When Did the Exodus Happen?

Summary

Every argument about whether the Exodus happened is really an argument about when.  Pick 1446 BC and the pharaoh is an 18th-Dynasty king and Jericho's walls should fall around 1400.  Pick 1260 BC and the pharaoh is Ramesses II and Jericho was already a ruin.  A skeptic who says "there is no evidence" and an apologist who says "the evidence is overwhelming" are often digging in different centuries.  This article works through the numbers, in public, and reports where they land.

• The sharpest disagreement is not between believers and skeptics. It is between Kenneth Kitchen and James Hoffmeier, who date the Exodus to about 1260 BC, and Bryant Wood and Douglas Petrovich, who date it to 1446 — four scholars who all believe it happened. Behind them stands a mainstream Egyptology that mostly says it cannot be dated at all.
• The one verse the early date rests on says something different in the Septuagint, and the defenders of the Hebrew reading admit it. 1 Kings 6:1 gives 480 years in the Hebrew and 440 in the Greek — Exodus 1446 against Exodus 1406.¹ Twenty-four Greek manuscripts read 440 against seven for 480, and the four oldest uncials are all in the 440 column.² Henry Smith, who argues for 480, concedes in print that "440th is most likely the original OG reading" and that choosing 480 "would by no means be a text critical slam dunk."² Almost nobody quotes that.
• But 440 makes the Bible's own arithmetic worse, and nobody has ever answered the objection. The explicit figures overshoot 480 by about a hundred years; they overshoot 440 by about a hundred and forty, and on the Greek reading the judges are still ruling fourteen years after Saul is king.² A tempting extra argument — that the Greek lengthens Judges, since Codex Vaticanus gives Jephthah sixty years against the Hebrew's six — turns out not to work, and the article shows why: the witnesses that carry 440 in Kings mostly carry the ordinary Judges figures.³ The only named advocate of 440, Steven Collins, adopts it because his Tall el-Hammam chronology requires it, not on any textual ground.⁴
• The Bible's own year-figures do not add up to 480 — they add up to about 580. Summed from the text, the explicit numbers in Judges come to 410, and adding the wilderness, Eli, Saul, David, and Solomon's first three years reaches 573 before a single year is assigned to the conquest, to Joshua's later life, or to Samuel. So a flat serial reading is impossible on anybody's chronology, and Judges 11:26 proves it from inside the book: Jephthah says 300 years where the preceding figures already total 319. Overlapping judgeships are not a rescue device invented by conservatives — a critical scholar, Antti Laato, derives Jephthah's 300 exactly by the same method.⁵
• The late date's best argument is a city that did not exist yet. Exodus 1:11 has the Israelites building "Rameses." Pi-Ramesse has been located at Qantir and its occupation runs from about 1279 to roughly 1130 BC; Tanis never bore the name.⁶ Against this, Wood observes that on the late-date reckoning the Israelites built "Rameses" before Ramesses II was born.⁷ Hoffmeier replies that construction at the site is now documented under Horemheb and Seti I.⁸
• Hoffmeier states the real crux better than either side's slogans. "Both texts cannot be treated literally. If the former text is taken literally, then the reference to Ramesses has to be dismissed as an editorial updating at a later date. Should the reference to Ramesses be treated at face value, then the 480 year figure has to be interpreted symbolically."⁸ That is the honest shape of the problem, and it comes from a conservative Egyptologist.
• Moving Egypt's dates is not an available escape. Velikovsky, Courville, James, and Rohl all propose shifting Egyptian chronology by centuries. The Assyrian eponym canon, anchored by the solar eclipse of 15 June 763 BC, blocks it; so do the Amarna synchronisms, Central European oak dendrochronology, and radiocarbon.⁹ The revisionists conceded the Assyrian framework themselves, calling it "as watertight as one could ask for any period of ancient history."¹⁰
• Egyptian chronology is softer than apologists think and much harder than revisionists need. Kitchen's own error bars run about 30 years at 2000 BC, 20 to 10 by 1500, and 11 years in the thirteenth century.¹¹ Sothic dating really does depend on an unknown observation latitude, and the assumed site has been moved south twice.¹² But 211 radiocarbon dates in 2010 broadly confirmed the conventional framework, and 48 more in 2025 pushed it earlier, not later.¹³
• The strongest concessions come from the defenders. Kitchen, in print: "No Egyptian records mention specifically Israelites working in the East Delta (or anywhere else), or a Moses who spoke for such a group, or an exodus by a group of this name (Israel). Nowhere in Sinai has a body of Late Bronze Age people passing through left explicit traces … That applies also to Qadesh-Barnea."⁶ He is right that pharaohs did not monumentalize defeats; he is also right about the silence.
• The best answer to the 580-year problem is one almost nobody uses, and it comes at a price its own publisher won't pay. David Austin reads the 480 as counting only years of sovereign theocracy, excluding the years of foreign servitude and usurpation. Sum those from the text — the six oppressions plus Abimelech — and it is exactly 114, which leaves 21 years of the 480 for the conquest, Joshua, the elders and Samuel.¹⁴ The overshoot dissolves without any appeal to overlapping judgeships. But it puts the Exodus at 1532, abandoning the Assyrian-anchored date for Solomon that everything else here depends on.
• Much of the popular chronology literature has not done the addition. Creation Ministries International holds 1446 and has published two incompatible solutions to the overshoot without ever adjudicating between them. Robert Carter, author of its fullest chronology paper, has no solution because he never states the problem — in that paper the whole Exodus-to-Solomon span is one link, the 480, and the Judges figures are never summed.¹⁵
• Honest weaknesses of the case made here: Jericho, the early date's flagship site, currently supports neither conservative date — the excavation now running there dates the destruction to 1550–1525 BC, and Titus Kennedy's own recalibration of the Kenyon samples returns seventeenth- and sixteenth-century dates;¹⁶ the Kadesh-Barnea excavation results are genuinely empty of Late Bronze material on either date, and no apologetic answer to that is very good; Wood's re-dating of two key sites rests on an unpublished dissertation; Hoffmeier's "limited window" argument rests on an unpublished remark by Kitchen; and the early-date camp is split three ways on the date, the pharaoh, whether he drowned, and whether the Amarna letters may be used at all.

Verdict:  the Exodus cannot presently be dated, and both conservative options survive only by reading one of the two anchor texts non-literally.  The early date has the better internal case — 1 Kings 6:1 read plainly, Jephthah's 300 years, and a Levite genealogy that needs about 23 years per generation on the early date and an impossible 12 on the late one.  The late date has the better external case — a Delta capital that only existed in the thirteenth century, a toponym cluster that first appears in 19th-Dynasty sources, and Merneptah's Israel in Canaan by 1208.  Neither has produced the artifact that would settle it.  Anyone who tells you this question is closed, in either direction, has not read the other side.

Why the date is the whole argument

"Did the Exodus happen?" cannot be answered until "when?" is answered, because every test depends on the date.

Pick 1446 BC and you are looking for Israelites in the 18th Dynasty, in the Amarna correspondence, and in a Jericho destroyed about 1400.  Pick 1260 BC and you are looking in the 19th Dynasty, in the reign of Ramesses II, and in Late Bronze destruction layers two centuries later.  The two searches do not overlap.  This is why the popular debate goes nowhere:  the evidence for one date is, by construction, evidence against the other.

Three positions are live.

The early date, about 1446 BC.  Takes 1 Kings 6:1 at face value.  Bryant Wood, Douglas Petrovich, Rodger Young, Eugene Merrill, and most of the conservative American tradition.  John Bimson belongs here too, though at about 1470 rather than 1446.

The late date, about 1260–1250 BC.  Reads the 480 years as schematic and follows the archaeology of the Delta.  Kenneth Kitchen and James Hoffmeier — the two scholars who have done the most sustained work defending the Exodus as a real event.

No datable Exodus.  A later national story with at most a small kernel.  Israel Finkelstein, Donald Redford, Niels Peter Lemche, Philip Davies, and the mainstream of Egyptology.  William Dever sits apart from these, holding a real but small thirteenth-century exodus.

Part 1 — The numbers in the Bible

Bible quotations here are from the NET Bible.  Where the Hebrew, Greek, or Samaritan text matters, it is quoted directly.

The verse everything hangs on

In the four hundred and eightieth year after the Israelites left Egypt, in the fourth year of Solomon's reign over Israel, during the month Ziv (the second month), he began building the Lord's temple. — 1 Kings 6:1

This is the only verse in the Bible that dates the Exodus against a later fixed point.  Solomon's fourth year can be tied to Assyrian records within a year or two.  Add 480 and the Exodus lands in the middle of the fifteenth century.

The whole early-date case runs through this one sentence, which is worth saying plainly, because the sentence has a textual problem.

Is 480 a symbolic number? The genre has now been tested

The commonest way to defuse 1 Kings 6:1 is to say it was never meant as a count of years — that 480 is twelve generations of forty, a schematic figure of the kind ancient scribes liked.  The people who argue this are mostly conservatives, not skeptics: Kenneth Kitchen, James Hoffmeier, and, from the critical side, Victor Hurowitz, Peter Machinist and Gary Rendsburg.¹⁷  Hoffmeier's own framing of the crux, quoted above, depends on it being available.

