The Genesis genealogies and gaps

[F]or aught we know instead of twenty generations and some two thousand years measuring the interval between the creation and the birth of Abraham, two hundred generations, and something like twenty thousand years, or even two thousand generations and something like two hundred thousand years may have intervened.¹

The "begat" indicates the birth of the person named after it;  and the date of that birth being given, it matters not how many un-named generations intervene.  The chronology is fixed and unchanged.²

We have to distinguish between literal intention and historical, factual truth.  The figures are not, to us, historically, scientifically or factually true, but they were literally intended.³

Summary

Genesis 5 and 11 list twenty men from Adam to Abraham and give each man's age when the next man in the list was born.  Add the ages and you get about 2,000 years from Adam to Abraham on the Hebrew text.  Since 1890 the standard evangelical escape from that number has been "gaps":  the lists skip generations, so the sum is not a chronology.  This article traces that argument to its source, William Henry Green's Bibliotheca Sacra article of April 1890, tests each of its steps against the Hebrew text and against Green's own critics, and reports what the surviving text-forms allow.

1. Green proved that biblical genealogies skip names, and that part of his case stands.  Matthew 1:8 says "Joram the father of Uzziah" when 1 Chronicles 3:11-12 puts Ahaziah, Joash and Amaziah between them.  Ezra 7:1-5 drops six consecutive names that 1 Chronicles 6 supplies.  Exodus 6:16-20 makes Moses the great-grandson of Levi, yet Numbers 3:27-28 counts 8,600 male Kohathites in Moses' lifetime, and the 430 years of Exodus 12:40 put Moses' birth about 350 years after Kohath entered Egypt, when Kohath lived 133 and Amram 137.  A name is missing there, and Green was right to say so.⁴

2. But Genesis 5 and 11 do something no other biblical genealogy does:  they date the birth of each named son.  "When Enosh had lived 90 years, he became the father of Kenan" (Gen 5:9).  Whether Kenan was Enosh's son, grandson or great-grandson, the text says Kenan was born when Enosh was 90.  A missing name changes the family tree.  It does not move the date.  Warfield, Green's strongest ally, conceded that this feature "differentiated" these lists "from all other genealogies in Scripture."¹

3. Green's whole case for missing years rests on one unargued clause.  He allowed that Genesis 5:9 could mean "one was born from whom Kenan sprang."⁴  That is, the verse dates the birth of an unnamed ancestor of Kenan rather than Kenan.  He offered no example of the verb wayyoled (וַיּוֹלֶד, "and he fathered") working that way, and none has been found since.  Smith Goodenow answered the point in 1893;  Bibliotheca Sacra accepted his article and never printed it.⁵

4. The verb dates the birth of its object, even when the object is a remote descendant.  Gerhard Hasel counted the phrase wayyoled 'et-PN fifteen times in the Old Testament, all in Genesis 5 and 11, and found every other occurrence of that form denoting a physical birth.⁶  Deuteronomy 4:25 and 2 Kings 20:18 use the same Hiphil of grandchildren and distant descendants, and in both the verb still describes the birth of the people it names.  Jeremy Sexton's 2018 article put the point in a sentence:  "Missing time does not follow from missing generations."⁵

5. The Bible itself supplies the control case, and it cuts against Green.  Exodus 6:16-20, Green's best example of a compressed genealogy, gives lifespans for Levi, Kohath and Amram and no begetting ages at all.  That is what a genealogy with numbers but no chronology looks like.  Genesis 5 and 11 look different because they are doing something different.⁷

6. The "ten and ten" symmetry is real, partial, and beside the point.  Genesis 5 has ten names.  Genesis 11 in the Hebrew has nine;  Green reached ten by counting Noah twice.⁷  The Sumerian King List, the usual parallel, has seven, eight, nine or ten antediluvian kings depending on the tablet.⁸  Even a deliberately selected list of ten would only show that names were chosen, which no one disputes.

7. Jude 14 calls Enoch "the seventh from Adam," and that is a modest witness.  It shows a first-century Jew counting Genesis 5 as written, with no missing name between Adam and Enoch.  It shows nothing about Genesis 11, and the phrase is already in 1 Enoch 60:8 and 93:3, so Jude may be repeating a tradition rather than doing arithmetic.⁹

8. Ussher's 4004 BC came from the text plus two secular facts, and the "nine in the morning" belongs to another man.  Ussher inherited the medieval tradition of four thousand years from creation to Christ, accepted Scaliger's dating of Herod's death to 4 BC, and found that the Hebrew numbers fit both.¹⁰  He put creation at the evening before 23 October because it was the first Sunday after the autumn equinox of that year.¹¹  The hour comes from John Lightfoot's 1642 notes on Genesis 1:26, and Lightfoot's own creation year was 3929 BC, not 4004.¹²

9. The three surviving text-forms put creation between about 4,200 and 5,600 BC.  On a 2166 BC birth of Abraham the Masoretic Text gives 4174 BC, the Samaritan 4415 BC and the Septuagint 5560 BC.  The most generous reading that keeps the numbers, Frederic Gardiner's 1873 proposal that each begetting age dates a firstborn rather than the named son, could add at most about 8,700 years, and Genesis 5:7 refutes it.¹³

10. The honest crux is not whether the numbers were meant, but whether they are true.  James Barr, no friend of the young-earth position, put it plainly in 1987:  the figures "were literally intended," and calling them non-literal is "a cowardly expedient" to avoid saying they are false.³  He called the gap theory "only a hidden way of saying that the biblical chronology is not true."³  Whether it is true is a question for the dating articles on this site, not for exegesis.

Verdict:  gaps in names are possible;  gaps in years are not what the text offers.  Green's 1890 argument establishes that a genealogy may skip generations and nothing more.  Every verse in Genesis 5 and 11 dates the birth of the son it names, so the sum of the begetting ages is a chronology whether or not names are missing.  The largest creation date any manuscript tradition supports is about 5,600 BC.  A reader who wants humanity to be 50,000 or 200,000 years old has to say the numbers are wrong.  Saying they contain gaps does not get there.

