Which Bible Text Gives the Right Date for Creation?

"According to the Septuagint, from Adam until the flood, there are 2,242 years… According to the Hebrew of the Jews, from Adam until the flood, there are 1,656 years… According to the Hebrew of the Samaritans, from Adam until the flood, there are 1,307 years." — Eusebius of Caesarea, Chronicon, c. AD 325¹

"The number of Genesis 5 or Genesis 11 begetting ages physically attested in ink in the entire Judaean Desert corpus is zero." — this article's own audit of every Dead Sea scroll of Genesis²

Summary

Open a Bible and add up the ages in Genesis 5 and 11 and you get a date for creation.  Ussher famously got 4004 BC.  What almost nobody is told is that the answer depends on which ancient copy of Genesis you add up, and the three surviving versions disagree by about 1,400 years.

The three are the Masoretic Text (the Hebrew Bible used for nearly every English translation), the Septuagint (the Greek translation made in Egypt in the third century BC), and the Samaritan Pentateuch (the Torah preserved by the Samaritan community).  The Dead Sea Scrolls are usually offered as the tiebreaker.  They are not, and this article will show exactly why.

The three dates

Anchoring Abraham's birth at 2166 BC and running each version's own arithmetic — using the Greek text that reads 187 for Methuselah, since the alternative reading of 167 moves the Septuagint's creation to 5540 BC:³

Move the anchor to Ussher's 1996 BC and the Masoretic row reproduces his 4004 / 2348 / 2247 exactly, which is the check that the arithmetic is right.⁴  The spread between the columns is 780 years at the Flood and about 1,386 years from Adam to Abraham.  The disputes after Abraham are much smaller — 215 years at Exodus 12:40 and 40 at 1 Kings 6:1 — and are handled in the article on the date of the Exodus.

One thing in that table is not like-for-like, and it is better said out loud than found by a critic.  The Masoretic and Septuagint rows put Abram's birth in Terah's 130th year, because those texts give Terah 205 years and Acts 7:4 then forces it.  The Samaritan row uses Terah's 70th year, because the Samaritan text gives Terah 145 and its own numbers only cohere that way.  That is each version being read on its own terms, which is the right thing to do — but it means the Samaritan column sits 60 years later than a like-for-like comparison would put it.  Measured purely as elapsed years from Adam to Abraham's birth, the Samaritan exceeds the Masoretic by 301 years, not by the 241 the BC columns show.  Part 2 gives both.

The four answers, up front

Which one gives the correct young Earth?  All three do.  The disagreement is entirely between about 4,200 and about 5,600 BC.  No manuscript of Genesis in any language supports an old Earth, and that is worth stating plainly before the argument starts.

Which one is original?  The honest verdict is that the Masoretic numbers have the better case, but not by the margin either side claims, and the strongest single argument on each side is one its own advocates rarely use.  A serious reader should hold the Masoretic figures with about 70% confidence, not 100%.

Do the Dead Sea Scrolls settle it?  No — and the reason is arithmetic, not interpretation.  Across every manuscript recovered from Qumran, Masada, Wadi Murabba'at, Wadi Sdeir and Nahal Hever, the surviving text of Genesis 5 is one fragment of one scroll carrying one word, the name "Kenan," with no legible numeral;  Genesis 7, 9 and 11 are missing from the biblical scrolls altogether, and Genesis 8 and 10 survive there only as questioned identifications carrying no numeral.⁵  Not one begetting age from Genesis 5 or 11 survives in ink anywhere in the Dead Sea Scrolls.  Anyone who tells you the scrolls confirm the Hebrew numbers — or the Greek ones — is asserting something the manuscripts cannot carry.

Which is reliable where?  This is the question that actually has a clean answer, and it is not "one text wins."  The Greek preserves a better text in Jeremiah and in Samuel;  the Hebrew preserves a better text in Esther, in Daniel, and against the Samaritan expansions throughout the Torah.  The map is in Part 8.

The finding that decides the shape of the argument

The popular versions of this debate all assume the split happened for a reason someone can name.  One side says the rabbis cut 1,400 years out of Genesis in the second century AD to undercut Christian claims about Jesus.  The other says Alexandrian Jews padded the numbers to make Israel older than Egypt.  Both stories are dead, and the same two facts kill them.

Demetrius the Chronographer, writing in Alexandria under Ptolemy IV (222–204 BC), already had the long chronology.  His surviving figures — 3,624 years from Adam to Jacob's descent into Egypt — sum forward from the Greek numbers exactly, and the match is so tight that it fingerprints the text he was reading:  it requires the extra Kainan, it requires Terah at 70, and it requires Methuselah at 187 rather than 167.^{6  7}

The Book of Jubilees, written in Hebrew about 150 BC, already had a short chronology.  It is usually cited as having the Samaritan figure exactly — the Flood at Anno Mundi 1307 — and that specific claim does not survive checking:  every manuscript reads a figure that yields 1062 at Jubilees 5:22 and 6:18's "till the days of Noah" implies about 1308 or 1309, and the 1307 is an editor's emendation that brings 5:22 into line with 6:18 and lands on the Samaritan figure by the editor's choice of two years.⁸  What survives is the part that matters here.  Either number is far shorter than the Greek, and both are in Hebrew, before Christ.

So a long chronology and a short chronology were both circulating among Jews before the Christian era, and the disputed extra patriarch with them.  No motive involving Christianity can explain a split that predates Christianity.  The question is thrown back onto the internal evidence, which is where it belongs.

What the young-earth side gets right

The systematic pattern in the Greek is real, and it is a scheme rather than an accident.  Compare the Greek Genesis 5 begetting ages against the Hebrew and one mechanical rule describes them: add 100 if the Hebrew age is below 160, otherwise leave it alone.  That rule is exact for seven of nine patriarchs, and the two misses are the two figures everyone already agrees are textually disturbed.⁹  Every one of those six raised figures is paired with a matching reduction of exactly 100 in the remaining years, so the stated total lifespan comes out unchanged.  Two numbers moved in step, six times, by the same round figure, with a checksum kept intact.  That is editing, not copying error.

The two "harmonising" witnesses harmonise differently, which is fatal to treating them as two votes.  Where the Samaritan and the Greek agree against the Hebrew they are usually repairing the same difficulty in visibly different ways.  Genesis 11 in the Hebrew has no total lifespans;  the Greek adds only "and he died" to each entry, while the Samaritan adds the full total-lifespan formula.¹⁰  Two separate completions of one perceived gap.  That is the signature of independent secondary repair, not of a shared original.

And the strongest support comes from a critical scholar with no stake in the outcome.  Ronald Hendel's Oxford critical edition of Genesis 1–11 reconstructs the Masoretic reading as original at every single figure in Genesis 11 except one, and writes that "M shows no signs of revision in this chapter."¹⁰

What does not survive

The Methuselah argument, in its popular form, has been abandoned by its own side.  The famous objection is that on the Greek numbers Methuselah outlives the Flood by fourteen years, which the flood narrative forbids.  The arithmetic is right.  But the current Masoretic-side paper of record — Answers Research Journal 18 (2025) — concedes in print that "Smith's paper gives a convincing argument for 187 being the original reading for Methuselah's begetting age," which means the impossibility is a corruption inside the Greek transmission and not a feature of the Greek chronology.¹¹  Codex Alexandrinus shows the correction physically:  the original scribe wrote 167, and a later hand scraped the parchment and wrote 187 over the erasure.¹²

The "they padded it to out-date Egypt" argument does not do what it claims.  No ancient source says it;  the idea is traced by its own leading advocate to a 1968 article.¹³  And it does not work on its own terms.  Larsson's Manetho puts the first "historical" pharaohs about 3,000 years before the third century BC — near 3250 BC — which the Greek Flood at 3298 BC barely clears rather than comfortably beats.  Against the figures the Alexandrians actually had to compete with, it does not compete at all:  the Old Egyptian Chronicle claims 36,525 years and Berossus 432,000.¹⁴  If the point of adding 100 years twelve times was to out-date Egypt, it clears Manetho's starting point by about fifty years and comes nowhere near Berossus.

The Cainan of Luke 3:36 is unresolved, and the standard answer to it is false.  The Greek Genesis 11 inserts a second Kainan;  Luke's genealogy has him;  the Masoretic Genesis does not.  The usual reply is that the earliest manuscript of Luke, P75, omits him.  That claim does not survive checking — Gordon Fee's 1966 dissertation says the omission is "not demonstrable from the extant text" — quoted by Smith, who also produces Papyrus 4, possibly older, with the name in ink.¹⁵  The most careful writer on the Masoretic side said so himself in 1998:  "the Cainan problem is still unsolved to my knowledge."¹⁶

The thing neither side wants

Run the standard text-critical method without a thumb on the scale and it does not return any of the three surviving chronologies.  It returns a shorter one.  Hendel's reconstructed archetype puts the Flood at Anno Mundi 1342 — 314 years shorter than the Masoretic Text.¹⁰  Remarkably, Cosner and Carter, arguing the Masoretic case from the opposite direction, reconstruct a "proto-Septuagint" that lands on 1342 as well.¹⁷  Cosner and Carter reach 1342 by the same method, not independently, so the agreement confirms the method rather than the number — and the article states the number rather than hiding it.  What that reconstruction costs is spelled out in Part 7.