Titus Kennedy tested the claim against the genre it appeals to.  1 Kings 6:1 belongs to a recognisable ancient Near Eastern type — a temple construction or dedication text that dates the building by counting years elapsed since an earlier significant event.  Kennedy assembles the surviving examples: Assyria, Babylon, Egypt, Phoenicia, Israel and Rome, spanning the thirteenth to the second century BC.¹⁷  His finding is that the symbolic reading has no support in the comparanda:

While it is often assumed or argued that the 480th year, or 479 elapsed years duration, in the Solomon temple construction and dedication text is a symbolic number referring to generations or an ambiguous but much shorter period of time, none of the similar texts of this type from the ancient Near East support that hypothesis.  On the contrary, the parallel examples from Mesopotamia, Egypt, and the Levant refute the idea of purely symbolic years in temple construction and dedication texts, and instead demonstrate that the intent of the authors was to relay an accurate number of years even if errors in calculations were made, while one cited example from Mesopotamia demonstrates that generations were specified when intended in these types of texts.¹⁷

Why this is a stronger form of argument than it looks.  It converts a claim about authorial intention — usually untestable — into a claim about a text type with surviving parallels, which can be checked.  And the last clause is the sharp one: where a Mesopotamian scribe meant generations, he wrote generations.  Kennedy also disposes of the nearest apparent counterexample, the Tanis stele's "year 400" with its fourth day and fourth month, by noting that a commemoration date chosen for its repeating fours does not make the year-count fictional.¹⁷

What it does not do.  It does not date the Exodus.  Kennedy says so himself: the authorial intent appears to be a real 479 years placing the departure near 1446 BC, "even if this year total and reference cannot be conclusively established at this time through archaeological data."¹⁷  Scribes can intend a real number and still be wrong — Kennedy's own wording allows for "errors in calculations."  Nor does it touch the Septuagint variant below, which is a textual question, not a genre one.  What it removes is the easy escape: you can no longer wave 480 away as a schematism without arguing against the parallels.

One caution about weighing it.  This is a single-author study by an early-date advocate, published in a general religious-studies journal rather than an Assyriological one.  It has not, so far as this project can find, been answered by Kitchen's or Hoffmeier's side.  Silence is not refutation, and neither is it endorsement.

The Septuagint says 440

The Greek Old Testament, translated by Jews in Egypt in the third to second centuries BC, reads forty years shorter:  "ἐν τῷ τεσσαρακοστῷ καὶ τετρακοσιοστῷ ἔτει" — in the four hundred and fortieth year.  The reading is in the Rahlfs-based critical text and in Swete's printed edition, whose apparatus records no manuscript variant on the number at all.¹

The manuscript count is worse for the early date than the bare split suggests.  Henry Smith, who argues for the Masoretic 480, reports the tally himself:  "In 1 Kgs 6:1, twenty-four Greek MSS preserve 440th, while seven preserve 480th."²  And the age distribution runs the same way — Vaticanus, Alexandrinus, Coislinianus, and Basiliano-Vaticanus all read 440.

Smith's case for 480 does not rest on manuscript weight.  It rests on the Jubilee cycles, which is a real argument and is examined below.  The point here is narrower:  the number at the base of the early-date case is disputed in the manuscripts, the dispute runs against it, and popular treatments do not mention it.

The Greek is not only different in that verse.  It restructures the whole chapter:  3 Kingdoms 6 has no equivalent of the Hebrew 6:11–14, no 6:18, and no 6:37–38, and the material of the Hebrew 6:37–38 has been moved to the front as 6:1c–1d.¹  This cuts both ways.  It weakens any claim that the Greek here is a careful witness to the Hebrew;  it also weakens any claim that its 440 is a simple copying slip.

How long in Egypt? 430 years, or 215

The second number is the sojourn, and here the witnesses split cleanly.

• Masoretic Hebrew (Exodus 12:40): 430 years, "in Egypt." Egypt only.
• Septuagint: "ἐν γῇ Αἰγύπτῳ καὶ ἐν γῇ Χανααν" — in the land of Egypt and in the land of Canaan.
• Samaritan Pentateuch: "ומושׁב בני ישׂראל ואבותם אשׁר ישׁבו בארץ כנען ובארץ מצרים" — "and their fathers," and Canaan first.¹⁸

If the 430 covers Canaan too, the time actually spent in Egypt is roughly half.  Josephus read it that way:  "four hundred and thirty years after our forefather Abraham came into Canaan, but two hundred and fifteen years only after Jacob removed into Egypt."¹⁹  Paul appears to read it that way as well, dating the law 430 years after the promise to Abraham (Galatians 3:17).

Two cautions belong here, and they point in opposite directions.

First, the two witnesses that relieve the difficulty relieve it differently.  The Samaritan puts Canaan first and adds "and their fathers";  the Greek puts Egypt first and does not.  That looks like two independent attempts to smooth a hard reading rather than two witnesses to a common original.  The qualification is that Codices Alexandrinus and Ambrosianus do share the Samaritan's "and their fathers," so the word-order difference is the real one and the plus is not.

Second, Swete's printed text reads 435, not 430, at both verses.  Do not make anything of this:  Swete's own apparatus records that a corrector of Vaticanus, together with Alexandrinus and Ambrosianus, delete πέντε and read 430, so the 435 survives only in the uncorrected first hand of one manuscript.²⁰  It is noted here because it is the kind of detail that gets quoted as "a third tradition" by someone who has not opened the apparatus.

Jephthah's three hundred years

Israel has been living in Heshbon and its nearby towns, in Aroer and its nearby towns, and in all the cities along the Arnon for 300 years!  Why did you not reclaim them during that time? — Judges 11:26

Israel took Heshbon at the end of the wilderness years, so this puts about 300 years between the Transjordan conquest and Jephthah — who is then followed by Ibzan, Elon, Abdon, Samson, Eli, Samuel, Saul, David, and three years of Solomon.  Stack those and the total runs well past the late date's roughly 250 years for the entire span.  This is the early date's second-best argument and it is independent of 1 Kings 6:1.

Kitchen's reply is often quoted in its rudest form.  Wood quotes him calling the number the "bold bluster" of an ignorant man and calls that an ad hominem argument.⁷  Two things should be said about that exchange.  Kitchen's actual argument on the same page is arithmetical:  placing Jephthah around 1070, his 300 years would put the Transjordan occupation about 1370, which fits no proposed exodus date — not 1447, not 1260, not anything between.  And his second move is a genre argument, not an insult:  he classes the datum with statements a biblical writer reports accurately without endorsing, like the speeches of Job's comforters or Sennacherib's propaganda in 2 Kings 18.  An evangelical reader is entitled to disagree with that, but should hear it in its own terms.

Adding up the Judges: the 580-year problem

Here honesty costs the early date something.

The books themselves supply explicit year-figures for the span between the Exodus and the temple.  Every figure below was pulled from the text, not from a chronology chart.

Samson's twenty years are stated twice and counted once.  Shamgar (Judges 3:31) carries no figure.

Now add the rest:  forty years in the wilderness (Numbers 14:33–34), forty for Eli (1 Samuel 4:18), forty for Saul (Acts 13:21), forty for David (2 Samuel 5:4), and three for Solomon up to the temple.  That is 573 years — with nothing yet assigned to the conquest, to Joshua's life after it, to the elders who outlived him (Judges 2:7), or to Samuel's judgeship.  Caleb was forty when sent from Kadesh and eighty-five when the land was divided (Joshua 14:7, 14:10), and thirty-eight of those years were the wilderness (Deuteronomy 2:14), leaving about seven for the conquest.  The floor is roughly 580.

1 Kings 6:1 says 480.

Three observations follow, and only one of them helps the early date.

A flat serial reading is impossible for everyone.  The figures overshoot the verse by at least 93 years.  Whatever 480 means, it does not mean "the sum of the periods this book records."

The text itself says so.  Sum the Judges figures that precede Jephthah's speech and you get 319 — and Jephthah says 300.  Judges 10:7 has the Philistines and Ammonites oppressing Israel in the same sentence, and Samson's twenty years fall inside the Philistines' forty.  Overlap is what the book describes.  This is worth stressing because overlapping judgeships are usually treated as a conservative dodge.  They are not.  Antti Laato, arguing the critical case, reaches Jephthah's figure exactly:  fold Abimelech's three years into Gideon's forty years of peace and treat Tola and Jair as contemporaries in Israel and Gilead, and "we have exactly 300 years."⁵

The number is a compression, not an inflation.  A late writer inventing a glorious deep past out of these books would have written a bigger number.  480 is smaller than the material in front of him.  That is a small but real point against the theory that 1 Kings 6:1 is late theological arithmetic.

One more caution:  three of the five non-Judges figures are secure, but Saul's forty years come from Acts 13:21, not from any Hebrew verse — the Masoretic text of 1 Samuel 13:1 is defective and literally reads that Saul was one year old when he became king.  And the Septuagint gives Eli twenty years rather than forty, which would take twenty more off the total.

Who has solved it, and at what cost

There are two families of solution, and they are genuinely different in kind.

Family one:  the periods overlap.  This is the majority answer and it has the text's own warrant.  Kitchen concedes overlap is unavoidable for everyone on any reading of the 480.  Antti Laato, a critical scholar, gets Jephthah's 300 years exactly by folding Abimelech's three into Gideon's forty years of peace and treating Tola and Jair as contemporaries in Israel and Gilead.⁵  Andrew Steinmann has the fullest published concurrency scheme, and Rodger Young — reviewing it sympathetically — calls one of its moves "unique," another "less plausible," and closes that "the furniture will continue to be moved around."²¹  That is a fair summary of the state of this family:  it works, and no particular arrangement of it commands agreement.

Family two:  the 480 is not counting what we assume.  Kitchen's own Table 28 belongs here — he reaches 480 by selecting only the non-oppression periods, and has to supply the last five years himself.⁶  But the cleanest version was published in the Journal of Creation by David Austin in 2008, and it deserves more attention than it has had.¹⁴

Austin reads the 480 as counting only years of sovereign theocracy — the years when Israel actually governed itself — and excluding the years of foreign servitude and usurpation.  Sum those excluded years from the text:  the six oppressions (8 + 18 + 20 + 7 + 18 + 40) plus Abimelech's three.  That is exactly 114.  And 573 − 114 = 459, which leaves 21 years of the 480 for the conquest, Joshua's later life, the elders, and Samuel — the four items the text never quantifies.