The lists date each birth, and the arithmetic is simple

Genesis 5 runs from Adam to Noah in ten steps.  Genesis 11:10-26 runs from Shem to Terah in nine.  Each step has the same shape:  A lived X years and fathered B;  A lived Y more years and fathered other sons and daughters;  the days of A were Z years.  Genesis 11 drops the total.  Gerhard Hasel of Andrews University coined a word for a list built this way.  "We suggest a designation that more accurately reflects the nature of the features of Genesis 5 and 11:10-26 by the choice of the name of chronogenealogy.  This compound term reflects both the repeated time elements by means of the word chrono- and the linear succession of generations by means of the term genealogy."¹⁴

The sum is short work.  On the Hebrew text Adam to the Flood is 1,656 years.  Arpachshad was born two years after the Flood (Gen 11:10), and the Flood to Abram's birth is 292 years if Abram was born in Terah's 70th year (Gen 11:26) or 352 if he was born in Terah's 130th, which Acts 7:4 and Genesis 11:32 and 12:4 together require.  Green himself gave the 292.⁴  Adam to Abram is therefore 1,948 or 2,008 years.  Barr, checking Ussher's arithmetic in 1985, wrote that "the chronology can be obtained by simple addition of the ages at which the first son was born, in the genealogies of Gen. v and xi," and that by the Masoretic text "the flood began in the year 1656 A.M."¹⁰

Ussher put Abram's birth at 1996 BC and reached 4004 BC.  A modern anchor of 2166 BC gives 4174 BC on the same numbers.  Both are under 6,200 years ago.  The whole "gaps" literature exists because those sums are too small for the human antiquity that geology and archaeology report, and because the Hebrew of Genesis 5 and 11 does not read like a list built to be skipped.

Green proved that biblical genealogies skip names

William Henry Green taught Old Testament at Princeton Theological Seminary for fifty years.  His "Primeval Chronology" appeared in Bibliotheca Sacra in April 1890, at the invitation of its editor, and it is the article every later gap argument cites.⁴  Its first half is a catalogue of biblical genealogies that omit names, and that half is sound.

Matthew 1.  "In verse 8 three names are dropped between Joram and Ozias (Uzziah), viz., Ahaziah (2 Kings viii. 25), Joash (2 Kings xii. 1), and Amaziah (2 Kings xiv. 1);  and in verse 11 Jehoiakim is omitted after Josiah."⁴  Anyone can check this against 1 Chronicles 3:11-12:  "Joram his son, Ahaziah his son, Joash his son, Amaziah his son, Azariah his son."  Matthew writes "Joram the father of Uzziah" (Matt 1:8) and skips four generations.  Matthew does it to get three sets of fourteen (Matt 1:17).

Ezra 7.  Ezra's own pedigree in Ezra 7:1-5 runs Aaron to Seraiah in sixteen names.  The same line in 1 Chronicles 6:3-14 has twenty-two.  Green set the two columns side by side and showed that Ezra drops six consecutive names, Amariah through Johanan, between Meraioth and Azariah.⁴  "The genealogy of Ezra is recorded in the book which bears his name;  but we learn from another passage, in which the same line of descent is given, that it has been abridged by the omission of six consecutive names."⁴

1 Chronicles 26.  Shebuel "the son of Gershom, the son of Moses" was treasurer under David (1 Chr 26:24).  Green:  "it is absurd to suppose that the author of Chronicles was so grossly ignorant as to suppose that the grandson of Moses could be living in the reign of David."⁴  "Son of" reaches across centuries here.

Exodus 6.  This is the example that matters, because it is the one genealogy with numbers in it that Green could show was compressed.  Exodus 6:16-20 runs Levi, Kohath, Amram, Moses, with lifespans of 137, 133 and 137 for the first three.  Green gave three proofs that generations are missing.

First, parallel lines of the same period are longer.  Bezalel's line in 1 Chronicles 2:18-20 has seven generations from Jacob;  Joshua's in 1 Chronicles 7:23-27 has eleven.  "Now it is scarcely conceivable that there should be eleven links in the line of descent from Jacob to Joshua, and only four from Jacob to Moses."⁴

Second, the census.  Numbers 3:27-28 lists the four Kohathite clans, the Amramites among them, and counts "every male from a month old and upward, there were 8,600."  Green:  "if no abridgment has taken place in the genealogy, the grandfather of Moses had, in the lifetime of the latter, 8,600 descendants of the male sex alone, 2,750 of them being between the ages of thirty and fifty (Num. iv. 36)."⁴  Amram, Moses' father on the plain reading, would be the ancestor of a quarter of them.  "Moses' father surely could not have been the ancestor of one-fourth of this number in Moses' own days."⁴

Third, the years.  Kohath went down to Egypt with Jacob (Gen 46:11).  The sojourn lasted 430 years (Exod 12:40).  Moses was 80 at the Exodus (Exod 7:7), so he was born in about the 350th year of the sojourn.  Kohath lived 133 years.  Even if Amram was born in Kohath's last year, and Moses in Amram's last, Amram's 137 years run out in about year 270.  Green:  Moses "must have been born more than 350 years after Kohath, who consequently could not have been his own grandfather."⁴  The arithmetic is reproduced in research/gap-bounds.ts.¹³

Green's conclusion from all this was that "in the usage of the Bible, 'to bear' and 'to beget' are used in a wide sense to indicate descent, without restriction to the immediate offspring."⁴  His examples were Genesis 46:18, where Zilpah "bore" sixteen souls who include her great-grandsons, and Genesis 10:15-18, where Canaan "fathered" the Jebusites and Amorites.  He was right about all of it.  Hebrew ben ("son") and yalad ("bear, beget") can and do reach across generations.  Jeremy Sexton, whose two articles on Green are the sharpest modern critique of him, writes:  "All of this is true, and yet none of it implies the possibility of chronological gaps in Genesis 5 and 11, a logical leap that Green made but failed to defend or even acknowledge."⁵

There is a cost to the young-earth side in conceding this, and it should be said.  A creationist who argues "biblical genealogies are always complete" is wrong, and Exodus 6 refutes him from inside the Pentateuch.  The 430 years of Exodus 12:40 force a missing generation between Kohath and Moses unless one adopts the Septuagint and Samaritan reading, which halves the time in Egypt to 215 years.  Either way, the question about Genesis 5 and 11 cannot be settled by a general rule about genealogies.  It has to be settled by what those two chapters say.

