That is correct, and it is the diagnostic-versus-consistent test stated for texts rather than for minerals. He also concedes the reverse in the same passage — that non-appearance of known historical entities should not automatically tell against historicity — and calls both arguments "silly."¹
• His substantive claim, by contrast, was checkable and has failed. He wrote that archaeology "has, of course, discovered new population centres in the early Iron Age, but it has no archaeological reason to call these 'Israel'," and that the northern kingdom, existing "between roughly the ninth and precisely the late eighth centuries BCE," is "the only 'Israel' which an archaeologist or historian … can encounter."¹ A year after he wrote it, the Tel Dan stele produced a ninth-century foreign king naming the House of David. That is a name, not a sherd, and it is exactly the kind of criterion Davies said was missing.
• The strong form of the position has been falsified point by point, and mostly by things found in the ground since 1993. A ninth-century Aramaic stele names "the House of David." A 2024 study with 103 radiocarbon dates finds occupation across every excavated area of Jerusalem in the twelfth to tenth centuries. A Babylonian chronicle dates the capture of Jerusalem to 16 March 597 BC. Ration tablets name the deported king of Judah and his five sons. These are treated in the sibling articles; the point here is that they are the kind of evidence that decides things, and it went one way.
• But the maximalist side has lost more arguments in the same period, and lost them worse. Three flagship claims have been withdrawn since 1976: the Ebla tablets naming Sodom and Gomorrah, the Nuzi tablets proving a second-millennium Abraham, and the Tunguska-scale airburst that destroyed Tall el-Hammam. A fourth, the pig bone as an Israelite ethnic marker, was retracted by the archaeologist who made it famous. Two of these were withdrawn by their own proponents, which is to the credit of the people involved and is a reason to be slow about the next flagship claim.
• The real methodological disease is asymmetry, and both sides have it. Kenneth Kitchen argues that the archaeological silence at Dibon and Heshbon "stands supplemented and corrected by the texts," and that the same move is illegitimate when Kathleen Kenyon makes it at Jericho.² Titus Kennedy discounts radiocarbon at Jericho as conflicting with ceramics, while the same camp relies on radiocarbon to rule out revised Egyptian chronologies.³ On the other side, Finkelstein's Low Chronology is treated by its users as an established result when it is a contested hypothesis, and the Copenhagen school's forgery argument against the Tel Dan stele was pressed on evidence nobody would accept against a non-biblical inscription.
• The two labels have definitions, and stated plainly both are bad rules. Lester Grabbe reports the standard ones: "'minimalist': one who accepts the biblical text only when it can be confirmed from other sources; 'maximalist': one who accepts the biblical text unless it can be proved wrong."⁴ The first would discard most of ancient history, since most ancient testimony is uncorroborated. The second makes the text unfalsifiable in practice. Grabbe adds that the maximalist category is nearly empty — the only history he knows that fits is Provan, Long and Longman's.⁴ Most people called maximalists are not applying that rule at all.
• The strongest maximalist argument is philosophical and this article partly accepts it. Provan, Long and Longman charge that critical scholarship cannot say where its own suspicion stops: "having adopted this principled stance of suspicion towards the tradition, none can agree with the other as to where suspicion should then be suspended and faith in the tradition reinvested."⁵ That is the reversal test aimed at the other side, and it lands. Their positive claim — that all knowledge of the past is testimony, archaeology included — is also true.⁵ What does not follow is that all testimony is evidentially alike. Testimonies differ in re-checkability, distance from the event, and interest, and those are properties of the evidence rather than prejudices about the author.
• The "same standard as other ancient history" argument is half right. It is true that no classicist would deny Alexander's existence on evidence comparable to what is available for David, and true that a uniquely sceptical standard applied to one corpus is not a standard but a prejudice. It is also true that the disputed claims are usually not about existence but about scale — the size of Solomon's kingdom, the extent of a conquest, the population of a city — and scale is exactly what archaeology tests well. The argument works against the strong minimalist position and does not rescue the maximalist one.
• Seven tests are proposed below. They are not novel; they are the ones that actually did the work in the cases where an argument was settled. Briefly: does the claim make a prediction that could fail?; is the evidence diagnostic or merely consistent?; is the absence you are reasoning from an absence of the thing, or of its usual trace?; what does the text itself predict should be findable?; is the source provenanced?; did the correspondence arise by accident?; and would you accept this move from the other side?
• The third one was added after this project got it wrong. Erez Ben-Yosef's radiocarbon work on the Aravah copper industry found a tenth-century polity that ran international-scale mining, controlled territory and imported feed across a hundred miles while leaving almost no stone architecture. His verdict is aimed at everyone: the "architectural bias" is a flaw that "cuts across all schools of biblical archaeology," because "both 'minimalists' and 'maximalists' have been using stone-built architectural remains as the key."⁶ It dissolves a whole class of negative argument — and, used carelessly, becomes unfalsifiability with a citation.
• The single most useful of the seven is the last. Most of the bad arguments in this field are moves that their own users would reject if the other side made them. Applying the reversal test to your own position removes more error than any amount of new evidence.
• A note on tone, from the person best placed to give it. Grabbe founded the European Seminar on Methodology in Israel's History in 1995 to get past this argument, and his description of what it had become is worth keeping in view of any scorecard: the labels "began to be pejorative," "nihilist" came into use, and "Some said the 'minimalists' were dangerous; others, that they could be safely ignored. The curious anomaly of dangerous people who can be safely ignored serves as a symbol of the unhelpful way in which the debate has moved."⁴ He notes the ugliness ran both ways.⁴
• Honest weaknesses of this article. It is written from a position — this project holds that the Bible is reliable — and a method article written by an interested party should be read with that in mind. The specific mitigation offered is that every case used against the minimalists here is documented in the sibling articles with primary sources, and so is every case used against the maximalists, and the second list is longer.