The overshoot vanishes, with no appeal to overlapping judgeships at all.  The arithmetic checks against the verified figures, and it is the most economical answer to this problem in print.

Its cost is severe, and Austin states it:  his Solomon's fourth year is 939 BC and his Exodus is 1532 BC.  That abandons Thiele's Assyrian-anchored 967 — which is the one date in this whole article that rests on a dated eclipse.  And it detonates the 1446 that his own publisher's other authors defend.  Keep Thiele's anchor and run Austin's mechanism and the Exodus lands around 1560 BC, in the late Second Intermediate Period, a couple of decades before Ahmose.  That is an interesting place for it to land — it is roughly where Josephus put it — but no one currently argues for it, and this article is not going to be the first on the strength of one Viewpoint article.

What is worth noticing is who has not done this arithmetic.  Creation Ministries International holds the 1446 date and has published both of the above solutions — Osgood's overlapping scheme in 1984 and Austin's theocracy reading in 2008 — without ever adjudicating between them, though they yield Exodus dates 86 years apart.  And Robert Carter, CMI's geneticist and the author of its fullest biblical-chronology paper, has no published solution to the overshoot because he has never stated the problem:  in that paper the entire span from the Exodus to Solomon is a single link, the 480 of 1 Kings 6:1, and the Judges figures are never summed.¹⁵  His only use of a Judges datum runs the other way — Jephthah's 300 years, cited to argue the period cannot be shortened much.

That silence matters more than it might seem, because his paper is explicitly a worst-case exercise designed to bracket every source of chronological slack in Scripture.  It is a fair measure of how much of the popular chronology literature has actually done the addition.

The generations problem

Two genealogies cover the same stretch of time and disagree about how long it was.

For a short sojourn.  Exodus 6:16–20 runs Levi → Kohath → Amram → Moses.  Moses is the third generation after Levi.  Four names do not stretch easily over 430 years.

For a long sojourn.  1 Chronicles 7:20–27 runs Ephraim → Shuthelah → Bered → Tahath → Eleadah → Tahath → Zabad → Shuthelah → Rephah → Resheph → Telah → Tahan → Ladan → Ammihud → Elishama → Nun → Joshua.  Ephraim and Levi belong to the same generation;  Joshua and Moses were contemporaries.  The same span is three generations in one line and roughly ten to sixteen in the other, depending on whether verse 25 lists brothers or successive sons — Resheph carries no "his son" in the Hebrew.

The usual resolution is that Hebrew genealogies skip generations and "son of" can mean "descendant of."  That is true and well documented.  Notice which way it cuts:  it rescues the long sojourn without shortening anything, and it rescues the short sojourn only by assuming the Chronicles list was padded, which is not how genealogies usually go wrong.

A control sits nearby and it is the best under-used argument on the early-date side.  1 Chronicles 6:33–38 gives twenty-two names from Levi to Heman, a musician in David's court — twenty-one generations.  On the early date that span is about 480 years, or roughly 23 years per generation, which is ordinary.  On the late date it is about 250 years, or roughly 12 years per generation, which is impossible.

Both sides have published on this and reached opposite verdicts from the same kind of list.  Laato uses the twelve-name priestly line from Aaron to Ahimaaz to argue that 480 is schematic — twelve generations of forty — while noting that twelve generations at about 25 years each is about 300 real years.  Kitchen uses the same line for the opposite purpose, arguing it covers "the 300 years from the exodus to early in Solomon's reign," and calls the short genealogies "clearly selective."⁶

What the ancient chronographers thought

They did not agree, and Josephus did not agree with himself.

In Antiquities 8.61 he puts 592 years between the Exodus and the temple.  In Antiquities 20.230 he puts 612, and repeats the 612 in Contra Apionem.  In Antiquities 2.318 he gives the 430/215 short sojourn.  In Contra Apionem 1.103 he identifies the Israelites with Manetho's expelled shepherds and dates their departure 393 years before Danaus came to Argos — a scheme no modern scholar accepts and which contradicts all three of his other figures.¹⁹  He is not a witness;  he is four witnesses who disagree.

Sulpicius Severus, about AD 400, reckoned about 588 years, and did something instructive with the discrepancy.  His Greek text read 440, and he rejected it:  "I do not doubt that the truth had been falsified by the carelessness of copyists, especially since so many ages intervened, rather than that the sacred writer erred."²²  He knew about the 440 problem sixteen centuries ago and reasoned about it openly.

Seder Olam Rabbah chapter 3 is the classic source for the rabbinic 210-years-in-Egypt figure, derived by starting Genesis 15:13's four-hundred-year clock at Isaac's birth.  Separately, it uses the Levite lifespans — Kohath 133, Amram 137, Moses 80, totalling 350 — to rule out a four-hundred-year stay in Egypt.²³  Chapter 15 preserves the 480 reading of 1 Kings 6:1.

The Book of Jubilees agrees with nobody:  the Exodus at 2410 from creation, Jacob's descent at 2172, a sojourn of 238 years, and the entry into Canaan at 2450 — which is exactly 50 × 49, i.e. a purely schematic scheme.²⁴

Part 2 — From Solomon backward: is 1446 BC a real number?

The Assyrian anchors

The chain from Solomon to an absolute BC date is the strongest link in the whole argument, and it is worth seeing how it is fastened.

Assyria dated its years by eponyms — one named official per year.  One of those years records a solar eclipse.  Thiele:  "the date 763 for the eponymy of Bur-Sagale has been established not only by the astronomical evidence of Assyria but also by that of Ptolemy's canon, and we thus have complete assurance that 763 is the correct date."²⁵  A continuous eponym list runs from 911 BC down to 648 BC, so one eclipse dates the whole series.¹¹

Two Assyrian monuments then tie Israelite kings into that list.  Shalmaneser III's Kurkh Monolith names the allies who fought him at Qarqar, including "2,000 chariots, 10,000 soldiers of Ahab, the Israelite."²⁶  His annals record tribute from "Iaua (Jehu), son of Omri" in his eighteenth year.²⁶  Qarqar falls in Shalmaneser's sixth year, 853 BC;  Jehu's tribute in his eighteenth, 841.  The twelve-year gap matches the twelve years between Ahab's death and Jehu's accession in Kings exactly.²⁵

From 853, Thiele sums the reigns of Jeroboam I through Ahab.  Israel used non-accession-year reckoning, which double-counts one year per reign, so 84 official years are 78 actual ones.  "With an interval of 78 years between the accession of Jeroboam I and the death of Ahab, and with the latter taking place in 853 B.C., we thus secure the date 931 B.C. as the year of Jeroboam's accession and of the schism."²⁵

The arithmetic to 1446

Solomon's fourth year comes out at 967 BC.  Kitchen, who holds a thirteenth-century Exodus, agrees on that figure — which is worth noting, because it means the dispute is not about the monarchy.⁶

Then Rodger Young adds a step most popular treatments skip.  1 Kings 6:1 gives an ordinal — the four hundred and eightieth year, not 480 years after — so 479 elapsed years are added, not 480.  "Since Solomon's fourth year was the 480th year of the Exodus Era (not 480 years after the Exodus…), the Exodus, according to this verse, occurred in 968t BC + 479 = 1447t BC according to a Tishri-based calendar.  But since the Bible — and the undeniable celebration of Passover in the spring — indicate that the Exodus took place in Nisan in the spring … we have the Exodus in Nisan of 1446 BC."²⁷

This ordinal reading is not a modern device built to rescue 1446.  Ussher stated it in 1658 in as many words:  "The period is not 480 whole years, but only 479 years and 16 days."²⁸  Ussher nonetheless reached 1491 BC, and his 45-year difference from Young comes entirely from his monarchy dates, not from his reading of 1 Kings 6:1.

Young also offers a second, independent route.  The Talmud places the sixteenth Jubilee in Josiah's eighteenth year;  counting back fifteen cycles of 49 years puts the first Jubilee cycle beginning in 1406 BC, which is the conquest, hence 1446 for the Exodus.²⁹  Smith runs the same argument from two directions and finds both converge on 1406, while the 440-based date of 1366 does not fall at the start of any cycle.²

How much weight this bears depends on how much one trusts a Talmudic datum from roughly a millennium after the fact, reasoning about a cycle whose observance the Old Testament itself never records.  It is a genuine convergence and it is not nothing.  It is also not independent evidence in the way a datable inscription would be.

Different scholars land on 1446, 1447, or 1445 depending on whether they follow Thiele's own numbers or Young's correction to them.  Young identifies the unstated assumption in Thiele that produces 1445.²⁷  These one-year differences do not matter for anything archaeological.

Part 3 — 480 or 440? The two readings weighed

If 1 Kings 6:1 is a real elapsed figure, there are exactly two candidates, because there are exactly two readings in the manuscripts.  The Masoretic 480 puts the Exodus at 1446 and the conquest at 1406.  The Septuagint's 440 puts them at 1406 and 1366.  Henry Smith, who defends the Masoretic reading, states the arithmetic in the same form:  the Greek reading "would therefore place the start of Joshua's conquest in 1366 BC (967 + 439 – 40), while 480th would place it in 1406 BC (967 + 479 – 40)."²

Most popular treatments never mention that there is a choice.  This section states the case for the shorter reading at its strongest, then tests it.

The case for 440, from its opponents

The strongest evidence for the Greek reading comes from the man arguing against it.