What the choice buys, and what it does not

The longer Greek chronology is not just a curiosity.  It clears Egypt's entire dynastic record, which the Masoretic Flood date does not.  It moves the first writing from 850 years before the Flood to about 20 years before it — which is the difference between writing that survives a global deluge unexplained and writing that begins on its eve.  It widens the window for population growth before Babel from 101 years to 531.  A young-earth researcher writing in Answers in Genesis's own journal concluded in 2024 that Michael Oard's post-Flood Ice Age model "most likely is not compatible with the Masoretic Text" but works on the Greek numbers without alteration.¹⁸

Then there is Göbekli Tepe.  Charcoal and plaster from Enclosure D date to roughly 9,700–9,300 BC.¹⁹  That is nearly 7,000 years before either Flood, and the Greek chronology's extra 780 years covers about 11% of the gap.  Egypt is the argument everyone has;  the Neolithic is the one that actually decides it.  The full accounting is in Part 9.

Part 1 — What the four witnesses actually are

Before comparing what they say, it is worth being concrete about what these objects are.  Three of the four survive only through medieval copies.  The fourth is ancient and overwhelmingly fragmentary.

The Masoretic Text

"The Masoretic Text" is not one manuscript.  It is the medieval end-product of a school of Jewish scribes, the Masoretes, working mainly at Tiberias on the Sea of Galilee, who added three layers to an already-fixed consonantal Hebrew text:  vowel signs, chanting marks, and a marginal apparatus of counting-notes called the Masorah.²⁰

Two things about it matter here.

The Masorah is a checksum, not a commentary.  The short notes between the columns record how many times a form occurs in the whole Bible.  A single letter ל against a word means "there is nothing like the word so marked" — this form is unique.  A note of ב means the form occurs twice, and the longer notes at the head and foot of the page spell out where.²¹  The system exists to make copying errors detectable.  That is the strongest concrete fact in favour of Masoretic transmission fidelity, and it is a mechanism rather than a devotional claim.

And it worked.  The two rival tenth-century Masoretic schools, ben Asher and ben Naphtali, were catalogued in the eleventh century by Mishael ben Uzziel.  His book lists 867 disagreements between them across the whole Hebrew Bible — roughly 305,000 words — and they are disagreements about vowels and accents, not about wording.²¹  About one disagreement per 350 words, none of them a change of text.

The honest counterweight:  the oldest complete Hebrew Bible is about 1,000 years younger than the New Testament and about 1,400 years younger than the latest Old Testament book.  The Leningrad Codex, the base of the standard scholarly edition, is dated 1008/1009.  The Aleppo Codex is about 930 but lost most of the Torah in the 1947 riot.  A reader who wants the Hebrew text as the Masoretes fixed it is reading a medieval book.

The Septuagint — and always ask which Septuagint

The Greek Old Testament began as a translation of the Torah alone, made in Alexandria in the third century BC;  the rest followed over the next two centuries, by different translators of wildly different quality.  The traditional account, in the Letter of Aristeas, has seventy-two elders producing it under Ptolemy II.  Nearly all scholars treat that letter as legend written to give the translation prestige, and it covers only the five books of Moses in any case.

The point that matters for this article is duller and more important:  "the Septuagint reads X" is close to meaningless without naming an edition or a manuscript.  Genesis 5 and 11 are precisely where the Greek manuscripts split.

• Rahlfs' 1935 hand edition prints Methuselah's begetting age as 167 and yields a Flood at Anno Mundi 2242.
• Swete and Brooke–McLean print 187 — the Masoretic figure — and yield 2262.²²

Wevers' Göttingen critical Genesis of 1974 prints 167 as well, and keeps the second Kainan at Genesis 11:12–13, so the shorter reading is the one a modern critical text gives you.²³

The same warning applies at Genesis 11:24, where Rahlfs and Swete give Nahor 79 years and the Sixtine text behind Brenton's English gives 179 — a hundred-year difference on a figure the whole Demetrius argument depends on.

The commonly quoted "2242" is therefore a choice of edition, not a fact about the Greek.  This article gives both.

One further trap, because it is repeated constantly:  Codex Vaticanus is not a witness to Genesis 5 at all.  Its Old Testament has lost its first thirty-one leaves, and the original hand begins at Genesis 46:28.²⁴  Anyone who tells you what Vaticanus reads at Genesis 5:25 is mistaken.  The principal uncial for Greek Genesis 5 is Alexandrinus.

The Samaritan Pentateuch

The Samaritans preserved only the five books of Moses, in their own script, in their own community, continuously.  Their oldest actual manuscripts are medieval, and critical study of the text began only when Pietro della Valle brought the first copy to Europe in 1616.  It went into the Paris Polyglot within about fifteen years, and into Walton's London Polyglot of 1654–1657, and it was the Polyglots' parallel columns that made the comparison possible at all.²⁰

The single most important fact about it is a distinction that popular treatments miss.  The Samaritan Pentateuch has two layers, and they are of very different dates:

• An expansionist layer — long harmonising insertions that repeat Torah text from elsewhere in the Torah. This layer is old, and it is not Samaritan. Dead Sea scrolls such as 4QpaleoExodᵐ and 4QNumᵇ carry these same expansions without the sectarian readings.
• A sectarian layer — the Mount Gerizim material, including an extra commandment after Exodus 20:17. This layer is late.

The sectarian layer has an archaeological floor.  Excavations on Mount Gerizim under Yitzhak Magen from 1982 to 2006 found a sacred precinct in continuous use from early in the fifth century BC, about 96 by 98 metres, enlarged around 200 BC to 212 by 136 metres, and destroyed by John Hyrcanus I in 111–110 BC, after which the city was never resettled.²⁵  That dates the sanctuary, not the manuscript tradition.  What it gives is a bracket:  a Torah edited to point at Gerizim is possible from the fifth century onward, and the scholars who date the sectarian recension put it at or after the Hasmonean destruction.

So the Samaritan Pentateuch is not "a sectarian forgery."  It is an old expansionist Hebrew text-type with a later sectarian edition laid on top of it, and the two have to be judged separately.

The Dead Sea Scrolls

Roughly nine hundred manuscripts from eleven caves at Qumran, plus finds from Masada, Wadi Murabba'at, Wadi Sdeir and Nahal Hever, dating from about 250 BC to AD 68.  Twenty Qumran manuscripts of Genesis are catalogued in Ulrich's standard edition, plus four from the other sites.⁵

They are not a fourth text.  They are a snapshot of the period in which the other three separated — and, at Qumran, that snapshot shows a plural text.

How plural?  Of the 46 Qumran Torah manuscripts extensive enough to classify, Tov counts 22 (48%) Masoretic-like, 19 (41%) non-aligned, 3 (6.5%) pre-Samaritan and 2 (4.5%) close to the Septuagint's Hebrew source;  in the rest of the Hebrew Bible, 33 (44%) Masoretic-like, 40 (53%) non-aligned, 2 (3%) Septuagint-like.²⁶  Two things about those numbers matter more than the numbers.  They apply only to Qumran — at Masada, Wadi Murabbaʿat, Nahal Hever and Wadi Sdeir every biblical manuscript is Masoretic.  And they moved between Tov's own editions:  the 2001 edition gave a quite different split, and it contains a sentence about the pre-Samaritan share that cannot be read consistently at all.²⁰  Anyone who quotes a Qumran percentage without saying which edition it comes from is quoting a moving target.

What they are not is a source of chronology numbers.  Ulrich's published transcriptions run, for this stretch of Genesis:  1:1–4:11, then a two-line fragment labelled "Gen 5:13 or 14" with no legible numeral, then 6:13–21, then 12:4–5.⁵  Not one begetting age from either chapter survives anywhere.  Two qualifications belong with that, because a universal negative deserves them:  Ulrich's index does list Genesis 8:20–21 and 10:6, 20 as questioned identifications, neither carrying a numeral;  and his index covers biblical manuscripts only, so the Genesis Apocryphon — which preserves Genesis 11:10's "two years after the flood" in ink — falls outside it.  A date interval is not a begetting age, and the count of begetting ages remains zero.

How much do they actually differ?

Before listing disputes it is worth stating the base rate, because the disputes are a small fraction of the text.

Comparing the Masoretic Torah word by word against the Samaritan Torah as transcribed for Schorch's critical edition — Dublin Chester Beatty Library MS 751, with Garizim 1 for the last three chapters — across all 5,841 verses:  the two share 89.1% of their words letter for letter, and 93.5% once plene/defective spelling is set aside.  52.4% of all Pentateuch verses are word-for-word identical on the same allowance, and 31.7% are identical letter for letter.²⁷  The traditional figure of "about 6,000 differences" is a tally from the 1657 London Polyglot;  the count made for this article gives 6,953 strict difference sites, which is close enough to say the old number was honest.

Whatever else is true, most of the Torah is not in dispute.

