The sign of Jonah

21 min read

They came for a sign. Scribes and Pharisees, and they said it politely: Teacher, we would like to see a sign from You. No hostile crowd — a request, with the proper words. And He refuses. An evil and adulterous generation asks for a sign, He says, and none will be given it but one. No miracle on demand. No proof for whoever requires it. One sign (Matthew 12:38–39).

And it is a sign of time.

For as Jonah was three days and three nights in the belly of the sea monster, so will the Son of Man be three days and three nights in the heart of the earth (Matthew 12:40).

Watch what He does with that sentence. He reaches back for the oldest image He can find — a prophet in the dark of the deep, buried alive and given back — and He lays His own grave beside it. As Jonah. So the Son of Man. It is no loose allusion; it is an equation, and an equation has two sides that must agree. What stands on Jonah's side must return on the side of the Son of Man, or there is no comparison. And what stands on both sides is, word for word, the same: three days and three nights.

Count the nights. He named them Himself.

And hold on to this, for it is the heart of the matter, and it is more than a question of language. A sign is not atmosphere. A sign is a guarantee — something God sets down so that a man can later check it and say: it came as it was said. That is what a sign exists for. And therefore a sign has one property it cannot do without: it must be measurable, and it must come out exactly. A sign that "comes true" in ten ways comes true in none — for then it proves nothing, and the speaker might as well have kept silent. Scripture itself lays down that standard. If a prophet speaks in the name of YHWH, it says, and the word does not happen and does not come to pass — then YHWH has not spoken that word (Deuteronomy 18:21–22). The authenticity of a divine word hangs on its exact fulfilment. Elasticity is the death of it.

Now see what that means for this sentence. The Lord gives an evil generation exactly one sign, and He stakes His whole authority on it. He does not then choose the loose form. Elsewhere, where the resurrection is merely announced, the supple ordinal sounds again and again: on the third day — a form that bends to more than one reckoning. But here, where the word must do duty as a sign, as the checkable promise to which He binds Himself, He gives a measure: three days and three nights. He names the days, and then He names the nights as well, as if to board up every way out. That is no stylistic ornament. That is the precision that makes an utterance a sign.

And therefore more is at stake here than one detail in the chronology of the passion. Whoever reads three days and three nights as a loose manner of speaking is not pruning away a twig — he is dismantling the sign itself. He turns the one guarantee the Lord gave that generation into a sentence "fulfilled" by any weekend whatever. That is the price, and it must be said aloud honestly: the Friday buys its reckoning with the exactness of the sign. It makes the sign unfalsifiable — and a sign that cannot fail is not a sign.

Expertsign, and the measure of a sign

The Greek word is sêmeion (σημεῖον), the Hebrew ot (אוֹת): a mark, a distinguishing signal that vouches for something — the rainbow (Genesis 9:12), the blood on the doorposts (Exodus 12:13), the sign Isaiah offers to Ahaz (Isaiah 7:11–14). An ot is by its nature testable: it is given in order to be recognised and checked. And the biblical test of a prophetic word is as hard as it is simple — "when the word does not happen and does not come to pass, that is a word YHWH has not spoken" (Deuteronomy 18:22). Exact fulfilment is not an extra added to a sign; it is its essence. This weight is an inference of faith, not a measurement — but it rests on what the text itself says about signs, and it weighs as heavily as you take that promise seriously.

So the matter stands. On the one axis of this book, the axis of the day, two formulas stand side by side. The one — on the third day — is an ordinal, and it is supple; it bends to the Friday without creaking. The other — three days and three nights — is a duration, and it is not. It names the nights. And the question of this whole chapter is narrow enough to fit on a single line: can "three days and three nights" ever mean less than three full twenty-four-hour days? For if it can, the sign says nothing, and you may lay it aside. If it cannot, then in His only sign He left behind a clock — and that clock does not point to Friday.

Expertthe two formulas in the Greek

The credal word is the ordinal: têi tritêi hêmerai, "on the third day" (τῇ τρίτῃ ἡμέρᾳ). So it stands in the resurrection predictions time after time — Matthew 16:21; 17:23; 20:19; Luke 9:22; 18:33; 24:7; 24:46; Acts 10:40; 1 Corinthians 15:4. Beside it two kindred forms: meta treis hêmeras, "after three days" (Mark 8:31), and en trisin hêmerais, "in three days" (the temple word, John 2:19). All these forms count days — the bare noun, without the nights.

Only once does the other form fall: Matthew 12:40, treis hêmeras kai treis nyktas (τρεῖς ἡμέρας καὶ τρεῖς νύκτας), "three days and three nights". And it falls there twice in one breath — once of Jonah, once of the Son of Man — because the comparative hôsper [...] houtôs ("as [...] so") demands the identical term on both sides. Nor is the phrase an invention of Matthew's: it is the literal Greek rendering of Jonah 2:1 (Eng. 1:17), where the Hebrew reads sheloshah yamim usheloshah lelot (שְׁלֹשָׁה יָמִים וּשְׁלֹשָׁה לֵילוֹת). Matthew quotes the prophet exactly.

Note precisely what that difference is. The ordinal "the third day" names an end point: the day on which something happens. How many full days lie before it, it leaves open. The duration form "three days and three nights" names no end point but a length, and it names that length twice — first with the days, then with the nights, as though the speaker wished to shut out every doubt. Whoever says "three days" may be speaking loosely. Whoever adds "and three nights" has already cut that looseness off himself.

And that is exactly the point at which the tradition ran up against it — and at which, to save the Friday, it deployed a counting rule fetched from somewhere else.

Countercasethe strongest case for the Friday, here given the floor in full

This is the objection at its strongest, and it deserves to be heard first and weighed only afterwards. It runs: the Jews counted part of a day as a whole day. There exists a rabbinic rule to that effect — no invention of embarrassed apologists, but a real halakhic rule. In the Jerusalem Talmud Rabbi Elazar ben Azariah says: "a day and a night make an onah, and part of an onah counts as the whole" (יוֹם וְלַיְלָה עוֹנָה וּמִקְצָת עוֹנָה כְּכוּלָּהּ; Jerusalem Talmud, Shabbat 9:3). Apply that to the time in the grave, and "three days" becomes: part of Friday, all of Saturday, part of Sunday — three days, with fewer than three full twenty-four-hour days inside them. The leading defender of the Friday in the twentieth century, Harold Hoehner, sums it up crisply: "it is a well-known fact that the Jews reckoned any part of a day as a whole day" (Hoehner, 1974, p. 242). And he stands in a long line: as early as the twelfth century the Syrian bishop Dionysius bar Salibi answers this same Jewish objection by saying that "three days and three nights" and "on the third day" mean the same, "according to common usage" (Williams, 1935, p. 108).

And Scripture itself seems to lend a hand. In 1 Samuel 30 a starving Egyptian is found who had eaten nothing for "three days and three nights" (v. 12) — and in the next sentence he says he fell ill "three days ago" (v. 13). Jerome rendered that last phrase in the Vulgate as nudiustertius, "the day before yesterday" — not yet two full twenty-four-hour days (Biblia Sacra Vulgata, 1 Samuel 30:13). The great church father, then, read inclusively here, in the same breath as the double construction. If the Friday has an anchor anywhere, it is here.

Whoever wants to read this chapter honestly must take this as the case to beat. Not as a straw man. The onah rule exists, it is old, and it is Jewish. That is not denied here — it is put up front.

So the opposing side stands at its strongest. And the rebuttal begins not by contradicting it, but by asking it something: what does that rule apply to? For a rule has a dwelling place, and this rule has its own — and it is not the duration form.

Look where the elasticity really sits. In Scripture the bare form, "three days", is indeed elastic — and you can read that off the text itself. In Genesis 42 Joseph puts his brothers in custody "three days" (v. 17) and releases them "on the third day" (v. 18): three days, counted inclusively. But in Ezra 8 the caravan arrives in Jerusalem and stays "three days" (v. 32) — and "on the fourth day" the silver is weighed (v. 33): three days, counted exclusively. Twice "three days", twice a different reckoning. The bare form, then, is genuinely ambiguous, and precisely for that reason a rule exists for it. The onah rule does real work — with the bare form, where there is something to decide.

And now back to the Egyptian of 1 Samuel 30, for he must not be left lying — the box above called him the anchor of the Friday, and the strongest card of the opposing side deserves an answer rather than silence.