Verdict: the strong minimalist programme of the early 1990s has been falsified, and this does not license the maximalist programme it replaced.  The correct posture is neither presumption of reliability nor presumption of invention, but the ordinary one: ask what each specific claim predicts, look, and report what was found — including when it is nothing.

Part 1 — What minimalism actually claimed

The term is used loosely, usually as an insult.  It is worth stating the position accurately before assessing it.

Davies's 1992 book made a distinction that is genuinely useful regardless of one's conclusions.  In his own words:

This is a book about history, though it is not another 'History of Israel': that genre is probably obsolete.  Previous scholarship has left us with an 'ancient Israel', which I shall designate with quotation marks in order to distinguish it from the biblical Israel and the historical Israel.  For I shall be dealing with three Israels: one is literary (the biblical), one is historical (the inhabitants of the northern Palestinian highlands during part of the Iron Age) and the third, 'ancient Israel', is what scholars have constructed out of an amalgamation of the two others.¹

His charge against the third is specific.  "The scholarly 'ancient Israel' … lies between literature and history—or rather, it straddles the two."  And more sharply: "this scholarly projection, this literary-historical hybrid, is also posited as the producer of the biblical literature—in other words as the creator of the biblical Israel."¹  That is, the reconstruction is used to explain the text that generated the reconstruction.

This warning is worth accepting.  A great deal of twentieth-century "history of Israel" writing did narrate the biblical story with archaeological illustrations attached and call the result history.  The warning has largely been absorbed;  historians of Israel now write much more carefully about which evidence is which.

The substantive claims are a different matter, and they were checkable.  Davies argued that archaeology has no independent warrant for the label:

Archaeologically, we cannot establish what the distinctive cultural characteristics of 'ancient Israel' are.  Indeed, it is a moot point whether, in the absence of the biblical literature, archaeology would have discovered any 'ancient Israel' at all.  It has, of course, discovered new population centres in the early Iron Age, but it has no archaeological reason to call these 'Israel'.¹

He is largely right about the sherds, and the failure of every proposed ethnic marker — pig bones, four-room houses, collared-rim pithoi — has vindicated him on that specific point.  See The Emergence of Early Israel.

