The Genre and Grammar of Genesis 1

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.¹¹

The interpretation of yom as aeon, a favourite resource of harmonists of science and revelation, is opposed to the plain sense of the passage, and has no warrant in Heb. usage (not even Ps. 90:4).²²

What is said here is intended to hold true entirely and exactly as it stands.  There is no trace of the hymnic element in the language, nor is anything said that needs to be understood symbolically.³³

Summary

This article asks one narrow question.  Read as Hebrew, with no science in the room, what does Genesis 1 claim about the length and order of its days?  The answer comes from counting words in an open morphological text, from the grammar of the first verse, from the verb chain that carries the chapter, and from what the scholars who reject a young earth say the author meant.  The findings:

1. The word yom ("day") occurs 2,304 times in the Hebrew Bible, and the circulating creationist figures are roughly right but not exact. A Deno script run over the Open Scriptures Hebrew Bible (Westminster Leningrad Codex 4.20, CC-BY) counts 2,304 tokens, of which 2,290 lie outside Genesis 1:1-2:3. Outside Genesis 1, yom stands next to a numeral 431 times in 354 verses; it shares a verse with "night" (laylah) 52 times; with "evening" (erev) 21 times; with "morning" (boqer) 24 times. Only four verses outside Genesis 1 put yom, erev and boqer together, but 38 verses pair erev with boqer, which is where the popular "38 times" figure comes from. The widely repeated "410 times with a number" is not reproducible by any rule tried; Stambaugh's 359 and this article's 354 verses are.⁴⁵⁶⁶⁷⁸
2. Every one of those numbered days is an ordinary day, with two exceptions that the young-earth literature itself concedes. Hosea 6:2 ("after two days ... on the third day") is an idiom for "soon," and Zechariah 14:7 ("one day, known to the LORD, not day and not night") is an eschatological day with no evening. Gerhard Hasel, the Adventist Hebraist who built the strongest form of the numeral argument, prints the Zechariah exception himself. Neither exception has an evening-and-morning refrain, and neither is a counted series.⁶⁴
3. The semantic-range argument does not transfer to Genesis 1. Yom can mean daylight (Gen 1:5a), a calendar day (Gen 1:5b), or a period ("in the day that the LORD God made," Gen 2:4). But Genesis 2:4 uses beyom with an infinitive, a Hebrew idiom for "when," and Psalm 90:4 and 2 Peter 3:8 are similes that compare God's time to human days. A simile needs its terms to keep their ordinary sense. John Skinner, the critical commentator of the 1910 International Critical Commentary, said this in one line: "no warrant in Heb. usage (not even Ps. 90:4)."²⁹
4. Exodus 20:11 and 31:17 are the earliest surviving interpretation of Genesis 1, and they read the days as the days of a working week. "For in six days (sheshet yamim) the LORD made heaven and earth ... and rested the seventh day. Therefore the LORD blessed the Sabbath day." The same word carries the six work-days of Israel and the six days of creation in one sentence. Skinner again: to read the days as aeons "destroys the analogy on which the sanction of the sabbath rests."²⁶
5. Genesis 1:1 can be read as an independent sentence or as a dependent clause, and the choice does not change the days. Rashi (d. 1105) and Ibn Ezra (d. 1167) read bereshit as a construct, "when God began to create," and the NJPS and NRSV follow them. Robert Holmstedt argued in Vetus Testamentum 2008 that the verse is a restrictive relative clause and that an "absolute beginning" "cannot be derived from the grammar." Cassuto, Westermann, and Wenham reply that verse 2 is a circumstantial clause and verse 1 is a main sentence, which Westermann calls "a cry of praise." The dispute touches creatio ex nihilo and whether 1:1 is a title or a first act (Waltke versus Poythress). It leaves the six numbered days exactly where they were.^{1011121213131414151516}
6. The verbs of Genesis 1 are the verbs of Hebrew narrative, and Steven Boyd's RATE II statistic reproduces exactly. Genesis 1:1-2:3 has 55 wayyiqtol ("and then he ...") forms among 84 finite verbs, a ratio of 0.65. This article's own count from the morphological text gives the same 55 and the same 84, and the same for Boyd's comparison texts (Exodus 14, Judges 4, Exodus 15, Judges 5, Psalm 104). Narrative prose runs 0.45-0.67; poetry runs 0.02-0.10. Genesis 1 sits with Exodus 14 and Judges 4.^17171718
7. But Boyd's statistic answers a question that serious critics do not ask. Von Rad, Westermann, Wenham and Collins all say Genesis 1 is prose, not poetry; Wenham calls it "elevated prose," Collins "exalted prose narrative." The live question is whether narrative form implies historical intent. Psalm 106, undisputed poetry, scores 0.675 on Boyd's ratio, higher than Genesis 1. Boyd's own sample flags the historical psalms as a "special case." Narrative syntax shows the author wrote a sequence of past events; it cannot show by itself that he meant them as history rather than parable.^31519171820
8. The scholars who reject a young earth agree, on the record, that the author meant ordinary consecutive days. James Barr (Oxford Regius Professor of Hebrew, no friend of creationism) said so in a 1984 letter whose provenance is imperfect, and again in print in 1987: "they were literally intended." Hugh Williamson, his successor in the Oxford chair, wrote in 2011 that Barr "was correct." Emanuel Tov: "For the biblical people this was history." Skinner in 1910, Marcus Dods in 1888, von Rad in 1961 and 1972, Charles Hodge in 1871, J. Barton Payne in 1962, Gleason Archer in 1985, and Pattle Pun in 1987 all concede that the plain sense is six ordinary days, and the last four then reject it on the authority of geology.^{2112222222232425262728}
9. The plain reading has real problem spots, and they are stated here rather than hidden. Light exists on day one and the sun on day four. The seventh day has no "evening and morning." Genesis 1:5 says "one day," not "the first day." Genesis 2:5 explains the absence of shrubs by the absence of rain and of a man, which Meredith Kline (1958) read as proof that ordinary providence governed the creation period, so that a 24-hour week becomes "ludicrous." Each has an answer from the young-earth side; none of the answers is as clean as the numeral argument.^{292929293031319}
10. The grammar establishes intent, not truth. Barr's distinction is the honest end of the argument. The Hebrew shows what the writer meant. Whether the writer was right is a question for the dating articles on this site, not for this one.¹

Verdict:  Judged by grammar and genre alone, the author of Genesis 1 meant six ordinary, consecutive days followed by a seventh, and the Sabbath commandment read him that way within the same Pentateuch.  The day-age, framework and analogical readings are rescues of the text from geology, and their own proponents say so.  The text does not, and cannot, settle whether its author was right.