On the count:  "In 1 Kgs 6:1, twenty-four Greek MSS preserve 440th, while seven preserve 480th."²  The age distribution runs the same way — Vaticanus, Alexandrinus, Coislinianus and Basiliano-Vaticanus all read 440.  And on the origin of the reading, Smith goes further than any 440 advocate has:

Despite all the complexities involved with the transmission of the Greek of 1 Kgs, the 440th year reading appears early and is spread across a wide array of MSS and other witnesses.  It can be definitively traced back as early as ca. AD 200 in the OL, pointing to a mid to late second century AD Greek exemplar.  440th is most likely the original OG reading, possibly originating in the first or second century BC.²

And on how firm the verdict is:

The totality of evidence would seem to somewhat favor 480th, especially since it has a greater number of the oldest witnesses attesting to it.  Even so, the choice of 480th as the original would by no means be a text critical slam dunk.²

That is the leading defender of 480 conceding that the Greek reading is probably the older one, and that the textual question is not closed.  Anyone who says the Septuagint reading here is obviously corrupt is claiming more than the best pro-Masoretic specialist claims.

The case against 440, and why the age argument is weaker than it looks

Smith's counterweight is that the oldest witnesses favour 480.  That needs a caveat he does not supply, and the standard critical apparatus supplies it for him.

Among the witnesses in the 480 column are Aquila and Symmachus — second-century Jewish revisers who corrected the Greek toward the Hebrew.  Their agreeing with the Masoretic text is what they exist to do.  And Brooke–McLean's hexaplaric band at 3 Kingdoms VI 1 records not merely that they read "eightieth" but why:  it notes that "likewise the Hebrew has basemoneim," glossing it with the numeral 80.³⁰  In other words, a scribe wrote down that the 480 reading here comes from the Hebrew.  Greek manuscripts reading 480 are not independent support for 480 when the apparatus says they are reading the Hebrew.

A defender can reverse the inference — perhaps Aquila and Symmachus preserve an original 480 that the Old Greek recalculated — and settling that needs the full recensional argument, which nobody has run for this verse.  But the burden has moved.

The better argument against 440 is about the character of Greek Kings.  3 Kingdoms 6 does not differ from the Hebrew at one number;  it restructures the chapter.  There is no equivalent of the Hebrew 6:11–14, no 6:18, no 6:37–38, and the material of the Hebrew 6:37–38 stands at the front as 6:1c–1d.¹  A translator or reviser who was already re-ordering this chapter chronologically is a translator who might also have re-timetabled its opening number.

That cuts both ways, and the article should say so.  It weakens any claim that the Greek here is a careful witness to a Hebrew original.  It equally weakens the claim that 440 is a simple copying slip — the numbers are written out in words in both languages, and τεσσαρακοστῷ and ὀγδοηκοστῷ are not confusable.  Whatever happened here was deliberate on one side or the other.

The decisive objection: 440 makes the Bible's own arithmetic worse

The explicit year-figures from the Exodus to the temple total about 573, with a floor near 580.  Against the Masoretic 479 elapsed years that is an overshoot of about a hundred.  Against the Greek 439 it is about a hundred and forty.  The reading that shortens the headline number lengthens the problem it has to solve.

Smith states one collision concretely:

The time from the conquest of the Transjordan to the end of the rule of Abdon would be approximately 300 + 6 + 7 + 10 + 8 = 331 years.  A 1446 BC Exodus ends this string of judges in about 1407 BC – 331 = 1076 BC, whereas a 1406 BC Exodus would end the judgeships forty years later, ca. 1036 BC.  However, 1036 BC extends well into the reign of Saul, whose kingship started ca. 1050 BC.²

On the Greek reading the judges are still ruling fourteen years after Israel has a king.  Smith's procedural challenge follows:  "A case for 440th should entail an exegetical analysis that internally reconciles the Judges data with the OG's reading.  In my judgment, it is difficult to foresee how this reconciliation might be accomplished."²

No such analysis appears to exist.  Seven lines of research failed to find one.  That absence is the most reportable fact about the 440 reading.

One reciprocal is owed, because Smith does not draw it.  On his own numbers the Masoretic date ends the judgeships at 1076 BC, leaving only twenty-six years before Saul for Samson's twenty, Eli's forty and the whole of Samuel.  480 does not escape the compression problem.  It survives it;  440 does not.

An objection that looks decisive and is not

Here is an argument against 440 that seems powerful, and it is worth walking through precisely because it fails.  It fails in a way that shows how easy it is to build a knockdown out of a manuscript apparatus one has not read to the end.

If the Greek tradition preserved a shorter chronology overall, its figures in Judges ought to be shorter too.  In the two great uncials, they are longer:

Vaticanus is the oldest witness to 440 and gives Jephthah fifty-four years more than the Hebrew;  Alexandrinus is the second oldest and gives Othniel's rest ten more.  Take either codex as a package and the Judges total rises from 410 to 464 or 420, which would make the overshoot worse still — a shorter headline number over a longer body.³

But the apparatus does not support the generalisation.  In Brooke–McLean the sixty at Judges 12:7 is confined to B and a small group, with Alexandrinus, the majority, and the Coptic, Ethiopic, Old Latin, Syro-Hexapla, Theophilus and Clement all reading six.  The fifty at Judges 3:11 is likewise confined to A and a few allies, with B and most others reading forty.  And the Old Latin, Syro-Hexapla, Armenian and Ethiopic — which carry 440 at 3 Kingdoms 6:1 — read the Masoretic figures in Judges.  So there is a single coherent line of witnesses that gives 440 in Kings together with the ordinary numbers in Judges, and a 440 advocate can simply stand in it.

The observation is therefore true of Codex B and Codex A individually and false of "the Greek."  What survives is much smaller:  the two oldest witnesses to 440 each contain an inflated judgeship, so anyone who wants to treat those codices as authorities on chronology inherits a problem elsewhere.  That is a caution about manuscript cherry-picking, not a refutation of the reading.

It is set out here at length because the tidy version of it — the Greek lengthens Judges, so 440 defeats itself — is exactly the kind of argument this article exists to check rather than repeat.

What 440 costs at Jericho

Bryant Wood dates the fall of Jericho City IV to about 1400 BC, which is why the early date holds it:  on the Masoretic reading the conquest begins in 1406 and Jericho falls six years in.  On the Greek reading the conquest begins in 1366, thirty-four years after Jericho fell.

This is not a marginal cost.  Titus Kennedy, an early-date advocate, sets out the reconstruction the argument depends on:  "the destruction of the final walled Bronze Age city of Jericho appears to have occurred in near the end of the Late Bronze I period around 1400 BC," with the site then abandoned "except for a palatial residence briefly occupied in Late Bronze IIA."³¹  Late Bronze I ends about 1400 both on the current excavators' periodisation and on Kennedy's own usage, so a 1366 conquest lands a third of the way into a phase in which Kennedy's Jericho is already a ruin.  Adopting 440 means giving up Wood's and Kennedy's Jericho.

The Amarna letters point the other way, and it is the strongest thing a 440 advocate has.  The correspondence opens about 1360.  On the Masoretic date the conquest is forty-six years earlier, so the ʿapiru in those letters would be Israel already settled;  on the Greek date the conquest begins six years before the letters start, so Canaanite kings would be writing to Egypt about invaders during the conquest itself.

Who actually holds it

Almost nobody.  The named advocate is Steven Collins, who dates the Exodus to a range of 1416–1386 BC — and who reaches the shorter number because his chronology for Tall el-Hammam requires it, not on any text-critical ground.⁴  Scott Stripling names him as the archaeologist who follows the Septuagint here.  The rest of the early-date camp does not engage the variant at all:  Joel Kramer never mentions the Septuagint, David Rohl never registers the reading, Creation Ministries International books it in a table and prefers the Masoretic without giving a reason, and Titus Kennedy disposes of it in a single footnote.

There is a third possibility worth naming, because it explains both numbers from one mechanism rather than making one of them a corruption:  Joshua Berman reads 480 as twelve generations of forty and 440 as eleven, keyed to the priestly line.³²  It is more economical than either "the Greek shortened it" or "the Hebrew padded it."  Two things must be said with it.  It dissolves the premise of this whole section, because on that reading neither number is an elapsed total and neither dates the Exodus.  And Berman is not a friendly witness for the early date:  he holds that it is "not possible [archaeologically] to reconcile a conquest of the land by Joshua on any scale before [the 13th century b.c.e.]", so the generational reading exists in his hands to support the late date.

Verdict on the variant.  440 is textually respectable and chronologically unworkable, and those two facts should be stated in that order.  It has the better manuscript count, the older uncials, and probably the older Greek reading;  it has no exegetical defence, no answer to the Judges arithmetic it makes worse, a Greek Judges tradition that runs against it, and no Jericho.  480 survives the same tests only by requiring substantial overlap among the judgeships — but requiring compression is a smaller problem than requiring compression plus forty more years of it.

Part 4 — The Egyptian side: how firm are Egyptian dates?

Every Exodus date is a claim about Egyptian chronology, so it is worth knowing what that chronology rests on.  This section is written to be fair in both directions, because the temptation in apologetics is to collect the uncertainties and stop, and the temptation in the textbooks is to present the framework as settled.

How the framework is built

The backbone is dead reckoning — adding up reign lengths from king lists and monuments.  The sources are worth taking one at a time.

Manetho, the third-century BC Egyptian priest whose dynasty scheme everyone still uses, survives only in quotations.  Kitchen:  "The list is incomplete (not naming all kings), and the figures for both reigns and dynastic totals show variants, some clearly wrong;  Egyptian names are often garbled in their Greek form."¹¹

The Turin King-list, the most important native document, is in poor shape.  Kim Ryholt, its specialist:  it "can be accurately described as a copy, drawn up in a somewhat careless manner on waste paper, from a damaged and imperfect original … It has since been reduced to more than 300 fragments, apparently due to rough handling … many of the very small scraps remain unpublished, and the position of many important fragments remains conjectural."¹²

The Royal Annals (Palermo Stone).  The Brill handbook's own verdict:  "The reliability of any sum is, however, open to question;  whereas the order of succession as preserved in the Annals conforms to contemporaneous data, its relevance for counting regnal years remains conjectural."¹²

Apis-bull burials show how the gaps get bridged in practice.  A gap between Shoshenq V and Bocchoris is closed because a bull buried in Bocchoris's sixth year succeeded one that died in Shoshenq V's thirty-seventh, and a bull's lifespan caps the interval at 26 years.¹²  That is real evidence.  It is also a reminder of how much of the structure is inference rather than a dated record.