The answer has two layers, and the second is the heaviest; it is therefore written out in full in the excursus at the end of this chapter, where the four old translations stand side by side. Here the main line, in brief.

The first layer is grammatical and already breaks the argument. Verse 12 gives the double construction — "three days and three nights", no bread and no water — and that is a precise statement. Verse 13 gives something else: hayyom sheloshah (הַיּוֹם שְׁלֹשָׁה), "today three", a vague range. Between a precise and a vague statement no contradiction arises: the precise one pins the vague one down, not the other way round. Nothing in this pericope is being compressed. Add to that that the narrator, eleven verses earlier, does use the ordinal — "on the third day" (v. 1, בַּיּוֹם הַשְּׁלִישִׁי) — and does not choose it here, and that 1 Samuel 9:20 has the same construction in full, hayyom sheloshet hayyamim, "three days ago". S. R. Driver, the first authority on Samuel, connects the two passages explicitly and reads both as elapsed time (Driver, 1913, ad 1 Samuel 9:20). [DOCUMENTED]

The second layer is the decisive one, and it has never before been traced: the four oldest translations do not render verse 12 and verse 13 with the same word. Targum, Peshitta, Septuagint and Vulgate all four do the same — verse 12 with the duration form, verse 13 with something else. That is the excursus below, with the texts alongside.

And there is a third layer, one that belongs to verse 12 itself and that I have seen adduced nowhere. The Egyptian's sentence is not a loose remark but a formula. The double construction together with "he ate no bread and drank no water" occurs four times in the whole Hebrew Bible: Exodus 34:28 and Deuteronomy 9:9 and 9:18 — three times Moses on the mountain — and 1 Samuel 30:12. What the formula does is make the statement binary. A fast is not a matter of degree: if he ate on the evening of the first day, the sentence is false. The number therefore stands there not as a rough impression but as the span over which nothing was eaten, and that is precisely why the form is used as it is. So it is to be read: as a literal statement in day-units. ⚠ And so it is not to be read: as a measurement in hours. The text nowhere reckons in hours, and "three days and three nights" is no seventy-two-hour clock. It counts three days and three nights, and the units are what they are. What the formula says is therefore something about the kind of sentence, not about a span measurable by the clock — and 1 Samuel 30:12 accordingly stays where it was put above: outside the confirming series. [DOCUMENTED — the four places fetched primarily]

And now the honesty of the other side, for it belongs in the same breath — and it asks for one distinction that must stay sharp. Two questions run past each other here. The first: do verse 12 and verse 13 measure the same span? The second: how long is the span of verse 13? To the first question the four old versions give the same answer — no, they keep the two verses apart — and Jerome is there no dissident but a witness for the prosecution against the idiom defence (see the excursus below). On the second question he does diverge: he renders verse 13 as nudiustertius, "the day before yesterday", and so measures the span short, inclusively. That is a real divergence, and it is not polished away here — but it touches the length of verse 13, not the equation on which the argument rests. And the Greek τριταῖος of the Septuagint is given in the lexicon first as an ordinal — "on the third day". But the same lexicon gives τριταῖος also as duration — "of three days' standing", "three days before" — and with that it decides nothing. Whoever deploys it as evidence, for or against, is not measuring; he is choosing.

What remains is this, and it has a consequence for the count further on: 1 Samuel 30 does not attest the shortened idiom, and it does not refute it either. It is, on close reading, no calibration point — and therefore, in the sweep below and in appendix B, it counts towards neither side. The anchor of the Friday holds no ground; not because it points the other way, but because it points nowhere.

And the double form? It needs no rule, and nowhere is it given one. "Three days and three nights" is not ambiguous; it names the nights precisely to kill the ambiguity. Go through the whole field where this form falls, and you find the same thing every time: full twenty-four-hour days. The flood rains "forty days and forty nights" — and that is not merely an image: in one of the Dead Sea Scrolls the flood calendar is worked out, and the forty come out at forty full days, closing to the day (4Q252, cols. I–II; García Martínez & Tigchelaar, 1997–1998). Moses is "forty days and forty nights" on the mountain, without bread and without water (Exodus 34:28) — no one has ever subtracted days from that. Job's friends keep silence "seven days and seven nights" (Job 2:13). Jonah lies "three days and three nights" in the deep (Jonah 1:17). Always full, always continuous, never shortened. (The starving man of 1 Samuel 30:12 carries the double construction too, but he remains here, as has just been shown, outside the count.)

And it is not only Scripture that counts this way; the earliest layer of the Jewish tradition does so too, and there it becomes recomputable. Seder Olam Rabbah — the tannaitic chronography — takes Moses' forty days and forty nights on the mountain, the double construction, and works it out: three times forty days from the seventh of Sivan. The sum is not decorative. It lands, at the end of the first series, on the seventeenth of Tammuz, the fast-day of the golden calf, and at the end of the third on the tenth of Tishri — Yom Kippur. Shift one day and both memorial days vanish (Seder Olam Rabbah 6). [CALCULATED — the three series recomputed by hand, with the customary month lengths; they close to the day.] Honesty marks the limit: Seder Olam counts whole days and includes the first day, and that is an inclusive count of end points — not the question at issue in the sign, whether half a day counts as a whole one. On that question the source is silent, for its days are whole. But the narrower thing it does show is enough: where the double construction, in the oldest layer of the tradition, falls within a counting context, it is counted in full and nowhere shortened. With that, the silence around Jonah is no longer an emptiness but an outcome — the tradition had the form in a count, and it never subtracted a night from it.

And the rule that lies beneath this is stated in so many words by a Jewish authority. Rashi notes, at the flood, why the first day falls outside the count: "אֵין יוֹם רִאשׁוֹן מִן הַמִּנְיָן, לְפִי שֶׁאֵין לֵילוֹ עִמּוֹ" — "the first day is not part of the count, because its night is not with it" (Rashi on Genesis 7:12). At Sinai he repeats the motif: the forty days are full — day with night — and the day of the ascent drops out, "for its night is not with it" (Rashi on Exodus 32:1). That is, from the mouth of a medieval Jewish expositor, precisely the measure of this chapter: a day counts only when its night belongs to it — and he applies it to both double constructions, the flood and Sinai, with the same motif. But even the friendly witness must not be made larger than he is. Rashi does not get his sharpness from the double construction alone: he must add to it the word shelemim, "full", to cut the partial day out — the form itself does not force that — and he himself notes that the people counted inclusively, and erred in doing so. He is therefore no neutral measurement but a Jewish master who chooses the strict reading and marks the inclusive one as an error. That is less than a proof and more than nothing: the rule the Friday deploys to count the nights away is shattered, by a Jewish authority, precisely on the nights. [DOCUMENTED — Seder Olam and both Rashi passages seen at first hand.]

Dossierthe count, and what you can recompute yourself

The formula inventory. Walk through the New Testament for the time of the resurrection and you find the ordinal family (on / after / in three days) at every turn — nine times in the form "on the third day" alone (see the Scholar above). That form admits the Friday. The duration form "three days and three nights" falls exactly once: in the sign of Jonah. The whole weight of the refutation of the Friday therefore hangs on the question whether that one duration form is as supple as the ninefold ordinal form — and the rest of this Workbench shows that it is not.

The corpus sweep. The double construction "N days and N nights" stands in the whole Hebrew Bible in thirteen places: Genesis 7:4 and 7:12 (the flood) · Exodus 24:18; 34:28 and Deuteronomy 9:9, 9:11, 9:18, 9:25; 10:10 (Sinai) · 1 Samuel 30:12 · 1 Kings 19:8 · Job 2:13 · Jonah 1:17. The count is machine-verified: the consonantal text of all thirty-nine books was swept for every verse in which a night-word stands anywhere beside a day-word — sixty-four verses — and those sixty-four were then read one by one. Thirteen carry a numeral before both nouns; the rest are the familiar "day and night" pairs (Genesis 1:5; Psalm 74:16) or the adverbial yomam walaylah, which is a different form. Deuteronomy 9:25 belongs in the list although it varies the form — et arba'im hayyom we'et arba'im hallaylah, with the article — and Esther 4:16 stays out of it, for the reason given below. In every case where the duration can be independently checked, it counts full twenty-four-hour days; not one passage counts it inclusively. The one exception to that checkability is 1 Samuel 30:12: its only internal measure, verse 13, is too vague to fix anything, and as shown above that passage therefore counts towards neither side — it neither confirms the full count nor refutes it. It falls, here and in appendix B, outside the confirming series.