The word yom, counted from an open text

The young-earth argument about yom is usually stated as a set of numbers.  Ken Ham's version, in the 2007 New Answers Book, reads:  "Outside Genesis 1, yom is used with a number 359 times, and each time it means an ordinary day ... yom is used with the word 'evening' or 'morning'" 23 times, "'Evening' and 'morning' appear in association, but without yom, 38 times," and "'night' is used with yom 53 times."⁷  The Creation Answers Book gives 410 for the numeral count and 52 for night.⁸  A Facebook post by Richard Swindell, the version that reached this project, has 2,301 total, 410 with a number, 38 with evening and morning, 23 with morning, 23 with evening, and 52 with night.  The numbers have been copied from source to source for thirty years without anyone publishing the rule that produced them.

They trace to two papers.  James Stambaugh, writing in the Creation Ex Nihilo Technical Journal in 1991, counted yom 2,291 times, and "with a number 359 times outside of Genesis 1," which he broke down as 189 plural-with-cardinal ("three days"), 162 singular-with-ordinal after a preposition ("on the seventh day"), and 7 others.⁵  Gerhard Hasel, in Origins 1994, counted 2,304 occurrences, 1,452 singular, and yom "together with a numeral" 150 times.⁶  Robert McCabe in 2000 repeated Hasel's 2,304 and 1,452, put the numeral count "in excess of 350" with about 150 singular, and counted "evening" and "morning" together with yom 19 times beyond Genesis 1 and without yom 38 times.³²³²  Ham's 359 is Stambaugh's; his 38 is McCabe's; nobody's is 410.

Rather than copy any of them, this article counted.  The corpus is the Open Scriptures Hebrew Bible (morphhb), a CC-BY morphological tagging of the Westminster Leningrad Codex 4.20, read from a clone of the GitHub repository at commit 3d15126 (27 August 2024).³³  The script, research/yom-count.ts, is a 150-line Deno program that walks all 23,213 verses of the 39 books, finds every word whose lemma is Strong's H3117 after stripping prefixes, and records for each token whether a cardinal or ordinal numeral stands immediately before or after it, whether the token is singular or plural, and whether erev (H6153), boqer (H1242) or laylah (H3915) occurs anywhere in the same verse.  Its full output, with every verse listed, is in research/yom-count-output.md.⁴  The rules are stated so that anyone can disagree with them.

Three things stand out.  The total, 2,304, is Hasel's and McCabe's figure to the token.  The "night" count, 52 verses, is the circulating figure exactly.  And the "38" is real, but it is not what Swindell's post says it is:  38 verses outside Genesis 1 contain both erev and boqer, and in 34 of them yom is absent.⁴  The phrase "evening and morning" with the word "day" in the same verse occurs only four times outside Genesis 1 (Lev 6:20, Num 9:15, Deut 16:4, Dan 8:26), and one of those, Daniel 8:26, is the disputed "vision of the evenings and mornings" whose 2,300 units some commentators halve to 1,150 days.  A careful young-earth writer should say "evening and morning mark a day 38 times" and not "yom with evening and morning occurs 38 times."

The numeral count is the interesting one.  Stambaugh's 359 and this article's 354 verses are the same count by different hands.  The 431 tokens are higher because some verses carry two numbered days ("on the first day ... on the seventh day").  The 410 that has spread through the tracts sits between the two and matches no rule tried here:  not tokens, not verses, not singulars, not any combination.  It should be retired in favour of a number with a published rule behind it.

The singular-plus-cardinal row, 114, deserves a note because the older tallies fold it away.  Hebrew uses a singular noun after the tens ("arba'im yom," forty day, Gen 7:12; Deut 9:18) and after "one" ("yom echad," Gen 1:5; 1 Sam 2:34; Ezra 10:17).  Every one of these is a plain day or a plain count of days.  The script prints the whole list.

What the numbers show, and what they do not

The claim the numbers support is narrow.  When a Hebrew writer put yom beside a numeral, he meant a day of the ordinary kind:  a day of Jericho's siege, a day of a feast, a day of a journey.  This holds across 354 verses with two arguable exceptions, and the young-earth literature knows both.

The first is Hosea 6:2, "After two days he will revive us; on the third day he will raise us up."  The construction is dual plus ordinal, and the sense is "very soon."  Ham's footnote concedes the objection and answers that the idiom only works "if 'day' is understood in its normal sense," which is fair:  "in a couple of days" means soon precisely because a day is short.⁷  The second is Zechariah 14:7, "there shall be one day (yom echad), known to the LORD, not day and not night."  Hasel, whose paper is the most rigorous statement of the numeral rule, writes:  "The only exception in numbers of 1-1,000 is found in an eschatological text in Zechariah 14:7."⁶  A day with no night is a day in which the ordinary rhythm has stopped, and the verse says so.

Neither exception looks like Genesis 1.  Neither is a counted series of seven.  Neither is closed by "and there was evening and there was morning."  Neither is picked up by a law that tells men to work six days and rest one because God did.  The honest statement of the rule is therefore not "always," but "always, outside one idiom for 'soon' and one prophecy of a day without night."  That is still a rule.

