When Was the Book of Daniel Written?

Summary

Daniel says it was written by a Jewish exile in Babylon in the sixth century BC.  Most academic scholars say it was written in Judea about 165 BC, during the persecution under Antiochus IV Epiphanes.  The gap is nearly four hundred years, and a great deal rides on it — though less than either side usually admits.

• The question is not one question. Daniel is two books stitched together, and the evidence for each half runs the opposite way. Chapters 1–6 are court tales set in Babylon; chapters 7–12 are visions. The book also switches language mid-sentence: Hebrew in 1:1–2:4a, Aramaic in 2:4b–7:28, Hebrew again in 8:1–12:13. Almost every argument for an early date is an argument about the Aramaic half. Almost every argument for a late date is an argument about the Hebrew visions. Both can be right at once, and most popular treatments on both sides never notice.
• Daniel 11 is the whole late-date case, and it is strong. Verses 2–39 track Hellenistic dynastic politics in detail found nowhere else in the Bible: the Ptolemy–Seleucus succession, Berenice's marriage and murder, Raphia, Panium, Cleopatra I given "to destroy the kingdom," Antiochus III's defeat by a Roman commander and his death robbing a temple, Seleucus IV's tax collector, then twenty verses on Antiochus IV. Then, at 11:40–45, it goes wrong. Antiochus does not die "between the seas and the beautiful holy mountain."¹ Polybius says he "died at Tabae in Persia, smitten with madness"; 1 Maccabees has him die of grief in Persia after failing to plunder a temple at Elymais.^{2 3} Accuracy that stops dead at a particular year, and then fails, is the signature of history written as prophecy.
• Porphyry made this argument in the third century AD, and the church's reply was not that he had the history wrong. Jerome preserves it: Porphyry held that Daniel "did not foretell the future so much as he related the past," that everything up to Antiochus "contained authentic history," and that anything beyond that point "was false."⁴ Jerome's answer was not to dispute the fit; he conceded it was so exact that Porphyry "could not deny that they had taken place," and argued instead that 11:36–45 jumps to a future Antichrist.⁴ Every conservative reading since has needed that jump, and the Hebrew gives no marker for it.
• The numbers look like a deadline being recalculated. Daniel gives four figures for the same crisis: 2,300 evenings and mornings (8:14), "a time, times, and half a time" (7:25), 1,290 days (12:11), and 1,335 days (12:12).¹ The desecration ran from 15 Kislev of Seleucid year 145 to the rededication on 25 Kislev of year 148 — three years and ten days, about 1,105 days.³ A prediction revised upward twice, in adjacent verses, is what happens when a date passes and the crisis has not ended.
• Belshazzar is the strongest thing the early date has, and it is very strong. Daniel 5 names a king of Babylon unknown to Herodotus, to Berossus, and to every classical source; his name was recovered from cuneiform only in the nineteenth century.⁵ A foundation cylinder Nabonidus buried at Ur around 540 BC prays for "Belshazzar, the eldest son — my offspring" beside the king himself;⁶ the Nabonidus Chronicle has the crown prince holding Babylonia while Nabonidus was away at Tema;⁷ the Verse Account says Nabonidus "entrusted the kingship to him" and left.⁸ Three independent documents. That combination explains the oddest detail in the chapter: Daniel is made "third ruler in the kingdom" (5:29) — the highest post available when the first two were taken.¹ No one has explained how a Judean writing in 165 BC would know any of this.
• But Daniel 5 also gets things wrong, and honesty requires saying so. Belshazzar's father was Nabonidus, not Nebuchadnezzar (5:2, 5:11), and Aramaic ab covering "ancestor" or "predecessor" is a real but unverified escape. No cuneiform text gives Belshazzar the title king; contracts continued to be dated by Nabonidus's regnal years, and Belshazzar never performed the New Year ritual.⁵ The chapter holds knowledge nobody in the second century had, and it also holds errors nobody in the sixth century should have made.
• Darius the Mede is the hardest problem in the book. Daniel has a Mede aged about sixty-two take Babylon after Belshazzar and reign before Cyrus (5:31, 9:1, 6:28).¹ The Nabonidus Chronicle leaves no room for him: Ugbaru of Gutium entered Babylon on 16 Tashritu, Cyrus arrived eighteen days later, Ugbaru died three weeks after that.⁷ Wiseman's own survey concedes the difficulty and offers Darius = Cyrus, reading 6:28 as "Darius, that is, Cyrus the Persian"; Whitcomb offered Gubaru the governor; Rowley concluded he was fictitious.⁹ Wiseman also concedes there is "no clear evidence" that Cyrus was ever called son of Xerxes or bore another name.⁹
• The one piece of hard physical evidence is a radiocarbon date, and both sides are misreading it. In 2025 a Groningen-led team dated Qumran scroll 4Q114 — which preserves Daniel 10:5–11:29, including Antiochus IV's Egyptian campaigns — to 2168 ± 15 BP.¹⁰ Calibrated, that is 355–285 BC (49.5%) or 230–160 BC (45.9%); the team rejects the older band on script grounds.¹⁰ Associates for Biblical Research reported it as "230–160 (mean 195) BC," which treats a two-humped distribution as if it had a mean.¹¹ Three critical scholars replying in February 2026 said the range carries "equal probability throughout," which the paper's own published intervals contradict: within the accepted band, 34.5 of 45.9 percentage points — three-quarters of the probability — sit in a thirty-year window, 205–175 BC.^{12 10} On the study's own numbers, the likeliest date for this copy of Daniel 11 is before Antiochus IV took the throne.
• Three Greek words, and not one Greek word of government. The Greek in Daniel is three musical instruments in chapter 3. Under a century and a half of Greek rule, an author writing in 165 BC produced no Greek administrative, military, or political vocabulary at all — while using around twenty Old Persian loanwords, four of which the Septuagint translators could no longer translate.¹³ Kitchen's honest conclusion is still the best statement of the linguistic evidence: the Aramaic "cannot be decided upon linguistic grounds alone," and it is "equally obscurantist" to rule out a sixth-century date or to treat one as proven.¹³ Even Mitchell and Joyce, arguing the conservative side on the instruments, concede that if sumponyah really is Greek symphōnia meaning an instrument, "then its use in Daniel is consonant with a late date."¹⁴
• Most of the apologetic weight does not rest here at all. The seventy weeks of Daniel 9 still predate Jesus by roughly two centuries at a 165 BC date. So do the "one like a son of man" of Daniel 7 and the bodily resurrection of Daniel 12. What does not survive a 165 BC date is Daniel 11 as predictive prophecy, the reading that makes the fourth kingdom Rome, and — the real stake — the book's own first-person claims: "I, Daniel, saw" (7:15, 8:1, 9:2, 10:2). If those are a literary convention, the argument is theological, not evidential, and should be conducted as such.