The check at Qumran. The flood text 4Q252 works the "forty days and forty nights" out explicitly in the calendar; the three independent fixed points in that computation close only if they are forty full days. This is the hardest support there is: here the double construction is not merely assumed but made recomputable — and that before the Christian era, outside any Christian interest (4Q252; García Martínez & Tigchelaar, 1997–1998).

The withheld test. To avoid choosing a corpus that happens to oblige, the prediction — "the double construction is never counted inclusively" — was first written down and only then tested on texts that had not been used in the original research: the Aramaic of the Babylonian Talmud and the plene-spelled Hebrew forms. Outcome: every narrative case — the slaughter at Tur Malka (b. Gittin 57a), the lament (b. Bava Metzia 86a), the ship between the fins of the fish (b. Bava Batra 73b) — full and continuous. No counterexample. The rule predicted in advance held on material that could have broken it. The full working stands in appendix B. [DOCUMENTED — the texts seen at first hand; the out-of-sample test CALCULATED]

There remains one passage where the opposing side believes it holds its best card: Esther. She fasts "three days, night and day" (Esther 4:16) and then goes "on the third day" to the king (Esther 5:1) — so, the argument runs, three days that are not three full twenty-four-hour days. It is the classic Friday argument, and it turns on its head the moment you read the words themselves.

For Esther does not use the double construction. It does not say "three days and three nights"; it says sheloshet yamim laylah vayom — "three days, night and day" (שְׁלֹשֶׁת יָמִים לַיְלָה וָיוֹם). An adjacent, emphatic form, but not the same one. And that is no ingenuity of mine — it is precisely what Jewish exegesis itself noted centuries ago. Rabbi Shlomo Alkabetz writes in the sixteenth century, in his Esther commentary, that if it had been a full three-day fast, the text would have had to say "three days and three nights"; it does not say that, it says "night and day" — and therefore, Alkabetz concludes, the fast was not a full seventy-two hours (Alkabetz, Menot HaLevi on Esther 4:16). The Maharal endorses it. And a later master even works the number out: three days and three nights are seventy-two hours — the number of the word chesed, grace (Yosef Chayyim, Ben Yehoyada on b. Bava Batra 73b). The tradition itself, then, uses the double construction as a measuring-rod: where it stands, the count is full; where it does not stand, it is not. Esther does not break the rule. Esther shows precisely that the rule existed, and that it was known.

And the limit belongs here at once, because it cuts my own way. Esther does not stand alone. Leviticus 8:35 carries the same build — yomam walaylah shiv'at yamim, "day and night, seven days" (וּפֶתַח אֹהֶל מוֹעֵד תֵּשְׁבוּ יוֹמָם וָלַיְלָה שִׁבְעַת יָמִים) — and it is on this very verse that the tradition does apply the part-for-whole rule: the seven days of mourning are derived in part from it (y. Moed Katan 3:5; Bereshit Rabbah 100:7), and to those seven days miktsat hayom kekhulo applies. Anyone claiming that any mention of night beside day makes a span exact is refuted by this single verse. I do not make that broad claim, and it appears nowhere in this book. The distinction that does hold is narrower, and it is grammatical: it turns on the numeral. Stated once, with "day and night" as an adverbial phrase saying how — then the span is not measured; so read Leviticus 8:35 and Esther 4:16, and those are the two cases there are, both counted inclusively. Stated twice — sheloshah yamim usheloshah leylot, three days and three nights — then a count is being made; those are the thirteen, and in none of the thirteen is the part-for-whole rule applied. It is not the night that makes the difference, but the number standing beside it. [DOCUMENTED — the texts seen primarily; the count of two against thirteen is CALCULATED and remains a claim about the corpus searched, not a measurement of the language]

Excursusthe hostile witness, and the limit of what this chapter shows

There is a witness who weighs more than any friend of the cause, precisely because he is an enemy. In a medieval Jewish polemic against Christianity — the Nizzachon Vetus, around 1260, in Hebrew — Matthew 12:40 comes under attack. The writer knows the onah rule through and through; it is his tradition. He names it by name, miktsat yom kekhulo, "part of the day as the whole". And watch what he does with it: he calls it the excuse of the Christians ("even according to their words"), he grants it in full — and then he computes it through. Even with the rule, he writes, you reach no more than "three days and two nights". In Berger's translation the passage runs: *"But it is impossible for Jesus to have been in the heart of the earth for three days and three nights since he was crucified on Friday evening and arose early Sunday morning, and even if they should argue that part of the day is considered like the whole day, there would still be only three days and two nights"* (Berger, 1979, §201, p. 202; already in Wagenseil, 1681, p. 236). [DOCUMENTED — Berger's critical edition seen at first hand] Six centuries before the modern debate, a Jewish native speaker, with every reason to crush the Christian reading, says exactly where the crack runs: the rule saves the days; it does not reach the nights.

And there is more, for the rule itself can be recomputed against its own sources. The application of the onah rule to Matthew 12:40 is not drawn from the Talmud but from one book: John Lightfoot's Horae Hebraicae et Talmudicae of 1658. He is the one who extends the rule from the bare days to the double construction, and he justifies that step by an appeal to the "Scholarum consensu", the agreement of the schools. Then open his three references. The first, Avodah Zarah 75a, defines the onah as twelve hours — half a full day, never a full day with its night. The second, Shabbat 9:3 in the Jerusalem Talmud, is not a rule but a dispute: Rabbi Akiva there in fact rejects "part of an onah counts as the whole", and the sages count whole days. And the third, the baraita in Shabbat 86a, uses the part-principle in the opposite direction: there one part supplements another part into a single whole — it does not shorten, it compensates. All three passages are moreover about status periods — purity, terms — and not about the duration of a narrative. The rabbinic cover on which the objection leans is, when its own references are followed up, not there. [DOCUMENTED; the three passages read at first hand; Lightfoot's Latin read in the text layer of the 1675 collected edition and collated against Gandell's translation]

And then there is one sentence which does not close the case but does set it sharply, and it comes from the Talmud itself. In Pesachim 55a the question is asked how a plant can take root within three days when the first and the last of those three days are only partly available, and Ravina answers: "חַד ״מִקְצָת הַיּוֹם כְּכוּלּוֹ״ — אָמְרִינַן, תְּרֵי ״מִקְצָת הַיּוֹם כְּכוּלּוֹ״ — לָא אָמְרִינַן" — "one 'part of the day counts as the whole' we say; two 'part of the day counts as the whole' we do not say". [DOCUMENTED — the Aramaic text seen at first hand] The structure is exactly that of this book's question: a span of three days of which both ends are partial, and the ruling is that one may round off one end and not two.

Lay that beside what the Friday needs. It must deploy the part-principle twice: once for the three hours between the dying and the sunset, and once for the night hours before the dawn of the Sunday. The Wednesday needs it not once.

And now the limitation, for it is obligatory here and it is threefold. First, this is Ravina, an Amora of around 400 — no tannaitic rule and certainly no first-century practice; it testifies to how the tradition thought the principle through, not to how one counted in the year 31. Second, the domain is agricultural (the taking of a cutting), not sepulchral; the transfer to this book's question is an analogy. Third, it is one of several proposed solutions in the same sugya — Rav Sheshet, Rav Ashi and Rava each give another — so that one may not say "the Talmud decides", but "Ravina decides thus". And then the objection that presents itself at once: does not the law of mourning apply miktsat hayom kekhulo at two ends? It has been measured out, and it does not hold. In Moed Katan 19b the rule is deployed at one point, and at the end: Abba Shaul teaches that part of the seventh day counts as the whole, "וְיוֹם שְׁבִיעִי עוֹלֶה לוֹ לְכָאן וּלְכָאן" — "and the seventh day counts for him this way and that way", that is to say: it closes the seven and opens the thirty. That is no double rounding but one rounding that serves two counts — precisely the reverse of what the Friday needs. That the day of burial counts as the first day rests, moreover, not on this rule but on the fact that the obligation of mourning begins then; the sugya does not adduce miktsat hayom for it. And even that one application is disputed: Rava — the same who appears in the Pesachim sugya — rules "הֲלָכָה כְּאַבָּא שָׁאוּל בְּיוֹם שְׁלֹשִׁים, וְאֵין הֲלָכָה כְּאַבָּא שָׁאוּל בְּיוֹם שִׁבְעָה", while Abaye draws the boundary the other way round and the Nehardeans allow both on the strength of Shmuel's rule of thumb "הֲלָכָה כְּדִבְרֵי הַמֵּיקֵל בְּאֵבֶל", "in matters of mourning one follows the more lenient". [DOCUMENTED — Moed Katan 19b seen at first hand] That last is telling: the law of mourning is by express rule the most lenient corner of the halakhah, and therefore precisely the wrong place from which to derive a strict counting rule. The objection therefore falls away; but the passage remains a weight and not a keystone, for the three reasons above. [CONSTRUED] That is no play on words but the core of the matter: the Friday is one onah short, and it must make it up with a double rounding that is disputed within rabbinic reasoning itself.