The rule does not show that the days were 24 hours of 3,600 seconds.  It shows that the writer used the word for a day the way every other Hebrew writer did when he numbered it.  Andrew Steinmann's study of echad in Genesis 1:5 adds a detail:  the first day is not "the first day" but "one day," with no article, and the second through fifth days also lack the article, so that the text reads "an evening and a morning:  one day," which is a definition.  "Evening + morning = one day."⁹  Steinmann is a Lutheran Hebraist and a young-earth creationist, but the grammatical observation is checkable by anyone with the morphological text:  in Genesis 1:5 the numeral is tagged as a cardinal (Acmsa), and in 1:8, 13, 19, 23 the ordinal carries no article.

The semantic-range argument does not transfer

The old-earth reply to the count is that yom has a wide range.  Gleason Archer, who defended inerrancy and rejected the plain reading, leaned on Genesis 2:4, where yom "is apparently used ... to refer to the whole creative process just described in Genesis 1 as taking up six days."²⁷  Progressive creationists add Psalm 90:4 and 2 Peter 3:8, "with the Lord one day is as a thousand years, and a thousand years as one day," a text Hasel says "has been invoked time and again" for long creation days.⁶  Charles Hodge made the fuller list in 1871:  yom is used "sometimes for the time the sun is above the horizon; sometimes for a period of twenty-four hours; sometimes for a year, as in Lev. xxv. 29 ... sometimes for an indefinite period, as in the phrases, 'the day of your calamity,' 'the day of salvation,' 'the day of the Lord.'"²⁵

All of this is true, and none of it reaches Genesis 1.  A word's range is the set of things it can mean somewhere.  What it means in a given sentence is fixed by the sentence.  English "day" can mean daylight ("it was still day"), a calendar unit ("three days"), or an era ("in Caesar's day"), and no English speaker hesitates over "he worked six days and rested on the seventh."  The Hebrew markers that fix the sense in Genesis 1 are the ones the count measured:  a numeral, a singular, and an evening-and-morning refrain.  Hasel called this the "triple interlocking connection."⁶

Genesis 2:4 is the strongest of the old-earth proof texts and the weakest on inspection.  The Hebrew is beyom asot YHWH elohim erets veshamayim, "in-day-of making-of the LORD God earth and heaven."  Yom is in the construct state and is followed by an infinitive construct (asot, morph Vqc).³³  That construction, beyom + infinitive, is a fixed idiom meaning "when," used dozens of times with no reference to a day's length (Gen 2:17, "in the day you eat of it"; Num 3:1; 2 Sam 22:1).  The ESV translates 2:4 "in the day that" and the NIV "when."  It is no more evidence that yom means an age than "in the day of King Herod" is evidence that Herod reigned for one day.

Psalm 90:4 is a simile.  "A thousand years in your sight are but as yesterday when it is past, or as a watch in the night."  The comparison only works if a day is short and a thousand years is long; it says that God is outside both.  2 Peter 3:8 quotes it and adds the reverse clause, "and a thousand years as one day," which no one takes to mean that a thousand-year period in Scripture is a 24-hour day.  Hasel's treatment of this verse fills a page and ends where Skinner ended eighty years earlier:  "no warrant in Heb. usage (not even Ps. 90:4)."⁶²

The semantic-range argument, then, proves that yom could have meant an age if the author had wanted it to.  The count shows how he signalled that he did not.  Hodge saw this in 1871 and said so:  "It is of course admitted that, taking this account by itself, it would be most natural to understand the word in its ordinary sense; but if that sense brings the Mosaic account into conflict with facts, and another sense avoids such conflict, then it is obligatory on us to adopt that other."²⁵  That is not a linguistic argument.  It is a decision about which authority wins, and Hodge was candid that geology was the authority.

Exodus 20:11 is the earliest commentary on Genesis 1

The Pentateuch interprets its own creation account twice, both times in the Sabbath law.  Exodus 20:9-11:  "Six days you shall labor, and do all your work, but the seventh day is a Sabbath to the LORD your God ... For in six days the LORD made heaven and earth, the sea, and all that is in them, and rested the seventh day.  Therefore the LORD blessed the Sabbath day and made it holy."  Exodus 31:17:  "in six days the LORD made heaven and earth, and on the seventh day he rested and was refreshed."

The Hebrew of 20:11 is ki sheshet-yamim asah YHWH, "for six days made the LORD," the same sheshet yamim that opens verse 9, "six days you shall labor."³³  One word, in one sentence, carries the six days of a farmer's week and the six days of creation, and the sentence's logic depends on their being the same kind of day.  The command is "work six, rest one, because God worked six and rested one."  If God's six were six ages, the command would be "work six ages."  Whatever Genesis 1 was, the writer of Exodus 20 read it as a week.

This is a datum, not an inference.  It is the earliest reading of Genesis 1 that survives, older than any rabbi, any Father, any lexicon.  Its weight does not depend on Mosaic authorship; on the standard critical dating the Decalogue and the Priestly creation account are both pre-exilic or exilic, and the Sabbath law shows how the creation week was understood by the people who kept it.  Skinner, no conservative, put the point negatively:  reading the days as aeons "not only destroys the analogy on which the sanction of the sabbath rests, but misconceives the character of the Priestly Code.  If the writer had had aeons in his mind, he would hardly have missed the opportunity of stating how many millenniums each embraced."²  Hasel, from the other side of the theological street, drew the same line:  the two Sabbath passages "inform the reader how the creation 'days' were understood by God."⁶

The old-earth readings all have to handle this text, and the handling is where the strain shows.  The day-age view must say the analogy is between six periods and six days, which works as an analogy but not as the verse's grammar.  The framework view must say that the "week" in Exodus is a metaphor for a metaphor.  The analogical-days view, C. John Collins's, says God's work-days are the pattern for ours without being the same length, which is a real reading but one that Exodus 20:11 does not signal.  None of these is impossible.  All of them are harder than the plain one.

Genesis 1:1: one sentence or a dependent clause

The first verse is the most argued sentence in the Hebrew Bible, and the argument is older than modern science.  The Hebrew is bereshit bara elohim et hashamayim ve'et ha'arets.³³  The problem is the first word.  Bereshit has no definite article, so a mimetic rendering is "in beginning," and a noun without the article in this position is often a construct, "in the beginning of ...".