Sothic dating and its problems

The Egyptian civil calendar had 365 days and no leap year, so it drifted against the seasons by one day every four years, returning to alignment after about 1,460 years.  A text that records the heliacal rising of the star Sothis (Sirius) on a given civil date therefore fixes the year — in principle.

In practice Kitchen names three problems:  "One is the rarity of such dates in our existing texts … Another is the fact that (on its own) such a date is only correct within four years … A third is that the date BC to be calculated varies with the latitude of the place of observation — itself a controverted matter.  The further south the observation was made, the later the date BC."¹¹

The latitude problem is not small.  Rolf Krauss, who wrote the standard treatment:  "Egypt's ancient boundaries extended from 24° (First Cataract, Elephantine) to 31.15° (Diospolis inferior) north latitude … the star rose 8 to 10 days later on the Mediterranean coast than at Egypt's southern border."¹²  Eight to ten days of drift is thirty-two to forty years.

Krauss also records, without embarrassment, that the assumed observation site has been moved south twice — from Memphis to Thebes around 1950, after new Assyrian and Babylonian king-list evidence shortened Near Eastern chronology by about 20 years;  and from Thebes to Elephantine in the 1970s, after a further shortening of at least 10 years.  Each move "compensated for" the chronological difference.¹²

That sequence deserves to be reported carefully rather than triumphantly.  Krauss is the proponent of the Elephantine reference point, and he gives independent reasons for it:  Elephantine is the southernmost point in Egypt where Sothis could first be seen, it is the mythical source of the Nile, Late Period texts tie the rising of Sothis to the inundation there, and inundation height was measured there.  Kitchen thinks that is "a very local theoretical plea, unsubstantiated by any decisive factual data."  Both views belong on the page.

Once the assumptions are allowed to vary, the windows are wide.  The Illahun Sothic date "corresponds to any year between ca. 1882 and 1830 BC" — a 52-year window.¹²  The Ebers date "cannot be earlier than ca. 1549 BC and not later than 1506 BC" — 43 years.¹²  And the Middle Kingdom star-clock coffin lids are internally inconsistent with the Illahun date by 50 to 75 years.¹²

The handbook's own editor draws the moral:  "Sothic dates with all of their countless parameters and uncertainties are likewise only of value as confirmations, and can never serve as the point of departure."¹²  Historical dates come first;  astronomy checks them.

What the error bars actually are

Kitchen's published figures, from 1991:

His prose summary:  "from about 30 years in 2000 BC, through 20/10 years by 1500 BC, to 11 years in the thirteenth century BC and less into the first millennium BC, dates come exact from 664 BC."¹¹

Two honest caveats about that table.  It is Kitchen's personal 1991 assessment, not a discipline-wide statement, and it is thirty-five years old.  The 2006 Brill handbook uses a different apparatus:  one-sided superscript alternatives (+25, +16, +6, +3 years) on Early Dynastic through early Middle Kingdom dates, which stop entirely at Senwosret II in 1845 BC.  Those are not symmetric error bars — they are discrete alternatives reflecting specific unresolved questions such as whether the Twelfth Dynasty coregencies are counted.

So the honest summary is:  around 20 years of play at the Exodus horizon, growing to a century or two in the third millennium, and zero after 664 BC.  Twenty years is enough to matter for matching a destruction layer.  It is nowhere near the two to six centuries a revised chronology needs.

Radiocarbon

In 2010 an Oxford team ran 211 radiocarbon measurements on short-lived plant material with a Bayesian model of reign lengths.  Their result:  the data "indicate that the New Kingdom started between 1570 and 1544 B.C.E., and the reign of Djoser in the Old Kingdom started between 2691 and 2625 B.C.E.;  both cases are earlier than some previous historical estimates."¹³  They also found a small regional offset — 19 radiocarbon years older — probably from the Nile Valley's different growing season.  The project's own summary is that radiocarbon "does give a chronology which is in broad agreement with historical estimates."³³

A 2025 study added 48 high-precision dates and concluded that "the Low Chronology is no longer empirically supported for the Old and Middle Kingdoms," with Middle Kingdom accessions peaking 10 to 15 years earlier than most High Chronology estimates.³⁴

Note the direction.  Both programmes push Egyptian dates slightly earlier, not later.  A creationist or apologist hoping radiocarbon would compress Egyptian chronology should read them and adjust.

Where Egyptology's chronology really is contested

It is contested — just not where an Exodus argument would want.

The Thera eruption is the live fight.  Radiocarbon has long pointed 50 to 100 years earlier than the archaeological date;  Sturt Manning's 2022 analysis gives "~1606–1589 BCE (68.3% probability), ~1609–1560 BCE (95.4% probability)."³⁵  But the radiocarbon camp's own newest evidence pulls the date down again:  a buried olive shrub on Therasia supports a mid-sixteenth-century date, and the authors say 1611 BCE "is marginalised by our dating results."³⁶  A 2025 study places the eruption before Ahmose and concludes that assessments linking the Tempest Stela to Thera "can now be considered incorrect."³⁷

The Venus Tablet of Ammisaduqa has been abandoned as a fixed point.  Jörg Klinger:  "This point has now been lost since the observations are no longer viewed as reliable, with the result that the long debate about the relative merits of the Middle or Low chronologies must now be abandoned."¹²

None of that helps.  Klinger's very next point is that the sequence survives even where the anchor does not, and the relative chronology of the first half of the second millennium "can be viewed as certain."  And the reason the second-millennium framework is robust is not astronomy at all but cross-linkage:  the Hattusha and Amarna correspondence documents synchronisms between Ramesses II and Kadashman-Turgu and Adad-nerari I, and between Akhenaten and Burnaburiash.  Krauss and Warburton:  "Since the anchors for the astronomical absolute dates and the historically relative dates for Mesopotamia and Egypt are fundamentally different, these synchronisms can be viewed as virtually certain."¹²

That is the honest bottom line of this section.  Egyptian chronology has genuine, documented softness — more than apologists who cite it as fixed usually admit.  It has nothing like the softness a revision needs.

Part 5 — The late date's best argument: the store cities

So they put foremen over the Israelites to oppress them with hard labor.  As a result they built Pithom and Rameses as store cities for Pharaoh. — Exodus 1:11

This is the single strongest external argument for a thirteenth-century Exodus, and the exchange over it — Wood 2005, Hoffmeier 2007, Wood 2007, all in the same journal — is the best short introduction to the whole dispute.

What is actually at Qantir

Pi-Ramesse is not at Tanis, where nineteenth-century explorers put it.  Kitchen:  "for miles around Khataana-Qantir, the deceptively flat terrain shrouds thousands of acres of the last remains of Pi-Ramesse/Raamses … When the city was largely abandoned from circa 1130 onward, and the new (Twenty-First) dynasty needed stone to build great new temples at its capital, Tanis, they simply removed the Ramesside temple stonework from Pi-Ramesse to Tanis for reuse — where its presence was later initially to deceive modern explorers."⁶

The occupation window is narrow:  a summer palace under Sethos I, the city built by Ramesses II in his own name, work continuing to Ramesses VI, abandoned as a royal residence about 1130.  Kitchen's conclusion:  "the occurrence of Raamses is an early (thirteenth/twelfth century) marker in the exodus tradition, and that fact must be accepted."  He also notes, against Lemche, that "Tanis never ever bore the name Pi-Ramesse."⁶

Kitchen's summary of why he holds the late date is one sentence:  "an exodus before 1320 would have no Delta capital to march from."⁶

Pithom, which nobody can agree on

The site identifications are genuinely unsettled, and running them as a smooth cumulative case would be dishonest.

Kitchen puts Pithom at Tell er-Retaba and Succoth at Tell el-Maskhuta.  His decisive evidence is a Roman milestone found at Maskhuta reading "9 [Roman] miles on the road from Ero [Pithom] to Clysma [Suez]" — and Retaba is almost exactly nine Roman miles from Maskhuta.⁶

Wood, who worked on the Maskhuta dig, puts Pithom there instead, and reports that "prior to ca. 610 B.C. (Saite period), the only occupation was during the Hyksos period … the only possible time period when the Israelites could have worked there as slaves was during the Hyksos period."³⁸  Kitchen calls the Maskhuta-was-only-Saite claim "palpably false."

The independent excavation report cuts against the early date at the other candidate.  The Polish-Slovak mission at Tell el-Retaba dates the earliest defence wall to phase E4, the 19th Dynasty, "already by Seti I ?", with the fortress proper built by Ramesses II and rebuilt larger by Ramesses III.³⁹

The one thing all three agree on is the thing that matters:  neither candidate site has a fifteenth-century Late Bronze building phase.  That is a real problem for 1446 and it should be stated as such.

"Rameses" as a modernized name

The early date's answer is that "Rameses" in Exodus 1:11 replaces an older name that had gone out of use.  Wood:  "Editorial updating of names that had gone out of use is not uncommon in the Hebrew Bible.  Other examples are Bethel … Dan … and Samaria."  And he presses an inconsistency:  "Kitchen allows for editorial updating of the name Rameses in Gen 47:11, and Dan in Gen 14:14, but not for Rameses in Exod 1:11."⁷

The Genesis 47:11 point is stronger than Wood's phrasing makes it sound, and it is the hinge of this whole section.  Genesis 47:11 says Joseph settled his family "in the land of Rameses."  On every dating scheme, no pharaoh named Ramesses existed in Joseph's lifetime.  So both sides already concede at least one modernized place name in the Pentateuch.  The question is not whether such updating happens.  It is whether it happened here.