He is wrong about the names.  His book restricts the only historian-accessible Israel to the northern kingdom "between roughly the ninth and precisely the late eighth centuries BCE."¹  Merneptah had named Israel in Canaan around 1208 BC since 1896, and within two years of publication the Tel Dan stele produced a ninth-century Aramaean king naming the House of David.  An inscription is not a sherd:  it supplies the criterion Davies said archaeology lacked.

Two claims, very different in strength.  It is possible to accept the methodological warning and reject the substantive conclusion, and most working scholars now do.

Part 1b — The definitions, and why the labels are a hindrance

Lester Grabbe founded the European Seminar on Methodology in Israel's History in 1995 precisely to get past this argument, and his survey Ancient Israel: What Do We Know and How Do We Know It? is the best statement of what the two labels actually mean.  The definitions he reports, which originate with Ernst Axel Knauf:

'minimalist': one who accepts the biblical text only when it can be confirmed from other sources;  'maximalist': one who accepts the biblical text unless it can be proved wrong.⁴

Stated that way, both are recognisably bad rules, and it is worth seeing why before taking sides.  The minimalist rule would discard most of ancient history, since most ancient testimony is uncorroborated.  The maximalist rule makes the text unfalsifiable in practice, since "proved wrong" is a very high bar for events that leave no trace.  Neither is how a historian handles Livy or Herodotus.

And Grabbe's next observation is the one conservatives should notice.

There are very few true maximalists, that is, those who accept the testimony of the Bible unless it be falsified.  Indeed, the only history known to me so far written that one could call 'maximalist' is that of Provan, Long and Longman (2003).⁴

The category is nearly empty.  Most people called maximalists — Kitchen, Dever, Mazar, Hoffmeier — are not applying that rule;  they are arguing for particular reconstructions on particular evidence, which is what everyone else is doing.

And the label is routinely applied to someone who rejects it.  Israel Finkelstein is the scholar most often called a minimalist in popular writing.  In the volume where he debates Mazar, the editors record that he "disagrees with both the maximalists and the minimalists as to the facts of history," and Finkelstein himself writes that the Tel Dan inscription "leaves no doubt that David and Solomon were historical figures" and that the Samuel bandit narratives "contain genuine, early historical memories."⁷⁷  A term that captures neither the Copenhagen school's position nor Finkelstein's is not doing much work.

His verdict on the conduct of the argument is worth having in full, because this article's Part 2 keeps a scorecard and a scorecard invites the wrong spirit:

The terms 'maximalist' and 'minimalist' began to be pejorative labels, and the term 'nihilist' came to be used of certain positions … Some said the 'minimalists' were dangerous;  others, that they could be safely ignored.  The curious anomaly of dangerous people who can be safely ignored serves as a symbol of the unhelpful way in which the debate has moved.⁴

He adds that the ugliness ran both ways — "there has been intimation that those who defended the status quo were nothing but biblical fundamentalists, which has not usually been the case" — and that when "some begin to use the term 'anti-Semitic' of particular academic positions, one begins to despair of any chance of a proper scholarly exchange."⁴

Part 2 — The score since 1976

This is the part of the argument that is usually conducted by assertion.  Here it is conducted by list.

Claims the minimalists lost

David and his dynasty.  The Tel Dan stele, found in 1993 and 1994, has a ninth-century Aramaean king naming the "House of David" alongside the king of Israel.  The objection about the missing word divider is real and is not enough.  See David, Solomon, and the United Monarchy.

Tenth-century Jerusalem as an empty village.  The 2024 PNAS study of 103 radiocarbon dates from five excavation areas reports "concrete evidence for a widespread human presence in the city during the 12th to 10th centuries BC, as short-lived material from all the areas studied was dated to these centuries."⁸  Note that the same study also dates the city's monumental expansion to the ninth century, which is a loss for the other side.