What the begetting ages do, and why Green needed them not to

Green saw the problem himself and stated it better than most of his critics.  "Why should the author be so particular to state, in every case, with unfailing regularity, the age of each patriarch at the birth of his son, unless it was his design thus to construct a chronology of this entire period, and to afford his readers the necessary elements for a computation of the interval from the creation to the deluge and from the deluge to Abraham?  And if this was his design, he must of course have aimed to make his list complete.  The omission of even a single name would create an error."⁴

Benjamin Warfield, who took Green's article up in 1911 and gave it the backing of Princeton's leading theologian, conceded the same point.  The appearance of a chronology "is due to a peculiarity of these special genealogies by which they are differentiated from all other genealogies in Scripture.  We refer to the regular attachment to each name in the lists, of the age of the father at the birth of his son."¹  "It certainly looks, at first sight, as if we needed only to add these one hundred and thirty, one hundred and five, and ninety years together in order to obtain the whole time which elapsed from the creation of Adam to the birth of Kenan."¹

Green's reply had two parts.  The first was that the author never sums the numbers.  "He nowhere sums these numbers, nor suggests their summation.  No chronological statement is deduced from these genealogies, either by him or by any inspired writer."⁴  The second was that the numbers have a different purpose.  "They merely afford us a conspectus of individual lives.  And for this reason doubtless they are recorded.  They exhibit in these selected examples the original term of human life.  They show what it was in the ages before the Flood.  They show how it was afterwards gradually narrowed down."⁴  Warfield expanded this into the claim that the numbers are parenthetical notes, and that their chronological look "is pure illusion, due wholly to the nature of the parenthetical insertions which are made."¹

Then comes the sentence on which the whole argument turns.  Green wrote that when the text says "Enosh lived ninety years and begat Kenan," the wide usage of "begat" "makes this statement equally true and equally accordant with analogy, whether Kenan was an immediate or a remote descendant of Enosh;  whether Kenan was himself born, when Enosh was ninety years of age or one was born from whom Kenan sprang."⁴

Read the two halves of that sentence separately.  The first half, that Kenan might be a remote descendant, is what the first half of the article proved.  The second half, that the verse might date the birth of someone other than Kenan, is a new claim, and it is the only claim in the article that creates missing time.  If Genesis 5:9 dates Kenan's birth, then it does not matter how many unnamed men stand between Enosh and Kenan.  Enosh was 90 when Kenan was born.  Seth was 105 when Enosh was born.  Adam was 130 when Seth was born.  The sum holds.  Only if the verse dates the birth of an ancestor of Kenan, and says nothing about when Kenan himself arrived, does time open up.

Green did not argue for the second half.  He asserted it once, in passing, and moved on.  Sexton:  "This assumption is the linchpin of Green's case.  It inserts B's anonymous ancestor into the formula, creating a gap of unknown length between the birth of this unnamed ancestor and the birth of B himself.  Green only mentioned this crucial premise once, almost in passing, in a comment on Gen 5:9."⁵

Green also conceded, in the next sentence, what his critics would later press.  "These genealogies may yield us the minimum length of time that it is possible to accept for the period that they cover."⁴  On his own reading the sum is a floor.  The dispute is only about whether there is a ceiling.

Goodenow answered in 1893, and the answer was never printed

The first person to see the flaw was Smith Bartlett Goodenow, a minister and chronologist who had already published in Bibliotheca Sacra.  His rebuttal was written soon after Green's article and appeared in his book Bible Chronology Carefully Unfolded in 1896.  The key paragraph:

There is no getting round this fundamental defect of the theory.  The "begat" indicates the birth of the person named after it;  and the date of that birth being given, it matters not how many un-named generations intervene.  The chronology is fixed and unchanged.  No such anomaly is known in Scripture, or in reason, as a dating given to an un-named ancestor's birth.  The chronology here added to the genealogy forbids any such fast-and-loose play upon omissions.²

Goodenow also saw that Green's "conspectus of individual lives" left the begetting ages without a job.  "This theory takes away all purpose on the part of the sacred writer in giving the birth-dates he has so carefully arranged."²  A list built to show longevity needs only the total.  Exodus 6 gives only the total.  Genesis 5 gives three numbers per man, and the first of the three is useless unless it is meant to be added.  His conclusion:  "no mode of speech could be contrived to give successive dates to Bible generations if those tables in Genesis be denied as such."²

What happened to the rebuttal is a small scandal of editorial history, recovered by Sexton from Goodenow's own book.  Goodenow sent two manuscripts to Bibliotheca Sacra, the reply to Green and a paper on pre-Adamic man.  The editor, G. Frederick Wright, a geologist and an advocate of an old humanity, wrote to him on 29 June 1893:  "The two Mss. which I have in hand ought to be published in the Bibliotheca, and I can say to you positively, that if you will let them remain in my hands, I will work them into the January and April numbers."⁵  The journal printed the pre-Adamite paper in 1894.  It never printed the reply to Green.  Wright went on to reprint Green's article in full in his own Origin and Antiquity of Man in 1912.⁵

Ronald Numbers, the historian of creationism, has documented what Green's paragraph meant to the men who received it.  Charles Hodge's son Alexander recalled his father "walking up and down in his study when he heard (about it) and saying, 'What a relief it is to me that he should have said that.'"  George Macloskie of Princeton called it in 1904 "the most important biblical discovery of our time."⁵  Relief is not an argument, but it explains why the argument was not examined for a century.