And then the anchor itself, 1 Samuel 30 — for there the objection stands at its strongest, and it can be tested in a way no one has tried. The standard argument runs that verse 12 ("three days and three nights") and verse 13 ("today three") measure the same span, so that the double construction here demonstrably counts inclusively. But see first what stands there. Verse 12 is a precise statement; verse 13 is a vague range. The precise pins the vague down, not the other way round — nothing in this pericope is being compressed, and the standard argument is in that form a non sequitur. To which comes that verse 13 is a known construction with ellipsis: 1 Samuel 9:20 has it in full, "הַיּוֹם שְׁלֹשֶׁת הַיָּמִים", and there it means "three days ago". Driver — the first authority on Samuel and co-author of the standard lexicon — connects the two passages explicitly and reads them both as elapsed time: "'to-day, three days,' i.e. [...] three days ago. Cf. 30, 13, where [הימים] is omitted." [DOCUMENTED]

But the decisive thing is this, and it has never been checked before: the ancient translators, independently of one another and centuries before this debate, did not render the two verses with the same word. Targum Jonathan renders verse 12 by *"תְּלָתָא יְמָמִין וּתְלָתָא לֵילַוָן"* — three yemamin and three nights — where yemama in Aramaic is precisely the day that stands over against the night; and he renders verse 13 by *"יוֹמָא דֵין דְּנָן תְּלָתָא יוֹמִין"*, with the ordinary word yomin, calendar days. The translator could have chosen one word and chose two. He read verse 12, then, as the full measure of day-and-night, and verse 13 as a loose indication of days elapsed. [DOCUMENTED — the Aramaic text itself]

And in this he does not stand alone. Open the other ancient versions, and all three do the same.

The Peshitta, the Syriac Bible, has in verse 12: "ܡܛܠ ܕܬܠܬܐ ܐܝܡ̈ܡܝܢ ܘܬܠܬܐ ܠܝ̈ܠܘܢ ܠܐ ܐܟܠ ܠܚܡܐ" — "for three imamin and three nights he had eaten no bread" — and in verse 13: "ܗܐ ܬܠܬܐ ܝܘܡ̈ܝܢ", "behold, three yawmin". The Syriac imama is the lexical twin of the Aramaic yemama; the standard lexicon gives it one sense, "daytime", the day in so far as it stands over against the night, while yawma is the general word "day" (Comprehensive Aramaic Lexicon, s.v. )ymm and ywm; text after the CAL edition of P 1S 30).

The Septuagint goes a step further, and it is the sharpest of the four. Verse 12 has the double construction in full: "τρεῖς ἡμέρας καὶ τρεῖς νύκτας". But verse 13 has no numeral "three days" — there stands *"ὅτι ἠνωχλήθην ἐγὼ σήμερον τριταῖος"*, with the adjective tritaios: "I am today a third-day man" (Swete, 1 Regn. 30:13). That is exactly the word-field with which John indicates the four days of Lazarus — tetartaios estin, "he is a fourth-day man" (John 11:39). The Greek, then, possesses its own form for the ordinal reckoning, one that the duration formula is not; and the translator of 1 Samuel uses in verse 12 the duration formula and in verse 13 that other form.

And the Vulgate closes the row. Verse 12: "non enim comederat panem, neque biberat aquam, tribus diebus et tribus noctibus". Verse 13: *"quia aegrotare coepi nudiustertius"* — no duration, but the Latin adverb of elapsed time, literally nunc dies tertius, "the day before yesterday". Jerome, then, reads verse 13 exactly as Driver reads it.

And the modern standard commentaries confirm exactly this reading — it is no longer an inference of this book's own. The two authoritative commentaries on Samuel both read verse 13 as elapsed time. Klein renders "I was sick for three days" and, in his textual note, supplies the dropped הַיָּמִים — on the authority of Hebrew manuscripts, the Septuagint, the Old Latin, the Peshitta and the Targum — with the cross-reference "see 1 Sam 9:20" (Klein, 1 Samuel, WBC 10, 1983, note 13.a). The link with 9:20 thus stands no longer as this book's reasoning but as a note read in the critical apparatus of a standard commentary — and the four ancient versions set side by side above are there named together, on precisely this word, by a modern editor. McCarter renders verse 13 "three days ago, because I was ill" and, unlike at ten other places in this chapter, leaves the Masoretic text untouched here: his textual notes pass from verse 12 straight to verse 14 (McCarter, I Samuel, AB 8, 1980, pp. 429–438). Where the field's most exhaustive textual critic would tick off the Greek apparatus, he is silent — a sign that the Greek witnesses do not diverge here. [DOCUMENTED on both commentaries' rendering; CONSTRUED on McCarter's silence] (The full Greek apparatus on Samuel stands in Brooke–McLean, not in the Göttingen series, which has no Samuel volume yet; that edition has not itself been seen.) And that is exactly what makes them independent witnesses: 1 Samuel 30 is the very passage the other side adduces for Matthew 12:40, and the reading of Samuel is here not bent toward that debate — it lands on elapsed time on Samuel's own grounds. Neither discusses Matthew 12:40 or the passion week; the application remains this book's, and Klein supplements the text where McCarter precisely does not emend, so that the two cover each other's weakness.

Add that all together, for the outcome is unusually unequivocal. Four independent ancient translations, in four languages, Jewish and Christian, spread over centuries, none of them with the slightest interest in a Wednesday — and not one of the four equates verse 12 and verse 13. All four render verse 12 with the duration form and verse 13 with something else: a different noun, an ordinal form, or an adverb of elapsed time. The standard argument of the idiom defence rests on the two verses measuring the same span. The whole ancient translation tradition reads them apart. That is no favourable reading of one witness; that is unanimity. [DOCUMENTED — all four texts seen at first hand]

Then draw the limit at once, too, for it belongs with this. This measures how the oldest readers understood the two verses, not what the Hebrew author meant; the Hebrew itself remains, on this point, underdetermined. What falls away is not the possibility of the equation, but the claim that it is self-evident. And with that the anchor falls away on which the objection most leaned — not because verse 13 is reasoned away, but because every ancient reader on record already kept the two verses apart himself, beginning in the language the Lord spoke.

One counter-witness must stand here in full, for he is the heaviest there is against this measurement, and he is Jewish. Bereshit Rabbah 56:1 lists the "third days" of deliverance — Abraham, the tribes, the spies, the law-giving, Jonah, the exiles, the resurrection, Esther — and in the received text Jonah's "three days and three nights" stands under that heading. Whoever reads that sees the tradition itself equate the duration form and the ordinal form, and that would undermine this whole measurement. Two things, however, stand against it, and both can be checked. First, the critical text of Theodor and Albeck reads the lemma bare: the list then consists of five ordinals (Abraham, the tribes, the law-giving, the resurrection, Esther) and three bare cardinals (the spies, the exiles, and Jonah), and contains not a single double construction — Jonah stands there with "on the third day", not with the duration formula. Second, the list classifies, it does not count: Ezra stands under "the third day" while the verse itself names the fourth day (Ezra 8:32–33). A list that sets its own example a day wrong is a register of a formula of salvation and not a chronology. [DOCUMENTED]

And now the limit, honestly drawn, for without it this chapter is mere display. This witness himself assumes the Friday — he fights the Christian on the Christian's own calendar; he is a witness to the idiom, not to the day. And what this whole chapter establishes is precisely that: something about the idiom, not about the day. It takes from the Friday its linguistic rescue — no more. It does not prove the Wednesday. And the earliest Christian reader cuts the same way, which honesty requires be said: Irenaeus (around 180), the oldest witness to cite Matthew 12:40, himself reads the sign as a triduum — "post triduum resurgens", in the same paragraph in which he quotes the verse — and so did not see the exact-three-nights reading (Irenaeus, Adversus Haereses 5.31). Whoever says the sign is deliberately exact must explain why the man who stood closest to it did not read it so.