Israel as a late construct.  Merneptah names Israel in Canaan about 1208 BC.  This has been known since 1896 and has never been seriously disputed.  See The Emergence of Early Israel.

The exile as a literary theme.  A Babylonian chronicle dates the capture of Jerusalem to the day;  ration tablets name Jehoiachin and his five sons in Babylon;  two of the four officials named in Jeremiah 38:1 have left seals in the City of David.  See Exile, Babylon, and Persia.

Claims the maximalists lost

The Ebla tablets name Sodom and Gomorrah.  Announced by Giovanni Pettinato at the Society of Biblical Literature on 29 October 1976, from tablets whose facsimiles had not been published for anyone else to check.  Alfonso Archi contested the readings;  Pettinato withdrew Zoar.  The standard survey verdict: "The present consensus is that Ebla has no bearing on the Minor Prophets, the historical accuracy of the biblical Patriarchs, Yahweh worship, or Sodom and Gomorrah."⁹  Still circulating in popular apologetics fifty years later.

The Nuzi tablets prove a second-millennium Abraham.  Withdrawn by conservatives as well as critics.  M. J. Selman, in the 1974 Tyndale Biblical Archaeology Lecture: "no special relationship exists between the Nuzi tablets and the patriarchal narratives."¹⁰  See The Patriarchs in Their Ancient Setting.

A Tunguska-scale airburst destroyed Sodom.  Scientific Reports retracted the 2021 paper on 24 April 2025: "the claims that an airburst event destroyed the Middle Bronze Age city of Tall el-Hammam appear to not be sufficiently supported by the data in the Article."¹¹  See Sodom and Gomorrah.

Pig bones mark Israelite ethnicity.  Retracted by the team that made the argument famous.  Sapir-Hen, Bar-Oz, Gadot and Finkelstein, 2013: "All this makes it difficult to use the presence/absence of pigs as an indicator of ethnic identity."¹²  See The Emergence of Early Israel.

Jericho fell about 1400 BC.  Bryant Wood's re-dating was rejected by name in the site's own radiocarbon publication, and the early-date advocate who recalibrated the samples with IntCal20 got seventeenth- and sixteenth-century ranges.¹³  See Jericho, Ai, and the Conquest Narratives.

How the losing side describes it, and how the winning side does

Two quotations set the terms.

Yosef Garfinkel, co-director at Khirbet Qeiyafa, in a 2011 Biblical Archaeology Review article titled "The Birth and Death of Biblical Minimalism":

The argument that Judah was an agrarian society until the end of tenth century B.C.E. and that David and Solomon could not have ruled over a centralized, institutionalized kingdom before then has now been blown to smithereens by our excavations at Khirbet Qeiyafa.¹⁴

His reasoning is that the site's fortifications required organised labour beyond a chiefdom's reach, and that a trained scribe stationed at a remote border town implies a state able to train and post scribes.  Finkelstein and Silberman had themselves written that "the power of the chief was limited … The economic and military capacity of a chiefdom was severely limited."¹⁵

But notice what the reply was.  Finkelstein and Fantalkin did not answer that Qeiyafa was a chiefdom's work.  They answered that it belonged to "an early north Israelite entity" rather than to Judah.¹⁶  The argument moved from "chiefdom or state" to "which state," and that is the honest description of what happened to minimalism.  The strong position lost.  What replaced it is not the maximalist position;  it is a contest over whose tenth-century state it was.

And Davies himself moved.  By 2008 he was writing that the Bible is indispensable to the historian even though its "stories may be false, true, or a mixture of fact and fiction."¹⁷  That is a long way from 1992, and it is closer to the ordinary handling of any ancient source than either camp's slogans suggest.