Part 2 — The numbers themselves

What a chronogenealogy is

Genesis 5 and 11 are not bare name lists.  They use a repeated formula that ties a father's age to a son's birth:

"When Adam had lived 130 years, he fathered a son in his own likeness, after his image, and named him Seth.  The days of Adam after he fathered Seth were 800 years… Thus all the days that Adam lived were 930 years, and he died."  (Genesis 5:3–5)

Three numbers per man:  age at the son's birth, years lived afterwards, and — in Genesis 5 — a stated total.  That third number is a checksum.  It constrains what a later scribe can do, and it is exactly why the alterations in the Greek and Samaritan traditions took the shapes they did.

Genesis 11 in the Masoretic Text gives only the first two numbers.  It has no totals and no death notices.  Remember that;  it is load-bearing.

Genesis 5, three ways

Age at the son's birth / years lived after / total lifespan.  Every figure below was read out of a text of that version rather than copied from a chart:  the Hebrew from the Westminster Leningrad Codex and Sefaria's "Miqra according to the Masorah," the Greek from the CATSS morphological Septuagint with the readings checked against Brooke–McLean's apparatus, and the Samaritan from the Text-Fabric edition of Schorch's critical text.²⁸

All twenty-seven triples add up correctly in all three versions — begetting age plus remainder equals the stated total, without exception.²⁹  That is itself informative:  whoever changed these numbers was careful.

Read the columns rather than the rows and the pattern jumps out.  Six of the Greek's begetting ages are exactly 100 higher than the Hebrew's, and in every one of those six the remainder is exactly 100 lower, so the total is untouched.  Three are not raised by 100 — Jared, Methuselah and Lamech, the only three whose Hebrew begetting age is 160 or more.  Jared is untouched;  the other two do differ, by −20 and +6, and those are the two figures the next part shows are disturbed.  The Samaritan leaves six patriarchs identical to the Hebrew and rewrites the same three.

Genesis 11, three ways

Three features to notice, because each will matter later.

The Greek has an extra man.  A second Kainan sits between Arpachshad and Shelah, with 130 years to his begetting.  His numbers duplicate his son Shelah's exactly.  He is in Luke 3:36.  He is not in 1 Chronicles 1:18 or 1:24 in either Hebrew or Greek.

Only the Samaritan states totals here — and it is the only version that does.  The Greek adds "and he died" to each entry;  the Hebrew has neither.

Terah dies at 145 in the Samaritan and 205 in the other two.  That difference is not decoration.  Genesis 11:26 says Terah was 70 when he fathered Abram;  Genesis 12:4 says Abram was 75 when he left Haran;  Acts 7:4 says he left after Terah died.  On a lifespan of 205 those cannot all be true unless Abram was born in Terah's 130th year, not his 70th.  On the Samaritan's 145 they all cohere:  70 + 75 = 145 exactly.

That is either the original reading or an extremely tidy repair.  Hendel judges it a repair, and gives the reason:  raising an original 145 to 205 has "no textual or exegetical motive and would have created an obvious narrative problem," whereas lowering 205 to 145 removes one.¹⁰

The totals, computed

Setting Adam at year 0, the Flood in Noah's 600th year, Arpachshad born two years after the Flood (Genesis 11:10), and Abram born when Terah is 70:³⁰

The Greek exceeds the Hebrew by 1,366 years to Abraham's birth on the Rahlfs reading and 1,386 on the Alexandrinus one;  the Samaritan by 301 on either.

On the Terah question, state your assumption or your arithmetic is silently wrong by 60 years.  The table above uses Terah at 70, which is what the text of Genesis 11:26 says.  Readers who follow Ussher in taking Abram's birth at Terah's 130th year — required if Terah lived 205 years and Acts 7:4 is to hold — should add 60 to every figure in the last two rows.  Ussher's own 4004 BC assumes 130.  Two of the research passes behind this article diverged by exactly this amount before the discrepancy was traced;  it is an easy error and a common one.

One more thing, and the Masoretic side says it first

At Genesis 11:17 the Masoretic Text gives Eber 430 remaining years.  The Greek gives 370.  The Samaritan gives 270 — and the Samaritan's stated total of 404 requires 370, not 430.  Two independent witnesses require 370 — the Greek states it, and the Samaritan's stated total of 404 implies it — and the Masoretic Text is the odd one out.³¹

Every serious treatment, including the pro-Masoretic ones, concludes that the Masoretic figure here is a copying error — most likely contamination from Eber's own begetting age of 34 in the previous verse, followed by a transposition.³²

This costs the Masoretic chronology nothing, because the remaining-years figures never enter the sum from Adam to Abraham;  only the begetting ages do.  But it is worth stating loudly, because a reconstruction that finds errors in every witness including its own favourite is doing textual criticism, and one that finds errors only in the others is doing advocacy.

Part 3 — Before any of that: can the list carry gaps at all?

This question is logically prior.  If Genesis 5 and 11 skip generations, then no version's numbers give a date for creation, and the whole argument is about the length of a chain whose number of links is unknown.

Start with what is not in dispute.  James Barr, Regius Professor of Hebrew at Oxford and no friend of young-earth creationism, put it in print in 1987:  the Genesis figures "were meant to be exact and to be taken literally.  They do not mean anything at all unless they mean actual numbers of years."  He was explicit that this is a claim about what the writers intended, not about what happened — "The figures are not, to us, historically, scientifically or factually true, but they were literally intended" — and he called the modern habit of saying they were "not to be taken literally" a "cowardly expedient."³³  The same judgement circulates in creationist literature as a private letter of 1984 to David C. C. Watson.  The letter is probably genuine and is regularly quoted with its hedges stripped out;  the published lecture says the same thing and needs no defending.³⁴

The classic statement of the gaps position is William Henry Green's "Primeval Chronology" in Bibliotheca Sacra in 1890, and B. B. Warfield's essay of 1911.  Both are usually quoted second-hand.  Read directly, the position is weaker in one specific way than either its friends or its enemies notice.

The verb does not settle it

The no-gaps argument usually runs through the Hebrew verb:  ילד in the hiphil, "begot."  The claim is that it must mean immediate fatherhood.

That claim fails.  Scanning every occurrence of the hiphil of ילד across all thirty-nine books — 120 verses — the verb does occur in demonstrably telescoped lists, including Ruth 4:18–22, and its Greek equivalent ἐγέννησεν is the verb Matthew uses in a genealogy that admits to skipping generations.³⁵  "X begot Y" alone is compatible with Y being a remote descendant.

But the formula is not "X begot Y"

The Genesis 5 and 11 formula is longer:  "X lived N years and begot Y;  and X lived after he begot Y, M years."  The exact Hebrew string — the wayyiqtol of the verb immediately followed by the direct-object marker and a personal name — occurs sixteen times in the whole Hebrew Bible, and all sixteen are in Genesis 5 and 11.³⁶  There is no other passage in Scripture built this way.

The distinction that actually decides it

Here is the point neither popular side makes, and it is the strongest thing in the entire literature.  It comes from Jeremy Sexton in the Journal of the Evangelical Theological Society.³⁷

A genealogical gap and a chronological gap are not the same thing.  Suppose names are missing.  "When Enosh had lived ninety years, he begot Kenan" still dates Kenan's birth to Enosh's ninetieth year — because Kenan is the grammatical object of the verb.  The chronology survives the missing generations intact.

To get missing time, Green had to read the formula as "he begot [the ancestor of] Kenan," inserting an unnamed man as the real recipient of the verb.  He asserts this once, in passing, and never defends it.

So the honest framing is:  the formula does not forbid missing generations.  It forbids missing years, unless you insert a person the text does not name.

The strongest counterexample, conceded in full

Exodus 6:16–20 runs Levi → Kohath → Amram → Moses.  On the Masoretic Text's 430 years in Egypt, that is impossible.  The absolute ceiling from the text's own lifespans — Kohath 133, Amram 137, Moses 80 at the Exodus, each fathering his son in his final year — is 350 years, eighty short.³⁸

The short-sojourn reading of Exodus 12:40 (215 years, which the Greek and Samaritan support, and which Galatians 3:17 appears to assume) rescues the arithmetic.  But it does not rescue the census.  Numbers 3:27–28 counts 8,600 Kohathite males a month old and upward.  At the family sizes the Pentateuch itself reports — the eleven Levite families named in Exodus 6 and 18 average 2.64 sons — reaching 22,000 Levite males takes seven to ten generations from Levi.  The genealogy gives three begettings.  Three generations of multiplication at three sons each yields about 117 males alive;  at four sons each, about 336.  The census figure is short by a factor of sixty-five to nearly two hundred.³⁹

There is no way out of that except to say Exodus 6 omits names.

What this does and does not do to Genesis 5 and 11.  It does not touch them directly, and the no-gaps side is entitled to say so:  Exodus 6 gives total lifespans but no ages at begetting, so on the no-gaps side's own criterion it is not a chronogenealogy and is free to telescope.  The force of Exodus 6 is negative.  It destroys any claim that Pentateuchal genealogies are presumptively complete, and it shows the same author compressing a line without warning the reader.

The Cainan problem, checked rather than repeated

The Greek Genesis 11 has a second Kainan between Arpachshad and Shelah.  Luke 3:36 has him.  The Masoretic Genesis does not.