The verb dates the birth of its object

The Hebrew of Genesis 5:9 is wayechi 'enosh tish'im shanah wayyoled 'et-qenan (וַיְחִי אֱנוֹשׁ תִּשְׁעִים שָׁנָה וַיּוֹלֶד אֶת־קֵינָן):  "and Enosh lived ninety years and he fathered Kenan."¹⁵  The verb wayyoled is the Hiphil, the causative stem, of yalad, "to bear."  Its object is marked by 'et, the particle that flags a definite direct object.  The object is Kenan.

Hasel made the first count.  "The repeated phrase 'and he fathered PN' (wayyôled 'et-PN) appears fifteen times in the OT, all of them in Genesis 5 and 11.  In two additional instances the names of three sons are provided (Gen 5:32;  11:26).  The same verbal form as in this phrase (i.e., wayyôled) is employed another sixteen times in the phrase 'and he fathered (other) sons and daughters.'"  He then checked every remaining Hiphil of the verb in Genesis and beyond:  "the expression 'and he fathered' (wayyôled) is used in the sense of a direct physical offspring (Gen 5:3;  6:10).  A direct physical offspring is evident in each of the remaining usages of the Hiphil of wayyôled, 'and he fathered,' in the OT (Judg 11:1;  1 Chron 8:9;  14:3;  2 Chron 11:21;  13:21;  24:3)."  His conclusion:  "Thus the phrase 'and he fathered PN' in Genesis 5 and 11 cannot mean Adam 'begat an ancestor of Seth.'"⁶

Hasel's count is right but his conclusion overreaches, and the overreach matters.  Green had already produced Hiphil uses of the same verb where the object is a remote descendant.  In Deuteronomy 4:25 Moses says "after you have produced (tolid) children and grandchildren."  In 2 Kings 20:18 Isaiah tells Hezekiah that "some of your very own descendants whom you father (tolid) will be taken away."  So the Hiphil can take a grandson or a distant descendant as its object.  Green was right about that, and a young-earth writer who denies it will lose the point.

What Sexton noticed is that these very verses settle the real question.  In both, the verb still describes the birth of the people it names.  Deuteronomy 4:25 dates the corruption of Israel to a time "after" the grandchildren are born, not after some intermediate act.  2 Kings 20:18 predicts the birth of the exiles themselves.  In neither case does the verb describe the birth of an ancestor of its object.  Sexton's 2015 article in the Westminster Theological Journal laid the grammar out in a table and drew the conclusion (Hebrew transliterated here):  "Who was born when Enosh was 90?  The untranslatable particle 'et marks qenan ('Kenan') as the direct-object accusative of wayyoled ('he had').  A 'direct-object accusative is the recipient of a transitive verb's action.'  The transitive verb wayyoled describes birth.  Therefore Gen 5:9 refers to Kenan's birth when Enosh was 90."⁷

He then wrote out Green's argument as a syllogism, which is the fairest way to show where it breaks.

Premise 1:  Gen 5 and 11 may contain genealogical gaps (that is, B may be a remote descendant of A in some cases). Premise 2:  Where B is a remote descendant of A, the formula "When A had lived X years, he had B" means "When A had lived X years, he had [the son from whom sprang] B." Conclusion:  Gen 5 and 11 may contain chronological gaps.⁷

"We grant P1," Sexton wrote.  "However, we find several problems with P2."⁷  The problems are that no Hebrew text uses the verb that way, that the direct object is the person born, and that Genesis has ordinary words for a father's part in conception (yada', "knew," in Gen 4:1, 17, 25) which the author did not use.

The clearest modern statement of Green's semantics is Kenneth Kitchen's, in On the Reliability of the Old Testament (2003).  Kitchen, an Egyptologist and a defender of the historical books against minimalism, wrote that in "the formal, representative genealogies of Gen. 4, 5, 11, 'A fathered B' can be literal;  or it may equally stand for 'A fathered [P, who fathered Q, who fathered R, who fathered S, who fathered T, who fathered ...] B.'  Or in short for 'A fathered (the line culminating in) B.'"¹⁶  That is Premise 2 stated without argument.  Kitchen's reason for wanting it follows at once:  "we can neither date the flood before Abraham nor the creation before Noah merely by counting the Genesis figures continuously as did the worthy Archbishop Ussher in the carefree days when no evidence from outside the Bible was even imagined," and "So an Ussherite solution is ruled out."¹⁶  The external evidence may be decisive.  It is not exegesis.

Gordon Wenham, whose Word commentary on Genesis 1-15 is the standard evangelical one, is more careful than Kitchen and reaches the opposite reading of the text.  "Another suggestion (W. H. Green, BSac [1890] 285-303) is that the genealogy is not intended to be complete, that generations have been omitted, and therefore it should not be used for chronological purposes.  However, the Hebrew gives no hint that there were large gaps between father and son in this genealogy.  4:25 makes it clear that Seth was Adam and Eve's third son.  At the other end of the genealogy, Lamek comments on Noah's birth, and Ham, Shem, and Japhet were contemporaries of their father.  It therefore requires special pleading to postulate long gaps elsewhere in the genealogy."¹⁷  Wenham then accepts an old humanity anyway, on archaeological grounds.  That is the honest form of the position:  the text says one thing, and he believes the evidence says another.

Steinmann's causation argument

Andrew Steinmann tried a new route in 2017.  Rather than deny that wayyoled dates a birth, he argued that a causative verb dates the causing action, so the begetting age marks when the father did something that eventually led to the named descendant's birth, not the birth itself.⁵  Sexton's reply in JETS the next year made two points.  First, in standard accounts of causation a Hiphil "always expresses an event that corresponds to the meaning of its root," so a time indicator on wayyoled "specifies the time of the caused event, not the time of the causing action."⁵  Second, Steinmann had said the opposite in 2014:  "These genealogies list the age of each ancestor [A] at the birth of his descendant [B]."  Sexton observed that Steinmann "apparently altered his interpretation after becoming aware recently that his earlier formulation created missing generations but not the desired missing time."⁵  Steinmann's own candidate for what the ages might mark, "the onset of virility," fails on Seth, who was not Adam's firstborn (Gen 4:25), and Steinmann conceded that the purpose of the ages might have "been irretrievably lost over time."⁵  The 2017 article itself is not free online and was not read for this article;  its argument is taken from Sexton's quotations of it, with page numbers.