That explanation exists, and it already stands in this book — with the same man, in the other direction. For the four Passovers of the ministry Irenaeus is this book's witness (ch. 9); but from that same count he himself drew the conclusion that Jesus was approaching fifty and taught for more than ten years (Adversus Haereses 2.22.5–6), and that conclusion no one has followed. There the rule is: his count is adopted, not his conclusion (ch. 14). Then the same measure must hold here too, now that it runs the other way: his text is adopted, not his reading. What Irenaeus attests is that Matthew 12:40 stood fixed in his text around 180, double duration form included — that carries, and it carries the age of the verse. What he does with it — equating the duration form with the triduum — is no memory of the week itself but a reading, a century and a half later, by a man whose church confessed the resurrection in the supple ordinal; he read the credal form back into the measured form, as later centuries read "Friday" back into paraskeuê (ch. 6). A witness is gold for what stood in his text and fallible in what he inferred from it — that holds for the Passovers, and so it holds here too; whoever follows him there and not here, or the other way round, measures with two measures. But then the price must be said just as loudly: the earliest reader who possessed the verse did not read the exact three nights in it, and before the modern era no reader demonstrably does. The measurement of this chapter, then, has no ally in the early reception; it rests on the corpus, and on nothing else. Whoever accepts it must know on what he is accepting it — and whoever takes the reception as the standard loses here. [CONSTRUED]

There remains one text which, with the other formula — the bare "third day" — genuinely presses against a Wednesday: the word of the Emmaus travellers, "it is now the third day" (Luke 24:21). That text receives its own chapter, and it is not reasoned away there. So say no more than what stands here. The sign of Jonah proves no day. It pulls a beam out from under the Friday — and it lays down a burden the Friday has not carried in seventeen centuries: one clear use of "three days and three nights" within the corpus itself — not a later expositor who argues it so, for such exist (Augustine; see above and appendix B §B.3c) — for demonstrably less than three full twenty-four-hour days. That burden has still not been taken up.

Countercase"his flesh saw no corruption"

There is one objection to the three full days that comes not from the calendar and not from the grammar, but from the body itself. Peter says on the day of Pentecost, with Psalm 16 on his lips, that the Holy One "saw no corruption" (Acts 2:27, 31; Paul repeats it, 13:35–37). And the objection is obvious: three full days in a grave are a long time to say that of; thirty-six hours are a short one. Put that way, this is the strongest material objection the Friday has, and I take it seriously.

Begin with the word. Peter cites the Septuagint, which reads diaphthora, "corruption." The Hebrew of Psalm 16:10 reads shachath, and that word stands there in parallel with Sheol — "the pit, the grave," as Job uses it five times in one chapter. The Greek choice is no slip: the Septuagint renders shachath by diaphthora as a matter of course, here and in seven of the nine places in the Psalms. But that is the point: neither sense — "pit" or "corruption" — carries a clock. The promise is about not-remaining-in-the-pit, or about a body that is kept; it names no hour.

And see, then, what Peter himself does with it. He names no hour and no day. He says: David died and was buried, "and his tomb is with us to this day." Paul makes it sharper: David saw corruption; He whom God raised did not. The contrast is a body that stayed in the grave against a body that was raised. The mechanism is the resurrection, not the brevity.

There is a reversal there, and it is decisive. Whoever reads "no corruption" as a natural fact — a body does not decay much in a day and a half — turns the promise into an open door: it would then hold for any dead man buried on a Friday afternoon and disinterred on Sunday morning, and it would prove nothing about anyone. But Peter's whole argument is that it distinguishes Jesus from David. It is precisely the long interval that gives it content: a body that goes the full distance and yet sees no corruption is a preservation; a body that lies there a day and a half is a Tuesday. [CONSTRUED]

And the biology does not supply the clock the objection needs. The body breaks down from the moment of death, without a threshold; the odour comes from bacteria and hangs on warmth and moisture. There is no hour at which corruption switches on. One threshold the sources do know, and it is one of recognizability, not of decay — and it lies on the fourth day, not the third. The Gospel supplies the measure itself: at Bethany Martha says, "Lord, by now he stinks, for he has been dead four days" (John 11:39). Four. Later rabbinic texts count the same three days, but they are late and speak of identifying a dead man, not of the onset of decay; I therefore carry them as background and not as proof. And even within that frame the Wednesday touches the boundary and does not cross it.

One thing more, in honesty: the objection has a single ancient ancestor, and I do not conceal him. Athanasius reasons, as early as around the year 318, that the body was "not to remain long and be wholly corrupted," lest the risen one be taken for a different body — "therefore not more than three days" (On the Incarnation 26). That is the same sensitivity as the objection. But see what he does exactly: he counts inclusively — "one intervening day" — so that his "three days" are not my seventy-two hours but the Friday three; and his whole point is that after the interval the body turned out to be incorruptible, the wonder, not the open door. His reasoning therefore presupposes a body that would have corrupted had it stayed long — precisely the preservation of which Peter speaks. The ancestor of the objection stands, rightly read, on this side.

So say no more than what stands here. This does not prove the Wednesday; it removes an objection. And it leaves one honest remainder, which I name rather than smooth away: if anyone shows that the first century placed the onset of corruption on the third day rather than the fourth, then the objection stands and must be priced as I price Luke 24:21. I have searched and not found such a source; John 11:39 is for now the only first-century witness, and it says four.

CountercaseLuke gives the same sign without a clock

There is an objection that touches the sign more directly than everything that went before, and I have nowhere seen it systematically answered in the Friday literature — all the more reason to set it out here myself at its strongest. Matthew is not the only one who transmits the sign of Jonah. Luke does so too, and in him no clock stands. "As Jonah became a sign to the Ninevites, so will the Son of Man be to this generation" (Luke 11:30), and Luke then explains the sign by the preaching and the repentance of Nineveh: the men of Nineveh will rise up in the judgement, for they repented at the preaching of Jonah (Luke 11:32). No three days, no three nights, no belly of the fish. The whole duration form on which the preceding pages rest is absent — and it is absent in the one evangelist who transmits precisely the same saying. Whoever says the duration is the sign must explain why the synoptist who stands beside Matthew leaves it out at the very place where the sign falls. This weighs more heavily than the objection from the body, for it touches not the length of the sign but the question whether the length is the sign at all. I take it as such.

Three lines of answer lie open, and I give them in order of strength — with the price attached, for it is real. The first is text-critical. Matthew's form is the longer and the harder: "three days and three nights" stands in the manuscripts of Matthew 12:40 without the standard critical editions noting a load-bearing variant here, and a longer, harder reading is sooner worn away than added. There is, then, no textual trace that Matthew received the nights by later insertion; if anything moved, it moved towards the shorter and the smoother. [DOCUMENTED — as far as the standard critical text reaches; no load-bearing variant recorded.]

And there belongs to it one observation that is seldom made, for it narrows the distance between the two evangelists considerably — and it hangs on the text one follows. In the received text Luke 11:29 reads: *"εἰ μὴ τὸ σημεῖον Ἰωνᾶ τοῦ προφήτου"* — "except the sign of Jonah the prophet". So read Stephanus 1550, Scrivener 1894, the Byzantine majority text and the text of the Greek Church; the critical editions — Westcott-Hort, Tischendorf, Nestle, SBL — all four strike τοῦ προφήτου. [DOCUMENTED — the eight editions laid side by side] That is precisely the designation with which Matthew introduces the sign (Matthew 12:39). In the received text, then, Matthew and Luke transmit the same formula and differ only in what they elaborate afterwards; it is the eclectic text that makes Luke's form barer than it is. Whoever wants the objection at its sharpest — "Luke gives a different, shorter sign" — must therefore first say which text he follows, for the objection becomes measurably weaker once one reads the received text. This decides nothing about the duration; it removes the suggestion that Luke would carry a divergent, reduced wording.

The second is redactional. Luke abbreviates more sayings than Matthew and gives the sign from a different motive — the generation that demands a sign and receives none but the preaching that calls to repentance. That explains what Luke leaves out: he tells the sign towards his own point, the summons, not the grave. It does not explain what Matthew would have invented — a duration formula that comes word for word from Jonah 1:17 and that puts his own passion chronology under pressure rather than relieving it. The third line is the soberest: an omission is no denial. Luke nowhere contradicts the duration; he does not mention it. Between "he did not say it" and "he denied it" lies the whole distance of this objection.