Rashi (Solomon ben Isaac, d. 1105) took it that way.  "If, however, you wish to explain it in its plain sense, explain it thus:  At the beginning of the Creation of heaven and earth when the earth was without form and void and there was darkness, God said, 'Let there be light'.  The text does not intend to point out the order of the acts of Creation ... Because, wherever the word reshit occurs in Scripture, it is in the construct state."¹⁰  He cites Jeremiah 26:1, "in the beginning of the reign of Jehoiakim," and Hosea 1:2, where a construct noun stands before a finite verb.  Ibn Ezra (d. 1167) agreed that bereshit is a construct, disagreed with Rashi on the main clause (for Ibn Ezra it is verse 2, "the earth was"), and added that Genesis 1:1 "does not tell us that God created heaven and earth on the first day; neither does it deal with the creation of matter."¹¹

The New Jewish Publication Society translation of 1985 follows Rashi:  "When God began to create heaven and earth," with the footnote "Others 'In the beginning God created.'"³⁴  The NRSV of 1989 has "In the beginning when God created the heavens and the earth, the earth was a formless void."³⁵  Robert Holmstedt gave the dependent reading its most careful modern defence in Vetus Testamentum 2008.  His claim is that Genesis 1:1 is an unmarked restrictive relative clause, "in the beginning (in which) God created," and that such clauses in Hebrew are always restrictive, so that "the traditional understanding of a reference to an 'absolute beginning' cannot be derived from the grammar of the verse."  His conclusion:  "the grammar of Gen. i 1 points forward only; it does not comment about whether this basic creative event was unique or whether there were others like it."¹²¹²

The independent reading has the majority of modern commentators.  Wenham lists four options and adopts the fourth, "V 1 is a main clause describing the first act of creation," noting that "on these grounds most modern commentators agree that v 1 is an independent main clause to be translated 'In the beginning God created.'"  His reasons:  the absence of the article does not make a noun construct, since "temporal phrases often lack the article (e.g., Isa 46:10; 40:21; 41:4, 26; Gen 3:22; 6:3, 4; Mic 5:1; Hab 1:12)," and reshit "may well have an absolute sense in Isa 46:10."¹⁵  Cassuto argued from the syntax of verse 2 rather than verse 1.  If verse 1 were a temporal clause, verse 2 would have the verb first, hayetah ha'arets, as in Jeremiah 26:1 and Hosea 1:2.  Instead it has subject first, veha'arets hayetah, which "proves ... that v. 2 begins a new subject.  It follows, therefore, that the first verse is an independent sentence that constitutes a formal introduction to the entire section."¹³  Westermann, the German critical commentator, reached the same place from tradition history:  verse 1 "has no parallel in the other creation stories," it was "prefixed" by the Priestly writer, and "the style of the chapter leaves us to expect that v. 1 is a main sentence."  He quotes Wellhausen on the dependent reading:  "this translation is desperate."¹⁴

So the grammar is genuinely two-sided.  The dependent reading has Rashi, Ibn Ezra, Holmstedt, the NJPS and the NRSV; the independent reading has Cassuto, von Rad, Westermann, Wenham, and most English versions.  A skeptic should notice that the dependent reading is not a modern invention to make room for pre-existing matter; it is a medieval Jewish reading made on Hebrew grounds.

Bara and creation from nothing

What hangs on the choice is creatio ex nihilo, not the days.  On the dependent reading, the story opens with an earth already "without form and void" and says nothing about where it came from.  On the independent reading, verse 1 states that God made the heavens and the earth and verse 2 describes their first state.  Wenham:  "apart from #4, the translations all presuppose the existence of chaotic pre-existent matter before the work of creation began."¹⁵

The verb does not settle it.  Bara is used only with God as subject, but Ibn Ezra already denied that it means creation from nothing:  "Most Biblical commentators explain that the word bara (created) indicates creation ex nihilo ... However, they have overlooked And God created (va-yivra) the great sea monsters (Gen. 1:21)," which were made from the waters, and Isaiah 45:7, "I form the light, and create darkness."  "Hence bara does not necessarily imply creation ex nihilo."¹¹  The doctrine rests on the whole sentence and on later texts (Ps 33:6, Heb 11:3, Rom 4:17), not on the lexicon.

Summary or first act

Among those who read verse 1 as independent, a further split.  The majority (Wenham names Driver, Gunkel, von Rad, Cassuto, Westermann, and others) take it as a title that summarizes verses 2-31, with the actual creating beginning at verse 3.¹⁵  Westermann:  "The sentence in 1:1 is not the beginning of an account of creation, but a heading that takes in everything in the narrative in one single sentence, and it is much more than a mere heading ... the first sentence itself is really a cry of praise."¹⁴  Bruce Waltke defended the summary view at length in Bibliotheca Sacra in 1975.  Vern Poythress answered him in the Westminster Theological Journal in 2017 with the argument that Hebrew headings carry a signal ("These are the generations of ...," "The words of ...") and 1:1 carries none; that 1:1 has a main verb in the perfect, which "can naturally be construed as describing the first event in a series"; and that the word "earth" links 1:1 to 1:2 as a circumstantial clause.  His verdict:  "all three of the main arguments for the summary view have superficial plausibility, but none has weight."¹⁶

Why does any of this matter to an article about the days?  Because the gap and "old universe, young earth" readings live in the space between 1:1 and 1:3.  If 1:1 is an act, the heavens and earth were made at the start of day one, and the whole cosmos is as old as the week.  If 1:1 is a title, or a dependent clause, the text is silent on when the raw material came to be, and only the ordering work is dated.  The text supports either; it does not adjudicate.  What it does not permit is stretching the six days themselves, which is where the numerals and the refrain sit.