Exodus 6 is the control case, and it cuts against Green

Green's argument needed Exodus 6 twice.  It was his proof that a Mosaic genealogy could skip generations, and it was his model for what Genesis 5 and 11 were doing:  "as they proceeded from the same pen, we may fairly infer from the similarity of construction a similarity of design."⁴  Exodus 6 gives lifespans and cannot be summed, so, Green argued, Genesis 5 gives lifespans and was not meant to be summed either.

But the similarity is the problem.  Exodus 6 gives lifespans and nothing else.  "The sons of Kohath:  Amram, Izhar, Hebron, and Uzziel.  And the years of the life of Kohath were 133" (Exod 6:18).  No age at Amram's birth.  Nothing that can be added.  Sexton:  "if the author of the genealogies had wanted to supply life spans, but not chronology, he could have accomplished this more efficiently without begetting ages.  The genealogy in Exod 6:16-20 gives the fathers' life spans without using begetting ages."  He then rewrote Genesis 5:9-11 in the Exodus 6 pattern to show the difference:  "Enosh had Kenan.  And Enosh had other sons and daughters.  And all the days of Enosh were 905 years, and he died."⁷  That would carry Green's "conspectus of individual lives" perfectly.  It is not what Genesis says.

A second control is inside Genesis 5 itself.  Each verse dates the son's birth twice.  Genesis 5:6 says Seth had Enosh at 105.  Genesis 5:7 says Seth lived 807 years "after he had Enosh" ('acharei holido 'et-'enosh).  The 807 years are counted from Enosh's birth, not from the birth of some ancestor of Enosh.  Sexton:  "Scripture thus bears a twofold witness to the year of each named son's birth."⁵  The same double witness sinks Gardiner's older proposal, discussed below, that the named son might have been born any time in the father's remaining years.

A third control is the Flood.  On the plain reading of the Masoretic numbers, Methuselah dies in the year the Flood comes.  Methuselah had Lamech at 187 and lived 782 more years, 969 in all (Gen 5:25-27).  Lamech had Noah at 182 (Gen 5:28).  The Flood came in Noah's 600th year (Gen 7:6).  187 + 182 + 600 = 969.  Sexton:  "According to Green, we cannot know when Methuselah died in relation to the flood, because we cannot know how big the chronological gap was between him and Noah.  Methuselah could have died thousands of years before the flood.  According to the chronological interpretation, however, Methuselah died in the very year of the deluge."  Either the author expected his readers to add the numbers, or the coincidence is an accident.⁷  The Samaritan text moves Methuselah's death to the same year by a different route, and the Septuagint's 167 reading has him outliving the Flood by fourteen years, which is one reason the Septuagint's Genesis 5 is suspect.  That dispute belongs to the text-forms article.

Ten and ten: the symmetry is real, partial, and beside the point

Green's fifth argument was structural.  "Each genealogy includes ten names, Noah being the tenth from Adam, and Terah the tenth from Noah.  And each ends with a father having three sons, as is likewise the case with the Cainite genealogy (iv. 17-22).  The Sethite genealogy (chap. v.) culminates in its seventh member, Enoch."  Since Matthew got his three fourteens by dropping names, "it seems in the highest degree probable that the symmetry of these primitive genealogies is artificial rather than natural."⁴

Umberto Cassuto, whose Hebrew University commentary is the fullest literary reading of Genesis 1-11, agreed that the ten is deliberate.  The list of generations from Adam to Noah "consists of ten paragraphs, one for each of the ten generations."¹⁸  Cassuto read the chapter as a quiet reply to the Mesopotamian tradition of ten long-lived kings before the Flood:  "It is correct—the Bible comes to tell us—that there lived before the Flood ten generations of notable personages;  but they were only ordinary mortals, not gods, or demi-gods, or even men transformed into divinities."¹⁸  His arithmetic on the numbers themselves, the multiples of five and seven, is handled in the longevity article and is not repeated here.

Three things limit what the symmetry can do.

The second list has nine names, not ten.  Shem, Arpachshad, Shelah, Eber, Peleg, Reu, Serug, Nahor, Terah.  Green reached ten by counting Noah at the head of a list in which he does not appear.  Sexton:  "it is worth noting that MT Gen 11 comprises only nine names.  To arrive at ten names in each genealogy, Green includes Noah in both genealogies, even though Noah does not appear in the second one.  In a later publication, Green acknowledges that the two genealogies have 'nearly the same number of links (one ten, the other nine).'"⁷  The Septuagint's Genesis 11 does have ten, because it adds a second Kainan between Arpachshad and Shelah, and Luke 3:36 has him too.  Whether that Kainan is original is the hardest single question in the text-form debate, and it is argued out in the text-forms article.  Note only that if he is original, the symmetry argument gains a name and the chronology gains 130 years;  neither outcome helps the case for unlimited gaps.

The Mesopotamian parallel has no fixed number.  Hasel checked the tablets.  "The major recension of the Sumerian King List (WB 444) contains only eight and not ten kings.  One text contains only seven kings (W) and another (UCBC 9-1819) either seven or eight, whereas a bilingual fragment from Ashurbanipal's library has but nine kings.  Berossos and only one ancient tablet (WB 62), i.e. only two texts (of which only one is a cuneiform document), give a total of ten antediluvian kings.  On the basis of the cuneiform data it can no longer be suggested that the Sumerian King List contained originally ten antediluvian kings after which the biblical genealogies were patterned."⁸  Marshall Johnson, whose Cambridge monograph on biblical genealogies is the standard literary study, allowed one real parallel:  the seventh king in the Sumerian lists, Enmeduranki, is the one "regarded as possessing special wisdom concerning the gods," as the seventh name in Genesis 5 is Enoch.¹⁹  But Johnson also found that "there is no borrowing of actual names or even translation of names involved."¹⁹

Symmetry of names would only ever prove selection of names.  Jack Sasson's 1978 argument that Hebrew genealogies move important figures to the fifth or seventh slot, relayed here from Travis Freeman's survey because the ZAW article itself is paywalled, would if true show that names were chosen and arranged.²⁰  Freeman also relays David Bryan's objection that Sasson counts inconsistently, starting after Enoch to make Eber seventh and starting with Eber to make Abraham seventh.²⁰  Suppose Sasson is right anyway.  A selected list of ten men, each verse dating the birth of the next, still sums to the same number of years.  Sexton:  "While an artificial ten-and-ten symmetry would only entail genealogical (not chronological) gaps, it is worth noting..."⁷  The structural argument and the semantic argument are different arguments, and only the semantic one could ever add time.