Verdict:  undecided, and the honest reading splits the book.  The court tales carry sixth-century Babylonian knowledge that a second-century Judean should not have had, in an Aramaic that sits with the Persian chancery rather than with Qumran.  The visions carry second-century Hebrew, second-century detail, and an ending that failed in 164 BC.  The two strongest facts on the table point in opposite directions:  Daniel 11:40–45 is wrong exactly where the ex-eventu theory predicts, and the only physically dated copy of Daniel 11 most likely predates Antiochus IV's reign.  Anyone who tells you this is settled — in either direction — is quoting one of those facts and suppressing the other.

Why the date matters

Set the date at 530 BC and Daniel 11 is a prediction of two centuries of Hellenistic history so detailed that a skeptic must either accept prophecy or find a way out.  Set it at 165 BC and Daniel 11 is a competent history of the recent past, and the apologetic argument evaporates.

That is the whole reason the question is fought so hard.  It is also why the fight is often bad-tempered.  Conservatives treat the sixth-century date as a test of orthodoxy;  the academy treats the second-century date as a test of competence.  Neither posture helps anyone weigh evidence.

The useful question is narrower:  which arguments actually depend on the answer?

So the answer to "what breaks?" is:  one prophecy argument, one prophetic-chart tradition, and the book's own voice.  Not the messianic material, which is what most people think is at issue.

The three positions

Sixth century, about 530 BC.  The book is what it says it is.  Held by Gleason Archer, E. J. Young, Stephen Miller, Andrew Steinmann, Joe Sprinkle, and J. Paul Tanner;  defended on historical and linguistic grounds by the 1965 Tyndale volume edited by D. J. Wiseman, which is still the best conservative case in print.⁹

Second century, about 165 BC.  The visions are vaticinium ex eventu — prophecy after the event — written during the Antiochene persecution.  This is the mainstream academic position and has been since Driver.  Notably, it is also held by a number of evangelical scholars:  Ernest Lucas in the Apollos commentary, John Goldingay in the revised Word commentary, and Aubrey Buster with John Walton in the 2025 NICOT volume.¹²

Composite.  The court tales of chapters 1–6 are older — third century or earlier, possibly much earlier — and the visions of 7–12 are Maccabean.  This is now the ordinary form of the critical position, and the February 2026 statement of it is blunt:  "The Book of Daniel as we know it is formed by a combination of the four visions focused on the crisis under Antiochus with an older story collection."¹²

Keep the third option in view.  Most of the shouting happens because each side attacks the other's weakest half.