What the score means

Not that the maximalists are wrong about everything.  It means that the maximalist camp has produced more announced discoveries than the minimalist camp, and announced discoveries fail more often than negative claims do.  A negative claim — "there is no evidence for X" — fails when someone finds X.  A positive claim — "this ruin is Sodom" — fails whenever any of its several supporting arguments fails.  Positive claims are more informative and more fragile, and a camp that makes them will have a worse retraction record even if it is right more often overall.

That is a structural point, not an excuse.  The lesson is that the announcement practice is the problem: the Ebla claim was made from unpublished tablets, the airburst claim was made with withheld data, and both reached the popular press before they reached scrutiny.

Part 2b — The maximalist methodology, stated properly

Grabbe names A Biblical History of Israel by Iain Provan, V. Philips Long and Tremper Longman as the only history he knows that deserves the label "maximalist."  It is therefore the book this article owes a hearing, and its argument is philosophical rather than archaeological.

Their charge against the field

Their central complaint is not that critical scholars are hostile.  It is that they are inconsistent, and cannot say where their own principle stops:

In one moment, biblical testimony about Israel's past is embraced as reflective in reality of that past.  In the next, such testimony is rejected for the most unconvincing of reasons, which in some cases comes down to little more than prejudice.  In one moment, extrabiblical evidence is apparently to be regarded as providing "knowledge" about the ancient past that is the solid rock upon which biblical claims founder.  In the next, such evidence is marginalized and relativized, and the biblical version of events retained regardless of what other sources of evidence have to tell us … having adopted this principled stance of suspicion towards the tradition, none can agree with the other as to where suspicion should then be suspended and faith in the tradition reinvested.⁵

This is Test 7 turned outward, and it lands.  It is the same criticism this article makes of Kitchen at Dibon and Kennedy at Jericho, aimed the other way, and a reader who accepts it in one direction has to accept it in the other.  The observation that nobody can state where suspicion is suspended is a genuine defect and it is not answered by pointing out that the authors have a confessional interest.

Their positive argument

Their epistemology rests on testimony:

We know about the past, to the extent that we know about it at all, primarily through the testimony of others.  Testimony lies at the very heart of our access to the past … Even if I am the person who digs up an artifact from the Palestinian soil, I am still entirely dependent upon the testimony of others who have gone before me when I try to make sense of its significance.⁵

The conclusion they draw is that treating biblical texts as guilty until corroborated is not a neutral default but an arbitrary one, since no other testimony is held to it.

Where this article agrees and where it does not

The premise is true.  All knowledge of the past is mediated, archaeological knowledge included.  An excavation report is testimony;  a radiocarbon date is testimony about what a machine did in a laboratory;  a site identification is an argument someone made.  Anyone who talks as though archaeology delivers unmediated fact and texts deliver only claims is confused.

The conclusion does not follow, and the reason is worth stating precisely.  That all evidence is testimony does not make all testimony evidentially alike, because testimonies differ in properties that matter:

• Re-checkability. A stratigraphic claim can be tested by re-excavating or by re-dating the samples. Regev's team could re-run the Jerusalem sequence; Kennedy could recalibrate Kenyon's samples; Manning could re-test Pearson's offset. A narrative composed centuries after the events it describes cannot be re-checked in that way — only compared against other testimony.
• Distance from the event. Contemporaneity is not everything, but a Babylonian scribe dating a siege in the month it happened stands differently from an account compiled generations later, and the difference is a fact about the testimony, not a prejudice about its author.
• Interest. Shalmaneser's inscriptions exaggerate his victories, and everyone discounts them accordingly. Applying the same discount to any interested source is consistency, not hostility.

So the honest position is neither "the text is guilty until corroborated" nor "the text is innocent until falsified."  It is that a text is testimony of a particular kind, whose weight depends on how close it stands to the event, whether it can be checked, and what it wanted.  Provan, Long and Longman are right that the field applies its scepticism unevenly.  They do not thereby establish that the even application should be at their setting.