Jude 14 counts seven names, and that is all it counts

"Now Enoch, the seventh in descent beginning with Adam, even prophesied of them" (Jude 14).  The Greek is hebdomos apo Adam Henoch (ἕβδομος ἀπὸ Ἀδὰμ Ἑνώχ), "seventh from Adam."²¹  Adam, Seth, Enosh, Kenan, Mahalalel, Jared, Enoch:  seven, counting Adam as the first.  Jonathan Sarfati's Journal of Creation article lists this among his reasons for continuity:  "Jude 14 says Enoch was seventh from Adam, which indicates straightforward father-son relationships from Adam to Enoch."²²  Freeman reports the same argument from the continuous-genealogy side.²⁰

The witness is real but small, for three reasons.

It covers six links out of nineteen.  Nothing in Jude touches Genesis 11, where the pressure for gaps is greatest because the Flood date is what collides with Egyptian chronology.

It is not independent.  The phrase is already in 1 Enoch, which Jude quotes in the next verse.  1 Enoch 60:8 in Charles's translation:  the garden "where my grandfather was taken up, the seventh from Adam, the first man whom the Lord of Spirits created."  1 Enoch 93:3:  "I was born the seventh in the first week."⁹  Jude is repeating how Genesis 5 was read in the second century BC, which is worth knowing.  He is not doing independent arithmetic.

And it settles the wrong question.  The strongest form of the gaps argument does not deny that Enoch is the seventh name.  It denies that the seven begetting ages sum to the years between Adam and Enoch.  Jude does not speak to that.  The rabbis who noted that "all sevenths are favourites in the world," and counted Enoch seventh from Adam and Moses seventh from Abraham, were counting names, not years.¹⁹

What Jude does show is that a first-century Jewish writer, working from a text of Genesis 5 substantially as we have it, counted the names as they stand and found no missing one before Enoch.  That is a fact about how the list was read.  Add Josephus, who sums the numbers, the rabbinic Seder Olam, which sums them, and the fathers from Theophilus onward, who sum them (all traced in the church-history article), and the reading Green called a "wrong interpretation" is the only reading anyone had for two thousand years.  Freeman:  "The most telling evidence on this issue is the fact that the latter interpretation was virtually unknown by Jews or Christians prior to A.D. 1800."²⁰  That is a historical point, not a grammatical one, and it is answered by "everyone was wrong," which Green in effect said.

How Ussher got 4004, and who said nine in the morning

The gaps argument was aimed at Ussher, so it is fair to ask what Ussher did.  James Barr, Regius Professor of Hebrew at Oxford and a determined critic of fundamentalism, wrote the definitive account in 1985, and it does not describe a man adding up numbers.

Three things fixed the year.  First, a tradition.  "Firstly, there was a long tradition, going back into the Middle Ages, and thus long before Ussher, that there were four thousand years between creation and the coming of the Christ."¹⁰  Second, a secular date.  "By Ussher's time, and especially since the work of the great scholar Scaliger (1583), it was clearly known and accepted that Herod the Great had died in 4 B.C.  Since it was clear from St. Matthew that he was alive when Jesus was born ... the birth of Jesus must go back to 5-4 B.C.  Hence, the date of 4004 B.C. for creation gave a more or less exact period of four thousand years from creation to Christ."¹⁰  Third, the numbers.  For the stretch from creation to Abram, "the chronology can be obtained by simple addition of the ages at which the first son was born, in the genealogies of Gen. v and xi."¹⁰  Ussher's contribution was to make the Hebrew numbers, read without gaps and with Abram born in Terah's 130th year, land on the traditional four thousand.  Barr:  "he succeeded in making it work out, and as a precise figure;  and that without any tampering - at least as he thought - with the biblical text itself."¹⁰

The day came from the calendar.  Ussher was sure creation was in autumn, because the Jewish civil year began then, and on a Sunday, because the first day of the week was the first day.  "Knowing that it was 4004 B.C., he looked in the astronomical tables for the Sunday which came first after the Autumnal equinox in that year ... He found that Sunday to be 23 October 4004 B.C."¹⁰  Ussher said so himself, in the preface to the 1658 English Annals:  the creation "did fall out upon the 710 year of the Julian Period, by placing its beginning in Autumn," and the Sunday nearest the autumnal equinox in that year "happened upon the 23 day of the Julian October."¹¹  The first entry of the Annals then reads that the beginning of time "fell upon the entrance of the night preceding the twenty third day of October, in the year of the Julian Calendar, 710."¹¹  Petavius, working from the same numbers with different assumptions, got Monday 26 October 3983 BC.¹⁰

There is no hour anywhere in Ussher.  The "nine o'clock in the morning" that every popular account attaches to him was traced by the text-forms article to Andrew Dickson White's Warfare of Science with Theology (1896), which twice attributes to Lightfoot the claim that "man was created by the Trinity on October 23, 4004 B. C., at nine o'clock in the morning."²³  This article went back to Lightfoot.  The sentence exists.  In A Few and New Observations upon the Book of Genesis, published in 1642, eight years before Ussher's Annals, Lightfoot's note on Genesis 1:26 reads:  "Man created by the Trinity about the third hour of the day, or nine of the clock in the morning."¹²  No date and no year appear with it.  Lightfoot's own year of creation, worked out elsewhere in his writings, put 1644 at the end of the world's 5,572nd year, which makes his creation autumn 3929 BC.¹²  White stitched Lightfoot's hour to Ussher's day and year and produced a date that neither man wrote.  It is a small error, but it is the one skeptics quote most, and it is worth having the pages.