And now the move that goes deeper than the three lines together, for it turns the objection round. The sign of Jonah is not one thing; it is double — and Matthew and Luke each give one half. Read Jonah himself and you see it. The call goes out from Jonah — the man given back out of the fish after three days and three nights, the resurrection figure, called back alive from the deep — and the same call goes out to the city: "Yet forty days, and Nineveh shall be overthrown" (Jonah 3:4). The figure and the preaching belong to one and the same prophet, in one and the same history. Matthew takes the figure: three days and three nights, the man from the deep. Luke takes the preaching: the summons that compels repentance. Two halves of one sign — not two signs contending for precedence.

And see, then, what Luke does with that second half, for it is not softer than Matthew but harder. Nineveh repents — and it does so under the lead of its own ruler, the king who rises from his throne, lays off his robe and himself sits down in sackcloth and ashes (Jonah 3:6). A heathen city, a heathen prince, at the bare preaching of one stranger. Set Israel beside it. "And behold, more than Jonah is here" (Luke 11:32) — and this generation does not repent. Worse: it kills Him. It calls Him, while He does not yet even lie in the grave, "that deceiver" who said that He would rise after three days (Matthew 27:63), and it then does everything to smother precisely that resurrection — the heart of the sign: the guard at the tomb, the watchmen's report brought to the chief priests, and the considerable sum of silver pieces with which the soldiers are bought off to keep silent (Matthew 28:11–15). The Ninevites believed the preaching of the figure-from-the-deep; Israel buried the truly risen One under bought lies. That is the direction Luke gives the sign.

And with that, Luke 11:29–32 is not narrower than Matthew 12:40 but broader. The duration, in Matthew, gives the figure of the sign — what happens, and how long. The preaching-and-repentance, in Luke, gives its direction — against whom it stands, and what it judges. Take them together and you have the whole sign: the call the risen One embodies, and the verdict on whoever rejects it. Luke does not leave the clock out because there is no clock; he tells the other half — the half in which the clock had already struck at Nineveh and the city had listened, and in which this generation hears the same clock and turns away. So the objection is not refuted by fighting the clock back into Luke, but by showing that Luke never was the place of the clock. He is the other half of the same sign, and precisely the sharpest half against unbelief. [CONSTRUED]

And then the price, honestly and without detour, for even softened it remains a price. The clock of the day-axis stands in Matthew and not in Luke; whoever wants to recompute the three nights reads them in the one evangelist and not in the other. That remains so, and I do not book it away. But it is no choice between two evangelists who contradict each other — and there the double reading changes everything. Luke carries the other half of the same sign, not a rival version without a clock. He is silent about the duration as Matthew is silent about the repentance of Nineveh: each tells his half in full and leaves the other to his fellow witness. The burden, then, is not that the day-axis rests on one of two rivals, but the narrower and honest fact that the duration is named in so many words by only one of the two. That the nights stand there and are nowhere contradicted, that Matthew's form is the harder one which text-critically carries no trace of addition, and that Luke's own half presses against unbelief more sharply rather than more softly — all that makes the burden light. But light is not gone, and I pass it on as it is. [CONSTRUED — that the two synoptic forms are complementary and not rival is an inference from the double figure of the Jonah sign, not a measurement.]

And then, behind all the foregoing — the body, the second evangelist — one question remains that cuts deeper than all the others, and it returns to the rule itself: where, then, does the crack sit? Not in the rule — that exists, it is old, and it is Jewish; this chapter has conceded that from the outset and it is not taken back here. The crack sits in the last step. Lay side by side what the rule speaks of and what the claim speaks of. The rule counts yamim, days, or onot, day-night units. The claim counts "three days and three nights". Who made that leap?

No primary text at all. The leap can be pointed to, and it has one author, one book and one year: John Lightfoot, Horae Hebraicae et Talmudicae, 1658. More precisely, for the difference counts: what he adduced were the three places traced above (Avodah Zarah 75a, y. Shabbat 9:3, b. Shabbat 86a), and none of them carries the extension. What has since been repeated, from hand to hand, for three hundred and fifty years, is an abbreviated standard citation with two references — and they are not the same two. One of Lightfoot's three has vanished on the way, and one has been added that was not in him.

Walk through those two, for they are what the reader gets to see today. The first, Jerusalem Talmud Shabbat 9:3, really exists and Lightfoot's reference is correct: Rabbi Elazar ben Azariah says there that day and night together form an onah and that part of an onah counts as the whole — and the Jerusalem Talmud even works out for itself that the third day is thereby reached after "a day and a little" (יוֹם וְכָל שֶׁהוּא), some twenty-six hours. That must be said honestly, and it stands here: whoever asserts that this rule does not exist is wrong. The second reference, Babylonian Talmud Pesachim 4a, does indeed contain the formula miqtsat hayyom kekhulo — "part of the day as the whole" — but strictly within the law of mourning: about the mourner, about the late report of a death, about the removing of shoes. Not a word about three days and three nights. Half of the standard citation, then, stands in the wrong place; the other half stands rightly, but does not carry the conclusion. [DOCUMENTED — both passages seen at first hand]

And two independent checks point the same way. Bereshit Rabbah 91:7 counts the three days of Joseph's brothers through to Hosea 6:2, by way of Jonah, Mordecai and David — and everywhere counts yamim, days, never days-and-nights. And where the Babylonian Talmud does reckon in onot, in Shabbat 86b, it says "six full onot" (שֵׁשׁ עוֹנוֹת שְׁלֵמוֹת) and precisely the miqtsat formula is absent. Both times the tradition counts in days or in full units, and both times the night stays out of the picture as a deduction. [DOCUMENTED — both texts seen at first hand]

And then the counter-witness who is most honestly put up front, for he is the heaviest there is — heavier than Esther, and not late. The same midrash on Genesis carries, beside the series at 91:7 just named, another list of rescues "on the third day", and in this one too stands Jonah. In the printed text, which most readers know, his lemma runs in full: "בַּיּוֹם הַשְּׁלִישִׁי שֶׁל יוֹנָה [...] וַיְהִי יוֹנָה בִּמְעֵי הַדָּגָה שְׁלשָׁה יָמִים וּשְׁלשָׁה לֵילוֹת" — "the third day of Jonah [...] and Jonah was in the belly of the fish three days and three nights" (Bereshit Rabbah 56:1). Here an early rabbinic source — amoraic, roughly fifth century — seems to shove the double construction itself under the rubric "the third day". That is exactly what the Friday is looking for, and it must stand here without softening.

Two things take this witness's weight from him, and the first is textual. The critical edition of the midrash — Theodor–Albeck, which weighs the oldest manuscripts against each other — reads the lemma not in full but bare: "וַיְהִי יוֹנָה בִּמְעֵי הַדָּג שְׁלשָׁה יָמִים", "three days", without the nights, and with the masculine haddag from Jonah 2:1 (Eng. 1:17) in place of the feminine haddagah from 2:2 (Eng. 2:1). The full form with the nights stands there only as a variant in the apparatus, and it is the reading of the printed editions — the most ordinary scribal movement there is: a copyist who knows the verse by heart completes a clipped citation into the whole Bible verse. And the list as a whole points the same way: eight members, five ordinals and three bare cardinals, and in the critical text not a single double construction. Jonah stands in the list precisely because he is cited bare there; put the nights in and he no longer fits under "the third day". [DOCUMENTED — the critical page, Theodor–Albeck vol. 2, p. 595, seen at first hand.]