Part 1 — The case for the second century

The prophecy that stops

Daniel 10–12 is a single vision, delivered in the third year of Cyrus, running from the Persian empire to the end.  Here is what chapter 11 says, beside what happened.

Nothing else in the Old Testament resembles this.  Isaiah names Cyrus;  Daniel 11 gives dynastic marriages, the fate of individual princesses, and the outcome of specific battles, in order, for roughly a hundred and sixty years — and gets essentially all of it right.

Two features of that accuracy are worth noticing before anyone decides what it proves.

It is lopsided.  The Persian empire lasted about two centuries and had ten kings after Cyrus.  Daniel 11:2 gives it one verse and four kings, then jumps to Alexander.  Jerome noticed the gap:  "after he has specified four kings of Persia after Cyrus, the author [i.e., Daniel] omits the nine others and passes right on to Alexander."  He explained it as selection:  "the Spirit of prophecy was not concerned about preserving historical detail but in summarizing only the most important matters."⁴  The Hellenistic period gets thirty-seven verses.  A revelation given in Cyrus's third year has no reason to be thin about the empire it was given in and dense about the empire two centuries later.  An author living in the second century has every reason.

It thins out at the end.  Verses 21–39 are still recognizable but increasingly generic.  Verses 40–45 stop matching anything.

Where it stops

Daniel 11:40–45 predicts a final campaign:  the king of the south attacks, the king of the north sweeps through Egypt with Libyans and Cushites in his train, is alarmed by news from the east and north, and pitches his royal tents "between the seas toward the beautiful holy mountain."  Then:  "he will come to his end, with no one to help him."¹

Antiochus IV did none of this.  He had already invaded Egypt twice, in 170–169 and 168, and Rome had thrown him out;  there was no third campaign, no Libyan and Cushite train, and no return to Judea.  He died in the opposite direction.  Polybius, writing within a generation:

In Syria King Antiochus, wishing to provide himself with money, decided to make an expedition against the sanctuary of Artemis in Elymaïs.  On reaching the spot he was foiled in his hopes, as the barbarian tribes who dwelt in the neighbourhood would not permit the outrage, and on his retreat he died at Tabae in Persia, smitten with madness. — Polybius, Histories 31.9²

1 Maccabees, independently, has him hear of his generals' defeat while in Persia, take to his bed, and say "I am perishing with great grief in a foreign land."  He died in Seleucid year 149.³

So the vision is exact for a century and a half and then fails on the one event it places at "the time of the end."  That is precisely the pattern a second-century composition predicts:  the author's own present is the seam.  Where he stops reporting and starts guessing, he is wrong.

The conservative answer has been the same since Jerome:  somewhere around verse 36 the subject changes from Antiochus to a future Antichrist.  That answer has one thing going for it and two against it.

For it:  Daniel does move into eschatology.  Chapter 12 opens with Michael arising, "a time of distress unlike any other," and the resurrection of the dead.  None of that happened in 164 BC.  The text plainly reaches past Antiochus somewhere.

Against it:  first, the Hebrew gives no break.  Verse 36 begins "the king will do as he pleases" with no new subject introduced;  the natural referent is the despicable person of verse 21.  Second, the transition conservatives need is not marked by "time of the end," because Daniel already uses that phrase for the Antiochus vision:  "the vision pertains to the time of the end" (8:17), of a chapter everyone — including Jerome — reads as Antiochus.¹

Jerome's own handling is instructive.  He does not argue that Porphyry's history was wrong.  He argues that Porphyry's identification of the holy mountain was strained, that Antiochus never held Libya or Ethiopia, and that in any case "even though he were able to prove that these statements applied to Antiochus instead of the Antichrist, what does that matter to us?"⁴  That is a theological answer, not a historical one.

The oldest readers already read it as Antiochus

The Antiochene reading of Daniel 11 is not a modern skeptical invention.  It is the oldest reading on record, and it is embedded in the text's own earliest translation.

The Old Greek of Daniel — the first translation, made in the later second century BC¹² — does not say "the king of the south" in chapter 11.  It says "the king of Egypt."  Verse 40:  "according to the time of consummation the king of Egypt will lock horns with him."  Verses 5, 8, 9, 11, 25 and 42 do the same.³  The translator was not guessing at a distant future;  he was reading a document about the Ptolemies and the Seleucids and rendering it accordingly, within a century of the crisis.