The genre: narrative verbs, elevated style

Hebrew has a verb form that carries the spine of a story.  The wayyiqtol ("and-then-he-did"), also called the waw-consecutive imperfect or the preterite, is the form that links one past event to the next:  "and God said ... and there was ... and God saw ... and God separated ... and God called."  Genesis 1 is built of it.  Poetry rarely uses it; a psalm says what God does, not what he did next.

Boyd's statistic, reproduced

Steven Boyd, then Associate Professor of Bible at The Master's College, turned this into a number for the RATE II volume in 2005.  He counted the four finite verb forms of Hebrew (preterite, imperfect, perfect and waw-perfect) in a stratified random sample drawn from a population of 295 narrative texts and 227 poetic texts, and fitted a logistic regression of "narrative or poetry" on the ratio of preterites to finite verbs.  The model reduced classification errors by more than 96% in the sample.  "For Genesis 1:1-2:3, this probability is between 0.999942 and 0.999987 at a 99.5% confidence level.  Thus, we conclude with statistical certainty that this text is narrative, not poetry."¹⁷¹⁷  His data table gives Genesis 1:1-2:3 55 preterites, 14 imperfects, 13 perfects and 2 waw-perfects, 84 finite verbs, ratio 0.65.¹⁷

Boyd's counts were made by hand with Bible software.  This article re-counted them from the morphological text with a second Deno script, research/finite-verb-ratio.ts, which reads every verb tag in a passage and sorts it by conjugation.¹⁸  The results match Boyd to the verb.

The counting rule is Boyd's:  jussives and cohortatives are folded into the imperfects (so Gen 1:3 yehi or, "let there be light," is an imperfect), imperatives and participles and infinitives are excluded.  Under that rule Genesis 1 gives 55 of 84, Exodus 14 gives 48 of 84, Judges 4 gives 44 of 74, the Song of the Sea gives 2 of 50, the Song of Deborah 3 of 65, Psalm 104 1 of 62, and Psalms 105-106 together 77 of 144, each one identical to Boyd's printed figure.¹⁷¹⁸  The exercise took an afternoon.  Anyone with the CC-BY text can repeat it.

So the first thing to say about Boyd's work is that it is right.  Genesis 1 uses the verb system of Hebrew narrative prose, and it does so at the high end of the narrative range.  Anyone who says Genesis 1 "is poetry" is wrong on the syntax, and Boyd showed it with numbers.

What the statistic cannot show

The second thing to say is that nobody with standing says Genesis 1 is poetry.  Gerhard von Rad, whose Priestly-source reading is the modern critical one, wrote in 1961:  "Nowhere at all is the text only allusive, 'symbolic,' or figuratively poetic," and in the 1972 revision:  "What is said here is intended to hold true entirely and exactly as it stands.  There is no trace of the hymnic element in the language, nor is anything said that needs to be understood symbolically."²⁴³  Westermann calls verse 1 "a cry of praise" but treats the chapter as Priestly narrative.¹⁴  Wenham:  "Gen 1 is not typical Hebrew poetry ... On the other hand, Gen 1 is not normal Hebrew prose either; its syntax is distinctively different from narrative prose ... In that it is elevated prose, not pure poetry, it seems unlikely that it was used as a song of praise as the psalms were."¹⁵  Cassuto explained the elevation as "the special importance of the subject."¹³  Collins, whose analogical-days reading the young-earth writers reject, coined the phrase that has stuck:  "the genre of this pericope is what we may call exalted prose narrative.  This name for the genre will serve us in several ways.  First, it acknowledges that we are dealing with prose narrative ... Finally, this name is far better than the term some use, 'poetry,' since that term is ambiguous; if we say Hebrew poetry we might go on to speak of parallelism, which is absent."¹⁹

Boyd's model separates narrative from poetry.  The readers he wants to refute (Waltke, Collins, the framework school, Walton) already grant that Genesis 1 is narrative.  What they deny is that narrative form fixes historical intent, and the statistic cannot reach that question.  Boyd's own table shows why.  Psalm 106 is poetry by every measure of Hebrew poetics, and it scores 54 preterites in 80 finite verbs, 0.675, above Genesis 1.  Psalm 105 scores 0.36.  Boyd knew this and set the historical psalms aside as "a special case, in which narrative clauses are juxtaposed in poetic bicolons."¹⁷  That is an honest note, but it concedes the point:  a high preterite ratio marks a text that narrates past events in sequence.  It does not mark whether the events happened.  A parable told in wayyiqtol chains (Nathan's story of the ewe lamb, 2 Sam 12:1-4; Jotham's trees, Judg 9:8-15) scores as narrative.  Kenneth Turner, an Old Testament professor whose doctorate was on Deuteronomy, made this objection in a 2022 discussion of Boyd's paper:  "History is not a literary genre, but a way of referring to the past ... A parable, for example, is written in prose narrative, full of wayyiqtol verbs.  Yet, Boyd's analysis would conclude these obviously-non-literal texts as 'historical narrative.'"²⁰  Sarfati's 2019 reply to a similar challenge grants the exceptions and answers that "rare exceptions should not override the vast majority of cases," which is true and beside the point; the question is not what most narratives are but what this one is.³⁶

So the syntax supplies one premise and not two.  Premise one:  the author wrote a sequence of past events, one after another, in the form Hebrew uses for such sequences.  Premise two, that he meant them as history rather than as a story told for another purpose, comes from elsewhere:  from the numbered days and the refrain, from the genealogies that run from Adam to Abraham without a seam, from Exodus 20:11, and from how every ancient reader took it.  Boyd's conclusion is right; his proof carries less of it than he claims.