The standard reply is that P75, the earliest manuscript of Luke, omits him.  The published transcription of every witness extant at that verse was pulled from the INTF's own database.⁴⁰  Here is what is actually there:

• P75 does read Luke 3:36 without Kainan — but every letter of the relevant stretch is editorial restoration. The only ink is a handful of letters; "Sala" and the "of" before "Arphaxad" are entirely inside a hole. The omission is inferred from the width of the hole, which fits nine letters rather than twenty. That is a real argument, and a factor-of-two one, but it is an inference from a lacuna, not a reading.
• The papyrus cannot be re-read. The 2007 study of new P75 fragments notes of this passage "kein Photo" and explains that it "can no longer be read today, because it is stuck to folio 11ᵛ."⁴¹
• P4, of the same century and possibly older, reads Kainan with ΚΑ in clear ink.⁴⁰
• Codex Bezae omits Kainan too — and omits Noah as well. A genealogy that drops Noah is defective, not a witness to a shorter original.

So "the earliest known copy of Luke omits the extra Cainan" is not sustainable as stated.  There are two third-century copies of Luke 3:36:  one restores the text without the name, and one has the name in ink.

Three options remain, and an honest article lays them out rather than picking one silently.  Either Luke used a Greek Genesis containing a man who never existed;  or Kainan is real and the Masoretic Genesis 11 has a demonstrable gap at exactly the point where the chronogenealogy argument is most confident;  or Kainan entered Luke's text earlier than P4, Sinaiticus, Vaticanus, Alexandrinus and every minuscule.  None is comfortable.

Verdict on gaps

The comparison in this article is worth doing.  Genesis 5 and 11 are not bare name lists, the ages at begetting are there for some reason, and nobody has produced a parallel anywhere in ancient literature of a list that gives ages at begetting and is known to omit generations.

But the certainty that popular young-earth literature projects is not available.  The Exodus 6 census problem has to be conceded in full, and the P75 claim should be dropped.

Part 4 — The case that the Greek is original

State it at full strength.  The claim is that the Septuagint preserves the original numbers and the Masoretic Text was shortened.  Its leading modern advocate is Henry B. Smith Jr. of Associates for Biblical Research, and the case does not depend on him.

Demetrius the Chronographer, about 220 BC

This is the strongest single item, and the published treatments undersell it.

Demetrius was a Jewish writer in Alexandria under Ptolemy IV Philopator.  Nothing of his survives directly;  he reaches the modern reader through Alexander Polyhistor, quoted by Eusebius.  What survives includes three summary totals:

"from Adam until Joseph's brethren came into Egypt there were three thousand six hundred and twenty-four years;  and from the Deluge until Jacob's coming into Egypt one thousand three hundred and sixty years;  and from the choice of Abraham from among the Gentiles and his coming from Charran into Canaan until Jacob and his family came into Egypt two hundred and fifteen years."⁶

The usual argument subtracts:  3,624 − 1,360 = 2,264 for Adam to the Flood, against the Masoretic 1,656.  That is already decisive on length.  But sum the Greek numbers forward instead and they land on 3,624 exactly — and the exactness identifies the text.⁷  Getting 3,624 requires the extra Kainan (without him you get 3,494), requires Terah at 70 (with 130 you get 3,684), and requires Methuselah at 187 rather than 167 (with 167 you get 3,604).  The Masoretic numbers cannot reach past 2,298 for the first stretch under any reading.

This is not a new observation, and that matters:  it was worked out and published in 1875, by a critical German scholar with no dog in this fight.  Jacob Freudenthal concluded that Demetrius "already knew all the great changes which the Septuagint underwent in the fifth and eleventh chapters of Genesis — including the second Kainan," and that he read Methuselah at 187 and Nahor at 79.⁴²

The honest limits.  The fragments give only summary totals;  not one individual begetting age of Demetrius survives.  The transmitted 1,360 leaves the first stretch two years long against the Greek chain.  Freudenthal closes the gap without changing a letter:  Demetrius, like the Greek text itself, counts to Arpachshad's birth two years after the Flood, not to the Flood.  Others emend ,ατξ΄ to ,ατξβ΄.  Either way the three totals still cohere.  And Greek letter-numerals corrupt easily.  The defence is internal:  three independent figures that cohere with each other and with the Greek chain to the year do not normally arise by corruption.

The Samaritan agrees with the Greek in Genesis 11

Seven consecutive begetting ages are shared by the Samaritan and the Greek against the Hebrew — 135, 130, 134, 130, 132, 130, and the anomalous 79 for Nahor — in two traditions separated by language and geography.  A Hebrew text in Palestine and a Greek text in Egypt agreeing on seven numbers is two witnesses against one.

Eusebius made this argument in about AD 325 and it has not improved with age or gone away.

The shape of the post-Flood ages

In the Masoretic Genesis 11, Arpachshad through Nahor father sons at 29 to 35 years old while living 148 to 464 years.  Every other begetting age in the Bible sits at a much higher fraction of the lifespan.  The Greek's 130 to 135 fits the declining-longevity curve smoothly;  the Hebrew's does not.

This has been quantified, and in a journal friendly to the Masoretic side.  Fitting exponential decay curves to the lifespans, the Greek numbers fit better than the Hebrew:  correlation coefficients of 0.984 for Vaticanus and 0.976 for Alexandrinus against 0.940–0.946 for the Masoretic.⁴³

The measurement cuts both ways.  The two Greek curves fit better, and both of them run the human lifespan below zero if extrapolated, which the Masoretic and Samaritan curves do not.

Nobody outlives Abraham

The related point is sharper.  On the Masoretic numbers with Abram born in Terah's 130th year, Noah dies two years before Abraham is born, Shem dies 25 years before Abraham does — and Eber outlives Abraham by four years, dying at 464, while Abraham is buried "old and full of years" at 175.⁴⁴  On the Greek numbers the same men predecease Abraham by 957, 805, 740, 580 and 406 years.

Eusebius pressed exactly this in about AD 325:

"It even suggests that Noah and Abraham were alive at the same time — which is something that is not mentioned in any history."¹

Rabbinic tradition felt the same strain and resolved it by having Isaac study Torah with Shem and Jacob with Eber.⁴⁵  Whether that overlap is a problem or a feature is a judgement call.  It is not nothing.

Where this case is weaker than advertised

The tampering charge has no ancient support in the form it is usually given.  The earliest Christian writer to accuse the Jews of shortening the numbers is Eusebius, and the motive he assigns is not anti-Christian polemic:

"Therefore we suspect that this was something which the Jews did:  that they ventured to compress and shorten the time before these sons were born, in order to encourage early marriages.  For if these most ancient of men, who lived such long lives, came quite soon to marriage and fatherhood, as their account declares, who would not want to imitate them and marry early?"¹

No rabbinic text admits any such change.  Every ancient source for the charge is Christian.  The argument is an inference to the best explanation, and the article labels it one.

And the leading critical scholar is not an ally.  Ronald Hendel's reconstruction of the original Genesis puts the Flood at Anno Mundi 1342 — shorter than the Masoretic Text, and 900 years shorter than the Greek.¹⁰  Four of his sub-rulings help the Greek side and those are the ones to cite;  his overall verdict does not.

Part 5 — The case that the Hebrew is original

Two traditions repairing the same problems in different ways

This is the strongest argument on either side, and it is structural rather than arithmetical.

Two witnesses agreeing against a third are worth two votes only if they are independent.  Where the Samaritan and the Greek disagree about how to fix the same difficulty, they are not two attestations of an older reading.  They are two separate repairs — and the thing being repaired is the Masoretic text.

Four cases, all read out of the primary texts:

The missing formula in Genesis 11.  The Hebrew entries end "and he fathered sons and daughters" — no total, no death notice.  Genesis 5's entries have both.  A scribe trained on Genesis 5 sees Genesis 11 as defective.  The Greek adds only "and he died."  The Samaritan adds the full Genesis 5 formula, total lifespan and death notice together.¹⁰  Two different completions of one perceived gap.  Neither is likely to be an original that the other abbreviated, because the Hebrew has neither element — which is precisely the reading that explains both.  And this is not only an inference from the printed texts.  Origen's Hexapla marked "and he died" with an obelus at every entry in Genesis 11 — the sign for material the Greek carries and the Hebrew lacks — and marked the second Kainan the same way.⁴⁶  The Hebrew of Genesis 11 was already short of those death notices in the third century.

The three antediluvians who should not have survived the Flood.  In the Samaritan, Jared, Methuselah and Lamech all die in exactly Anno Mundi 1307 — the Flood year, to the year, all three.  In the Greek, Jared and Lamech die before it and Methuselah survives it by fourteen years.  In the Hebrew, Methuselah dies in the Flood year and Lamech five years before.  Three different outcomes from three different sets of numbers.  Three simultaneous deaths in one year is not a chronology anybody wrote;  it is a chronology somebody adjusted.

Terah's lifespan.  The Hebrew and Greek both give 205.  The Samaritan gives 145, which makes Genesis 11:26, Genesis 11:32, Genesis 12:4 and Acts 7:4 all cohere.  A chronological number altered for a harmonising reason, in a manuscript tradition, demonstrably.

Exodus 12:40, the same signature outside Genesis.  The Hebrew says Israel was in Egypt 430 years.  The Greek adds "and in the land of Canaan."  The Samaritan adds "and their fathers" and puts Canaan before Egypt.⁴⁷  Same difficulty, two different fixes, neither matching the other.