The second takes the list's weight from it whichever reading one chooses, for it shows that the list does not count but classifies. It sets Ezra's caravan under "the third day" — "they stayed there three days" (Ezra 8:32) — while the next verse dates the event expressly "on the fourth day" (Ezra 8:33). Were the midrash counting, a flat error would stand here: a fourth-day event under the third day. Is it ordering a theme — God never leaves the righteous in distress longer than three days — then it fits without friction. The list, then, does not measure; it orders, and therewith carries no chronometric weight for Jonah either: it is a homiletic heading, not a clock. That is precisely the distinction this chapter holds to throughout — a list classifies, a count counts — and it separates this counter-witness from Seder Olam and Rashi, who do count the double construction, and then count it in full. And the remainder that honestly stands: the received text reads Jonah here with the nights, so a real textual tradition places the full form under "the third day". Theodor–Albeck judges that form secondary, and the apparatus shows why: the oldest witnesses — the London codex (British Museum Add. 27169, mid-twelfth century) and the Paris one — read the bare form, and the nights appear only in the printed editions (Constantinople 1512, Venice 1545) and a few medieval manuscripts. That makes "secondary" a defensible text-critical judgement, but not on its own yet a measurement: the same copyists abbreviate biblical verses, so the bare form could be an abbreviation of the full, familiar verse. One old witness settled this, and it has now been weighed. The oldest complete codex, which Theodor did not use here — the Vatican one (ebr. 30, tenth or eleventh century) — reads the lemma bare at the page itself: "and Jonah was in the belly of the fish", after which the next member begins at once ("the third day of the exiles who went up", Ezra). Not merely without the nights, but without any duration phrase at all. So the three oldest witnesses — Vatican 30, London, Paris — all read the short form, and the full form with the nights is demonstrably a later accretion: "secondary" is no longer an editor's judgement but the converging reading of the oldest manuscripts. [DOCUMENTED — Vatican ebr. 30, fol. 86r, read at first hand via DigiVatLib.] What remains, then, is this, and no more: a homiletic list could archive the double construction under an ordinal — but in a text that demonstrably classifies rather than counts, and in a citation the critical edition reads bare. It is not a count that saves the Friday; it is a theme that happens to name a prophet. [CALCULATED/CONSTRUED — Ezra 8:32–33 seen at first hand; that the list orders follows from its own fourth-day member.]

So that is how the matter stands. The rule dwells with the days. The extension to the nights is to be found in no source, but in one scholar in 1658, and half of what has since been handed on as proof-texts stands, moreover, cited in the wrong place. And Lightfoot's own text, read in the Latin, shows the leap undisguised. At Matthew 12:40 he first concedes the shortfall himself: the span between the cross and the rising, he writes, "is nevertheless called by him three days and three nights, when only two nights had intervened, and but one complete day" (cùm duæ tantùm noctes intercesserint, et dies completus tantum unus). He then adduces a single rabbinic source — the onah rule of the Jerusalem Talmud, Shabbat 9:3 — and licenses the reckoning not by a text but by usage: Christ may truly be said to have lain in the grave three onoth, though "the greatest part of the first day was wanting, and the night altogether, and by far the greatest part of the third day also", the consent of the schools and the dialect of the nation agreeing thereunto (suffragante Scholarum consensu, dialecto gentis). The bridge is there, in his own words; but it rests on usage and on a rule that counts days, not on a source that counts the nights too. That is not yet a refutation of the Friday — one may grant Lightfoot right even without a source, and that must be said. It is something more modest and something more precise: the burden of proof the Friday has not taken up for seventeen centuries turns out, on closer inspection, never to have been carried by a source, but by an inference.

Where does the investigation stand, then, at the end of the first axis? Not at a proof — that word does not belong here, and it will belong nowhere in this book. It stands at a shift of the burden of proof, and it is sharp. The literal reading of "three days and three nights" is the ordinary, running language of Scripture: attested everywhere, recomputable all the way to Qumran, contradicted nowhere. The inclusive reading of precisely this form is the exception — introduced to save one text, and from the earliest centuries until today always asserted, never measured. Bar Salibi said "common usage" and named not a single example. Hoehner said "a well-known fact" and named none either. And those who do name examples keep naming the same two. David Baron offers three places for the idiom, and his own word for them is "used, or implied": of Jonah he grants that "we have no means of accurately ascertaining" how long it actually was; 1 Samuel 30 requires, on his reading, that verses 12 and 13 measure one span, and his ground for that is "the impression left on the mind"; and Esther 4:16 he writes out himself as "three days—night and day" — he puts the difference on the page and walks past it. E. W. Bullinger, defending the same case in 1898, rests on Esther 4:16 and 1 Samuel 30:13 — the same two. So this is not a series of witnesses each producing his own evidence; it is one basis of two verses, handed on, and one of the two does not carry the construction. That observation covers the defenders I have read myself, and not one more. [DOCUMENTED — Baron and Bullinger read primarily; appendix B §B.2c] The rule they invoke truly exists — but, as has just been shown, it dwells with the days, and the leap from that rabbinic rule to the nights stands in no rabbinic source; it has only an author and a year. One qualification belongs here, and it is my own. That does not mean that no one before 1658 ever counted the double construction inclusively: Augustine does so in the fourth century, and unmistakably (De consensu evangelistarum III.24.66). But he does it as an expositor who wants to reconcile two Gospel passages, in Latin, by a piece of reasoning — not as a witness to an existing Hebrew or Aramaic usage, and not with an example from the corpus. Appendix B §B.3c weighs him in full. The claim this chapter makes is therefore the narrower one: there is no source that uses the idiom so; there are only expositors who argue that it was used so. That is not proof. It is convergence, and it weighs as heavily as the burden of proof weighs which the other side has not taken up in seventeen centuries.

And with that, at the end of the day-axis, a concrete demand lies on the table. If the sign speaks true — if the Son of Man was three days and three nights in the heart of the earth — then the week in which that happened must be able to carry those three nights. Three nights between dying and rising. That does not fit between a Friday evening and a Sunday morning; two fit there. It asks for more room in the week itself — for a day that stands between the day of death and the weekly sabbath. For, in other words, a week with two sabbaths.

Does that room exist in the text? That is the next question.

HeartHEART

He could have given them a sign that split the heavens. He chose the image of a man in the dark — sunk away, given up, three nights long beyond the reach of any hand, and yet given back to the light. As Jonah. So He Himself. The sign is no trick to convince the curious; it is a promise, wrapped in a grave. Whoever checks it counts nights. But whoever believes it hears something else beneath the numbers: that there is no depth from which He cannot be called back — and that when He announced His own depth, He counted the nights and named them one by one, so that no one could say afterwards that it was only a manner of speaking.

Note on the sources

The Greek and Hebrew material has been checked in the source itself; the Bible translation is my own working translation from the Hebrew (BHS) and the Greek of the received text (Scrivener 1894 / Stephanus 1550, collated against the critical editions). The Targum Jonathan and the Peshitta on 1 Samuel 30:12–13 have both been read with the Aramaic itself (the Targum via the Mikraot Gedolot text, with the Brill rendering of Van Staalduine-Sulman; the Peshitta via the Comprehensive Aramaic Lexicon). The corpus sweep and the Qumran computation are reproducible and stand worked out in the appendices. The rabbinic and medieval witnesses were consulted in the source itself or through a named scholarly rendering, with location. The two rabbinic passages coupled to "three days and three nights" — Jerusalem Talmud Shabbat 9:3 and Babylonian Talmud Pesachim 4a — have been checked with the Aramaic/Hebrew itself, as have the two checks (Bereshit Rabbah 91:7; Babylonian Talmud Shabbat 86b); Lightfoot's own text ad Matthew 12:40 has been read in the Latin — in the collected edition Horae Hebraicae et Talmudicae in quatuor evangelistas (1675), in the text layer of the scan rather than on the page image, and collated word for word against Gandell's English translation of the same note (1859, vol. II, pp. 209–211): he concedes the shortfall of nights himself, adduces the onah rule of the Jerusalem Talmud as his only rabbinic source, and covers the extension by usage — suffragante Scholarum consensu, dialecto gentis, the consent of the schools and the dialect of the nation. No earlier source applies the rule to the days-and-nights form; the attribution of the leap to him therefore rests on his own words. The inclusive reading of the double construction has nowhere been found in a primary source; that is the outcome of a directed search, not a claim that it could not exist. Where this chapter says "convergence", it means convergence — a weighed surmise, not a proof.

The witnesses for the positive count have likewise been checked in the source itself: Seder Olam Rabbah 6 (the forty days and nights of Sinai, recomputable by hand onto Yom Kippur — a count of whole end points, not of partial days, and marked as such) and Rashi on Genesis 7:12 and Exodus 32:1, where the partial day is excluded "because its night is not with it". The heaviest rabbinic counter-witness, Bereshit Rabbah 56:1, has been read in two guises: the received (Vilna/printed) text, which places Jonah's double construction in full under "the third day", and the critical edition of Theodor–Albeck (vol. 2, p. 595, read by me from the page itself), which reads the lemma bare and registers the full form as an apparatus variant. The apparatus shows that the oldest witnesses (London, BM Add. 27169, mid-12th century; Paris) read the bare form and that the nights appear only in the printed editions (Constantinople 1512, Venice 1545) and a few medieval manuscripts; "secondary" is thus text-critically defensible but not a measurement (the same copyists abbreviate verses), so the variant stays undecided at the archetype. The very oldest complete codex, Vatican ebr. 30 (10th–11th century), not used by Theodor here, has now been read from the page itself (DigiVatLib, fol. 86r): it reads the Jonah lemma bare, "and Jonah was in the belly of the fish", with no duration phrase — so the three oldest witnesses (Vat. 30, London, Paris) all read the short form, and the full form is a later accretion. For Luke 11:29–32 the Greek text has been checked; the weighing of the two synoptic forms against each other is an inference and marked as such. The reading that the sign of Jonah is double — the figure-from-the-deep in Matthew, the preaching-and-repentance in Luke, both going back to one prophet (Jonah 3:4, 6) — is exegesis and is marked CONSTRUED; the verses cited (Jonah 3:4, 6; Luke 11:32; Matthew 27:63; 28:11–15) have been checked in the original text.

