Thus far the case has argued for. Now I turn it round. An argument that shows only its own strong points is not an argument but a plea, and the reader has a right to the opposite — to the strongest text that speaks against this date, not tucked away in a footnote but up front, at its strongest. There is such a text. There is exactly one that truly hurts, and I am not going to smooth it over.
It is a sentence from the mouths of two disillusioned men on the road to Emmaus. It is Sunday. They are walking with a stranger — it is the Risen One, but they do not know it — and they tell Him, with the despondency of people whose hope has died, what has happened. And then they say: "and besides all this, it is now the third day since these things happened" (Luke 24:21).
The third day. On Sunday. Since the crucifixion.
Do the arithmetic, and the grip is felt at once. From Friday to Sunday is, counted Jewish-inclusively, the third day: Friday one, Saturday two, Sunday three. Exactly. But from Wednesday to Sunday? Thursday, Friday, Saturday, Sunday — that is the fourth day, and counted inclusively even the fifth. On no method of counting is Sunday the third day after a Wednesday. This is the mirror image of the sign of Jonah, and it weighs exactly as much. Where Matthew 12:40 asks for three nights the Friday cannot supply, Luke 24:21 names a third day the Wednesday cannot reach. Even Hoehner, the defender of the Friday, concedes how much the whole Wednesday case hangs on that one verse of Matthew: had there been no Matthew 12:40, he writes, the Wednesday theory would probably never have been proposed (Hoehner, 1974, pp. 242–243). Fair is fair: that point bites. And it bites here.
So let me not pretend that this verse resolves neatly. It does not. What follows is not a defusing but a weighing — and it ends in a trade, not a knock-out: this verse can be absorbed, at a real price, and the question is which side gets the better bargain. Not to be reasoned away, but to be weighed honestly. I shall show that in three layers, and each layer yields less ground than the one before.
The first layer is the grammar, and it falls largely the right way — but it does not save the case. There is an old discussion about the verb in that sentence, agei (ἄγει). Read personally, the subject is "he" or "one"; read impersonally it means roughly the Latin agitur, "it is passing, one is spending" — "the third day is under way". The defenders of the Wednesday have laid great weight on that impersonal reading, and linguistically they have a point: the impersonal reading is not marginal; it stands in Thayer, in Buttmann, in Plummer. But look closely at what it delivers, and what it does not. It shifts the subject of the sentence; it does not shift the number. "The third day is passing" is still the third day. Whether you read "one is living through the third day" or "it is the third day" — it remains three, and three since the crucifixion remains Friday. The grammatical gain is real, but it touches the wrong variable. Whoever hoped that agei would turn a three into a five is disappointed; no reading of that verb does that.
Expertthe verb, and why it does not move the count
The sentence runs: tritên tautên hêmeran agei aph' hou tauta egeneto (τρίτην ταύτην ἡμέραν ἄγει ἀφ' οὗ ταῦτα ἐγένετο).
And here something must be set right at once that tells against this book. That is the critical text. The received text reads one word more: *"τρίτην ταύτην ἡμέραν ἄγει σήμερον ἀφ' οὗ ταῦτα ἐγένετο"* — "today he is spending this third day". So read Stephanus 1550, Scrivener 1894, the Byzantine majority text and the text of the Greek church; Westcott-Hort, Tischendorf, Nestle and SBL all four omit σήμερον. [DOCUMENTED — the eight editions laid side by side] This book holds to the received text, and then it ought to cite the heavier reading: σήμερον binds "the third day" expressly to the day of speaking, and that is Sunday. The received text thus makes the objection harder, not softer. I take that as it comes and do not conceal it; whoever chooses a form of the text chooses it also where it tells against him. The impersonal reading of agei (≈ Latin agitur, "is being spent") is well attested — Thayer s.v. ἄγω; Buttmann (1873, p. 118), who expressly gives "ἡ ἡμέρα ἄγεται" as "the day is passing"; and Plummer (1896) at this very place: "we are at the third day". But it is and remains a present — one is in the third day; it is not over (that would require a perfect tense). And Luke elsewhere counts inclusively in exactly this way: in Luke 13:32 "today" is day one and "the third day" is the day after tomorrow; in Acts 10:30 the same author reckons "ἀπὸ τετάρτης ἡμέρας" inclusively. The demonstrative tautên ("this, the third day") binds the day of speaking — Sunday — to the ordinal three, just as "this, the second letter" in 2 Peter 3:1. In short: impersonal or personal, the count comes out at three, and three since the crucifixion is Friday. The grammar does not relieve the Wednesday.
The second layer is the wider "third day" field, and it changes the picture — not by refuting the verse, but by isolating it. For "the third day" is not the only way the New Testament designates the time between cross and resurrection. There are seven different constructions, and if you lay all seven along the Wednesday chronology (with the death late on Wednesday and the resurrection towards the end of the third full-day unit, Saturday towards sunset), it turns out that they do not stand against the Wednesday en bloc. One of them favours the Wednesday outright — "three days and three nights" (Matthew 12:40). Four constructions can be reconciled with either chronology — among them "after three days" (meta treis hêmeras), which sits somewhat less smoothly on a Friday but does not exclude it (see the Workbench). One is neutral. And one — exactly one — favours the Friday: Luke 24:21. So the wider field does not stand against the Wednesday; it isolates the objection to this one verse. That is no escape, but it is the right measure of the matter: not "the third day refutes the Wednesday", but "one of the seven third-day forms refutes it, one supports it, and the rest permit both chronologies".
Dossierthe seven constructions and the reckoning-core
The reckoning-core. Sunday is called "the third day since X" if and only if X, counted inclusively, is Friday (exclusively: Thursday). From a Wednesday, Sunday is inclusively the fifth day, exclusively the fourth. No single re-anchoring makes Sunday "day three" under a Wednesday; every attempt stacks at least two assumptions foreign to the text.
A way out that suggests itself, and does not work. One often points out here — and rightly — that the Jewish day begins not at midnight but at sunset, so that "the first day of the week" already begins on our Saturday evening. That is correct, and it is load-bearing elsewhere in this book: Matthew 28:1 sets the walk to the tomb precisely on that seam — "Ὀψὲ δὲ σαββάτων, τῇ ἐπιφωσκούσῃ εἰς μίαν σαββάτων", "late on the sabbath, as it grew light towards the first day of the week" — a reading in which the eight printed editions do not diverge, the two textus recepti and the majority text included. [DOCUMENTED] But for this verse it does not help, and I would rather say so myself than have a reader discover it. The conversation on the Emmaus road falls in the daylight of that first day — the travellers later urge the Stranger to stay "for it is towards evening and the day is far spent" (24:29). Wherever one places the day-boundary, the span remains the same: from the fourteenth of Nisan to the eighteenth is inclusively the fifth and exclusively the fourth day. Shifting the boundary by a few hours turns no five into a three. And neither does it disadvantage the opposing side: on a Friday one counts 14, 15, 16 Nisan — the third in Jewish days as well. The day-boundary is thus a real correction to the formulation "the first day of the week is our Sunday", but it is no argument in this count. [CALCULATED]
The seven forms, laid along the Wednesday chronology (death late Wednesday, resurrection Saturday evening):
#
construction
places
outcome
1
"three days and three nights"
Matt. 12:40
favours Wednesday
2
"after three days" (meta treis hêmeras)
Mark 8:31; 9:31; 10:34; Matt. 27:63
reconcilable (permits Wednesday)
3
"on the third day" (têi tritêi hêmerai)
Matt. 16:21 and elsewhere (9×)
reconcilable (Friday more natural)
4
"in/within three days"
John 2:19 and elsewhere
reconcilable (both)
5
"after a lapse of three days"
Mark 14:58 and elsewhere
reconcilable (both)
6
"until the third day"
Matt. 27:64
neutral
7
"this is the third day since"
Luke 24:21
favours Friday
One pair deserves separate treatment, for it is the sharpest form of the objection. The chief priests say to Pilate: "after three days I will rise" (27:63), and in the very next sentence they ask for the tomb to be secured "until the third day" (27:64). Two forms, one span, two verses — the same speakers, consecutive verses. That is no proof that the forms are interchangeable — both belong here to the flexible family, the ordinal form and the loose duration; the double construction (days and nights) does not appear — but it does show that first-century speakers switched within that family without friction. I therefore withdraw the claim that "after three days" as such favours the Wednesday: it does not; it permits it.
And beneath that pair lies the synoptic test, which confirms it. That the two forms designate the same end point is not to be seen in those two verses alone; it stands in the parallels themselves. Mark has the Lord foretell His resurrection three times, and the first time it runs "after three days" (Mark 8:31) — none of the printed text editions diverges there, the Byzantine majority text and the textus receptus included. In that same pericope, about that same prediction, Matthew (16:21) and Luke (9:22) write "on the third day", likewise without divergence. Two evangelists thus render, independently of each other, in the ordinal form what the third gives in the duration form — and that is possible only if both designate the same end point. At the second and third predictions that picture repeats itself in the critical text (Mark 9:31 and 10:34 against Matthew 17:23 and 20:19 and Luke 18:33); in the Byzantine tradition those two places in Mark themselves read the ordinal form, and the apparatus marks that as assimilation to the parallel place — which makes them unusable as an independent witness. But it does not touch the first prediction, and that one alone is enough. [DOCUMENTED]
What that yields, and what it does not. It takes one thing away: the thought that the nine "on the third day" places would measure something other than the sign of Jonah, and would therefore stand against the Wednesday. A stay of three full days ends within — at the close of — the third day; the duration form and the ordinal form interlock instead of colliding. But it yields no Wednesday, and it may not be read so: if the forms are interchangeable, then they are so for both chronologies.