The +100 is a rule, not noise

Compare the Greek's Genesis 5 begetting ages against the Hebrew's and one mechanical rule describes them:  add 100 if the Hebrew age is below 160;  otherwise leave it alone.

Seven of nine, and the two misses are the two figures everyone already agrees are disturbed.  The threshold of 160 is not chosen to fit — the raised set {130, 105, 90, 70, 65, 65} and the untouched set {162, 187, 182} do not overlap at all.⁹

Genesis 11 shows the same scheme with a visible constraint.  Every begetting age from Arpachshad to Serug rises by exactly 100 in both the Samaritan and the Greek — except Nahor, who rises by 50 in both.  Why 50?  Because Nahor is the short-lived one.  As a fraction of the lifespan the Samaritan records for him, the inflated figures land in a consistent band of 30% to 56%.  A +100 would put Nahor at 87% of his 148-year life, far outside it;  a +50 puts him at 53%, inside it.⁴⁸  The editor was working from a text in which Nahor lived 148 years — which is the Masoretic figure.

The Samaritan's totals are arithmetic done on a Hebrew text

The Samaritan is the only witness that states total lifespans in Genesis 11.  Compare each stated Samaritan total against the total implied by the Masoretic Text's two numbers:

Seven of eight exact — and the eighth matches too if the Masoretic remainder for Eber was originally 370, which is what everyone independently concludes it was.⁴⁹

The sequence this implies is concrete.  A scribe took a text with the Masoretic numbers, computed the totals, wrote them in to make Genesis 11 look like Genesis 5, and then inflated the begetting ages — at which point the totals he had just written forced him to cut every remainder by the same amount.  That is exactly what the Samaritan numbers do.  The Greek, having written no totals, was under no such constraint, and its Genesis 11 inflation is correspondingly careless:  Peleg 130/209, Reu 132/207, Serug 130/200, with the begetting age raised and the remainder left alone, so the implied lifespan simply grows by 100 years.

Be careful what this proves.  A total is invariant under a +100/−100 swap in either direction, so the totals alone do not fix which way the change ran.  What they establish is shared ancestry with a Masoretic-type text.  Direction is fixed by the next argument.

There is a motive to lengthen and none to shorten

Adding 100 years to each begetting age from Arpachshad to Serug pushes Abraham's birth 650 years later and dissolves the awkward overlap in which Shem outlives Abraham.  Subtracting 100 from each creates that overlap.

Hendel states the asymmetry outright:

"Conversely, there is no identifiable motive for scribes in the proto-M tradition to reduce the numbers for b in this series, which would create the problem of contemporaneous generations at the time of Abraham."¹⁰

His reconstructed archetype of Genesis 11:10–32 is then the Masoretic reading at every single figure except Eber's remainder, with no Kainan — and he concludes:  "M shows no signs of revision in this chapter."¹⁰

That is the strongest single sentence available to the Masoretic case, and it comes from the Oxford critical edition of Genesis 1–11, not from a creationist journal.

Origen's Hexapla shows the split already open in the third century

This is hard external evidence and it is almost never used.

Three Jewish revisions of the Greek Old Testament — Aquila, Symmachus and Theodotion — were made independently from the Hebrew between roughly AD 130 and 200.  Their readings survive in the margins of Hexaplaric manuscripts.  Brooke and McLean print those found in the manuscripts they collate;  Field's Hexapla collects them from a wider field, and the two agree at every point that matters here.  In Genesis 5 the three give the Masoretic numbers at twelve points:  130 for Adam against the Greek's 230, 800 against 700, 105 against 205, 807 against 707, 90 against 190, 815 against 715, 840 against 740, 65 for Mahalalel against 165, 830 against 730, 65 for Enoch against 165, 300 against 200, and 187 for Methuselah against 167.  Ten of those twelve are single, clean readings.  Two are not, and saying "throughout" would overstate it:  at Genesis 5:6 one manuscript gives 135 where another gives 105, and at 5:9 one gives 95 where another gives 90.  Field records the same two splits from a different pair of manuscripts, and prefers the Masoretic figure in both.¹²

The cleanest line is at Genesis 5:26, where Brooke–McLean record that "the Three" read like the printed lemma — 782 — while "οἱ ο΄", the Seventy, read 802.¹²

The note is a Hexaplaric one, which Field records from the margin of the Syro-hexaplar.  It says that the Seventy read 802 and the Three read 782, and its own running text reads 782.  Whichever column carried which figure, the two chronologies were being compared side by side in the third century.  The split between the two chronologies is not a medieval development.  It is on the page in the 200s, with the Hebrew side reading what the Leningrad Codex reads seven hundred years later.

Report the exception:  at Genesis 5:19–20 the same marginalia give Jared's figures as 785 and 847 — the Samaritan numbers, not the Masoretic.  Only the 847 carries the "three translators" sigil in Brooke–McLean;  the 785 stands there as a bare variant, and Field notes that the manuscript gives it sine titulo.  Field nonetheless assigns it to the same translators, printing "(Οἱ λοιποί) ψπε."  So the exception may be worse than the apparatus alone shows.  The evidence is overwhelmingly Masoretic but not uniformly so, and Jared is exactly where all three traditions part company.

Two independent translations carry the Masoretic numbers

The Hexaplaric marginalia show a Hebrew text with the Masoretic numbers in the second century.  Two whole translations show the same thing, and they are independent of each other.

The Latin Vulgate.  Jerome translated Genesis from the Hebrew around AD 400, in deliberate preference to the Greek.  All nineteen begetting ages of Genesis 5 and 11, Noah's 500 included, are the Masoretic figures — including all four of the places where the Samaritan and the Greek agree against the Hebrew.  Methuselah is 187, Lamech 182, Arphaxad 35, Eber 34, Nahor 29.⁵⁰

And he says why, in a commentary written a decade earlier.  At Genesis 5:3 he tells the reader that all the way down to the Flood, "where in our codices someone is said to have begotten at two hundred and more years, in the Hebrew it has one hundred years and the remainder that follows," and that the Greek then "put seven hundred here, because it had erred in the two hundred."  He treats the whole +100/−100 pattern as one mistake with one correction, and he calls it a mistake.  On Methuselah he is blunter still:  the fourteen years past the Flood are "a famous question, aired in the debates of all the churches," and his answer is that "as in very many places, so here too, there is an error in the number."⁵¹

The Syriac Peshitta.  A separate translation, made in a different language, in a different church, from a Hebrew text — and not descended from the Vulgate.  Seventeen of its eighteen begetting ages from Adam to Lamech and from Shem to Terah are Masoretic too;  the exception is Terah, and it is treated below.⁵²

Two exceptions are worth reporting rather than smoothing over, and one of them does touch the chronology.  The Vulgate gives Arphaxad's remaining years as 303 where the Masoretic Text has 403 — which happens to be the Samaritan figure, so the Vulgate is not a pure Masoretic witness even here.  And the Peshitta gives Terah 75 years at Genesis 11:26 where all three Hebrew and Greek witnesses read 70.  That one is a chronological number, and it adds five years to any total computed from the Syriac.

What this is worth.  It does not prove the Masoretic numbers original — a translation can only show what its exemplar read.  What it shows is that by about AD 400 the Masoretic figures were the reading of the Hebrew text in Palestine as Jerome found it, and independently the reading behind the Syriac.  Together with Aquila, Symmachus and Theodotion in the second century, that is four independent lines reaching the Masoretic numbers five to six hundred years before the earliest surviving Hebrew manuscript of Genesis.  The gap between the oldest complete Hebrew Bible and the text it preserves is real;  this is how it gets bridged.

A ±100 change is a word, not a digit, in every surviving manuscript

In all three texts the numbers are spelled out as words.

• Masoretic, Genesis 11:14: ושלח חי שלשים שנה — "and Shelah lived thirty years."
• Samaritan, Genesis 11:14: ויחי שלח שלשים שנה ומאת שנה — "and Shelah lived thirty years and a hundred years."

The difference is the insertion of two whole words.  Six such changes run through Samaritan Genesis 11 — three of them inserting one word, three inserting two — plus a word substitution at Nahor, whose change is 50 rather than 100.  The Greek's six Genesis 5 changes are substitutions of the same kind, two hundred for hundred.  Twelve word-level edits, each paired with a matching subtraction elsewhere in the same entry, is not a class of error a copyist commits.  Whoever says the Masoretic Text lost 100 years from each of thirteen entries is alleging deliberate editing and carries the burden that goes with it.

State the limit honestly.  This establishes what the surviving manuscripts do, not what a pre-exilic original did.  Hebrew scribes had numerals and used them.  Yohanan Aharoni showed in 1966 that the marks on the Judaean shekel weights are the Egyptian hieratic numerals for 5, 10, 20 and 30, and that hieratic numerals appear in the ostraca from Samaria, Arad, Lachish, Wadi Murabbaʿat and Mesad Hashavyahu.  His conclusion:  "hieratic numerals were used in the royal administration both in Israel and in Judah."⁵³  Hieratic gives each hundred its own sign, so in a document numbered that way a ±100 change really would be one stroke of a pen.