Wood's other rebuttal is chronological, and it is his best line:  Exodus 1 has the store cities built, then the population increase, then the order to kill the newborn boys, and only then Moses' birth.  Moses was eighty at the Exodus.  "Since Rameses II was 25 years of age when he began his rule, the Israelites built the store city called 'Rameses' before Rameses II was even born!"⁷

Hoffmeier answers directly:  "construction at Tell el-Dab'a-Qantir is now documented under the previous reigns of Horemheb (1323–1295 bc) and Seti I (1294–1279 bc).  This means that the oppression of the Hebrews could have begun decades before the reign of Ramesses II and culminated with the construction of Pi-Ramesses."⁸

He then makes a grammatical argument against the updating theory.  When the Hebrew Bible updates a toponym it normally shows its work with a formula — old name, then "that is," then the new name, as in Genesis 14:2's "Bela, (that is Zoar)."  "The toponym Rameses occurs five times in the OT, in Gen 47:11;  Exod 1:11;  12:37;  and Num 33:3, 5.  In none of these cases is the formula … used."⁸  He adds that the name had a short life — the last datable inscription from Qantir is from Ramesses VIII — so the glossing would have to have happened in a window of about 150 years;  and he asks why, if the text was being updated, it was never updated again to Zoan/Tanis, which was the Delta capital from about 1070 into Roman times.

Wood's counter is that Rameses was the last name the site bore, so it is the natural one to retain, and that Zoan/Tanis was not where the Israelites lived but 19 km to the north-northeast.⁴⁰

Readers should notice that Hoffmeier's "limited window" argument rests on a footnote crediting "verbal communication from Professor Kitchen," which is not a checkable source.  Readers should also notice that Hoffmeier's formula argument is strong, and that Wood does not fully answer it.

Part 6 — The early date's other arguments, and their state

The Amarna Habiru

The Amarna letters, from the fourteenth century, have Canaanite rulers begging Egypt for help against people called Habiru or 'apiru.  Wood reads them as the Israelites of the conquest — a natural fit if the conquest happened around 1400.

This argument is weaker than its popularity suggests, and the best reason to think so comes from the early-date camp.  John Bimson, the most technically serious early-date archaeologist, declines to use it:  "I would certainly not wish to use the appearance of the Habiru in the Amarna letters to support the early date for the Exodus and Conquest;  rather I would simply assert that nothing in our knowledge of the Amarna period tells against the view that it belongs after, rather than before, the Conquest."⁴¹  That is a much more modest claim, and it is the defensible one.

Petrovich's related argument is more specific:  3,600 'Apiru appear on the booty list of Amenhotep II's second Asiatic campaign, listed first among the non-royal ethnic groups despite being fewer than the groups after them.  He argues that prominence "is easily explainable if these were Hebrews, and the exodus had occurred not [quite] seven months before."⁴²

Jericho

Bryant Wood re-dated the destruction of Jericho City IV from Kathleen Kenyon's roughly 1550 BC to about 1400, on ceramic grounds primarily, with three supporting lines.  His own ranking is honest:  "Of these four lines of evidence, the ceramic data are first and foremost.  The other three are merely supportive and are not in and of themselves sufficient to compel a revision of Kenyon's date."⁴³

Radiocarbon has since gone against him, and against Bimson, by name.  Bruins and van der Plicht dated short-lived cereal grain from the destruction:  "The 14C dates do not support a low 15th century BC chronology for the end of the MBA, as Bimson (1978) proposed.  Further, the fortified Bronze Age city at Tell es-Sultan was not destroyed by ca. 1400 BC, as Wood (1990) suggested."⁴⁴

Young's response is to compare uncalibrated BP values against Thera rather than trusting calibration.⁴⁵  This should be treated with caution and the article will not lean on it:  it is a self-hosted conference slide deck, never peer reviewed;  a difference in raw BP values is a difference in radiocarbon years, which is exactly what a calibration curve exists to convert;  and the curve wiggles hard in this interval, which is why a very tight measurement of 3306 ± 7 BP explodes into two separate 35-year calendar windows.  Young's own notes say the talk was not intended as a balanced presentation.

Two more recent items make this worse, and both come from inside the early-date camp or from the current excavators.

The excavation now running at Tell es-Sultan dates the destruction to the sixteenth century.  Lorenzo Nigro and Jehad Yasine, reporting the Italian-Palestinian expedition's 2019–2023 seasons:  "The latest 14C samples from House A341 allow to date the final destruction of the Middle Bronze city in the third quarter of the 16th century BC (1550-1525 BC)."¹⁶  That is Kenyon's date, not Wood's.  Their own phase table puts Late Bronze IB at 1470–1400 and Late Bronze IIA at 1400–1300 — so a 1406 conquest and a 1366 conquest fall on opposite sides of a boundary drawn at a round number.  What the expedition does report on the Spring Hill for the fifteenth century is a reused building in a ruined palace, dated by pottery including a Cypriot White Slip II bowl — a building, not a destroyed walled city.

And an early-date advocate has published the radiocarbon against himself.  Titus Kennedy, arguing for a c. 1400 destruction, concedes that the dates "appear to only rule out a ca. 1230 BC destruction hypothesis but have not been able to solve the dispute between a Middle Bronze III or Late Bronze I destruction of the city."³¹  Then he recalibrates the Kenyon samples himself:  "Using OxCal with IntCal20 and no adjustments for geographical location, these samples would yield calibrated dates ranging from 1699 to 1623 calBC 70.3% probability (GrN-18368) to 1617–1531 calBC 95.4% probability (GrN-18539)" — and concludes they are "significantly higher than the historical dates."³¹

Two things follow, and the second is the one that matters.  First, if radiocarbon cannot decide a 150-year question at this site, it certainly cannot decide the 40 years between 1406 and 1366.  Second, Kennedy discounts the result as a known conflict between Levantine radiocarbon and ceramics, and holds that pottery "continues to be the primary and most reliable method."  That move is available to him — but if it is accepted here, it also dissolves the radiocarbon that constrains Egyptian chronology, which the same camp relies on in Part 8 to rule out the revised chronologies.  A method cannot be decisive against Rohl and unreliable at Jericho.

Jericho belongs to article 6 of this series and is treated properly there.  What can be said here is narrower and firmer:  the site currently supports neither conservative date, and the strongest evidence for saying so has been published by the early-date camp itself.

The early-date camp does not agree with itself

This is worth stating plainly, because it is usually presented as one position.

• Date: Bimson says about 1470; Wood, Petrovich, and Young say 1446; Young allows 1445 under a different reading of Thiele.
• Method: Bimson treats 1 Kings 6:1 as a rough guide and thinks the real interval was longer than 480 years; Wood argues the numeral's form makes it technically precise.
• Pharaoh: Bimson says Thutmose III;⁴¹ Wood and Petrovich say Amenhotep II.
• Did he drown: Wood argues the text implies the Exodus pharaoh died in the sea and uses that against the late date; Petrovich concedes he did not — "No, he died in usual fashion, and his mummified body is still preserved. Yet this does not conflict with the Bible, since no biblical text explicitly states that the exodus-pharaoh died there with his army."⁴²
• Amarna: Wood makes the 'apiru a pillar; Bimson refuses to use them.
• Archaeological horizon: Wood dates Jericho's fall to the start of Late Bronze; Bimson dates the same event to a re-dated end of Middle Bronze.

None of that makes the early date wrong.  It does mean "the biblical date" is not one thing.

Part 7 — The late date's other arguments

Merneptah

The stele of Merneptah, from about 1208 BC, contains the earliest known mention of Israel outside the Bible.  Breasted's translation of the closing lines:  "Yenoam is made as a thing not existing.  /  Israel is desolated, his seed is not;  /  Palestine has become a widow for Egypt."⁴⁶

The detail that matters is grammatical.  Kitchen:  "The Israel of Merenptah's stela was, by its perfectly clear determinative, a people (= tribal) grouping, not a territory or city-state;  rare statements to the contrary are perverse nonsense, especially given the very high level of scribal accuracy shown by this particular monument."⁶

This is a terminus ante quem, not a date for the Exodus.  It establishes that a people called Israel was in Canaan by 1208 and therefore that the Exodus, if it happened, happened before then.  It is compatible with both conservative dates and with neither minimalist reconstruction that puts Israel's emergence later.  Merneptah is treated fully in article 6.

Transjordan

Numbers 20–21 has Israel negotiating with organised kingdoms of Edom and Moab.  Nelson Glueck's surface surveys concluded these did not exist before the thirteenth century:  "Had the Exodus through southern Transjordan taken place before the 13th century B.C., the Israelites would have found neither Edomite nor Moabite kingdoms, well organized and well fortified, whose rulers could have given or withheld permission to go through their territories."⁴⁷

Glueck should be cited as history, not as evidence.  His occupational gap was an artefact of surface survey;  Late Bronze material has since been found;  and later excavation moved the Edomite state to the eighth and seventh centuries.  His specific claim is now held by nobody, late-daters included.  An article that cites him as current evidence will be discarded by anyone who knows the Transjordan literature.

The modern version of the argument is different and weaker for both sides.  High-precision radiocarbon at Khirbat en-Nahas, the largest Iron Age copper-production centre in the southern Levant, dates occupation from the end of the twelfth century, with major smelting in the tenth and ninth.⁴⁸  Levy and colleagues are arguing that Edomite complexity is earlier than the prevailing low chronology, and their dating is contested by Bienkowski and by Finkelstein.