Parallelism and the absence of poetic markers

Hebrew poetry is marked by parallelism (a second line that restates, extends or reverses the first), by terseness, by ellipsis of the article and the object marker, and by a low rate of the preterite.  Genesis 1 has the article and the object marker in nearly every clause (et hashamayim ve'et ha'arets), has no parallel bicola except at 1:27 ("in the image of God he created him; male and female he created them") and 2:23 in the next chapter, and has the preterite ratio of a battle narrative.¹⁵¹⁸  What it has instead is repetition:  "and God said" ten times, "and it was so" six, "and God saw that it was good" seven, "and there was evening and there was morning" six, and a first verse of exactly seven words.  That is structure, and Wenham and Cassuto and Collins all call it deliberate.  Structure is not poetry, and a structured account is not a non-historical one; the Flood narrative is a palistrophe and the book of Kings is a formula.  The framework readers are right that the chapter is patterned and wrong that pattern excludes sequence.  The seven-day pattern is the sequence.

What the scholars who reject a young earth say the text means

The strongest evidence for the plain reading is not a creationist's count.  It is the concession of the people who do not want it to be true.

James Barr.  Barr was Oriel Professor of the Interpretation of Holy Scripture at Oxford from 1976 and Regius Professor of Hebrew there from 1978 to 1989, the author of Fundamentalism (1977), and a lifelong critic of conservative evangelicalism.  On 23 April 1984 he answered a letter from David C. C. Watson, a creationist, and the answer has been quoted by every young-earth ministry since.  The full text, as transcribed by Stephen E. Jones from a copy supplied by Answers in Genesis Australia, reads:  "probably, so far as I know, there is no professor of Hebrew or Old Testament at any world-class university who does not believe that the writer(s) of Genesis 1-11 intended to convey to their readers the ideas that (a) creation took place in a series of six days which were the same as the days of 24 hours we now experience (b) the figures contained in the Genesis genealogies provided by simple addition a chronology from the beginning of the world up to later stages in the biblical story (c) Noah's flood was understood to be world-wide and extinguish all human and animal life except for those in the ark."  Barr then qualified:  most professors "may avoid much involvement in that sort of argument," one might "find one or two people who would take the contrary point of view," and "it's really not so much a matter of technical linguistic competence, as of appreciation of the sort of text that Genesis is."²¹

The letter's provenance is imperfect and this article says so.  No facsimile has been published.  The only full text is Jones's transcription, and Jones, who is not a young-earth creationist, posted it to argue that the quotation is over-read.²¹  The creationist quote pages reproduce the three points and drop every qualification.³⁷  James McGrath, a critic, asked in 2012 whether anyone could "trace it to a verifiable source," and then added:  "I think that the quote from Barr is correct and would be correct today."³⁸  The letter's internal details check:  it mentions "another book coming out soon, Escaping from Fundamentalism, SCM Press London," and that book appeared from SCM in 1984 with an American edition under a different title, as the letter predicts.

The letter is also dispensable, because Barr said the same thing in print.  His Ethel M. Wood Lecture of 4 March 1987, Biblical Chronology:  Legend or Science?, is about the genealogies, but its central distinction is the one this article turns on:  "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."  And:  "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."¹

Hugh Williamson, Emanuel Tov, Peter Williams.  In 2010-2011 Jud Davis, a Greek professor at Bryan College, sent Barr's letter to living Hebraists and asked whether it was still true.  Hugh Williamson, Regius Professor of Hebrew at Oxford from 1992 to 2014, replied by email on 7 January 2011:  "So far as the days of Genesis 1 are concerned, I am sure that Professor Barr was correct ... I have not met any Hebrew professors who had the slightest doubt about this unless they were already committed to some alternative by other considerations that do not arise from a straightforward reading of the Hebrew text as it stands."²²  Emanuel Tov of the Hebrew University, editor-in-chief of the Dead Sea Scrolls publication project, replied on 28 December 2010:  "For the biblical people this was history, difficult as it is for us to accept this view."²²  Peter Williams, then Warden of Tyndale House, Cambridge, sent Davis a 2008 conference paper with the sentence:  "Although the Young Universe Creationist position is not widely held within secular academia, the position, that the author of Genesis 1 maintained that the world was created in six literal days, is nearly universally held."²²  These are emails and an unpublished paper, reported by a creationist, and a skeptic is entitled to weigh them accordingly.  They are also on the record with the names attached, and no one named has disowned them.

The critical commentators.  The point is older than Barr.  John Skinner, Principal of Westminster College, Cambridge, wrote the ICC Genesis in 1910 from a fully source-critical position, and on 1:5 he wrote the note quoted at the head of this article:  reading yom as aeon "is opposed to the plain sense of the passage, and has no warrant in Heb. usage."²  Marcus Dods, the Free Church liberal who became professor of New Testament at New College, Edinburgh, the next year, wrote in the Expositor's Bible in 1888 that the harmonizing schemes "do violence to Scripture, foster a style of interpretation by which the text is forced to say whatever the interpreter desires ... If, for example, the word 'day' in these chapters, does not mean a period of twenty-four hours, the interpretation of Scripture is hopeless."²³  Von Rad's two sentences are above.³  None of these men thought Genesis 1 was true as history.  All of them thought it meant what it said.

The evangelicals who took the old earth anyway.  A second class of witness is the conservative who admits the plain sense and then leaves it.  Charles Hodge of Princeton, 1871:  "It is of course admitted that, taking this account by itself, it would be most natural to understand the word in its ordinary sense; but if that sense brings the Mosaic account into conflict with facts, and another sense avoids such conflict, then it is obligatory on us to adopt that other."²⁵  J. Barton Payne, 1962, on the non-consecutive-days theory:  "But confessedly, it would not have been as readily deduced from the Genesis text had it not been for the evidences advanced by secular science."²⁶  Gleason Archer, 1985:  "From a superficial reading of Genesis 1, the impression received is that the entire creative process took place in six twenty-four-hour days.  If this was the true intent of the Hebrew author (a questionable deduction, as will be presently shown), this seems to run counter to modern scientific research, which indicates that the planet Earth was created several billion years ago."²⁷  Pattle Pun, a Wheaton biologist, 1987:  "It is apparent that the most straightforward understanding of the Genesis record, without regard to all of the hermeneutical considerations suggested by science, is that God created heaven and earth in six solar days, that man was created in the sixth day, that death and chaos entered the world after the Fall of Adam and Eve, that all of the fossils were the result of the catastrophic universal deluge which spared only Noah's family and the animals therewith."²⁸

Every one of these was checked against its printed page for this article, and one that is often added to the list was not.  James Montgomery Boice's Genesis commentary (1982) is quoted by Sarfati and by Mortenson as saying "the exegetical basis of the creationists is strong."  The only copy reachable was the secondary quotation, and so the sentence is held in this article's quarantine.md rather than printed as Boice's.