Two things keep that from settling the question.  The hieratic evidence is administrative — weights, tax dockets, delivery notes, day-of-the-month dates — not literary.  And where scribes of biblical books can be observed directly, they do not use numerals at all.  Tov's catalogue of the Judean Desert manuscripts that write numbers with signs runs to twenty-five items — the Copper Scroll, calendars, deeds, accounts, the New Jerusalem text — and not one of them is a biblical scroll.  The same text copied by different hands shows the split:  4QNJᵃ and 4QNJᵇ use number signs above ten, while 2QNJ, 5QNJ and 11QNJ "write the numbers in full."⁵⁴  Whether some lost copy of Genesis carried hieratic numerals is unknown, and this article does not pretend otherwise.  What can be said is that no manuscript of Genesis in any language, at any date, writes these figures as anything but words.

Part 6 — What the ancient chronographers actually used

If a chronology is a late invention, the writers closest to its supposed invention should show it.  They do not.  Here is who computed what, from which text.⁵⁵

Josephus is the noisiest witness in the table, and the noise is worth seeing.  His text as Niese prints it gives 2262 years from Adam to the Flood.  His manuscripts do not agree:  three of them and the Old Latin give 2656, the Epitome of the Antiquities and Zonaras' manuscripts carry the Masoretic 1656 — the Epitome as printed reads 1056, which Niese takes as a corruption of 1656 — and the printed text of Zonaras, who worked from the Epitome, gives the Septuagint's 2242.⁵⁶  He also tells his reader how to use the list:  "Let no one examine the deaths of these men … but let him look only at their births."⁵⁶  The chronology is built from the begetting ages, not the lifespans, and Josephus says so.

One example shows how these numbers were sometimes reached.  Hippolytus, writing in the early third century, proves the Incarnation fell in Anno Mundi 5500 from the dimensions of the ark of the covenant:  two and a half cubits long by one and a half wide by one and a half high adds to five and a half, "so that the five thousand five hundred years might be shown."⁵⁷  A chronological figure derived from the measurements of a piece of furniture is not a computation, and the early chronographers' clustering around 5500 is worth less than it looks.

Eusebius wrote this article in about AD 325

Book 1 of the Chronicon lays the three versions side by side — the earliest surviving three-way comparison.  The complete book survives only in an Armenian translation, but the Greek of these very passages survives as excerpts in George Syncellus, and Schoene prints it facing the Armenian:

"According to the Septuagint, from Adam until the flood, there are 2,242 years.  From the flood until the first year of Abraham, 942 years.  In total, 3,184 years.

According to the Hebrew of the Jews, from Adam until the flood, there are 1,656 years.  From the flood until the first year of Abraham, 292 years.  In total, 1,948 years.

According to the Hebrew of the Samaritans, from Adam until the flood, there are 1,307 years.  From the flood until the first year of Abraham, 942 years.  In total, 2,249 years."¹

Every one of those six figures reproduces exactly from the version texts.⁵⁸  Note that his Greek Genesis 11 has no second Kainan — 942 is only obtainable without him — and that his Samaritan and Greek post-Flood totals are identical.

His verdict:

"Therefore, because it has been definitely established that the Septuagint version was translated from the original, unadulterated Hebrew scriptures, it is reasonable for us to use that version for this chronicle."¹

Augustine, a century later, goes the other way

Augustine treats the Methuselah problem at length in City of God 15.10–14, notes that the alteration "does not smell of accident, but of design," and reports a manuscript split in which three Greek codices, one Latin and one Syriac make Methuselah die six years before the Flood.⁵⁹  That is an independent confirmation of the arithmetic:  substitute 187 for 167 in the Greek Genesis 5 and Methuselah dies six years before a Flood at 2262, exactly as he says.

He then states the rule that eventually won in the West:

"when something different is found in the two sets of manuscripts… we do well to believe in preference that language out of which the translation was made by translators into another."⁵⁹

And — this is worth noticing — he names an alternative rather than merely doubting.  The divergence, he says, is to be blamed "neither on the malice of the Jews nor on the diligence or wisdom of the seventy translators, but on the error of the copyist who first received the manuscript from that king's library to be transcribed."⁵⁹

Two warnings for anyone quoting these passages.  Marcus Dods's widely-reprinted English mistranslates a numeral at 15.10:  where the Latin reads "non centum minus, sed uiginti amplius" — twenty more — Dods prints "not 100 years less, but 100 years more."^{60  59}  And Codex Vaticanus cannot be cited here at all, since its Genesis begins at 46:28.

Choosing the Hebrew numbers was not a safe choice.  In 703 Bede published De temporibus, laying out the years "juxta Hebraicam veritatem" — by the Hebrew truth — which put the Incarnation short of Anno Mundi 5000 instead of at the traditional 5199.  Five years later he learned he was being sung about as a heretic over drink by "rowdy rustics," on the charge that he denied Christ had come in the sixth age of the world.  He wrote back to the monk Plegwin, naming the cause:  the chronology in his book, "where it is far shorter than in the Seventy translators, so that up to the coming of the Saviour in the flesh not even five thousand years are complete."⁶¹  Preferring the Hebrew text in the eighth century cost something.

Ussher, Lightfoot, and an hour that nobody wrote

Ussher's Annals opens with the creation date and contains no hour at all, in either the 1650 Latin or the 1658 English.⁶²  The familiar "9 a.m. on 23 October 4004 BC" is a conflation of two different chronologists, and it is provable from the pages its own source cites:  Andrew Dickson White footnotes Lightfoot's Works, and those pages say creation was in Tishri, "about our September," and place 1644 at Anno Mundi 5573 — which makes Lightfoot's epoch September 3929 BC, not 4004.⁶³  The date is Ussher's;  the hour belongs to neither.

This is a small thing, but it is the kind of small thing that a critic uses to discredit a whole position, and it is cheap to get right.

Part 7 — The verdict on the chronology

Both chronologies are older than the argument about them

The two motive stories that dominate popular treatments both require the split to have happened at a particular moment for a particular reason.  Neither can survive the dates.

Demetrius, about 220 BC, has the long chronology — with the extra Kainan, with Terah at 70, with Methuselah at 187.  That is three centuries before the oldest surviving Greek manuscript and four centuries before the earliest Christian chronographer.

Jubilees, about 150 BC and written in Hebrew, is the usual second half of this argument — its Flood is placed at Anno Mundi 1307/1308, the Samaritan figure exactly, and fragments of Jubilees were found at Qumran.  But that figure will not bear the weight, and the article has to say so.  Every manuscript of Jubilees reads "twenty-second jubilee" at 5:22, which gives Anno Mundi 1062.  But Jubilees 6:18, which nobody emends, says the feast of weeks was kept in heaven "from the day of creation till the days of Noah — twenty-six jubilees and five weeks of years," Anno Mundi 1309, and that Noah then kept it until his death — the 350 years Genesis gives him after the Flood.  So the book's own arithmetic puts the Flood at about 1308 or 1309 and contradicts its 5:22, and Charles emends 5:22 to bring it into line.  What the Samaritan Chronicle supplies is the last two years, which is what turns 1309 into 1307;  so the exact match with the Samaritan figure is Charles's choice, not the text's.  Two of the most diagnostic birth-years, Lamech's and Noah's, are not in the text at all and were supplied by Charles from the Samaritan Chronicle.⁸  So Jubilees is evidence that a chronology shorter than the Greek was in Hebrew circulation before Christ, which is real and worth having.  It is not evidence that the Samaritan figures specifically were.

And the Hexapla shows the Greek and Hebrew columns already split in the third century AD, with Aquila, Symmachus and Theodotion reading the Masoretic numbers from the Hebrew.

So a long chronology and a short chronology were both in Jewish circulation before Christ, and the extra Kainan with them.  "The rabbis cut it to spite Christians" is dead.  "The Christians padded it to reach Anno Mundi 5500" is dead.  Whatever happened, happened earlier, and for reasons no surviving source states.

The third possibility, which nobody wants

Run standard text-critical method without a preferred outcome and it does not return any of the three surviving chronologies.

Hendel's reconstructed archetype takes Jared's and Methuselah's begetting ages from the Samaritan, conjectures Lamech's at 88 — a figure in no manuscript, which he marks with an asterisk — and lands the Flood at Anno Mundi 1342, 314 years shorter than the Masoretic Text and 900 shorter than the Greek.¹⁰

That conjecture is load-bearing, and it is worth seeing how load-bearing.  Take Lamech's attested Greek reading of 188 instead and the archetype gives 1442, not 1342.  In fact 1342 requires the nine Genesis 5 begetting ages to sum to 742, and the lowest sum obtainable from any combination of readings actually attested in the three versions is 707, with every available increment at least 100.  Anno Mundi 1342 is not reachable from any manuscript reading at all.⁶⁴  It exists only because Hendel supplied a number for Lamech that no copy of Genesis has ever contained.

Cosner and Carter print the same figure, and it is sometimes offered as independent confirmation.  It is not.  Their Genesis 5 table gives Hendel's values figure for figure, conjectured 88 and 665 included, because both reconstructions apply the same rule from the same source:  strip the systematic Greek inflation, which turns Lamech's 188 into 88 and his 565 into 665.  Cosner and Carter cite Klein for it, and Hendel credits Klein for it too.¹⁷  Two reconstructions running one method are one result, not two, and this article said otherwise in an earlier draft.