A suggestion from outside, named here rather than used. Steven Law has put in discussion a sharper reconciliation of forms 2 and 3 than the one above: a rising in the twilight after sundown falls at the seam where, in rabbinic reckoning, two days touch, and such a moment could be described as "on the third day" and "after three days" alike. The underlying text exists, and it says less than the proposal needs. The baraita at Shabbat 34b runs: "בֵּין הַשְּׁמָשׁוֹת סָפֵק מִן הַיּוֹם וּמִן הַלַּיְלָה, סָפֵק כּוּלּוֹ מִן הַיּוֹם, סָפֵק כּוּלּוֹ מִן הַלַּיְלָה — מְטִילִין אוֹתוֹ לְחוֹמֶר שְׁנֵי יָמִים" — twilight is a doubt whether it belongs to the day or to the night, and the ruling is to impose the stringencies of both days upon it. That is an undecidability with a stringency attached, not a free assignment to whichever day one needs; and in the same baraita Rabbi Yehuda, Rabbi Neḥemiah and Rabbi Yosei dispute its very length, from three quarters of a mil down to the blink of an eye. [DOCUMENTED — the Aramaic seen at first hand; Sefaria's William Davidson text, a digital and not a critical edition, without apparatus]
And this book may not lean on it, for the reason with which it refuses the onah rule to the other side.Bein ha-shemashot is a rule about status periods — the onset of the sabbath, the days of a zav — precisely the class of which Chapter 3 says that it does not carry over to the duration of a narrative. Whoever denies the Friday that passage may not make it himself the moment it relieves him. The suggestion is therefore recorded and credited, and it is not counted: the reconciliation of the two forms rests where the synoptic parallels above put it, and on nothing else. [CONSTRUED]
The honest remainder, and it falls out otherwise than it seems. Construction 3 — the largest group — is usually called more natural for the Friday: there the crucifixion day itself is day one, and Sunday without more ado day three; the Wednesday would need an anchor of its own, and that looks like an ad hoc move. But count up what each reading really has to assume for it. On the Friday the count begins with the three hours that lie between the dying and sunset, and it ends in the night hours before the dawn of the Sunday: two parts of a day, both counted as a full day. On the Wednesday it begins at the fifteenth of Nisan and ends at the close of the seventeenth: three complete days, and not a single part that must pass for a whole. What the Wednesday does here is therefore not to choose an anchor of its own, but to apply the same rule this book lays down at the sign of Jonah — that a part of a day is not a day. Whoever rejects that rule there rejects it here as well, and then the count begins of itself at the first full day in the tomb. [CALCULATED]
And the place that contradicts it, for there is one. The only text where the New Testament itself counts days in a tomb points the other way: Lazarus has "four days in the tomb" (John 11:17; cf. tetartaios, "a man of four days", 11:39), and those four count from the day on which he died and was buried. That argues against a count that skips the day of burial, and it must be said — although with him, unlike here, the burial did not fall in the last minutes of the day, so that there that day was indeed a day in the tomb.
Where this comes out. Not at a decision, and I do not pretend otherwise. The other side answers that the ancient count simply was inclusive, and that a part of a day in that world simply counted; then the Friday is normal and the Wednesday the deviation. That objection stands, and it is the same objection the sign of Jonah already carries. But precisely there lies the gain of this section: the anchor point is no extra assumption on top of the thesis. It is the same question as the double construction, in another guise — and that question is decided not here but in Chapter 3, where it has been measured. Whoever reads the sign of Jonah literally counts here of himself from the first full day; whoever reads it idiomatically counts of himself from the day of death. The two stand or fall together, and that is precisely one measure. The "third day" corpus thus gives the Wednesday no preponderance: apart from the one exception (Matt. 12:40, which measures a duration and names no ordinal day), the third-day forms stand neutral or permit both chronologies — with the largest group in fact more natural for the Friday — and exactly one, Luke 24:21, favours the Friday hard. The corpus does not refute the Wednesday, but neither does it support it; it isolates the objection to this one verse and leaves the Wednesday's preponderance resting on the sign of Jonah, not on the third-day forms themselves. [CALCULATED]
The third layer is the question whether, after everything, a positive Wednesday reading of Luke 24:21 itself exists. And the honest answer is: no. I have exhausted the possibilities — reading the verb as "has elapsed" (no Greek parallel counts past the running unit like that); seeking another subject for agei; shifting the anchor point "these things" to the tomb guard (which does not even appear in Luke); reading the day as the prophetic "third day"; leaving the sabbaths out of the count. Every route breaks, or requires at least two assumptions the text does not give. There is no sentence in which a Wednesday chronologist can make Luke 24:21 say what he needs. What remains is one, and only one, tenable move — and it solves nothing; it absorbs. Namely: Luke is here reporting the words of confused disciples (a moment later he himself calls them "slow of understanding", anoêtoi), and he does not endorse their count as his own; and the framework within which their words fall is carried by the two-sabbaths reading (ch. 4). On that reading the verse stands not as a stone that crushes the thesis, but as a burden it carries: the thesis must assume that Luke puts into his characters' mouths a Friday-compatible sentence which he leaves uncontradicted.
That is thin, and I say it as it is: thin, but not incoherent. And yet "thin" is not the last word, for the question is not whether this move is elegant — it is not — but whether it is the better bargain. For this is a trade, and both camps make one. The Friday model reads this sigh cleanly, and breaks on the sign of Jonah: Friday to Sunday yields two nights, not the three the sign counts. The Wednesday model reads the sign cleanly — three days, three nights, literally — and pays with a strained reading of this sigh. The decisive point lies in the ranking of the witnesses. The sign of Jonah is the only sign the Lord deliberately left that generation — solemn, as a touchstone, with the nights counted in on purpose (ch. 3). Luke 24:21 is the sigh of two grieving travellers who do not yet even recognise Him. A strained reading of a passing complaint is a lighter price than dissolving the only deliberate sign. The better bargain is the Wednesday's.
And the sigh itself, once absorbed in this way, can also be honestly understood without any counting trick — not as yet another escape route (those have all been broken above), but as the tone in which that absorption sounds. "The third day" in the mouth of Cleopas is no calendar calculation but an echo — of Jesus' own promise, "on the third day He will rise", which the disciples had heard and were now anxiously counting down. He is saying, in effect: it is running towards the third day, the day of which He spoke, and still nothing for us but confused reports from women. It is plain, prophecy-coloured everyday speech — the voice of disappointment, not a chronological reference point. Whoever hears it so need not force it; it was never meant as a clock.
And here the verse turns round, from stumbling block to hinge — for look at what follows immediately. Cleopas seizes, without realising it, precisely the formula in which the Old Testament announces salvation — "on the third day" (Hosea 6:2) — at the very hour it is fulfilled; he has it on his lips and does not see it. And the Stranger beside him answers: "Ὦ foolish ones, and slow of heart to believe all that the prophets have spoken" — and He opens the Scriptures to them, "beginning at Moses and all the prophets" (Luke 24:25–27). What He unlocks there He utters Himself at the close: "thus it is written [...] that the Christ [...] should rise from the dead on the third day" (24:46). The third-day formula is thus not a point in time that Cleopas is computing, but the prophetic signal he brings forward and does not recognise — and precisely that blindness is the occasion on which the Scriptures open. Read as a reference point, his complaint points to Friday; as the prophetic signal it is, it points to the Risen One. That is no escape from the objection, but the scene as Luke stages it: steeped in irony.
Expertthe echo, and who heard it before this book
That "the third day" here echoes the promise and counts no calendar is no find of this book; it is the constant reading of the commentary tradition. Bengel noted in 1742 the bitter irony: the disciples lay down their hope "on the very day on which it is fulfilled" (Gnomon ad Lucam 24:21). John Gill read the number as the memory of "the day on which He had said He would rise", and set beside it — almost three hundred years before this book — the midrash that classes "the third day" as God's day of salvation. That midrash is worth reading at first hand, for it shows how charged the number was in Jewish ears. Bereshit Rabba opens with Hosea 6:2 — "after two days He will revive us; on the third day He will raise us up" — and then lists the third days of salvation: of the tribes, of the spies, of the giving of the Law, of Jonah, of the exiles, of the resurrection of the dead, of Esther (Bereshit Rabba 56:1). "The third day" there is no neutral date but a salvation formula — and, remarkably, the same formula under which the received text of the midrash also places Jonah's "three days and three nights" (the critical edition reads the lemma bare — ch. 3): the two signs that in this chapter seem to stand opposed are, in Judaism, two faces of one promise — salvation at the end of the third day. That same formula also had its dark reverse: the third day is the climax of mourning, the threshold on which hope dies — "for three days the soul returns to the grave, thinking it will return; when it sees that the face has changed, it departs" (Bereshit Rabba 100:7). A mourner who utters "the third day" is reaching for the formula of his hope and of its end, not for an abacus.