Kitchen's own handling of the Transjordan negatives is instructive, because he concedes the archaeology and answers it with texts.  Dibon has "no recovery of Late Bronze/Iron I-IIA remains" — but Ramesses II names Moab in a topographical list and depicts himself capturing five named forts there, and the reading of Dibon "is beyond doubt."  Heshbon "shows (at present) clear urban occupation from the twelfth century onward … but not Late Bronze II before it."⁶  His conclusion is that the archaeology is incomplete and "stands supplemented and corrected by the texts."  That is a legitimate move.  It is also the same move he denies to the early date at Jericho.

The toponym cluster

Hoffmeier's argument is that "Rameses" does not stand alone.  Pithom (Exodus 1:11), Migdol (14:2), and p3 twfy = yam suph (10:19;  13:18;  15:4, 22) are "attested beginning in 19th Dynasty sources, but are not found prior to the 13th century," and the closest Egyptian parallels to Pi-hahiroth and Baal-Zaphon are likewise thirteenth-century.⁸

This is the most serious version of the late-date case, because it does not depend on any single word.  It is also the argument the early-date literature answers least well.

Part 8 — Moving Egypt instead: the revised chronologies

If Egyptian dates were centuries too high, the whole problem would dissolve.  Several people have proposed exactly that.

Immanuel Velikovsky, Ages in Chaos (1952), removed some six centuries, identifying the Hyksos with the Amalekites, Hatshepsut with the Queen of Sheba, and Thutmose III's Palestinian campaign with Shishak's sack of Jerusalem — with Kadesh identified as Jerusalem.⁴⁹  His own analogy for the scale of the alleged error:  imagine "the year would be 1941 while in Britain it would be 1341."

Donovan Courville, The Exodus Problem and Its Ramifications (1971), proposed a downward revision "of the magnitude of more than 600 years at the time of the Conquest," largely by making dynasties parallel, borrowing the Hyksos-Amalekite identification from Velikovsky, and placing the Exodus at the end of the Early Bronze Age.⁵⁰

Peter James et al., Centuries of Darkness (1991), argued the Late Bronze chronology is up to 250 years too high, moving the end of the New Kingdom from 1070 to about 825 BC and Ramesses II into the eleventh and tenth centuries.⁹  Notably, James and his colleagues call Velikovsky's scheme "disastrously extreme."¹⁰

David Rohl, A Test of Time (1995), put the Exodus in the late 13th Dynasty under Dudimose:  "The biblical date for the Exodus is 1447 BC (Thiele's foundation date of the temple in 968 BC plus 479 years) and so Dudimose was in all likelihood the Pharaoh of the Exodus."⁵¹  Rohl also attacks the Shishak = Shoshenq equation on philological grounds, noting that Shoshenq appears in cuneiform as Susink(u), so "the 'obvious similarity' between the names Shoshenk and Shishak may be more apparent than real."⁵¹

What breaks

Four independent arguments block a large down-dating, and they are ranked here by how hard they are to answer.

1.  The Assyrian eponym canon and king list.  Nicholas Postgate, in the Cambridge Archaeological Journal review symposium:  "The Assyrian King List presents a major stumbling-block … The last well-known king of the Middle Assyrian era is Tiglath-pileser I:  if his dates are dropped by 250 years his reign runs from 865-827 BC, which is patently absurd … Thus the Assyrian evidence remains an insurmountable obstacle composed of precise data, one which cannot be removed by vague assertions."⁹  And the revisionists conceded the framework themselves:  "The resulting picture … is as watertight as one could ask for any period of ancient history."¹⁰

2.  A single small inscription.  Kitchen's coup de grâce is a West-Theban graffito recording a Nile inundation in Year 1 of Merenptah, whose date can be checked without any chronology-internal assumption:  "there can be no doubt that Year 1 of Merenptah in this little text fell in the 13th century BC — and not 250/350 years later … All such 'revisions' are excluded … The game is up."⁵²

3.  Dendrochronology.  Anthony Snodgrass conceded he could not refute the compression on Aegean pottery grounds — "We cannot" — and then said the scientific dating settles it, citing Kuniholm's Anatolian tree-ring series matched to a Hittite palace at one end and Gordion at the other.⁹  Andrew and Susan Sherratt on the Central European oak sequence:  "Not much scope here for a down-dating of two and a half centuries."⁹

4.  Radiocarbon.  Barry Kemp:  "We are faced with a choice between two significantly different chronologies.  The radiocarbon evidence favours the traditional chronology;  it does not favour the new one … with the one independent means of verification — radiocarbon dating — supporting the old, why should we accept the new?"⁹  Sturt Manning, writing in the Brill handbook, says radiocarbon "offers the independent means to test and reject the several publications of the last decade which have argued that conventional Egyptian … historical chronology is incorrect."¹²

On Velikovsky specifically, Kitchen names an impossibility rather than an improbability:  because Velikovsky made Ramesses II into Necho II of about 600 BC, "Shoshenq III (c. 800 BC) would then have cut up a colossus of Ramesses II (still 200 years into the future!) to build his gateway at Tanis."⁹

On Shishak, the dispute is philological, which is what makes it immune to re-dating.  Kitchen:  "For purely linguistic, not chronological, reasons, the equation of Shishak (Shushaq) and Shoshenq is, frankly, unassailable."⁹  His technical claim is that Egyptian s is always reproduced as s, never sh, and sh as sh, never s.  Rohl's counter is above.  Readers can weigh them.

Chris Bennett's assessment of Rohl is the fairest one available, and it is worth quoting because it is not dismissive:  "there is a world of difference between his intellectual standing and that of Velikovsky, or even Peter James … Rohl has a considerable mastery of his material … Even if the debate is ultimately not resolved in his favour, it is worthwhile because it forces a re-examination of long-held assumptions."  And then the obstacle:  "the difficulties of applying the New Chronology to the Assyrian kinglist — present major, if not insuperable, objections to the theory."  On Rohl's twenty-year average generation length:  "the number of methodological flaws in this analysis is almost too large to count."⁵³

Creationists are not unanimous, and that is worth reporting

Revised chronology is often presented as the conservative position.  It is not.

Associates for Biblical Research — from an early-date-Exodus position — rejects Rohl outright, and names the Philistines as fatal:  "The Philistines are scarcely mentioned by Rohl.  And for good reason – they utterly destroy his reconstruction … So we are left with a situation where Saul and David are fighting an enemy who does not appear in the historical or archaeological record until 300 years later!"⁵⁴

Creation Ministries International says "Rohl is most certainly incorrect," and warns that Velikovsky-derived identifications such as Hatshepsut as the Queen of Sheba are a liability.⁵⁵

Answers in Genesis publishes the other way, arguing the Sothic theory has collapsed.⁵⁶  And the Journal of Creation published a broadly sympathetic assessment of Rohl six years after CMI's rejection — so the split runs inside single organisations.⁵⁷

Part 9 — The case that there is no date to find

The mainstream position deserves to be stated at full strength, from its own books, before it is answered.

Finkelstein and Silberman on the Egyptian record:  "we have no clue, not even a single word, about early Israelites in Egypt:  neither in monumental inscriptions on walls of temples, nor in tomb inscriptions, nor in papyri.  Israel is absent — as a possible foe of Egypt, as a friend, or as an enslaved nation."⁵⁸

Their strongest argument is not silence but positive evidence — the Egyptian fort system across northern Sinai:  "one can hardly accept the idea of a flight of a large group of slaves from Egypt through the heavily guarded border fortifications into the desert and then into Canaan in the time of such a formidable Egyptian presence."⁵⁸

On Kadesh-Barnea:  "repeated excavations and surveys throughout the entire area have not provided even the slightest evidence for activity in the Late Bronze Age, not even a single sherd left by a tiny fleeing band of frightened refugees."⁵⁸  Their verdict is that the narrative reached final form in the seventh to sixth centuries, mirroring Josiah against Necho — "neither historical truth nor literary fiction."⁵⁸

Donald Redford goes further, and his point is that the archaeology is worse for the traditional account than the silence is:  "Far more damaging, however, than this argument from silence is the archaeological record.  Sites such as Hormah, Arad, Jericho, 'Ai, and Jarmuth had indeed suffered violent destruction, but this had been during the Early Bronze Age or at the end of Middle Bronze … others such as Kadesh Barnea, Heshbon, and Gibeon were not to be settled until the Iron Age."⁵⁹  His positive reconstruction is that the Exodus is a Canaanite memory of the Hyksos expulsion, which the Hebrews "appropriated from the earlier culture they were copying."⁵⁹

Niels Peter Lemche:  the exodus is "from a historian's point of view highly unlikely" — and he rejects the Hyksos link too.⁶⁰  Philip Davies:  "we now know that very many biblical images of the past (patriarchs, exodus, wandering, conquest, perhaps David and Solomon, too) are not authentic ones."⁶¹

Two more positions belong here because they are routinely mislabelled.

William Dever is not a minimalist and does not belong in this section except to be distinguished from it.  He holds a real thirteenth-century exodus of a small group:  "behind these stories there may be authentic memories of a small 'exodus group,' essentially the two southern tribes of Judah and Benjamin … But as far as we know archaeologically, the ancestors of most Israelites and Judeans had never been in Egypt.  The biblical narrative of the exodus is best understood, therefore, as metaphorically true."⁶²

Baruch Halpern makes the same shape of argument more sharply:  a mass exodus is "completely unthinkable," but "the very modest beginnings of a cult of YHWH associated with an exodus from Egypt can thus be divined in some incident, or series of incidents, that would be invisible to us archaeologically and historically — as the Exodus is."⁶³

What can honestly be said back

The papyrus argument is real.  Kitchen:  "99 percent of all New Kingdom papyri are irrevocably lost (administrative and otherwise), the more so in the sopping mud of the Delta;  the few survivors hail from the dry sands of Saqqara and Upper Egypt, far away from Pi-Ramesse's brickfields.  A handful of wine-vintage dockets from broken jars is the sum total of our administrative texts so far recovered from Pi-Ramesse!"⁶

The defeat argument is also real, with a qualification.  Pharaohs did not monumentalize defeats.  But Egyptian practice was to reframe awkward engagements rather than omit them:  Ramesses II published Qadesh across five temple complexes as a triumph, and modern scholars judge it a draw.⁶⁴  So total silence is not the only outcome one should expect.  This qualifies Kitchen's rule;  it does not refute it, and Qadesh is in fact an illustration of it, since the Egyptian scribes recorded a victory, not a loss.