Part 3 — Seven tests

Test 1: Does the claim make a prediction that could fail?

A good example.  Alan Millard's argument that the patriarchal narratives are old because they give the patriarchs no horses.¹⁸  This makes a checkable prediction about the text: an Iron Age writer inflating a rich man's holdings would use horses, the prestige animal of his own period.  The text either has horses or it does not.  It does not.

A bad example.  "The archaeology is incomplete, so absence of evidence at Dibon and Heshbon does not count."  This is sometimes true and it can never fail, because any absence can be attributed to incomplete excavation.  Kitchen's version — the archaeology "stands supplemented and corrected by the texts" — is a legitimate historiographical move only if one is willing to let the other side make it too.²

How to apply it.  Ask what observation would make the claimant give up the claim.  If there is none, the claim is not doing evidential work, whatever else it is doing.

Test 2: Is the evidence diagnostic, or merely consistent?

This is the test that killed the airburst paper, and Jaret and Harris state it better than any methodologist:

The presence of only non-diagnostic features—even if these same features sometimes occur in shock materials—is not sufficient to claim impact.¹⁹

The logic generalises completely.  Melted pottery is consistent with an airburst;  it is also consistent with a kiln, a house fire, and an army.  Only shocked quartz with genuine planar deformation features, or a full platinum-group-element signature matching a meteorite class, is diagnostic.  Their deepest point is that a criterion can be diagnostic in one setting and worthless in another: ultra-high-temperature melt identifies impact glass in a desert and identifies nothing at all on an archaeological site, "where our ancestors had access to ceramic and smelting technologies."¹⁹

The same test destroys the ethnic-marker arguments.  Pig avoidance is consistent with Israelite identity and is also the general lowland pattern.  The four-room house is consistent with Israelite identity and is also Moabite and Ammonite.  A collared-rim pithos is consistent with Israelite identity and is also a sensible way to store grain.

And it is the strongest thing that can be said for the inscriptions.  A ninth-century Aramaean king writing "House of David" is diagnostic in a way no pot is, because a name is not a technology.

Test 2b: Is the absence you are reasoning from an absence of the thing, or of its usual trace?

A corollary of Test 2, and the one this project got wrong before reading Erez Ben-Yosef.

Most negative arguments in this field run: no monumental architecture in tenth-century Jerusalem, therefore no tenth-century state;  no Late Bronze wall at Jericho, therefore no city to conquer;  no camel bones in the second-millennium Levant, therefore no camels.  Each step assumes that the thing, if present, would have left the trace one is looking for.

Ben-Yosef's Aravah work is the clean counter-example, and it is not an apologetic one.  The copper industry at Timna and Faynan was radiocarbon-dated to the tenth century BC, ran at industrial scale, controlled territory and imported animal feed from more than a hundred miles away — and left almost no stone architecture.  His conclusion is aimed at both camps:

the total reliance on stone-built archaeological features in the identification of social complexity in the vast majority of recent studies has resulted in skewed historical reconstructions.  Recognizing this "architectural bias" … [has] important implications on core issues in biblical archaeology today, as both "minimalists" and "maximalists" have been using stone-built architectural remains as the key to solving debated issues.⁶

How to apply it.  Before treating an absence as evidence, ask what the presence would have deposited, whether that deposit survives in this environment, and whether anyone has looked where it would be.  Then apply the answer to both sides.  Alan Millard's version of this for camels — "the camel is not an urban creature;  it is kept outside settlements," so "the likelihood of camel bones being found is small" — is the same move, and a conservative who accepts it there owes it to Finkelstein at Jerusalem.

The limit, which matters.  This test dissolves negative arguments;  it does not create positive ones.  "The evidence could be invisible" is compatible with the thing existing and with it not existing, and a claim that can only be defended this way is not being supported, merely protected.  Used honestly it narrows what can be concluded.  Used dishonestly it is unfalsifiability with a citation.