What the list shows is not that these men were wrong to believe in an old earth.  It shows where the old-earth readings of Genesis 1 came from.  They came from geology, and their authors said so.  That is the historical fact that the day-age, gap and framework readings have to carry, and it is the reason Barr called the "not to be taken literally" move "cowardly."

Where the plain reading has real problems

A reading that is right about the numerals can still have loose ends, and a skeptic will find them.  Here are the four that matter, with the young-earth answers stated and graded.

Light on day one, the sun on day four

"And God said, 'Let there be light,' and there was light" (Gen 1:3).  "And God made the two great lights" (Gen 1:16).  The days are counted by evening and morning from the first day, three days before the sun.  Kline used this against the literal week:  the "evening" and "morning" of the first three days "must be figurative for they are mentioned as features of the three 'days' before the text records the creation of those lights in the firmament of heaven which were to divide the day from the night."²⁹

The young-earth answer has two parts.  First, the text itself defines day and night in verse 5 before the sun exists, and nothing in the definition requires the sun; the sun is appointed in verse 14 "to rule" a day-night cycle that already runs.  Sarfati's phrase is "directional light and a rotating earth," and he quotes Calvin on 1:3:  "the Lord, by the very order of the creation, bears witness that he holds in his hand the light, which he is able to impart to us without the sun and the moon."³¹  Second, the reading is the ancient one.  Basil of Caesarea, preaching on Genesis 1:5 about AD 370, took "one day" as a definition, "twenty-four hours fill up the space of one day, we mean of a day and of a night," three homilies before he reached the sun.³⁹  The answer is coherent.  It is also an answer that the text forces on the reader rather than one the text offers, and a skeptic is entitled to say that a writer who wanted to describe ordinary solar days chose a strange place to start.  Grade:  the objection does not touch the meaning of yom, but it does show the author was not writing a naturalistic sequence, and both sides should admit it.

The seventh day has no evening and morning

Six days end with the refrain.  The seventh (Gen 2:1-3) does not.  Hebrews 4:3-11 treats God's rest as still open to be entered.  The framework and day-age readers infer that the seventh day is unending, and that if one day of the week is an age, the others may be.  Kline:  "The word 'day' must be figurative because it is used for the eternity during which God rests from his creative labors."²⁹

The young-earth answer:  the refrain marks the close of a work-day, and the seventh had no work to close.  Sarfati adds that if the missing refrain proves the day never ended, "to be consistent, it would follow that the seventh day had not begun either," since the refrain marks both ends; and that God's rest from creating continues without the day continuing, "the Father still works, but He is not creating."³¹  The seventh day is also numbered, three times (Gen 2:2, 2, 3), with the same ordinal the sixth day carries, and Exodus 20:11 makes it the pattern for a 24-hour Sabbath.  Grade:  the answer is adequate, but the day-age reader has a real foothold here, and the honest statement is that the text leaves the seventh day's end unmarked.

"Because it had not rained" (Genesis 2:5-6)

"When no bush of the field was yet in the land and no small plant of the field had yet sprung up, for the LORD God had not caused it to rain on the land, and there was no man to work the ground, and a mist was going up from the land and was watering the whole face of the ground" (Gen 2:5-6).  Meredith Kline's 1958 article built the framework interpretation on this verse.  His argument:  the verse gives "a perfectly natural explanation" for the absence of plants, namely no rain and no farmer, and so "the unargued presupposition of Gen. 2:5 is clearly that the divine providence was operating during the creation period through processes which any reader would recognize as normal in the natural world of his day."²⁹  If ordinary providence ran the creation period, a 24-hour third day on which "continents just emerged from under the seas" dry out and sprout is impossible:  "the twenty-four-hour day theorist must think of the Almighty as hesitant to put in the plants on 'Tuesday' morning because it would not rain until later in the day!"  Kline called the result "ludicrous" and the literal week "incompatible with belief in the self-consistency of the Scriptures."²⁹  From this he concluded that the days are "a figurative chronological framework" and "the literalness of the sequence is no more sacrosanct than the literalness of the duration."²⁹  Mark Futato extended the argument in 1998 by reading the ed of verse 6 as "rain cloud," so that rain arrives in verse 6 and the day-four luminaries are what make rain possible.⁴⁰

This is the best old-earth argument from the text itself, and it has been answered rather than ignored.  Robert McCabe's reply, in the Detroit Baptist Seminary Journal and on the Answers in Genesis site, has three parts.  First, verse 5 gives two reasons, no rain and no man, and Kline's reading "marginalizes the last clause of v. 5 to a parenthetical remark," because if the whole earth lacked plants until rain and a farmer, then man was made before vegetation, which Kline himself rejects.  Second, "shrub of the field" and "plant of the field" are not "vegetation" in general (that is Gen 1:11's "grass, herb, tree") but the cultivated and the thorny plants that appear after the Fall (Gen 3:18), so the verse is about the land of Eden on day six, not the planet on day three.  Third, Kline's "unargued presupposition" is exactly that:  the verse says the plants were not there because it had not rained; it does not say God could not have made them otherwise.³⁰  Sarfati puts the last point sharply:  "Note that Kline admits that this alleged presupposition is not argued in the text.  This would explain why no exegete saw this for thousands of years."³¹  Grade:  the young-earth reply is good, and Kline's own concession that the verse's principle is "unargued" is the weakest joint in his case.  But Kline's article remains a serious piece of exegesis that the plain reading must engage every time, and the restrictive reading of "shrub of the field" is a reading, not a demonstration.