Josephus, around AD 93, says the same about the parallel vision in chapter 8:  "it so came to pass, that our nation suffered these things under Antiochus Epiphanes, according to Daniel's vision, and what he wrote many years before they came to pass."¹⁵  Josephus is a hostile witness to the late date — he thinks Daniel wrote centuries earlier and counts him among "the greatest of the prophets"¹⁵ — and he still identifies the villain as Antiochus.  He then adds:  "In the very same manner Daniel also wrote concerning the Roman government."  Both readings, in one paragraph, from a first-century Jew.

So the sequence runs:  the Old Greek translator reads Egypt and Syria;  Josephus reads Antiochus and Rome;  Porphyry reads Antiochus and concludes the book is a forgery;  Jerome concedes the Antiochus fit and moves the ending to a future Antichrist.  No ancient interpreter on record read Daniel 11:21–35 as anything but Antiochus IV.  The dispute has always been about where he stops.

The numbers get revised

Daniel gives four separate figures for how long the desecration lasts.

The actual interval is fixed by 1 Maccabees.  The abomination went up on the altar on 15 Kislev of Seleucid year 145;  the first pagan sacrifice on the 25th;  the rededication on 25 Kislev of year 148.³  That is three years and ten days from the abomination, or exactly three years from the first sacrifice — roughly 1,100 to 1,105 days.

The 1,150 of Daniel 8:14 lands within about seven weeks of the true figure.  Then 12:11 raises it to 1,290, and 12:12 raises it again to 1,335, in consecutive verses, with a blessing attached to whoever "waits" for the longer number.  A revelation that gives one number is a prophecy.  A revelation that gives four, ascending, with the last two side by side, reads like a community recalculating a deadline that keeps not arriving.

This is the argument that carries the most weight with people who have actually read the end of the book, and it has no good conservative answer.  The usual reply — that the extra 45 days are a further period after the Second Coming — is possible, but it is a reading imposed to save the numbers, not one the text asks for.

The Persian period is too short

Daniel 11:2 gives Persia four kings.  Daniel 9 gives seventy weeks — 490 years — from a word about restoring Jerusalem to the end of the crisis.¹

Run that on the critical reading, where the starting word is Jeremiah's prophecy that Daniel 9:2 has just been reading, and the endpoint is the murder of the high priest Onias III in 171 BC.  From 605 BC, 483 years lands in 122 BC.  From 587 BC, in 104 BC.  Either way the scheme overshoots the real interval by fifty to seventy years.

That error is diagnostic.  Second-temple Jewish chronology badly compressed the Persian period — the later rabbinic Seder Olam allowed it fifty-two years, thirty-four of them after the Temple was rebuilt, against the true two hundred and seven (see Traditional Jewish Chronology).  An author in 165 BC counting backwards through a shortened Persian period would produce exactly this kind of overshoot.  A sixth-century prophet given a number by an angel should not.

Conservatives have a reply, and it is the one the site treats separately:  the seventy weeks are not aimed at Antiochus at all but at the Messiah, and the arithmetic works to the first century on a 360-day year from a Persian decree.  That is a live argument.  But notice what it costs:  it requires Daniel 9 to be about something entirely different from Daniel 8, 11, and 12, which are all about Antiochus, and it leaves 11:2's four Persian kings unexplained.

The historical problems in chapters 1–6

Darius the Mede.  This is the sharpest historical difficulty in the book, and conservative scholarship says so.  Wiseman opens his own defense by quoting Rowley:  the references to Darius the Mede "have long been recognised as providing the most serious historical problem in the book."⁹

The problem is not that Darius is obscure.  It is that the sequence is occupied.  The Nabonidus Chronicle is a contemporary Babylonian document:  Cyrus defeated the Akkadian army at Opis;  Sippar fell on the 14th of Tashritu;  Nabonidus fled;  on the 16th, Ugbaru governor of Gutium entered Babylon with Cyrus's army without a battle;  on the 3rd of Arahsamna Cyrus himself entered;  Gubaru appointed the district officers;  on the night of the 11th of Arahsamna Ugbaru died.⁷  There is no gap for a Median king aged sixty-two to receive the kingdom and reign a year or more before Cyrus.

Three rescues have been tried.