But it comes at a price that has to be named.  Hendel's archetype requires the original Genesis to have contained a chronology that contradicts its own flood narrative — Jared, Methuselah and Lamech all outliving the Flood.  He supplies a reason:  a redactor incorporated a separate document, the "Book of the Generations of Adam," into Genesis "without harmonizing the chronological data of the two documents."¹⁰  That is a source-critical hypothesis about how Genesis was composed, not a text-critical inference from manuscripts.  A reader who does not grant it has no reason to prefer an archetype that fails arithmetic over one that does not, and Cosner and Carter put the objection in the right form:  his reconstruction "requires all three extant texts to be recensions of a flawed original."¹⁷

Both things are true at once.  The reconstruction is the most rigorous one available, and its premise is one most readers of this site will reject.  It belongs in the article either way.

Where this article comes down

On the balance of evidence, the Masoretic numbers are more likely original than the Greek ones — but hold that at about 70%, not at 100%.

What carries the verdict is not any single knockout.  It is the shape of the evidence:

1. Two of the three witnesses show a scheme — round hundreds, applied by rule, checksums maintained where a checksum existed and abandoned where it did not.
2. Those two witnesses repair the same difficulties in visibly different ways, four separate times, which is what independent secondary correction looks like.
3. There is a motive to lengthen the post-Flood chronology and none to shorten it, and the critical edition of Genesis 1–11 says so in as many words.
4. Four independent witnesses reach the Masoretic numbers centuries before the oldest Hebrew manuscript: Aquila, Symmachus and Theodotion in the second century, and the Latin and Syriac translations by about AD 400.

What keeps it at 70% rather than higher:

1. Demetrius has the long chronology in 220 BC, and nobody has explained him away.
2. The Samaritan and Greek agree on seven consecutive Genesis 11 begetting ages in traditions separated by language and geography.
3. The Masoretic post-Flood begetting ages of 29 to 35 are genuinely anomalous, and the Greek fits the longevity curve better on a published measurement.
4. The Cainan of Luke 3:36 is unresolved, and the standard reply to it is false.
5. No manuscript evidence exists at all. Not one Genesis 5 or 11 number survives from antiquity in any Hebrew manuscript older than the Middle Ages.

Anyone who claims more certainty than that in either direction is claiming more than the evidence holds.

And the range is bounded

This matters more than which end of it is right.  The Masoretic Text is the shortest chronology anyone defends and the Septuagint the longest.  As one young-earth researcher puts it, the two "pose the youngest (MT) and the oldest (LXX) earth respectively that the Bible allows without modification."¹⁸

So the maximum available relief on any biblical reading is 780 years at the Flood and about 1,400 years at creation.  That is the whole size of the prize, and Part 9 measures it against what it would have to buy.

Part 8 — Reliable where: the book-by-book map

The chronology question is narrow.  The broader one — which text should I trust? — has a better answer, and it is not "one of them."

Where the Greek preserves a better text

Jeremiah is the strongest case in the Bible.  The Greek Jeremiah is about a seventh shorter than the Hebrew and puts the Oracles Against the Nations in a different place.  Before 1947 one could say the translator abridged.  Then Cave 4 produced Hebrew copies of both editions:  4QJerᵇ and 4QJerᵈ reflect the short text and the different arrangement;  4QJerᵃ and 4QJerᶜ reflect the long one.²⁰

It was measured independently here rather than taken on report.  Divide each book's Greek word count by its Hebrew word count — translation inflates length fairly consistently — and Jeremiah comes out at 1.32 against a median of 1.584 across 32 books, which is 3.8 standard deviations below the median and the lowest ratio in the Bible.⁶⁵  The method validates itself:  the second and third lowest are Joshua and Exodus, the other two books where a shorter Greek edition is the standard finding, and it finds them without being told to.

Two honest caveats.  The familiar "2,700 words, about one-eighth" figure is from 1889, sixty years before the scrolls, and the count behind it was never published.⁶⁶  Tov himself gives one-seventh in one place and one-sixth in another.⁶⁷  The estimate made for this article is 16.9% ± 4.  And the order is less settled than the length — Tov's own footnote names two scholars who argue the Greek arrangement is secondary.

1–2 Samuel.  The scroll 4QSamᵃ agrees with the Greek against the Hebrew repeatedly, and it contains a whole paragraph before 1 Samuel 11:1 that the Masoretic Text lacks, explaining why Nahash the Ammonite demanded that every man of Jabesh lose his right eye.  Josephus knew the paragraph.  The NRSV prints it.  Tov's decisive reason for accepting it is worth quoting, because it becomes the criterion for everything else:  the plus "is not based on any other passage in the Bible, and has no defined purpose, so that its originality seems plausible."²⁰

The Masoretic Samuel is also defective in a way anyone can see:  1 Samuel 13:1 literally reads "Saul was a year old when he became king, and he reigned two years."  The number slots are empty.  No witness supplies them.

Isaiah 53:11, where four pre-Christian witnesses — including three Qumran Isaiah manuscripts — read "he shall see light," which the Masoretic Text lacks.⁶⁸

Joshua and Ezekiel, both about 4–5% shorter in Greek on Tov's count, with the Hebrew pluses forming a recognisable literary layer rather than scattered glosses.²⁰

Where the Hebrew preserves a better text

Against the Samaritan expansions, throughout the Torah — and this is where the symmetry pays off.  Take the 44 places where the Samaritan text has ten or more extra words and probe the extra material against the whole Masoretic Pentateuch.  36 of the 44 repeat wording found verbatim in the Torah, and 34 if the plus's own verse is not allowed to count as the parallel.⁶⁹  Six of the misses are the Genesis 11 total-lifespan formulae, which are modelled on Genesis 5 rather than copied from it.

That is the exact opposite of the Qumran Samuel plus.  The criterion that vindicates 4QSamᵃ condemns the Samaritan expansions, and it is the same criterion applied the same way.  No special pleading in either direction.

Esther.  The Greek Esther has 5,843 words against the Hebrew's 3,057 — 91% more, with the lettered Additions alone making up 35.7% of the Greek book.  A control worth running:  the Hebrew Esther contains no divine name at all, while Ruth — run as a check that the counting code worked — returns eighteen.⁷⁰  The Greek additions supply the God that the Hebrew book conspicuously withholds.  That is an addition, not a loss.

Daniel.  Theodotion's Greek adds 1,096 words at chapter 3:24–90 alone — the Prayer of Azariah and the Song of the Three.⁷¹

What the New Testament actually does

This is usually asserted and rarely counted.  The only tabulation that answers the question in the form it is asked is D. M. Turpie's, published in 1868 and never superseded in design.  He sorted 275 New Testament citations into five mutually exclusive boxes:⁷²

Only 47 of 275 citations are evidential at all — and there the Greek wins about 3.7 to 1.

Two things follow.  The apostles clearly used a Greek Bible freely, which is awkward for any strict Hebrew-only position.  And the figure most often quoted for this is smaller than people think:  83% of the citations cannot tell you anything either way.  A widely used seminary textbook states that "about a third" of New Testament citations agree with the Greek against the Hebrew;  against Turpie's tabulation that is wrong by a factor of two to two and a half.⁷³

The map

The general rule that falls out:  the Masoretic Text is the best base text, and the other witnesses are the best check on it.  Treating any one of them as automatically right is the mistake — and, as Part 2 showed, the Masoretic Text does carry at least one demonstrable error in the very chapter under dispute.

Part 9 — What the choice costs downstream

If the difference were only 1,400 years on a chart it would not matter much.  It matters because the two dates land in different parts of the archaeological record.

What the longer chronology genuinely buys

Egypt.  This is the largest clean gain.  Egypt's dynastic sequence is now dated independently by radiocarbon:  a Bayesian model on 186 dates puts the accession of Aha, the foundation of the Egyptian state, at 3111–3045 BC;  a 2025 study on 48 new high-precision dates puts the start of the Old Kingdom at 2679–2642 BC and the start of the First Intermediate Period at 2267–2228 BC.^{74  75}  The Masoretic Flood — 2518 BC on the anchor used here, 2348 BC on Ussher's — lands inside the pyramid age.  The Greek Flood at 3298 BC clears the entire dynastic record.

Writing.  The earliest Egyptian writing, from Tomb U-j at Abydos, dates to about 3320 BC;  proto-cuneiform at Uruk runs 3350–3100 BC.⁷⁶  On the Masoretic chronology, writing predates the Flood by 800–850 years in both civilisations, and the writing on either side of the Flood is continuous.  On the Greek chronology the Flood lands about twenty years after Tomb U-j — and Dee's modelled Naqada IID/IIIA transition, the phase in which writing first appears in Egypt, is 3352–3297 BC, which brackets 3298 BC almost exactly.

That coincidence is striking, and it is also the sharpest statement of the Greek chronology's residual problem:  the invention of writing lands on the Flood.