One thing belongs here in honesty, and it is the price. All these witnesses — from Bengel to the modern commentaries — assume the Friday, and for them the echo works precisely because it really is the third day. What this book does — hearing the echo while letting go of the count — none of them does; that uncoupling is attested in no source. The witnesses lift the sigh-reading from "a construal of this book" to "the reading of the tradition", but they do not take the Friday count away. So much, and no more, is what this verse yields for the Wednesday — and that is better said honestly than kept silent.
Excursuswhy this does not undermine the measurement of Matthew 12:40
One thing must be kept sharp, or what was rejected at the sign of Jonah comes in through the back door. The same midrash that makes "the third day" a salvation formula (Bereshit Rabba 56:1) also places — in the received text — Jonah's "three days and three nights" beneath it. Whoever reads that as a chronological argument — "see, the tradition equates the two" — collapses the measured three-full-days reading of Matthew 12:40 (ch. 3) into the loose ordinal "third day" that this book expressly rejected. That is not allowed, and it is not needed, for there are two different things here, on two levels. "Three days and three nights" is a duration — a span that counts the nights in, and that nowhere in the whole corpus searched ever shortens (ch. 3, appendix B). "The third day" is a moment of salvation — the day on which God saves, somewhere on that day, up to its very end. Hosea 6:2 itself sets the two side by side in a single breath: "after two days [the span] He will revive us; on the third day [the moment] He will raise us up." A stay of three full days therefore ends within — at the close of — the third day; the formula and the measurement collide nowhere, they interlock. This book therefore accepts the echo of Luke 24:21 solely for its typological charge — the third day as the signal of salvation — and emphatically not as a chronological equation; it is salvation-history language, not a calendar note (Lehmann, 1968; McArthur, 1971). And the measure cuts honestly both ways: the same reason why a late midrashic equation may not save the Wednesday is the reason why it may not support the Friday. Whoever wants to use it to deflate the sign of Jonah into "just the third day" is measuring with the measure this book lets pass nowhere — not even against itself.
Countercasethe verse at its strongest, without softening
Let the Friday make its best move here, unhindered. Luke 24:21 is not just a loose time reference; it is a casual time reference, and that makes it heavier, not lighter. People do not lie, and do not construct, in a complaint; they count the way they are used to counting. Two disciples who say "it is now the third day" reproduce, without thinking about it, the self-evident count of their time and tongue — and that count makes crucifixion day to resurrection day three days. That is exactly what a Friday delivers. Add to this that the passion narrative designates the day of death with a technical term for the Friday: paraskeuê (Mark even glosses it as prosabbaton, "the day before the sabbath", Mark 15:42) — and that term is attested outside the gospels, at its strongest in an Augustan decree, older than the gospels, which exempts the Jews from legal proceedings "on the sabbath, or on the day of preparation before it, from the ninth hour" (Josephus, Ant. 16.163): paraskeuê in an official formula, with a time of day attached, long before there was a church to shift the word (later also Martyrdom of Polycarp 7.1). The two-sabbaths reading contests that equation (ch. 4) — but it contests it; it does not refute it. Whoever weighs honestly admits: taken at its strongest, the text simply says what it seems to say, and that is Friday. And the earliest readers who heard "the third day" as a formula (see the Scholar box) heard it precisely as confirmation of that Friday — for them the formula and the count were one and the same.
One element of this countercase deserves a frontal answer, here and not in a subordinate clause, for it is its strongest linguistic card: the appeal to paraskeuê. Take the decree at its strongest, for that is how it ought to be taken. It is Augustan, older than the gospels, official, and it sets the word in a formula with a clock in it — if paraskeuê is called "Friday" anywhere before Christianity, it is here. But now read the formula itself. It does not say "on the sixth day"; it says "on the sabbath, or on the day of preparation before it" — and that "before it" is the whole point. In the decree the word stands not as the name of a weekday but as a function: the preparation day before the rest day named beside it, and without that rest day it points to nothing. [DOCUMENTED] It stands exactly so in the passion narrative itself. Mark has to gloss the word — prosabbaton, "the day before the sabbath" (15:42) — and a gloss explains a function word for readers who do not know it; a fixed weekday name needs no gloss. And John binds the same word to the feast: paraskeuê tou pascha, "the preparation of the Passover" (19:14), with, as the day that follows, a sabbath that was "high" (19:31). In the first century, then, the word means "eve of a rest day", and which rest day the context must supply; the fixed equation with the sixth weekday is the later, Christian narrowing — down to modern Greek, where paraskevi simply is "Friday" (ch. 6). And then the boundary, honest both ways. This answer takes nothing from the Friday that belongs to it: in the decree the rest day that follows is the weekly sabbath, and in any week without a feast day the preparation day therefore simply fell on the sixth day — the Friday reading of Mark 15:42 is linguistically entirely natural, and that stands. What it does show is that the word cannot compel the Friday: in a week with a festival sabbath it points just as naturally to the eve of the fifteenth of Nisan, and John demonstrably uses it so. The term does not decide the weekday; it repeats the question — which sabbath followed? — and that is precisely the question of ch. 4. What looked like a name turns out to be a function, and a function follows the rest day one reads behind it. [CONSTRUED — the weighing; the decree and the gloss DOCUMENTED] The rest of the countercase — the casual count itself — is not taken away here; it is precisely what this chapter has weighed above, and that weighing stands as it stands.
Excursusthe mirrored pair, and why the mirror is not symmetric
Set the two heaviest texts side by side and you see the true shape of the day-axis. Matthew 12:40 fits the Wednesday perfectly and forces the Friday into a strained reading of "three days and three nights" — a reading that, as Chapter 3 showed, is attested nowhere in the whole corpus. Luke 24:21 fits the Friday perfectly and forces the Wednesday into a strained reading of "the third day" — a reading equally unattested for Luke's own way of counting. They look like mirror images: each camp fits one text flawlessly and must bend the other. But the mirror is not symmetric, and in two ways it tilts towards the Wednesday. First, the nature of the texts: on one side a deliberately given sign, on the other a casual sigh — and the deliberate weighs more than the casual. Second, the nature of the forcing: the Friday's runs against a measured zero (the double construction counts inclusively nowhere in the corpus searched; the late interpreter who argues it so stands weighed in appendix B §B.3c), the Wednesday's against a single authorial convention. Neither is a proof. But together they point out the better bargain, and it lies on the Wednesday's side.
What the textual transmission adds to this — and it does add something real — stands gathered in the next excursus; it belongs there and not here, for this box is about the form of the argument and not about the manuscripts. One piece of honesty belongs here, however, for it touches the mirror itself, and it is directed against myself: that the old interpreters did tie themselves in knots over Matthew's three nights and yet read Luke 24:21 never in favour of a Wednesday is partly an artefact — the Wednesday thesis is modern; no church father defended it, so every ancient Wednesday reading of this verse is missing as a matter of course. What remains standing is only this: the natural, ancient reading of Luke 24:21 is Friday-compatible.
Excursusfive lines that lower the price, and none that pays it off
Here the primary material is gathered that presses down the price of this verse — on condition that one says honestly where each line stops. Five lines, each at its own strength, and none of the five pays the bill.
First, the manuscripts.The verb agei is absent from the oldest hand of Codex Sinaiticus and was added above the line by a corrector; on the facsimile this can be established with the eye — in the running text stands "[...] ἡμέραν ἀφ' οὗ ταῦτα ἐγένετο", and ΑΓΙ stands above it. That is more than a foothold: the word from which the objection derives its fixed form was unknown to the first hand of the oldest complete Bible. But it does not cancel the price, and I say at once why: even without that verb there remains τρίτην ταύτην ἡμέραν — "this third day" — and that is still three, and three since the crucifixion remains Friday. The omission makes the sentence elliptical, not different in number; it is related to the Syriac reading, but not identical with it. There is besides the claim that Codex Bezae omits the demonstrative tautên; that could not be verified against the manuscripts themselves and thus remains open, not load-bearing. [DOCUMENTED — ℵ\* seen at first hand; Bezae not]
Second, the grammar in Merx. Adalbert Merx — a textual critic of rank, and emphatically himself a defender of the Friday — calls the Greek sentence outright a "künstlich verschrobene Konstruktion" (artificially contorted), and asks bluntly: "denn wer ist Subject zu ἄγει? [...] Ist ἡμέραν ἄγει griechisch?" (Merx, 1905, pp. 526–528). The objection rests on agei as a fixed, well-formed "it is the third day"; Merx shows that this text is anything but smooth — a sentence of which a friend of the Friday asks whether it is Greek at all. That does not remove the grip, but it strips it of the appearance of an unshakable formulation. [DOCUMENTED]
Third, the Peshitta. The oldest Syriac version reads here not the ordinal but the cardinal "three days": ûhā tlātā yawmīn (ܘܗܳܐ ܬ݁ܠܳܬ݂ܳܐ ܝܰܘܡܺܝܢ) — "and behold, three days [...] since all these things happened", a verbless nominal sentence, without agei. And Merx's Latin rendering shows that the three Syriac witnesses — Sinaiticus, Curetonianus and the Peshitta — are unanimous in this: "ecce tres dies [...] ex quo haec omnia facta sunt" (p. 527). An old and broad version tradition thus eases the tension that the Greek "the third day" imposes. [DOCUMENTED] But here at once lies the boundary, and I draw it myself: "three days since" is still counted inclusively — Friday, Saturday, Sunday is three as well. The Peshitta removes an obstacle; it does not demonstrate the Wednesday. (And the simplest explanation for its divergence is not a different Vorlage but a translator smoothing contorted Greek — which argues against giving it too much weight.)