What the surviving text-forms allow

The three ancient text-forms of Genesis disagree about the begetting ages by a total of about 1,400 years.  The Septuagint adds 100 to most antediluvian and postdiluvian begetting ages;  the Samaritan subtracts 100 from some and adds a second pattern of its own.  Green himself gave the spread for the Flood to Abraham's birth as "292 (Hebrew), 942 (Samaritan), or 1172 years (Septuagint)," and then refused the Septuagint's help:  "the superior accuracy of the Hebrew text in this instance, as well as generally elsewhere, can be incontrovertibly established.  This resource, then, is a broken reed."⁴  Modern scholarship is less sure than Green was.  The argument is in the text-forms article, whose verified dates are used here.

The dates on the 2166 BC anchor for the first four rows are the ones verified for the sibling article;  the Gardiner ceiling is computed in research/gap-bounds.ts by letting each named son be born in his father's last year of life, with Noah to Shem and Shem to Arpachshad held fixed by Genesis 5:32, 7:6 and 11:10.¹³  Gardiner's reading is the only one on the list, apart from Green's, that adds real time while keeping every number, so it is the one a careful old-earth reader would reach for.  Sexton summarised it fairly, quoting Gardiner:  the begetting age dates the firstborn, so Seth "might have begun to be a father at 105, but might have actually begotten Enos[h] at any reasonable time during the 807 years which he afterwards lived."⁵  It fails on Genesis 5:7, where the 807 years are counted from Enosh's birth, and it fails on Genesis 4:25, where Seth is Adam's third son and not his firstborn.  And even if it stood, 12,800 BC is not the 200,000 years Warfield reached for.

So the ceiling is about 5,600 BC for creation and about 3,300 BC for the Flood, and it is set by the Septuagint, not by gaps.  Anyone who needs Adam earlier than that has left the numbers behind.

Green's motive, in his own words

It is fair to ask why a man who had defended the chronology found gaps in it, and Green answered the question himself.  The argument first appeared as a footnote on page 128 of his 1863 reply to Bishop Colenso, and the footnote begins by naming its purpose:  "if scientific research should ever demonstrate what it cannot be said to have done as yet, that the race of man has existed upon the earth for a longer period than the ordinary Hebrew Chronology will allow, we would be disposed to seek the solution in this frequent, if not pervading, characteristic of the Scriptural genealogies."  It continues:  "Our current chronology is based upon the prima facie impression of these genealogies.  This we shall adhere to, until we see good reason for giving it up."²⁴

When Green reprinted the footnote as the opening of the 1890 article, that last sentence was gone.  The 1890 text reads "..... Our current chronology is based upon the prima facie impression of these genealogies. .... . But if these recently discovered indications of the antiquity of man ..."⁴  The dots mark the place where "This we shall adhere to" had been.  Sexton, who noticed the omission:  "Green left no trace of his earlier commitment to the chronological interpretation."⁵  Numbers found the earlier Green in his letters.  In 1849 Green wrote that any monuments "reaching back for millions of years or even for ten thousand years" would be "in flat contradiction to the Mosaic record."⁵

The 1890 article is candid about which way the evidence was pushing.  Its fourth argument is Egypt.  "If, then, Egyptian antiquity contradicts the current chronology, it simply shows that this chronology is based upon an unfounded assumption."⁴  And its closing sentence is not a claim about Hebrew grammar but about what Scripture does not do:  "we conclude that the Scriptures furnish no data for a chronological computation prior to the life of Abraham;  and that the Mosaic records do not fix and were not intended to fix the precise date either of the Flood or of the creation of the world."⁴  Freeman's summary is blunt and, on the record above, accurate:  "Green and Warfield, the source of the new interpretation, admitted their purpose was to save the credibility of the OT in the face of the new science."²⁰

None of this makes Green wrong.  Motive is not evidence.  But it explains why an argument with an unargued premise at its centre was received as a discovery, and why the man who exposed the premise could not get into print.

The honest crux: literally intended is not the same as true

This article has argued that the author of Genesis 5 and 11 meant his numbers to be added.  James Barr, who thought the numbers were legend, agreed, and his statement of the point is the one to end on.

All discussion of this matter has been bedevilled by the assertion that the chronological data of the Bible, and especially those of the earlier chapters of Genesis, are 'not to be taken literally'.  According to this argument, when we read that Methuselah lived to the high age of 969 years, we are to suppose that the writer did not mean 969 years but something different.  Now I submit that this is obviously false.  The biblical writers worked seriously on these figures, and they meant 969 years for Methuselah.³

We have to distinguish between literal intention and historical, factual truth.  The figures are not, to us, historically, scientifically or factually true, but they were literally intended.  A year to them was the same period as it still is to us. ... When, in modern times, people began to say that these passages were 'not to be taken literally', this was really a cowardly expedient which enabled them to avoid saying that, though they were literally intended, they were not literally true.³

On the gap theory in particular, Barr was as hard on conservative commentators as on anyone.  Of the Scofield note on Genesis 1:1 he said "this is only a hidden way of saying that the biblical chronology is not true."  Of the NIV Study Bible's note that "Scripture does not reveal the exact date of Adam's creation," he said:  "These statements, coming from 'Bible-believers', are straightforward falsities:  because they no longer believe the chronology of the Bible, the editors assure their readers that the Bible tells them nothing about it."³

That is where the exegetical question ends and a different question begins.  Genesis 5 and 11 claim that Adam lived about 2,000 years before Abraham, with a margin of about 1,400 years depending on the text-form.  Whether that claim is true is not settled by Hebrew.  It is settled, if at all, by radiocarbon, by tree rings, by ice cores, by human genetics, and by the Egyptian and Mesopotamian king lists that worried Green in 1890, and this site argues those out in their own places.  What this article establishes is narrower and, for that reason, firmer.  The escape route through the text is closed.  Whoever wants an old humanity and an inspired Genesis has to hold, with Barr, that the numbers were meant and are wrong, or with Green, that Moses wrote a chronology he did not intend anyone to read.