Test 3: What does the text itself predict should be findable?

This is the test almost nobody applies, and it is the one that changes the Joshua debate most.

Joshua 11:13 says "Israel did not burn any of the cities located on mounds except for Hazor;  it was the only one Joshua burned."  Joshua 24:13 has Israel occupying "cities you did not build."  So a conquest as the text describes it should leave three burned cities, not a horizon of destruction layers.  An archaeologist looking for a swath of destruction is testing a claim the text does not make.

The test cuts both ways, and that is what makes it a test rather than a device.  Applied honestly, it also says: the three cities the text does name should show it, at a consistent date — and Jericho and Ai do not.

How to apply it.  Before looking for evidence, write down what the text actually asserts and what physical trace that assertion should leave.  Do this before consulting the archaeology, not after.

Test 4: Is the source provenanced?

An object from a controlled excavation carries its own context: stratum, associated finds, date.  An object from the antiquities market carries nothing but itself and a story.

This is not pedantry.  The Tel Dan stele's forgery accusation failed largely because the fragments came out of a controlled dig.²⁰  The Āl-Yahūdu tablets — about 200 cuneiform documents recording a Judean community in Babylonia, and superb corroboration of the exile — are unprovenanced, entered the market by unknown means "possibly looting," and sit mostly in private collections.²¹  They are probably genuine.  They are still weaker evidence than one excavated tablet would be, and anyone citing them owes the reader that sentence.

The bullae of Jeremiah's officials are usable precisely because they came from an excavation;  the antiquities market is full of Hebrew bullae and a great many are forged.

How to apply it.  For any artefact-based claim, ask: who dug it, out of what layer, and was it published before or after it entered a collection?

Test 5: Did the correspondence arise by accident?

The strongest correspondences in this whole field are the ones nobody was trying to create.

Lachish Letter 4, an ostracon from a Judean officer, closes: "we are watching for the fire signals of Lachish … because we cannot see Azeqah."²²  Jeremiah 34:7 says the Babylonians were attacking "Jerusalem and the cities of Lachish and Azekah … because they were the only fortified cities of Judah that were still holding out."²³  The officer is reporting a signalling problem.  The prophet is listing cities still holding out.  Neither is corroborating the other, and they intersect anyway.

Compare the deliberate correspondence: a paper that says the Genesis account "is consistent with having been an eyewitness account of a cosmic airburst" and then lists six matching features.²⁴  Six features were chosen because they matched.  That is not corroboration;  it is fit, and fit can be manufactured.

How to apply it.  Ask whether the two sources could have been coordinated, and whether the matching features were selected after both were known.  Accidental intersection at an incidental detail is worth more than deliberate alignment at a central one.

Test 6: Would you accept this move from the other side?

The reversal test, and the most productive of the six.

• Kitchen supplements the archaeology with texts at Dibon and rejects Kenyon doing the same at Jericho.²
• Kennedy demotes radiocarbon at Jericho as conflicting with ceramics, while his camp relies on radiocarbon to exclude revised Egyptian chronologies.³
• Conservatives accept "Rameses" in Exodus 1:11 as later editorial updating and resist the same explanation for "Philistines" in Genesis 26 — or the reverse, depending on which date they hold.
• Minimalists accepted a forgery argument against the Tel Dan stele that rested on chisel marks, for an object recovered in a controlled excavation.
• Both sides treat a contested chronology — Finkelstein's Low Chronology, or Wood's Jericho re-dating — as established when its consequences are convenient.

How to apply it.  Take the inferential move, not the conclusion, and imagine it deployed against you.  If you would object, you have found a weakness in your own argument rather than a strength.

Part 4 — Where the "same standard" argument works and where it does not

The most common conservative rejoinder to minimalism is that no other field of ancient history would apply such scepticism.  Craig Evans puts it in the strong form: if other historians did this "we would have nothing to say about Alexander the Great … That kind of historical minimalism only occurs in biblical studies."