Fourth, Buttmann's own concession. The same Buttmann whom the Scholar box above calls as chief witness for the impersonal reading concedes in the same breath that it is grammatically an exception: the impersonal agei is here a ἅπαξ ῥηθέν — a thing said once — and, in his own words, "this supposition seems no less than the other to have the ordinary usage against it" (Buttmann, 1873, p. 134 [118]). The lesson cuts symmetrically, and precisely that is the gain: neither reading — impersonal or "supply Jesus" — has the idiom on its side. The verse is grammatically hard for everyone; the consensus may not act here as if it owns the null hypothesis.
Fifth, the lexical support. The impersonal "one counts / it is the so-many-eth day" reading, which the Scholar box above acknowledges as well attested, is not invented. LSJ classes ἄγω "of time, pass" and gives the classical parallel τέταρτον ἔτος ἄγων καὶ τριακοστόν (Galen, De libris propriis 1) — "being in his thirty-fourth year" — precisely the construction ordinal + ἄγω of τρίτην ταύτην ἡμέραν ἄγει (LSJ, 1940, s.v. ἄγω IV.4; cf. Sophocles, Electra 266, ποίας ἡμέρας δοκεῖς μ' ἄγειν;). But note where the parallel points: Galen has completed thirty-three years and is in his thirty-fourth — "one is in the third day" is therefore day three, and day three since the crucifixion remains Friday. The lexica support the reading; they do not move the number. [DOCUMENTED]
So it stands, without embellishment: five lines that turn the verse from a wall into a wobbly witness — a verb the first hand of the oldest Bible did not have, contorted Greek where a defender of the Friday asks whether it is Greek at all, a cardinal-reading version tradition, a chief witness who calls his own reading a hapax, and a reading that is lexically covered but does not move the count. Together they lower the price of the absorption; they do not pay it off. For ℵ\ still reads τρίτην [...] ἡμέραν* (the third day), the Peshitta counts its three days inclusively to Friday, and Galen's idiom places the speaker in day three. On its most natural reading Luke 24:21 continues to point to Friday. This is, and remains, the honest residue of the book — the one verse that weighs heaviest against the thesis. The four philological points are [DOCUMENTED]; that they lower the price and do not remit it is [CONSTRUED].
Where does all this come out, then? At an honest, incomplete balance, and that is exactly what this chapter ought to deliver. Luke 24:21 is the irreducible residue of the whole case — the one verse the Wednesday thesis cannot positively reconcile, only absorb. It is not a defeater: the reading in which Luke reports the count of confused disciples without endorsing it is logically admissible, and the wider "third day" field isolates the objection to this single point instead of strengthening it. But it is the strongest standing objection in the whole book, and the thesis pays a real price for it. I have not smoothed it over, and I do not declare it solved. But the honest final score is not a draw: it is a trade, and the Wednesday gets the better bargain. It pays with a strained reading of a casual sigh and wins with it the only deliberate sign; the Friday wins the sigh and loses the sign. That is the softest link in the book, honestly borne — and yet, weighed, on the right side of the trade.
And with that the heaviest text lies on the table. There remains the heaviest rival: not AD 33 with its shaky Good Friday, but AD 30 — the year that astronomy carries just as firmly as AD 31. That year deserves, like this verse, to be built up at its strongest before it is weighed.
HeartHEART
Two men are walking home, and the worst of it is not that their Master is dead, but that hope is dead: "we had hoped" — past tense. They count back the days the way mourners do, to be sure it really happened. And beside them walks, unrecognised, the only one who can settle the account. There is something disarming about a gospel that preserves this sentence — the sentence a defender of this date would most like to strike. Luke lets it stand, uncorrected, with all its discomfort attached. That is the hallmark of a witness who does not smooth things over. And it lays an obligation on anyone who builds his case on that same gospel: let the text that hurts stand too, and say honestly what it costs.
Note on the sources
The Greek material has been checked in the source itself; the reckoning-core (Sunday = "third day since" ⟺ Friday) can be recomputed. The impersonal reading of agei has been consulted through the standard works named (Thayer, Buttmann, Plummer); the exhaustion of the Wednesday routes is an inference from those primary data and is marked as such. The four price-lowering lines of the additional excursus have each been read in the primary text — Merx's German (pp. 526–528), the Syriac Peshitta text itself (with morphological parsing), Buttmann's Greek (p. 134/118), and the LSJ entry on ἄγω with the Galen parallel — partly via OCR and digital editions of those primary texts respectively; the German, Greek and Syriac carry reliably there. The text-critical footholds (ℵ\*, Codex Bezae, Old Syriac) could in part not be verified against the manuscripts themselves; they are therefore expressly presented as price-lowering, not as decisive. The Augustan decree (Josephus, Ant. 16.163) has been read at first hand in translation ("[...]nor on the day of the preparation to it, after the ninth hour"); the Greek formula itself remains a control point still to be closed, and the answer to the paraskeuê card leans solely on what the translation too carries — the relative "before it" and the ninth hour. The baraita on twilight (Shabbat 34b) has been read in the Aramaic, but in a digital and not a critical edition, so it carries without an apparatus; the delimitation drawn there — that a status rule does not transfer to a narrative duration — is the same one Chapter 3 applies to the onah rule, and it is applied here against this book's own convenience. The final judgement — "absorbable, at a price, no defeater" — is a weighing, not a measurement, and it is a proof for neither side.
Literature (APA)
Bengel, J. A. (1742). Gnomon Novi Testamenti (commentary on Luke 24:21). Tübingen.
Buttmann, A. (1873). A grammar of the New Testament Greek (J. H. Thayer, Trans.; p. 118). Warren F. Draper.
Danker, F. W. (Ed.). (2000). A Greek-English lexicon of the New Testament and other early Christian literature (3rd ed., s.v. ἄγω). University of Chicago Press.
Genesis Rabbah 56:1; 100:7 (The Midrash Rabbah, with the Hebrew source text). Sefaria. https://www.sefaria.org/Bereshit_Rabbah.56.1
Gill, J. (1763). An exposition of the New Testament (commentary on Luke 24:21). George Keith.
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. (pp. 242–243)
Lehmann, K. (1968). Auferweckt am dritten Tag nach der Schrift (Quaestiones Disputatae 38). Herder. (The "third day" as a salvation-history formula, not a calendar note; in Luke, however, Lehmann does read the evangelist himself as counting — the formula reading applies to the tradition/speaker.)
Liddell, H. G., Scott, R., & Jones, H. S. (1940). A Greek-English lexicon (9th ed., s.v. ἄγω). Clarendon Press.
McArthur, H. K. (1971). "On the third day". New Testament Studies, 18(1), 81–86. (Traces the formula back to Hosea 6:2 in Jewish-exegetical reading; the equation of "after three days" and "on the third day" is assumed by him and disputed by Walker (1960).)
Merx, A. (1905). Die vier kanonischen Evangelien nach ihrem ältesten bekannten Texte, Teil II/2: Erläuterung — Markus und Lukas (pp. 526–528). Georg Reimer. (Greek of Luke 24:21 as a "künstlich verschrobene Konstruktion"; the three Syriac witnesses read "ecce tres dies". German read at first hand; Merx himself holds to the Friday.)
Babylonian Talmud, Shabbat 34b [the baraita on bein ha-shemashot]. Consulted with the Aramaic, via Sefaria's William Davidson text — a digital edition, without critical apparatus. https://www.sefaria.org/Shabbat.34b
Peshitta (Vetus Syra), Luke 24:21 — Syriac text with SEDRA/CAL morphology. Comprehensive Aramaic Lexicon / dukhrana.com. (ܘܗܳܐ ܬ݁ܠܳܬ݂ܳܐ ܝܰܘܡܺܝܢ, cardinal "three days", verbless; parsed at first hand.)
Walker, N. (1960). "After three days". Novum Testamentum, 4(4), 261–262.
Plummer, A. (1896). A critical and exegetical commentary on the Gospel according to S. Luke (International Critical Commentary; ad Lucam 24:21). T&T Clark.
Thayer, J. H. (1889). A Greek-English lexicon of the New Testament (s.v. ἄγω). Harper & Brothers.
Primary texts: Hosea 6:2 · Luke 13:32; 24:7, 21, 25–27, 46 · Acts 10:30 · Mark 8:31; 9:31; 10:34; 14:58; 15:42 · Matthew 12:40; 16:21; 27:63–64 · John 2:19; 19:14, 19:31 · 2 Peter 3:1 · Josephus, Antiquitates Judaicae 16.163 · Martyrium Polycarpi 7.1 · Galen, De libris propriis 1 (τέταρτον ἔτος ἄγων καὶ τριακοστόν) · Sophocles, Electra 266 · Codex Sinaiticus (ℵ) ad Lucam 24:21. Greek: this book holds to the received text (Scrivener 1894 and Stephanus 1550). Every Greek reading that carries weight in this chapter has been collated against eight printed editions — both textus recepti, RP-Byz, the text of the Greek church, Westcott-Hort, Tischendorf⁸, Nestle 1904 and SBL — and where they diverge that is stated in the text. Author's own working translation.