Carol Meyers states the centrist position fairly:  "the lack of direct evidence — from the land of Canaan as well as from Egypt and Sinai — does not mean that the exodus and Sinai are pure inventions.  What is presented as a series of momentous events … would in their own time have been little more than a tiny blip in the shifting movements of peoples in the east Mediterranean."⁶⁵

Ronald Hendel denies that cultural memory supports minimalism:  the biblical images "are indeed 'authentic' as 'biblical images of the past' … Hence I disagree with Philip's contention that the concept of cultural memory supports what he calls 'the minimalist option' in biblical studies.  It does no such thing."⁶⁶

And Kadesh-Barnea is a genuine problem that no apologetic answer handles well.  The peer-reviewed reevaluation of the excavation data concludes that "Substratum 4c represents the earliest occupation, which dates to the Iron I in the 12th to 10th centuries BCE."⁶⁷  Dever, who holds a real exodus, reports the same:  "there is some Iron I pottery (ca. twelfth-eleventh century) but no thirteenth-century material."⁶²  Iron I is after the Late Bronze Age, so this does not rescue either date.

The strongest concession is Kitchen's own, and it should be read in full because it contains both the admission and the answer:

No Egyptian records mention specifically Israelites working in the East Delta (or anywhere else), or a Moses who spoke for such a group, or an exodus by a group of this name (Israel).  Nowhere in Sinai has a body of Late Bronze Age people passing through left explicit traces, still less traces that are labeled as Israelite.  That applies also to Qadesh-Barnea.  And so on.  Modern complaints about lack of evidence are often heard.  But they usually come from folk who have not done their homework or thought things through with sufficient rigor.⁶

Common objections and responses

"1 Kings 6:1 settles it.  480 years, done." It would, if the text were undisputed and if the rest of the Bible's numbers agreed with it.  Neither is true.  The Septuagint reads 440, with better manuscript support;  and the explicit figures in Judges and Samuel sum to about 580, not 480, so the verse already requires interpretation.  The early date can be defended.  It cannot be defended by pretending the verse stands alone.

"The 480 is obviously twelve generations of forty, so it means about 300 years." Kitchen offers this, but he does not commit to it.  He calls it an interpretation that "should not be lightly dismissed," and on the next page offers a second, quite different derivation and calls the two "equally workable options."⁶  Anyone citing Kitchen as a witness for the 12 × 40 reading is overstating him.

"Egyptian chronology is guesswork, so the dates can move." Egyptian chronology has genuine softness, documented above, of roughly 20 years at the Exodus horizon and a century or two in the third millennium.  Revised chronologies need 250 to 600 years.  The Assyrian eponym canon, the Amarna synchronisms, the Merenptah inundation graffito, dendrochronology, and radiocarbon each independently forbid that.  And the revisionists themselves conceded the Assyrian framework is "as watertight as one could ask."

"There is no evidence for the Exodus at all." There is no direct evidence, which is a different claim.  Whether a fifteenth- or thirteenth-century exodus of the size the text describes should have left evidence, and where, is the subject of articles 2 and 7 of this series.  What can be said here is that a date must be fixed before an absence can be measured, and the two conservative dates point at different dirt.

"Egyptians never recorded defeats, so silence proves nothing." Mostly true, and Kitchen states it well.  But the sharper version is that Egyptian scribes reframed rather than omitted — Qadesh was published as a triumph.  So the expected trace of a national humiliation is not necessarily zero;  it is a victory notice about something else.  This makes the argument from silence weaker than skeptics claim and weaker than apologists claim.

Arguments that should be retired

The "rags and tatters" quotation, as normally used.  The line — "What is proudly advertised as Egyptian history is merely a collection of rags and tatters" — is quoted constantly as evidence that Egyptian chronology is worthless.  It is Alan Gardiner's, from Egypt of the Pharaohs p. 53, not from Centuries of Darkness, and it is not about chronology at all.⁶⁸  In context it is about the survival of source material:  the preceding sentences say only tiny remnants of a civilization thousands of years old have survived, and the following sentence gives the cause — nine-tenths of Egyptian excavation has been in desert cemeteries, so the finds are overwhelmingly funerary.  Answers in Genesis prints "As a result" in front of it after a paragraph about Sothic dates, which is Mackey's inference, not Gardiner's claim — and Mackey concedes two sentences later that Gardiner defended the Sothic anchor.⁵⁶  Gardiner should be quoted against the misuse, not for it.

Colin Renfrew's "ramshackle" as an endorsement of revision.  Renfrew's foreword to Centuries of Darkness does call the chronological structure "ramshackle" and says a chronological revolution is on its way.  The next sentence is the one that gets cut:  "Whether their proposed solution is the correct one we shall know only when tree-ring dates and radiocarbon sequences based on 'wiggle matching' become more widely available."⁹  He named the test in 1991.  Kuniholm's Anatolian tree-ring sequence and Manning's Uluburun work are that test, and both came back against the revision.  Renfrew endorsed the critique and suspended judgement on the replacement.

Acts 13:20 as a 450-year judges period.  The Byzantine text and the KJV attach "about 450 years" to the judges;  the critical text attaches it to what precedes — the wilderness and the distribution of the land.  Metzger's committee rates the variant {D}, its lowest confidence level, and explains the Textus Receptus reading as a scribal transposition.⁶⁹  Cite it only with the variant disclosed.

"Josephus dates the Exodus to X."  Cite a passage, never "Josephus."  He gives 592 years from Exodus to temple in one book, 612 in another, 430/215 for the sojourn in a third, and a Hyksos identification in a fourth that contradicts all of them.

"No drowned pharaoh's mummy has been found, so the Bible is confirmed."  This is used both ways and works neither way.  Exodus 14 does not say the pharaoh entered the sea.  Petrovich, an early-date advocate, concedes the point:  "no biblical text explicitly states that the exodus-pharaoh died there with his army."⁴²  The argument should be dropped by everyone.

Glueck's Transjordan survey as current evidence.  See Part 7.  It was decisive in its day and it has been overturned.

Verdict

The date of the Exodus cannot be settled on present evidence, and both conservative positions survive only by reading one of the two anchor texts non-literally.  Hoffmeier said it plainly and no one has improved on it:  "both texts cannot be treated literally."⁸

Of the two readings that keep the verse a real number, 480 wins — but not on the manuscripts.  440 has the better count, the older uncials, and probably the older Greek reading, and the leading defender of 480 concedes as much.  It loses on everything else:  it deepens an overshoot that is already the biblical text's hardest internal problem, it puts the judges fourteen years past Saul's accession, the Greek Judges figures run against it, and it strands the early date's flagship archaeology at Jericho by forty years.  In two thousand years nobody has published an exegetical reconciliation of the Judges data with a 1406 Exodus.  That is the decisive fact, and it is an argument from the Bible's own numbers rather than from a manuscript census.

The early date has the better internal case.  1 Kings 6:1 read plainly gives 1446, and the ordinal reading behind it is Ussher's, not a modern rescue.  Jephthah's 300 years points the same way independently.  The Levite line from Levi to Heman needs about 23 years per generation on the early date and an impossible 12 on the late one, and that argument deserves more attention than it gets.  The Jubilee convergence is real, if only as good as a Talmudic datum can be.

The late date has the better external case.  Pi-Ramesse existed only from about 1279, Tanis never bore the name, and Exodus 1:11's "Rameses" is embedded in a cluster of toponyms — Pithom, Migdol, yam suph, Pi-hahiroth, Baal-Zaphon — that first appear in 19th-Dynasty sources.  Merneptah puts Israel in Canaan by 1208.  Neither candidate site for Pithom has a fifteenth-century building phase.

Neither has produced what would settle it, and the honest report is that the archaeological silence in Sinai and at Kadesh-Barnea is unexplained on either date — and that Jericho, the site the early date has leaned on for thirty-five years, currently supports neither 1406 nor 1366 on the excavators' own radiocarbon.  Kitchen's papyrus-survival and never-record-defeats arguments are good ones for the Delta.  They do not explain a desert.

What is settled is narrower but not trivial.  Egyptian chronology cannot be moved by the centuries a revised scheme requires;  the Assyrian eponym canon, the Amarna synchronisms, the Merenptah graffito, dendrochronology, and radiocarbon each block it independently, and the revisionists conceded the first of those themselves.  Anyone building an apologetic on Velikovsky, Courville, or Rohl is building on ground that four separate lines of evidence have already tested and rejected.

What would change the verdict.  A fifteenth-century building phase at Tell er-Retaba or Tell el-Maskhuta would move the case sharply toward 1446.  A Late Bronze occupation layer at Tell el-Qudeirat would help both conservative dates and hurt the minimalists.  An Egyptian administrative text naming a Semitic labour group at Pi-Ramesse, or an unambiguous fifteenth-century destruction at Jericho on short-lived-sample radiocarbon, would each be decisive for one side.  A Dead Sea Scrolls witness at Exodus 12:40 would settle the sojourn question, and to this article's knowledge no one has checked whether one exists.

Sources

Bible quotations are from the NET Bible unless otherwise noted;  the Hebrew is the Masoretic text as published by Sefaria, the Greek Old Testament is quoted from the editions named in each note, and the Samaritan Pentateuch from Stefan Schorch's critical edition as digitised by the DT-UCPH text-fabric project.