"One day," not "the first day"

Genesis 1:5 ends yom echad, "one day," a cardinal; days two to five are anarthrous ordinals, "a second day"; day six has the article on the ordinal only, "day the-sixth."  Payne cited John Urquhart's use of this pattern for a non-consecutive-days theory:  the days are literal but separated by ages.²⁶  Steinmann's 2002 study answers that the pattern is a definition, "Evening + morning = one day," which is how Basil read it in the fourth century,³⁹ followed by "another evening and morning constituted 'a' (not 'the') second day," and that "it would appear as if the text is very carefully crafted so that an alert reader cannot read it as 'the first day.'"⁹  Grade:  the grammar is unusual, and Steinmann's reading of it as a definition is the most economical one on the table.  But the non-consecutive theory is not refuted by grammar alone; it is refuted by Exodus 20:11, which treats the six as a week, and Payne conceded that the theory "would not have been as readily deduced from the Genesis text" without geology.²⁶

There is a fifth item that belongs to a sibling article:  the order of Genesis 2 (man, then plants, then animals, then woman) against Genesis 1, which turns on whether wayyitser in 2:19 is a pluperfect.  It is treated in Ancient Cosmology and Genesis.

Concordism, accommodation, and where the argument stops

Two hermeneutics compete for Genesis 1 among people who accept an old earth.  Concordism holds that the text and the science describe the same events and can be made to agree, so the "days" become ages, or the light of day one becomes the clearing of a cloud cover, or 1:1 becomes the Big Bang.  Accommodation holds that God spoke to ancient Israel in ancient Israel's terms, so the text describes a real creation in the world-picture of its time and makes no claim about the age of anything.  Concordism is Hodge, Archer, Ross.  Accommodation is Calvin on Genesis 1:16, Seely, Walton, Enns, and, at the far end, Barr.

The grammar of Genesis 1 is a problem for the first and not for the second.  Every concordist reading has to make yom mean something the numerals and the refrain do not let it mean, and every concordist since Hodge has admitted that he would not have read it so without geology.  Accommodation needs no such surgery.  It can grant everything in this article, that the author meant six ordinary days and a young world, and then say the author was writing what his age believed and God let him.  That is Barr's position exactly:  "they were literally intended," and "the figures are not, to us, historically, scientifically or factually true."¹  It is also McGrath's:  the quotation "is correct and would be correct today," and "the young-earth creationist approach simply isn't an option."³⁸

This is where the argument of this article stops, and it should be said plainly.  The grammar establishes what the writer meant.  It cannot establish that he was right.  A young-earth creationist who wins every point above has shown that the Bible teaches a recent creation in six days; he has not shown that the creation was recent.  That case is made, or not made, by the physical evidence in Dating the Earth, Radiometric Concordance, Radiocarbon, and their siblings, and by the theological cost of the alternative in Adam, the Fall, and Death.  What the grammar does settle is a narrower and older quarrel.  The reading that Christians held for eighteen centuries was not a misreading, and the readings that replaced it were not discovered in the Hebrew.  They were brought to it.

Common objections

Objection 1: "Yom has several meanings, so the days could be ages."

It has several meanings, and the sentence picks one.  The markers that pick the ordinary sense (a numeral, the singular, "evening and morning," a counted series, "night") are all present in Genesis 1 and are the markers that pick it in the other 354 verses.  The one text where yom covers the whole week, Genesis 2:4, uses a construct-plus-infinitive idiom that means "when" and says nothing about length.  A word's range is not a licence to choose.

Objection 2: "The Barr letter is a private letter that creationists have never let anyone verify."

Correct, and this article says so.  The transcription is by a non-creationist who posted it to argue that it is over-read, its internal details check, and its most hostile critic grants that its content "is correct and would be correct today."  More to the point, Barr published the same distinction under his own name in 1987, and the Regius Professor who succeeded him wrote in 2011 that Barr "was correct."  Drop the letter and nothing changes.

Objection 3: "Boyd's statistics were published by a creationist institute in a creationist volume."

They were, and they reproduce.  Every count in Boyd's table that this article re-ran from the public morphological text came out identical, and the script is in this article's research/ folder.  The criticism that sticks is different:  the statistic shows narrative form, which the critics already grant, and cannot show historical intent, which is what they deny.  Both halves of that are stated above.

Objection 4: "Genesis 1 is obviously structured and patterned, and patterned texts are not history."

Genesis 1 is patterned.  So are the Flood narrative (a palistrophe), the book of Kings (a regnal formula), and the genealogies (a fixed clause).  Hebrew historians patterned their history.  The pattern in Genesis 1 is a seven-day week, and the pattern is the sequence, not a substitute for it.  Wenham, who calls the chapter "elevated prose," also calls verse 1 the first act of creation and says "most modern commentators agree."

Objection 5: "Every ancient reader was pre-scientific; of course they read it literally. That says nothing about what a modern reader should do."

It says everything about what the author meant, which is this article's only question.  What a modern reader should do with an author who meant six days is a further question, and there are three answers on the table:  he was right (this site's position, argued elsewhere on physical evidence), he was wrong but accommodated (Barr, McGrath, Walton in part), or he did not mean it (day-age, framework).  The third is the one the grammar rules out.

Objection 6: "Kline's Genesis 2:5 argument shows that the text itself assumes ordinary providence during creation, so the plain reading is self-contradictory."

Kline called the assumption "unargued," which is a strange word for a premise that is supposed to overturn a chapter.  The verse gives two reasons for the absence of the "shrub of the field," no rain and no man, and the second cannot be a global reason without putting man before plants, which Kline himself refuses.  Read of Eden on day six, the verse is background to Genesis 2:7-8 and conflicts with nothing.  This is the best textual argument against the plain reading, and it has an answer; it does not have a knock-down.

Objection 7: "Even if the author meant six days, that only tells us what an Iron Age writer believed."

Yes.  That is the point at which this article stops and says so.  Grammar establishes intent.  Whether the intent was true is a question of evidence, and it is argued on this site where the evidence is.

Sources