Literature (APA)

Alkabetz, S. H. (16th century). Menot HaLevi on Esther [Commentary]. Sefaria. https://www.sefaria.org/Menot_HaLevi_on_Esther.4.16

Athanasius of Alexandria. (ca. 318). De Incarnatione Verbi 26 (SC 199, ed. C. Kannengiesser). Greek text consulted via the First1KGreek edition. [seen in the original] — the decisive chapter checked in the Greek itself (μίαν [...] μέσην, "one intervening day," counted inclusively; "οὐ πλείω τῶν τριῶν ἡμερῶν").

Babylonian Talmud, Pesachim 4a. Consulted via Sefaria, with the Aramaic. https://www.sefaria.org/Pesachim.4a

Babylonian Talmud, Shabbat 86b. Consulted via Sefaria, with the Aramaic. https://www.sefaria.org/Shabbat.86b

Baron, D. (n.d.). The actual day of the crucifixion and the question of the "three days and the three nights". Reprint (N. Battey, 22 October 2015), christianresearcher.com. [full text read primarily; ⛔ the original printing has not been identified, hence the citation without page]

Bullinger, E. W. (1898). Figures of speech used in the Bible. Eyre & Spottiswoode / E. & J. B. Young. [pp. 845–847 read primarily in the facsimile of the Princeton copy; ⚠ via the OCR text layer, not the page image]

Bullinger, E. W. (1922). The Companion Bible, appendix 144. [⚠ seen as a transcription, not in the printed edition; the posthumous editorial history of the appendices has not been established]

Berger, D. (1979). The Jewish-Christian debate in the High Middle Ages: A critical edition of the Nizzahon Vetus (§201, p. 202). Jewish Publication Society.

Bereshit Rabbah 56:1 [Midrash]. Received (Vilna) text consulted via Sefaria, with the Hebrew original; the critical text via Theodor & Albeck (vol. 2, p. 595). https://www.sefaria.org/Bereshit_Rabbah.56.1

Bereshit Rabbah 91:7 [Midrash]. Consulted via Sefaria, with the Hebrew original. https://www.sefaria.org/Bereshit_Rabbah.91.7

Biblia Sacra iuxta Vulgatam versionem (R. Weber & R. Gryson, Eds.). (1994). Deutsche Bibelgesellschaft. [1 Samuel 30:13]

Comprehensive Aramaic Lexicon [Peshitta, 1 Samuel 30, file 62008 (P 1S)]. Hebrew Union College. https://cal.huc.edu

Driver, S. R. (1913). Notes on the Hebrew text and the topography of the books of Samuel (2nd ed., ad 1 Samuel 9:20 and 30:13). Clarendon Press.

Klein, R. W. (1983). 1 Samuel (Word Biblical Commentary 10, pp. 279–288; esp. note 13.a). Word Books. [1 Samuel 30:13]

McCarter, P. K. (1980). I Samuel (Anchor Bible 8, pp. 429–438). Doubleday. [1 Samuel 30:12–13]

García Martínez, F., & Tigchelaar, E. J. C. (Eds. & Trans.). (1997–1998). The Dead Sea Scrolls study edition (2 vols.). Brill. [4Q252, cols. I–II]

Hoehner, H. W. (1974). Chronological aspects of the life of Christ, Part IV: The day of Christ's crucifixion. Bibliotheca Sacra, 131(523), 241–264. [Also included in Hoehner, H. W. (1977). Chronological aspects of the life of Christ. Zondervan.]

Irenaeus of Lyons. (1885). Against heresies (Book 5, ch. 31; Book 2, ch. 22, §5–6; Ante-Nicene Fathers, Vol. I). Christian Literature Company.

Jerusalem Talmud, Shabbat 9:3 (H. W. Guggenheimer, Trans.). Sefaria. https://www.sefaria.org/Jerusalem_Talmud_Shabbat.9.3

Yosef Chayyim of Baghdad. Ben Yehoyada on Bava Batra [Commentary]. Sefaria. https://www.sefaria.org/Ben_Yehoyada_on_Bava_Batra

Lightfoot, J. (1658). Horae Hebraicae et Talmudicae (ad Matthew 12:40). Cambridge. The note ad Mt 12:40 read in the Latin of the collected edition, Horae Hebraicae et Talmudicae in quatuor evangelistas (1675), In Evangelium Matthaei cap. XII, vers. 40 — read in the text layer of the scan, not on the page image, and collated throughout against Gandell's English translation of the same note (Lightfoot, J., Horae Hebraicae et Talmudicae, ed. R. Gandell, 4 vols., Oxford University Press, 1859, vol. II, pp. 209–211). The two witnesses agree word for word. The key sentences read "cùm duæ tantùm noctes intercesserint, et dies completus tantum unus" and "suffragante Scholarum consensu, dialecto gentis". The onah rule of y. Shabbat 9:3 is the only rabbinic source he adduces for the extension.

Rashi (Shlomo Yitschaki). Commentary on the Torah (ad Genesis 7:12 and Exodus 32:1) [Commentary]. Consulted via Sefaria, with the Hebrew original. https://www.sefaria.org/Rashi_on_Genesis.7.12 ; https://www.sefaria.org/Rashi_on_Exodus.32.1

Seder Olam Rabbah 6 [Tannaitic chronography]. Consulted via Sefaria, with the Hebrew original; the threefold forty-day count (7 Sivan → 17 Tammuz; → 10 Tishri = Yom Kippur) recomputed by hand. https://www.sefaria.org/Seder_Olam_Rabbah.6

Theodor, J., & Albeck, Ch. (Eds.). (1912–1936). Bereschit Rabba: Midrasch Bereschit Rabba mit kritischem Apparat und Kommentar (Vol. 2, p. 595). [critical edition; the printed page seen at first hand]

Van Staalduine-Sulman, E. (2002). The Targum of Samuel (ad 1 Samuel 30:12–13). Brill.

Wagenseil, J. C. (1681). Tela ignea Satanae: Liber Nizzachon Vetus (p. 236). Altdorf.

Williams, A. L. (1935). Adversus Judaeos: A bird's-eye view of Christian apologiae until the Renaissance (p. 108). Cambridge University Press.

Bible texts: Genesis 7:4, 7:12; 9:12; 42:17–18 · Exodus 12:13; 24:18; 32:1; 34:28 · Deuteronomy 9:9, 9:11, 9:18, 9:25; 10:10; 18:21–22 · 1 Samuel 9:20; 30:1, 30:12–13 · 1 Kings 19:8 · Ezra 8:32–33 · Job 2:13 · Isaiah 7:11–14 · Hosea 6:2 · Jonah 1:17; 3:4, 6 · Matthew 12:38–40; 16:21; 17:23; 20:19; 27:63; 28:11–15 · Mark 8:31 · Luke 9:22; 11:30, 11:32; 18:33; 24:7, 21, 46 · John 2:19 · Acts 10:40 · 1 Corinthians 15:4. The author's own working translation from BHS; Greek after the received text (Scrivener 1894 / Stephanus 1550), collated against eight printed editions. On the cornerstone of this chapter that collation yields an outcome worth recording: Matthew 12:40 reads word for word alike in seven of the eight editions — both textus recepti, RP-Byz, Westcott-Hort, Tischendorf⁸, Nestle 1904 and SBL — and the only divergence, in the text of the Greek Church ("ἐγένετο Ἰωνᾶς ὁ προφήτης" for "ἦν Ἰωνᾶς"), does not touch the duration formula: "τρεῖς ἡμέρας καὶ τρεῖς νύκτας", twice, stands there in all eight. [DOCUMENTED]