Thus far the case has argued for. Now I turn it round. An argument that shows only its own strong points is not an argument but a plea, and the reader has a right to the opposite — to the strongest text that speaks against this date, not tucked away in a footnote but up front, at its strongest. There is such a text. There is exactly one that truly hurts, and I am not going to smooth it over.
It is a sentence from the mouths of two disillusioned men on the road to Emmaus. It is Sunday. They are walking with a stranger — it is the Risen One, but they do not know it — and they tell Him, with the despondency of people whose hope has died, what has happened. And then they say: "and besides all this, it is now the third day since these things happened" (Luke 24:21).
The third day. On Sunday. Since the crucifixion.
Do the arithmetic, and the grip is felt at once. From Friday to Sunday is, counted Jewish-inclusively, the third day: Friday one, Saturday two, Sunday three. Exactly. But from Wednesday to Sunday? Thursday, Friday, Saturday, Sunday — that is the fourth day, and counted inclusively even the fifth. On no method of counting is Sunday the third day after a Wednesday. This is the mirror image of the sign of Jonah, and it weighs exactly as much. Where Matthew 12:40 asks for three nights the Friday cannot supply, Luke 24:21 names a third day the Wednesday cannot reach. Even Hoehner, the defender of the Friday, concedes how much the whole Wednesday case hangs on that one verse of Matthew: had there been no Matthew 12:40, he writes, the Wednesday theory would probably never have been proposed (Hoehner, 1974, pp. 242–243). Fair is fair: that point bites. And it bites here.
So let me not pretend that this verse resolves neatly. It does not. What follows is not a defusing but a weighing — and it ends in a trade, not a knock-out: this verse can be absorbed, at a real price, and the question is which side gets the better bargain. Not to be reasoned away, but to be weighed honestly. I shall show that in three layers, and each layer yields less ground than the one before.
The first layer is the grammar, and it falls largely the right way — but it does not save the case. There is an old discussion about the verb in that sentence, agei (ἄγει). Read personally, the subject is "he" or "one"; read impersonally it means roughly the Latin agitur, "it is passing, one is spending" — "the third day is under way". The defenders of the Wednesday have laid great weight on that impersonal reading, and linguistically they have a point: the impersonal reading is not marginal; it stands in Thayer, in Buttmann, in Plummer. But look closely at what it delivers, and what it does not. It shifts the subject of the sentence; it does not shift the number. "The third day is passing" is still the third day. Whether you read "one is living through the third day" or "it is the third day" — it remains three, and three since the crucifixion remains Friday. The grammatical gain is real, but it touches the wrong variable. Whoever hoped that agei would turn a three into a five is disappointed; no reading of that verb does that.
Expertthe verb, and why it does not move the count
The sentence runs: tritên tautên hêmeran agei aph' hou tauta egeneto (τρίτην ταύτην ἡμέραν ἄγει ἀφ' οὗ ταῦτα ἐγένετο).
And here something must be set right at once that tells against this book. That is the critical text. The received text reads one word more: *"τρίτην ταύτην ἡμέραν ἄγει σήμερον ἀφ' οὗ ταῦτα ἐγένετο"* — "today he is spending this third day". So read Stephanus 1550, Scrivener 1894, the Byzantine majority text and the text of the Greek church; Westcott-Hort, Tischendorf, Nestle and SBL all four omit σήμερον. [DOCUMENTED — the eight editions laid side by side] This book holds to the received text, and then it ought to cite the heavier reading: σήμερον binds "the third day" expressly to the day of speaking, and that is Sunday. The received text thus makes the objection harder, not softer. I take that as it comes and do not conceal it; whoever chooses a form of the text chooses it also where it tells against him. The impersonal reading of agei (≈ Latin agitur, "is being spent") is well attested — Thayer s.v. ἄγω; Buttmann (1873, p. 118), who expressly gives "ἡ ἡμέρα ἄγεται" as "the day is passing"; and Plummer (1896) at this very place: "we are at the third day". But it is and remains a present — one is in the third day; it is not over (that would require a perfect tense). And Luke elsewhere counts inclusively in exactly this way: in Luke 13:32 "today" is day one and "the third day" is the day after tomorrow; in Acts 10:30 the same author reckons "ἀπὸ τετάρτης ἡμέρας" inclusively. The demonstrative tautên ("this, the third day") binds the day of speaking — Sunday — to the ordinal three, just as "this, the second letter" in 2 Peter 3:1. In short: impersonal or personal, the count comes out at three, and three since the crucifixion is Friday. The grammar does not relieve the Wednesday.
The second layer is the wider "third day" field, and it changes the picture — not by refuting the verse, but by isolating it. For "the third day" is not the only way the New Testament designates the time between cross and resurrection. There are seven different constructions, and if you lay all seven along the Wednesday chronology (with the death late on Wednesday and the resurrection towards the end of the third full-day unit, Saturday towards sunset), it turns out that they do not stand against the Wednesday en bloc. One of them favours the Wednesday outright — "three days and three nights" (Matthew 12:40). Four constructions can be reconciled with either chronology — among them "after three days" (meta treis hêmeras), which sits somewhat less smoothly on a Friday but does not exclude it (see the Workbench). One is neutral. And one — exactly one — favours the Friday: Luke 24:21. So the wider field does not stand against the Wednesday; it isolates the objection to this one verse. That is no escape, but it is the right measure of the matter: not "the third day refutes the Wednesday", but "one of the seven third-day forms refutes it, one supports it, and the rest permit both chronologies".
Dossierthe seven constructions and the reckoning-core
The reckoning-core. Sunday is called "the third day since X" if and only if X, counted inclusively, is Friday (exclusively: Thursday). From a Wednesday, Sunday is inclusively the fifth day, exclusively the fourth. No single re-anchoring makes Sunday "day three" under a Wednesday; every attempt stacks at least two assumptions foreign to the text.
A way out that suggests itself, and does not work. One often points out here — and rightly — that the Jewish day begins not at midnight but at sunset, so that "the first day of the week" already begins on our Saturday evening. That is correct, and it is load-bearing elsewhere in this book: Matthew 28:1 sets the walk to the tomb precisely on that seam — "Ὀψὲ δὲ σαββάτων, τῇ ἐπιφωσκούσῃ εἰς μίαν σαββάτων", "late on the sabbath, as it grew light towards the first day of the week" — a reading in which the eight printed editions do not diverge, the two textus recepti and the majority text included. [DOCUMENTED] But for this verse it does not help, and I would rather say so myself than have a reader discover it. The conversation on the Emmaus road falls in the daylight of that first day — the travellers later urge the Stranger to stay "for it is towards evening and the day is far spent" (24:29). Wherever one places the day-boundary, the span remains the same: from the fourteenth of Nisan to the eighteenth is inclusively the fifth and exclusively the fourth day. Shifting the boundary by a few hours turns no five into a three. And neither does it disadvantage the opposing side: on a Friday one counts 14, 15, 16 Nisan — the third in Jewish days as well. The day-boundary is thus a real correction to the formulation "the first day of the week is our Sunday", but it is no argument in this count. [CALCULATED]
The seven forms, laid along the Wednesday chronology (death late Wednesday, resurrection Saturday evening):
One pair deserves separate treatment, for it is the sharpest form of the objection. The chief priests say to Pilate: "after three days I will rise" (27:63), and in the very next sentence they ask for the tomb to be secured "until the third day" (27:64). Two forms, one span, two verses — the same speakers, consecutive verses. That is no proof that the forms are interchangeable — both belong here to the flexible family, the ordinal form and the loose duration; the double construction (days and nights) does not appear — but it does show that first-century speakers switched within that family without friction. I therefore withdraw the claim that "after three days" as such favours the Wednesday: it does not; it permits it.
And beneath that pair lies the synoptic test, which confirms it. That the two forms designate the same end point is not to be seen in those two verses alone; it stands in the parallels themselves. Mark has the Lord foretell His resurrection three times, and the first time it runs "after three days" (Mark 8:31) — none of the printed text editions diverges there, the Byzantine majority text and the textus receptus included. In that same pericope, about that same prediction, Matthew (16:21) and Luke (9:22) write "on the third day", likewise without divergence. Two evangelists thus render, independently of each other, in the ordinal form what the third gives in the duration form — and that is possible only if both designate the same end point. At the second and third predictions that picture repeats itself in the critical text (Mark 9:31 and 10:34 against Matthew 17:23 and 20:19 and Luke 18:33); in the Byzantine tradition those two places in Mark themselves read the ordinal form, and the apparatus marks that as assimilation to the parallel place — which makes them unusable as an independent witness. But it does not touch the first prediction, and that one alone is enough. [DOCUMENTED]
What that yields, and what it does not. It takes one thing away: the thought that the nine "on the third day" places would measure something other than the sign of Jonah, and would therefore stand against the Wednesday. A stay of three full days ends within — at the close of — the third day; the duration form and the ordinal form interlock instead of colliding. But it yields no Wednesday, and it may not be read so: if the forms are interchangeable, then they are so for both chronologies.