Where it works.  Against the strong claim that a text is worthless as evidence until independently corroborated.  No ancient historian operates that way, because almost nothing in ancient history is independently corroborated.  A standard applied to one corpus and no other is a prejudice with a methodology attached.

Where it does not work, and this needs saying.  Three limits.

First, the disputed claims are usually about scale, not existence.  Nobody now argues that David did not exist.  The argument is whether he ruled a highland chiefdom or an empire to the Euphrates.  Alexander's existence is not comparable to that;  the extent of Alexander's empire is, and it is attested by coins, city foundations, administrative documents and the records of the states he conquered.  The right comparison makes the conservative case harder, not easier.

Second, the Hebrew Bible is a compiled and edited corpus in a way Arrian is not.  Asking when a tradition was written down is not scepticism;  it is the same question a classicist asks about Livy's sources for early Rome, and classicists are extremely sceptical about early Rome.

Third, the argument proves too much if pressed.  Ancient historians do routinely discount royal inscriptions' numbers, dismiss claims of universal dominion as convention, and treat foundation legends as legend.  Applying the same standard to 1 Kings 4:21's empire from the Euphrates to Egypt yields the conclusion that it is conventional royal hyperbole — which is this project's own conclusion, reached by exactly that route.

The honest form of the argument is therefore narrower and still strong: apply the ordinary standards of ancient history, in both directions, and the strong minimalist position fails while a good deal of the maximalist position fails with it.

Common objections

"Minimalism is just anti-religious bias."  Sometimes, and the motives do not matter.  Davies's distinction between biblical, historical and reconstructed Israel is a good one whoever made it, and the Tel Dan stele refutes his conclusion whatever he believed.  Arguments about motive are a way of avoiding arguments about evidence, and both sides use them.

"Davies said the Bible is fiction, so his method doesn't deserve a hearing."  He said something more careful than that, and the careful version is right:  authentic settings do not guarantee historicity, and the same holds in reverse.¹  A conservative reader should notice that this is the same principle by which this project rejects the Tall el-Hammam airburst's six-point Genesis match.

"Maximalism is just apologetics."  Also sometimes.  Note that the two cleanest retractions in this article's list — Selman on Nuzi and the Sapir-Hen team on pigs — were made by scholars withdrawing their own side's arguments.  That is what distinguishes a discipline from a campaign, and it happens on both sides.

"You can't test the Bible archaeologically at all."  You can test specific claims about specific places at specific dates, which is most of what this section does.  You cannot test theological claims, and nobody should pretend otherwise.

"These tests are just common sense."  They are.  The point is that they are not applied.  Every one of the seven is drawn from a case where a published, peer-reviewed argument failed for want of it — including one this project itself had been getting wrong until it read Ben-Yosef.

"This article is written by someone with a position."  Yes, and it says so.  The mitigation on offer is that the list of maximalist failures above is longer than the list of minimalist ones, that every item on both lists is documented with primary sources in the sibling articles, and that the reader can check.

What would improve the method

• Pre-registration of identification claims. A site identification announced before excavation — as at Tall el-Hammam and at the Large Stone Structure — should state in advance what finding would falsify it. Neither did.
• Data availability. The airburst paper's authors "withheld most of the data that they claimed provided the basis for many of their graphs, maps, and age interpretations."²⁵ Journals in this field should require deposited data as a condition of publication, as most sciences now do.
• A convention for citing unprovenanced material. Something like the convention for retracted papers: cite it, use it, and mark it, every time.
• Radiocarbon at the disputed sites, done the way Jerusalem was done in 2024 — dense stratified sampling with microarchaeological context verification and historical anchors. That method beat the Hallstatt Plateau down from 300-year ranges to decadal precision. Applied at Megiddo, Hazor and Gezer it would settle the Solomonic-gates question; applied at Jericho it would settle the destruction date.

Sources and further reading

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