Babel's population window.  On the Masoretic numbers, Peleg is born 101 years after the Flood.  On the Greek, 531.  Robert Carter and Chris Hardy's own agent-based population model — built by young-earth researchers — reports that at 100 years the world population can exceed 1,000 under favourable settings, and reaches 10,000 or 40,000 only under settings they call "not likely,"⁷⁷ while Uruk alone held 20,000 to 50,000 people.⁷⁶  At 531 years the same model is in a completely different regime.  This is the cleanest case where the longer chronology turns a strained result into a routine one, and it is demographic rather than archaeological.

The Ice Age.  Eric J. Tweedy, writing in Answers in Genesis's own journal in 2024, found that Michael Oard's post-Flood Ice Age model and the northern-European settlement chronology cannot both be fitted onto the Masoretic timeline, while on the Greek numbers they can be aligned without alteration.¹⁸  State what actually fails, because it is not the text.  What Tweedy shows is incompatible is a combination — Oard's model parameters plus a particular reading of the settlement evidence — and he offers four ways out, three of which rescue the Masoretic timeline without touching Genesis.  He is also honest about the price of the Greek option:  moving the Flood to 3298 BC forces a rethink of the peopling of the Americas.

What it does not buy, and this is decisive

Göbekli Tepe.  Wall plaster from the inner ring wall of Enclosure D gives 9745–9314 cal BC, and three charcoal samples from beside that wall cluster at 9664–9311 cal BC.¹⁹  Jericho and Jarmo sit inside the Pre-Pottery Neolithic band of roughly 9600–7000 BC.⁷⁸  Çatalhöyük is later and is not a Pre-Pottery Neolithic site;  its occupation ends around 5975–5865 BC.⁷⁹

Against Göbekli Tepe the Greek chronology's extra 780 years covers about 11% of the gap.  Against Çatalhöyük's abandonment, about 23%.  The Neolithic tells are untouched by this dispute, and no textual choice will touch them.

Henry Smith, the strongest advocate of the Greek chronology, does not claim otherwise.  His ask is narrower:  that any re-dating of pre-Abrahamic archaeological periods "should be done within the context of the LXX's post-Flood chronology, not the MT's."¹⁵  That is an argument for a smaller correction, not for none.

The honest accounting

So the answer to "which one gives the correct young Earth chronology" has two halves.  Textually, the Masoretic numbers have the better case.  Archaeologically, the Greek numbers relieve real pressure — and not nearly enough of it to end the argument.  A reader who adopts the Greek chronology because it helps has bought 780 years against a problem measured in millennia, and has done so on textual grounds that are, on balance, weaker.

One more thing, and it is uncomfortable for both camps

Neither side's rule is as clean as it looks.  Henry Smith defends the Masoretic "480 years" at 1 Kings 6:1 while arguing that the Greek preserves the original in Genesis 5 and 11.  Robert Carter is consistent — he prefers the Masoretic in both places, saying so in print at 1 Kings 6:1 — but the argument that carries his Genesis case, the visible +100 scheme, has no analogue in Kings, so the extension is analogy rather than evidence.⁸⁰

Objections and responses

"The Dead Sea Scrolls prove the Bible hasn't changed."  For the Hebrew text as a whole, the scrolls are genuinely impressive, and the non-Qumran Judaean Desert scrolls from Masada, Murabba'at and Nahal Hever are essentially pure proto-Masoretic.⁸¹  But at Qumran itself the biblical text was plural, not uniform.  And on the specific question of the Genesis chronology the scrolls are silent:  one word of Genesis 5 survives, and none of Genesis 11.  The popular "95% identical" claim is Gleason Archer's wording, in A Survey of Old Testament Introduction, and it is about the two Cave 1 Isaiah scrolls rather than about the Hebrew Bible as a whole.  It is often credited to Millar Burrows;  Burrows publishes no percentage of textual agreement anywhere in The Dead Sea Scrolls.⁸²  The figure has no stated derivation, and should not be repeated as though it did.

"So the Bible has been changed after all."  Not in the way the objection implies.  Across the whole Torah the Masoretic and Samaritan texts share 89.1% of their words letter for letter, 93.5% once plene and defective spelling is set aside, and 52.4% of verses are word-for-word identical on that same allowance.  The disputed material is a small, identifiable set of numbers and expansions, and the disagreements are visible precisely because independent witnesses exist to compare.  A tradition with only one manuscript line has no such check.

"If Christians can't even agree on the date of creation, why should I take any of it seriously?"  The disagreement is bounded and it is about arithmetic in two chapters.  All three versions describe the same events in the same order with the same people, and all three give an Earth a few thousand years old.  Compare that to the range of dates a critical historian would assign to, say, the composition of the Pentateuch, and the textual variation here is small.

"The Septuagint must be right, because the apostles used it."  They used it, and this article counts how often:  of 275 Old Testament citations in the New, only 47 can discriminate between the Greek and the Hebrew, and there the Greek wins 37 to 10.  But quoting a translation is not endorsing every reading in it — an English-speaking preacher quoting the ESV is not ruling on Hebrew textual criticism.  And no New Testament writer cites a Genesis 5 or 11 number.

"The Masoretic Text must be right, because it's the Hebrew."  The Samaritan Pentateuch is also Hebrew, and 4QJerᵇ is Hebrew, and it supports the short Greek Jeremiah.  "Hebrew" is not a text;  it is a language, and there was more than one Hebrew text.

"Methuselah outlives the Flood in the Septuagint, so the Septuagint is disqualified."  In the edition Rahlfs printed, yes, by fourteen years.  But the Masoretic side's own current paper concedes that 187 — the reading that removes the problem — was probably the original Greek, and Codex Alexandrinus physically shows a corrector scraping 167 and writing 187.  What survives of the argument is narrower:  the Greek text was unstable exactly where the chronology breaks.

"Ussher said 9 a.m. on October 23rd, 4004 BC."  He said no such thing.  His Annals gives no hour, in either the Latin of 1650 or the English of 1658.  The hour comes from a conflation with John Lightfoot, whose own creation epoch was September 3929 BC — a different year entirely.

"The genealogies have gaps, so the whole exercise is pointless."  Genealogical gaps and chronological gaps are different things.  Even if names are missing, "when Enosh had lived ninety years he begot Kenan" still dates Kenan's birth, because Kenan is the object of the verb.  To lose years you have to insert an unnamed person the text does not mention — which the founding article for the gaps position asserts once and never defends.

"Then why does Exodus 6 obviously have gaps?"  It does, and this article concedes it in full:  four generations cannot produce 8,600 Kohathite males.  But Exodus 6 gives lifespans without ages at begetting, so it is not a chronogenealogy and is free to telescope.  What Exodus 6 destroys is the claim that biblical genealogies are presumptively complete — which is a real loss for the young-earth side, and is why this article does not make that claim.

A note on method

Every version number in this article was read out of a text of that version during the research for it, never copied from a chart.  The Hebrew came from the Westminster Leningrad Codex (the openscriptures/morphhb OSIS XML) and from Sefaria's "Miqra according to the Masorah";  the Greek from the CATSS/CCAT morphological Septuagint on the Rahlfs base, with disputed readings checked against Swete and against Brooke and McLean's apparatus read off page images;  the Samaritan from the Text-Fabric transcription of the base manuscripts of Stefan Schorch's critical editio maior.  All arithmetic is in scripts kept beside the research dossiers.

Two normalisation traps were hit and fixed, and are recorded here because anyone repeating the work will hit them too.  The Samaritan dataset writes shin as the precomposed Unicode presentation form, so a naive comparison scores every word containing shin as a difference and drives the apparent agreement rate down to 72%.  And that dataset splits prefixes into separate word nodes while the Hebrew one splits both prefixes and suffixes, so a comparison made at morpheme level compares unlike things.

No quotation, page number or numeral in this article was established through an AI summarising fetch.  Where a source's text layer was unusable — Brooke–McLean's Greek apparatus, Niese's Josephus, Migne's Patrologia Graeca, Holladay's correlation coefficients — the pages were rendered as images and read directly.

What is still open, and stated rather than hidden.  Wevers' Göttingen critical Genesis could not be obtained, and its readings are known here only through NETS, the edition's own English translation:  it prints 167 at Genesis 5:25 and keeps the second Kainan at Genesis 11:12–13.  What the Göttingen apparatus attributes to which manuscripts could not be checked, so the Hexaplaric attributions in Part 5 rest on Brooke–McLean 1906 and on Field.  Klein's 1974 article was obtained on a fourth attempt and is cited first-hand.  Aharoni's study of hieratic numerals was obtained, and it cost this article an argument:  Hebrew scribes did have numerals, so the ±100 point is now stated as a fact about the surviving manuscripts rather than about a lost original.  The Mount Gerizim dates are Magen's, taken from Kartveit 2020, because Magen's own 2007 chapter is a paywalled ebook that the free JSTOR tier will not open.  The Samaritan Pentateuch's rediscovery is verified from Tov 2001, except for the familiar claim that della Valle bought the manuscript in Damascus, which no source consulted here supports and which has been dropped.

The article has had two independent adversarial verification passes, on 2026-08-16 and 2026-08-22.  Each used separate verifiers for arithmetic, ancient sources, modern scholarship and archaeology, each instructed to refute rather than confirm.

Sources