A suggestion from outside, named here rather than used. Steven Law has put in discussion a sharper reconciliation of forms 2 and 3 than the one above: a rising in the twilight after sundown falls at the seam where, in rabbinic reckoning, two days touch, and such a moment could be described as "on the third day" and "after three days" alike. The underlying text exists, and it says less than the proposal needs. The baraita at Shabbat 34b runs: "בֵּין הַשְּׁמָשׁוֹת סָפֵק מִן הַיּוֹם וּמִן הַלַּיְלָה, סָפֵק כּוּלּוֹ מִן הַיּוֹם, סָפֵק כּוּלּוֹ מִן הַלַּיְלָה — מְטִילִין אוֹתוֹ לְחוֹמֶר שְׁנֵי יָמִים" — twilight is a doubt whether it belongs to the day or to the night, and the ruling is to impose the stringencies of both days upon it. That is an undecidability with a stringency attached, not a free assignment to whichever day one needs; and in the same baraita Rabbi Yehuda, Rabbi Neḥemiah and Rabbi Yosei dispute its very length, from three quarters of a mil down to the blink of an eye. [DOCUMENTED — the Aramaic seen at first hand; Sefaria's William Davidson text, a digital and not a critical edition, without apparatus]
And this book may not lean on it, for the reason with which it refuses the onah rule to the other side. Bein ha-shemashot is a rule about status periods — the onset of the sabbath, the days of a zav — precisely the class of which Chapter 3 says that it does not carry over to the duration of a narrative. Whoever denies the Friday that passage may not make it himself the moment it relieves him. The suggestion is therefore recorded and credited, and it is not counted: the reconciliation of the two forms rests where the synoptic parallels above put it, and on nothing else. [CONSTRUED]
The honest remainder, and it falls out otherwise than it seems. Construction 3 — the largest group — is usually called more natural for the Friday: there the crucifixion day itself is day one, and Sunday without more ado day three; the Wednesday would need an anchor of its own, and that looks like an ad hoc move. But count up what each reading really has to assume for it. On the Friday the count begins with the three hours that lie between the dying and sunset, and it ends in the night hours before the dawn of the Sunday: two parts of a day, both counted as a full day. On the Wednesday it begins at the fifteenth of Nisan and ends at the close of the seventeenth: three complete days, and not a single part that must pass for a whole. What the Wednesday does here is therefore not to choose an anchor of its own, but to apply the same rule this book lays down at the sign of Jonah — that a part of a day is not a day. Whoever rejects that rule there rejects it here as well, and then the count begins of itself at the first full day in the tomb. [CALCULATED]
And the place that contradicts it, for there is one. The only text where the New Testament itself counts days in a tomb points the other way: Lazarus has "four days in the tomb" (John 11:17; cf. tetartaios, "a man of four days", 11:39), and those four count from the day on which he died and was buried. That argues against a count that skips the day of burial, and it must be said — although with him, unlike here, the burial did not fall in the last minutes of the day, so that there that day was indeed a day in the tomb.
Where this comes out. Not at a decision, and I do not pretend otherwise. The other side answers that the ancient count simply was inclusive, and that a part of a day in that world simply counted; then the Friday is normal and the Wednesday the deviation. That objection stands, and it is the same objection the sign of Jonah already carries. But precisely there lies the gain of this section: the anchor point is no extra assumption on top of the thesis. It is the same question as the double construction, in another guise — and that question is decided not here but in Chapter 3, where it has been measured. Whoever reads the sign of Jonah literally counts here of himself from the first full day; whoever reads it idiomatically counts of himself from the day of death. The two stand or fall together, and that is precisely one measure. The "third day" corpus thus gives the Wednesday no preponderance: apart from the one exception (Matt. 12:40, which measures a duration and names no ordinal day), the third-day forms stand neutral or permit both chronologies — with the largest group in fact more natural for the Friday — and exactly one, Luke 24:21, favours the Friday hard. The corpus does not refute the Wednesday, but neither does it support it; it isolates the objection to this one verse and leaves the Wednesday's preponderance resting on the sign of Jonah, not on the third-day forms themselves. [CALCULATED]
The third layer is the question whether, after everything, a positive Wednesday reading of Luke 24:21 itself exists. And the honest answer is: no. I have exhausted the possibilities — reading the verb as "has elapsed" (no Greek parallel counts past the running unit like that); seeking another subject for agei; shifting the anchor point "these things" to the tomb guard (which does not even appear in Luke); reading the day as the prophetic "third day"; leaving the sabbaths out of the count. Every route breaks, or requires at least two assumptions the text does not give. There is no sentence in which a Wednesday chronologist can make Luke 24:21 say what he needs. What remains is one, and only one, tenable move — and it solves nothing; it absorbs. Namely: Luke is here reporting the words of confused disciples (a moment later he himself calls them "slow of understanding", anoêtoi), and he does not endorse their count as his own; and the framework within which their words fall is carried by the two-sabbaths reading (ch. 4). On that reading the verse stands not as a stone that crushes the thesis, but as a burden it carries: the thesis must assume that Luke puts into his characters' mouths a Friday-compatible sentence which he leaves uncontradicted.
That is thin, and I say it as it is: thin, but not incoherent. And yet "thin" is not the last word, for the question is not whether this move is elegant — it is not — but whether it is the better bargain. For this is a trade, and both camps make one. The Friday model reads this sigh cleanly, and breaks on the sign of Jonah: Friday to Sunday yields two nights, not the three the sign counts. The Wednesday model reads the sign cleanly — three days, three nights, literally — and pays with a strained reading of this sigh. The decisive point lies in the ranking of the witnesses. The sign of Jonah is the only sign the Lord deliberately left that generation — solemn, as a touchstone, with the nights counted in on purpose (ch. 3). Luke 24:21 is the sigh of two grieving travellers who do not yet even recognise Him. A strained reading of a passing complaint is a lighter price than dissolving the only deliberate sign. The better bargain is the Wednesday's.
And the sigh itself, once absorbed in this way, can also be honestly understood without any counting trick — not as yet another escape route (those have all been broken above), but as the tone in which that absorption sounds. "The third day" in the mouth of Cleopas is no calendar calculation but an echo — of Jesus' own promise, "on the third day He will rise", which the disciples had heard and were now anxiously counting down. He is saying, in effect: it is running towards the third day, the day of which He spoke, and still nothing for us but confused reports from women. It is plain, prophecy-coloured everyday speech — the voice of disappointment, not a chronological reference point. Whoever hears it so need not force it; it was never meant as a clock.
And here the verse turns round, from stumbling block to hinge — for look at what follows immediately. Cleopas seizes, without realising it, precisely the formula in which the Old Testament announces salvation — "on the third day" (Hosea 6:2) — at the very hour it is fulfilled; he has it on his lips and does not see it. And the Stranger beside him answers: "Ὦ foolish ones, and slow of heart to believe all that the prophets have spoken" — and He opens the Scriptures to them, "beginning at Moses and all the prophets" (Luke 24:25–27). What He unlocks there He utters Himself at the close: "thus it is written [...] that the Christ [...] should rise from the dead on the third day" (24:46). The third-day formula is thus not a point in time that Cleopas is computing, but the prophetic signal he brings forward and does not recognise — and precisely that blindness is the occasion on which the Scriptures open. Read as a reference point, his complaint points to Friday; as the prophetic signal it is, it points to the Risen One. That is no escape from the objection, but the scene as Luke stages it: steeped in irony.
Expertthe echo, and who heard it before this book
That "the third day" here echoes the promise and counts no calendar is no find of this book; it is the constant reading of the commentary tradition. Bengel noted in 1742 the bitter irony: the disciples lay down their hope "on the very day on which it is fulfilled" (Gnomon ad Lucam 24:21). John Gill read the number as the memory of "the day on which He had said He would rise", and set beside it — almost three hundred years before this book — the midrash that classes "the third day" as God's day of salvation. That midrash is worth reading at first hand, for it shows how charged the number was in Jewish ears. Bereshit Rabba opens with Hosea 6:2 — "after two days He will revive us; on the third day He will raise us up" — and then lists the third days of salvation: of the tribes, of the spies, of the giving of the Law, of Jonah, of the exiles, of the resurrection of the dead, of Esther (Bereshit Rabba 56:1). "The third day" there is no neutral date but a salvation formula — and, remarkably, the same formula under which the received text of the midrash also places Jonah's "three days and three nights" (the critical edition reads the lemma bare — ch. 3): the two signs that in this chapter seem to stand opposed are, in Judaism, two faces of one promise — salvation at the end of the third day. That same formula also had its dark reverse: the third day is the climax of mourning, the threshold on which hope dies — "for three days the soul returns to the grave, thinking it will return; when it sees that the face has changed, it departs" (Bereshit Rabba 100:7). A mourner who utters "the third day" is reaching for the formula of his hope and of its end, not for an abacus.
One thing belongs here in honesty, and it is the price. All these witnesses — from Bengel to the modern commentaries — assume the Friday, and for them the echo works precisely because it really is the third day. What this book does — hearing the echo while letting go of the count — none of them does; that uncoupling is attested in no source. The witnesses lift the sigh-reading from "a construal of this book" to "the reading of the tradition", but they do not take the Friday count away. So much, and no more, is what this verse yields for the Wednesday — and that is better said honestly than kept silent.