Common objections

Objection 1: "Matthew 1 skips four kings. Genesis 5 could skip forty."

It could skip names.  Matthew 1:8 gives no ages, so dropping Ahaziah, Joash and Amaziah costs Matthew nothing but three names.  Genesis 5:9 gives Enosh's age when Kenan was born.  Drop forty men between Enosh and Kenan and Enosh is still 90 when Kenan is born, because the verse says so.  The example proves what Green used it to prove, that "father of" can mean "ancestor of."  It does not touch the years.  Goodenow saw this in 1893:  "the date of that birth being given, it matters not how many un-named generations intervene."²

Objection 2: "Exodus 6 has numbers and is compressed. So numbers do not prove completeness."

Correct, and this article concedes it.  But look at which numbers Exodus 6 has.  It gives Levi 137, Kohath 133, Amram 137, and no age for any of them at his son's birth.  That is exactly the shape a genealogy takes when the author wants lifespans and not a chronology, and it is not the shape of Genesis 5 and 11.  Sexton's rewriting of Genesis 5:9-11 in the Exodus 6 pattern shows how easy it would have been.⁷  The author of Genesis had the Exodus 6 form available and did not use it.

Objection 3: "Genesis never adds the numbers up. A chronology would have a total."

Green made this point and it is his best.⁴  Two replies.  First, Genesis never totals the postdiluvian lifespans either, and Green himself added begetting age to remaining years to get them, so that he could say the term of life was "narrowed down" after the Flood.  He used the numbers for one computation and forbade the other.⁷  Second, the text does add two of them.  Methuselah's 187, Lamech's 182 and Noah's 600 sum to Methuselah's 969, so that he dies in the Flood year.  If that is an accident, it is a remarkable one.  If it is not, the author expected addition.⁷

Objection 4: "The ten-and-ten pattern proves the lists are schematic, like the Sumerian King List."

Genesis 11 has nine names in the Hebrew.⁷  The Sumerian King List has seven, eight, nine or ten antediluvian kings depending on the tablet, and no tablet shares a name with Genesis.⁸  And a schematic list of names with dated births still sums to the same years.  The objection, if it worked, would show that names were chosen.  Green already had that.  It was never what he needed.

Objection 5: "Hebrew 'begat' can mean 'became the ancestor of.' The NIV margin says so."

It can, and Deuteronomy 4:25 and 2 Kings 20:18 prove it.  But "became the ancestor of" happens when the descendant is born.  A man does not become the ancestor of his great-grandson on the day his son is born;  he becomes it on the day the great-grandson is born.  So "When Enosh had lived 90 years he became the ancestor of Kenan" still dates Kenan's birth to Enosh's 90th year.  Sexton made the point against C. John Collins, who used that translation in 1994 and then switched in 2003 to "fathered an ancestor of," which is Green's Premise 2 again.⁷

Objection 6: "Green was a conservative who believed in inerrancy. He would not have invented a reading to escape science."

He said that he did.  The 1863 footnote opens with "if scientific research should ever demonstrate ... that the race of man has existed upon the earth for a longer period than the ordinary Hebrew Chronology will allow, we would be disposed to seek the solution in this ... characteristic of the Scriptural genealogies."²⁴  The 1890 article argues from Egyptian antiquity that the chronology "is based upon an unfounded assumption."⁴  Green was a conservative.  He was also frank about why he was looking.  Wenham, who also holds an old humanity, is at least as frank and reaches the opposite reading of the Hebrew.¹⁷

Objection 7: "Even if the numbers are a chronology, the Septuagint's are original, and that adds 1,400 years."

It might.  This article does not decide the text-form question and points to the article that does.  What it points out is that 1,400 years is the whole prize.  On the Septuagint's numbers creation is about 5,560 BC and the Flood about 3,300 BC.  That is still a young humanity by any anthropological standard, and it still requires the numbers to be summed, which is the point at issue here.  Sexton, who favours the Septuagint, says it exactly:  "We may question where the original chronology exists text-critically, but not whether it exists lexico-grammatically."⁷

Objection 8: "On the plain reading, Shem outlives Abraham. That is absurd, so the reading must be wrong."

Green raised this too.⁴  On the Masoretic numbers with Abram born in Terah's 70th year, Shem dies 35 years after Abraham does, and Eber later still.  Two replies, and neither is a full answer.  First, the absurdity is a judgement about what the author would have tolerated, not about what the words mean;  Genesis 25:8 says Abraham died "an old man and full of years," and the rabbis, who read the Masoretic numbers as a chronology, simply accepted that Shem was alive and had Isaac study at his academy.⁷  Second, the overlap is a real argument in the text-form debate, since the Septuagint and Samaritan numbers remove it, and Ronald Hendel and others think the shorter Masoretic chronology in Genesis 11 may be secondary for that reason.  That is an argument about which numbers are original, not about whether they are a chronology.

What this article does not settle

Three things are left open on purpose.

Whether the Masoretic, Samaritan or Septuagint numbers are original.  The spread is about 1,400 years and the arguments are two-sided.  See Old Testament text and chronology.

Whether Abram was born in Terah's 70th or 130th year.  The difference is 60 years and every table in this article says which it uses.

Whether the numbers are true.  Green's Egyptian argument has been replaced by radiocarbon, dendrochronology, ice cores and genetics, each of which this site examines on its own evidence.  If those clocks are right, the numbers are wrong, and no reading of wayyoled will rescue them.  This article claims only that the text means what Ussher thought it meant, that Green's gaps do not change that, and that a reader who rejects the chronology should say so, as Barr did, rather than say the Bible does not have one.

Sources