Excursuswhy this does not undermine the measurement of Matthew 12:40
One thing must be kept sharp, or what was rejected at the sign of Jonah comes in through the back door. The same midrash that makes "the third day" a salvation formula (Bereshit Rabba 56:1) also places — in the received text — Jonah's "three days and three nights" beneath it. Whoever reads that as a chronological argument — "see, the tradition equates the two" — collapses the measured three-full-days reading of Matthew 12:40 (ch. 3) into the loose ordinal "third day" that this book expressly rejected. That is not allowed, and it is not needed, for there are two different things here, on two levels. "Three days and three nights" is a duration — a span that counts the nights in, and that nowhere in the whole corpus searched ever shortens (ch. 3, appendix B). "The third day" is a moment of salvation — the day on which God saves, somewhere on that day, up to its very end. Hosea 6:2 itself sets the two side by side in a single breath: "after two days [the span] He will revive us; on the third day [the moment] He will raise us up." A stay of three full days therefore ends within — at the close of — the third day; the formula and the measurement collide nowhere, they interlock. This book therefore accepts the echo of Luke 24:21 solely for its typological charge — the third day as the signal of salvation — and emphatically not as a chronological equation; it is salvation-history language, not a calendar note (Lehmann, 1968; McArthur, 1971). And the measure cuts honestly both ways: the same reason why a late midrashic equation may not save the Wednesday is the reason why it may not support the Friday. Whoever wants to use it to deflate the sign of Jonah into "just the third day" is measuring with the measure this book lets pass nowhere — not even against itself.
Countercasethe verse at its strongest, without softening
Let the Friday make its best move here, unhindered. Luke 24:21 is not just a loose time reference; it is a casual time reference, and that makes it heavier, not lighter. People do not lie, and do not construct, in a complaint; they count the way they are used to counting. Two disciples who say "it is now the third day" reproduce, without thinking about it, the self-evident count of their time and tongue — and that count makes crucifixion day to resurrection day three days. That is exactly what a Friday delivers. Add to this that the passion narrative designates the day of death with a technical term for the Friday: paraskeuê (Mark even glosses it as prosabbaton, "the day before the sabbath", Mark 15:42) — and that term is attested outside the gospels, at its strongest in an Augustan decree, older than the gospels, which exempts the Jews from legal proceedings "on the sabbath, or on the day of preparation before it, from the ninth hour" (Josephus, Ant. 16.163): paraskeuê in an official formula, with a time of day attached, long before there was a church to shift the word (later also Martyrdom of Polycarp 7.1). The two-sabbaths reading contests that equation (ch. 4) — but it contests it; it does not refute it. Whoever weighs honestly admits: taken at its strongest, the text simply says what it seems to say, and that is Friday. And the earliest readers who heard "the third day" as a formula (see the Scholar box) heard it precisely as confirmation of that Friday — for them the formula and the count were one and the same.
One element of this countercase deserves a frontal answer, here and not in a subordinate clause, for it is its strongest linguistic card: the appeal to paraskeuê. Take the decree at its strongest, for that is how it ought to be taken. It is Augustan, older than the gospels, official, and it sets the word in a formula with a clock in it — if paraskeuê is called "Friday" anywhere before Christianity, it is here. But now read the formula itself. It does not say "on the sixth day"; it says "on the sabbath, or on the day of preparation before it" — and that "before it" is the whole point. In the decree the word stands not as the name of a weekday but as a function: the preparation day before the rest day named beside it, and without that rest day it points to nothing. [DOCUMENTED] It stands exactly so in the passion narrative itself. Mark has to gloss the word — prosabbaton, "the day before the sabbath" (15:42) — and a gloss explains a function word for readers who do not know it; a fixed weekday name needs no gloss. And John binds the same word to the feast: paraskeuê tou pascha, "the preparation of the Passover" (19:14), with, as the day that follows, a sabbath that was "high" (19:31). In the first century, then, the word means "eve of a rest day", and which rest day the context must supply; the fixed equation with the sixth weekday is the later, Christian narrowing — down to modern Greek, where paraskevi simply is "Friday" (ch. 6). And then the boundary, honest both ways. This answer takes nothing from the Friday that belongs to it: in the decree the rest day that follows is the weekly sabbath, and in any week without a feast day the preparation day therefore simply fell on the sixth day — the Friday reading of Mark 15:42 is linguistically entirely natural, and that stands. What it does show is that the word cannot compel the Friday: in a week with a festival sabbath it points just as naturally to the eve of the fifteenth of Nisan, and John demonstrably uses it so. The term does not decide the weekday; it repeats the question — which sabbath followed? — and that is precisely the question of ch. 4. What looked like a name turns out to be a function, and a function follows the rest day one reads behind it. [CONSTRUED — the weighing; the decree and the gloss DOCUMENTED] The rest of the countercase — the casual count itself — is not taken away here; it is precisely what this chapter has weighed above, and that weighing stands as it stands.
Excursusthe mirrored pair, and why the mirror is not symmetric
Set the two heaviest texts side by side and you see the true shape of the day-axis. Matthew 12:40 fits the Wednesday perfectly and forces the Friday into a strained reading of "three days and three nights" — a reading that, as Chapter 3 showed, is attested nowhere in the whole corpus. Luke 24:21 fits the Friday perfectly and forces the Wednesday into a strained reading of "the third day" — a reading equally unattested for Luke's own way of counting. They look like mirror images: each camp fits one text flawlessly and must bend the other. But the mirror is not symmetric, and in two ways it tilts towards the Wednesday. First, the nature of the texts: on one side a deliberately given sign, on the other a casual sigh — and the deliberate weighs more than the casual. Second, the nature of the forcing: the Friday's runs against a measured zero (the double construction counts inclusively nowhere in the corpus searched; the late interpreter who argues it so stands weighed in appendix B §B.3c), the Wednesday's against a single authorial convention. Neither is a proof. But together they point out the better bargain, and it lies on the Wednesday's side.
What the textual transmission adds to this — and it does add something real — stands gathered in the next excursus; it belongs there and not here, for this box is about the form of the argument and not about the manuscripts. One piece of honesty belongs here, however, for it touches the mirror itself, and it is directed against myself: that the old interpreters did tie themselves in knots over Matthew's three nights and yet read Luke 24:21 never in favour of a Wednesday is partly an artefact — the Wednesday thesis is modern; no church father defended it, so every ancient Wednesday reading of this verse is missing as a matter of course. What remains standing is only this: the natural, ancient reading of Luke 24:21 is Friday-compatible.
Excursusfive lines that lower the price, and none that pays it off
Here the primary material is gathered that presses down the price of this verse — on condition that one says honestly where each line stops. Five lines, each at its own strength, and none of the five pays the bill.
First, the manuscripts. The verb agei is absent from the oldest hand of Codex Sinaiticus and was added above the line by a corrector; on the facsimile this can be established with the eye — in the running text stands "[...] ἡμέραν ἀφ' οὗ ταῦτα ἐγένετο", and ΑΓΙ stands above it. That is more than a foothold: the word from which the objection derives its fixed form was unknown to the first hand of the oldest complete Bible. But it does not cancel the price, and I say at once why: even without that verb there remains τρίτην ταύτην ἡμέραν — "this third day" — and that is still three, and three since the crucifixion remains Friday. The omission makes the sentence elliptical, not different in number; it is related to the Syriac reading, but not identical with it. There is besides the claim that Codex Bezae omits the demonstrative tautên; that could not be verified against the manuscripts themselves and thus remains open, not load-bearing. [DOCUMENTED — ℵ\* seen at first hand; Bezae not]
Second, the grammar in Merx. Adalbert Merx — a textual critic of rank, and emphatically himself a defender of the Friday — calls the Greek sentence outright a "künstlich verschrobene Konstruktion" (artificially contorted), and asks bluntly: "denn wer ist Subject zu ἄγει? [...] Ist ἡμέραν ἄγει griechisch?" (Merx, 1905, pp. 526–528). The objection rests on agei as a fixed, well-formed "it is the third day"; Merx shows that this text is anything but smooth — a sentence of which a friend of the Friday asks whether it is Greek at all. That does not remove the grip, but it strips it of the appearance of an unshakable formulation. [DOCUMENTED]
Third, the Peshitta. The oldest Syriac version reads here not the ordinal but the cardinal "three days": ûhā tlātā yawmīn (ܘܗܳܐ ܬ݁ܠܳܬ݂ܳܐ ܝܰܘܡܺܝܢ) — "and behold, three days [...] since all these things happened", a verbless nominal sentence, without agei. And Merx's Latin rendering shows that the three Syriac witnesses — Sinaiticus, Curetonianus and the Peshitta — are unanimous in this: "ecce tres dies [...] ex quo haec omnia facta sunt" (p. 527). An old and broad version tradition thus eases the tension that the Greek "the third day" imposes. [DOCUMENTED] But here at once lies the boundary, and I draw it myself: "three days since" is still counted inclusively — Friday, Saturday, Sunday is three as well. The Peshitta removes an obstacle; it does not demonstrate the Wednesday. (And the simplest explanation for its divergence is not a different Vorlage but a translator smoothing contorted Greek — which argues against giving it too much weight.)
Fourth, Buttmann's own concession. The same Buttmann whom the Scholar box above calls as chief witness for the impersonal reading concedes in the same breath that it is grammatically an exception: the impersonal agei is here a ἅπαξ ῥηθέν — a thing said once — and, in his own words, "this supposition seems no less than the other to have the ordinary usage against it" (Buttmann, 1873, p. 134 [118]). The lesson cuts symmetrically, and precisely that is the gain: neither reading — impersonal or "supply Jesus" — has the idiom on its side. The verse is grammatically hard for everyone; the consensus may not act here as if it owns the null hypothesis.
Fifth, the lexical support. The impersonal "one counts / it is the so-many-eth day" reading, which the Scholar box above acknowledges as well attested, is not invented. LSJ classes ἄγω "of time, pass" and gives the classical parallel τέταρτον ἔτος ἄγων καὶ τριακοστόν (Galen, De libris propriis 1) — "being in his thirty-fourth year" — precisely the construction ordinal + ἄγω of τρίτην ταύτην ἡμέραν ἄγει (LSJ, 1940, s.v. ἄγω IV.4; cf. Sophocles, Electra 266, ποίας ἡμέρας δοκεῖς μ' ἄγειν;). But note where the parallel points: Galen has completed thirty-three years and is in his thirty-fourth — "one is in the third day" is therefore day three, and day three since the crucifixion remains Friday. The lexica support the reading; they do not move the number. [DOCUMENTED]
So it stands, without embellishment: five lines that turn the verse from a wall into a wobbly witness — a verb the first hand of the oldest Bible did not have, contorted Greek where a defender of the Friday asks whether it is Greek at all, a cardinal-reading version tradition, a chief witness who calls his own reading a hapax, and a reading that is lexically covered but does not move the count. Together they lower the price of the absorption; they do not pay it off. For ℵ\ still reads τρίτην [...] ἡμέραν* (the third day), the Peshitta counts its three days inclusively to Friday, and Galen's idiom places the speaker in day three. On its most natural reading Luke 24:21 continues to point to Friday. This is, and remains, the honest residue of the book — the one verse that weighs heaviest against the thesis. The four philological points are [DOCUMENTED]; that they lower the price and do not remit it is [CONSTRUED].
Where does all this come out, then? At an honest, incomplete balance, and that is exactly what this chapter ought to deliver. Luke 24:21 is the irreducible residue of the whole case — the one verse the Wednesday thesis cannot positively reconcile, only absorb. It is not a defeater: the reading in which Luke reports the count of confused disciples without endorsing it is logically admissible, and the wider "third day" field isolates the objection to this single point instead of strengthening it. But it is the strongest standing objection in the whole book, and the thesis pays a real price for it. I have not smoothed it over, and I do not declare it solved. But the honest final score is not a draw: it is a trade, and the Wednesday gets the better bargain. It pays with a strained reading of a casual sigh and wins with it the only deliberate sign; the Friday wins the sigh and loses the sign. That is the softest link in the book, honestly borne — and yet, weighed, on the right side of the trade.
And with that the heaviest text lies on the table. There remains the heaviest rival: not AD 33 with its shaky Good Friday, but AD 30 — the year that astronomy carries just as firmly as AD 31. That year deserves, like this verse, to be built up at its strongest before it is weighed.
HeartHEART
Two men are walking home, and the worst of it is not that their Master is dead, but that hope is dead: "we had hoped" — past tense. They count back the days the way mourners do, to be sure it really happened. And beside them walks, unrecognised, the only one who can settle the account. There is something disarming about a gospel that preserves this sentence — the sentence a defender of this date would most like to strike. Luke lets it stand, uncorrected, with all its discomfort attached. That is the hallmark of a witness who does not smooth things over. And it lays an obligation on anyone who builds his case on that same gospel: let the text that hurts stand too, and say honestly what it costs.
Note on the sources
The Greek material has been checked in the source itself; the reckoning-core (Sunday = "third day since" ⟺ Friday) can be recomputed. The impersonal reading of agei has been consulted through the standard works named (Thayer, Buttmann, Plummer); the exhaustion of the Wednesday routes is an inference from those primary data and is marked as such. The four price-lowering lines of the additional excursus have each been read in the primary text — Merx's German (pp. 526–528), the Syriac Peshitta text itself (with morphological parsing), Buttmann's Greek (p. 134/118), and the LSJ entry on ἄγω with the Galen parallel — partly via OCR and digital editions of those primary texts respectively; the German, Greek and Syriac carry reliably there. The text-critical footholds (ℵ\*, Codex Bezae, Old Syriac) could in part not be verified against the manuscripts themselves; they are therefore expressly presented as price-lowering, not as decisive. The Augustan decree (Josephus, Ant. 16.163) has been read at first hand in translation ("[...]nor on the day of the preparation to it, after the ninth hour"); the Greek formula itself remains a control point still to be closed, and the answer to the paraskeuê card leans solely on what the translation too carries — the relative "before it" and the ninth hour. The baraita on twilight (Shabbat 34b) has been read in the Aramaic, but in a digital and not a critical edition, so it carries without an apparatus; the delimitation drawn there — that a status rule does not transfer to a narrative duration — is the same one Chapter 3 applies to the onah rule, and it is applied here against this book's own convenience. The final judgement — "absorbable, at a price, no defeater" — is a weighing, not a measurement, and it is a proof for neither side.
Literature (APA)
Bengel, J. A. (1742). Gnomon Novi Testamenti (commentary on Luke 24:21). Tübingen.
Buttmann, A. (1873). A grammar of the New Testament Greek (J. H. Thayer, Trans.; p. 118). Warren F. Draper.
Danker, F. W. (Ed.). (2000). A Greek-English lexicon of the New Testament and other early Christian literature (3rd ed., s.v. ἄγω). University of Chicago Press.
Genesis Rabbah 56:1; 100:7 (The Midrash Rabbah, with the Hebrew source text). Sefaria. https://www.sefaria.org/Bereshit_Rabbah.56.1
Gill, J. (1763). An exposition of the New Testament (commentary on Luke 24:21). George Keith.
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. (pp. 242–243)
Lehmann, K. (1968). Auferweckt am dritten Tag nach der Schrift (Quaestiones Disputatae 38). Herder. (The "third day" as a salvation-history formula, not a calendar note; in Luke, however, Lehmann does read the evangelist himself as counting — the formula reading applies to the tradition/speaker.)
Liddell, H. G., Scott, R., & Jones, H. S. (1940). A Greek-English lexicon (9th ed., s.v. ἄγω). Clarendon Press.
McArthur, H. K. (1971). "On the third day". New Testament Studies, 18(1), 81–86. (Traces the formula back to Hosea 6:2 in Jewish-exegetical reading; the equation of "after three days" and "on the third day" is assumed by him and disputed by Walker (1960).)
Merx, A. (1905). Die vier kanonischen Evangelien nach ihrem ältesten bekannten Texte, Teil II/2: Erläuterung — Markus und Lukas (pp. 526–528). Georg Reimer. (Greek of Luke 24:21 as a "künstlich verschrobene Konstruktion"; the three Syriac witnesses read "ecce tres dies". German read at first hand; Merx himself holds to the Friday.)
Babylonian Talmud, Shabbat 34b [the baraita on bein ha-shemashot]. Consulted with the Aramaic, via Sefaria's William Davidson text — a digital edition, without critical apparatus. https://www.sefaria.org/Shabbat.34b
Peshitta (Vetus Syra), Luke 24:21 — Syriac text with SEDRA/CAL morphology. Comprehensive Aramaic Lexicon / dukhrana.com. (ܘܗܳܐ ܬ݁ܠܳܬ݂ܳܐ ܝܰܘܡܺܝܢ, cardinal "three days", verbless; parsed at first hand.)
Walker, N. (1960). "After three days". Novum Testamentum, 4(4), 261–262.
Plummer, A. (1896). A critical and exegetical commentary on the Gospel according to S. Luke (International Critical Commentary; ad Lucam 24:21). T&T Clark.
Thayer, J. H. (1889). A Greek-English lexicon of the New Testament (s.v. ἄγω). Harper & Brothers.
Primary texts: Hosea 6:2 · Luke 13:32; 24:7, 21, 25–27, 46 · Acts 10:30 · Mark 8:31; 9:31; 10:34; 14:58; 15:42 · Matthew 12:40; 16:21; 27:63–64 · John 2:19; 19:14, 19:31 · 2 Peter 3:1 · Josephus, Antiquitates Judaicae 16.163 · Martyrium Polycarpi 7.1 · Galen, De libris propriis 1 (τέταρτον ἔτος ἄγων καὶ τριακοστόν) · Sophocles, Electra 266 · Codex Sinaiticus (ℵ) ad Lucam 24:21. Greek: this book holds to the received text (Scrivener 1894 and Stephanus 1550). Every Greek reading that carries weight in this chapter has been collated against eight printed editions — both textus recepti, RP-Byz, the text of the Greek church, Westcott-Hort, Tischendorf⁸, Nestle 1904 and SBL — and where they diverge that is stated in the text. Author's own working translation.