Gubaru the governor (Whitcomb, 1959).  Wiseman's own objection stands:  Gubaru is nowhere called king, is nowhere called a Mede, is nowhere called son of Ahasuerus, and "there is no known instance of a governor of Babylon, other than a usurper, in the first millennium bearing a royal title and claiming a throne."⁹

Cyrus himself (Wiseman, 1957).  Daniel 6:28 is read "Daniel prospered in the reign of Darius, even the reign of Cyrus the Persian" — a known Hebrew construction, as in 1 Chronicles 5:26's Pul/Tiglath-pileser.  Cyrus was about sixty-two in 539, was of Median descent through his mother, and appointed governors.  Wiseman is careful to call it "only a theory," and concedes there is "no clear evidence for or against the claim that Cyrus was a 'son of Xerxes', or that he bore another name."⁹

Cyaxares II (Anderson, 2014).  A revival of the pre-1870s reading, resting on Xenophon's Cyropaedia, which has Cyrus serve a Median uncle-king.  Xenophon is the problem:  the Cyropaedia is a didactic romance, and Herodotus and the Babylonian records both have Astyages as the last Median king, deposed by Cyrus around 550.

None of these is nonsense.  None of them has an inscription.  A reader who is not already committed should record this as a genuine strike against the book's historical footing.

The four kingdoms follow from Darius.  This is the part of the Darius problem that carries the most weight, and it is usually discussed separately as though it were unrelated.

Daniel 2 and 7 each present four successive kingdoms.  Everyone agrees the first is Babylon.  The question is what the second and third are — and the answer decides what the fourth is.  Chapter 8 supplies the control:  its little horn grows out of the Greek goat and is unmistakably Antiochus IV, and it does what chapter 7's little horn does — magnifies itself, stops the sacrifice, persecutes the holy ones.  If the two little horns are the same figure, chapter 7's fourth kingdom is Greece.

If Daniel really believed a Median empire ruled Babylon before Persia — which is what "Darius the Mede received the kingdom" after Belshazzar amounts to — then the sequence is Babylon, Media, Persia, Greece, and the fourth kingdom is Antiochus IV's.  The book ends where the crisis of 167 BC ends, and there is no Rome in it.  Wiseman's own survey states the opposing view in the words of a critic he is answering:  Daniel 6:28 is "the clearest evidence of the book's belief in a Median empire between the Babylonian and the Persian," and Darius is "a figure constructed purely on the basis of a theory of the Median empire."⁹

If Darius is Cyrus, as Wiseman argues, the sequence is Babylon, Medo-Persia, Greece, Rome — and Wiseman says so explicitly:  "This new theory should lead to the re-examination of any interpretation of the four world empires which required a separate Median and Persian empire, since the writer could now be shown to maintain the view of a single Medo-Persian realm throughout."⁹

That is the honest structure of the argument, and it should be stated where both sides can see it:  the identity of Darius the Mede and the identity of the fourth kingdom stand or fall together.  A conservative who cannot produce Darius cannot produce Rome.  Daniel 8:20's "the ram that you saw with the two horns stands for the kings of Media and Persia" — one animal, two horns, one kingdom — is the best evidence on the conservative side, and it is a real point.¹  But it sits against Daniel 6, where a Median king reigns and a Persian succeeds him.

Daniel 1:1.  Nebuchadnezzar besieges Jerusalem "in the third year of the reign of King Jehoiakim of Judah."¹  Jeremiah 25:1 and 46:2 put Nebuchadnezzar's accession and Carchemish in Jehoiakim's fourth year.  The standard conservative answer — Daniel counts by the Babylonian accession-year system, Jeremiah by the Judean system — is reasonable and is probably right.⁹  The residual problem is the siege itself:  the Babylonian Chronicle records the Carchemish campaign but no siege of Jerusalem in 605.  Wiseman states the position fairly:  "The argument against a specific Babylonian siege rests on silence just as must, at present, any defence of it."⁹

Nebuchadnezzar's madness.  Daniel 4 has Nebuchadnezzar driven from men for seven periods and restored.  No Babylonian source knows of it.  A Qumran text does know of something very like it, about a different king:  the Prayer of Nabonidus (4Q242) has Nabonidus "afflicted with an evil ulcer … for seven years," driven far from men at Tema, forgiven by a Jewish exorcist.¹⁶  Nabonidus really did spend about ten years at Tema, and the Verse Account really does accuse him of being deranged.⁸

The critical reading is that Daniel 4 has moved a Nabonidus tradition onto the more famous Nebuchadnezzar.  The conservative reading is that two kings' illnesses circulated and Daniel preserves the older one.  The tie-breaker most people miss:  the combination of Daniel 4 with Daniel 5 is what makes the transfer suspicious.  Chapter 5 shows the author knows Belshazzar, Nabonidus's son.  Chapter 4 shows him attaching Nabonidus's story to Nebuchadnezzar.  Whoever wrote these chapters had good Babylonian material and did not always attach it to the right king.

The languages

The classic statement is Driver's, and it is worth quoting exactly, because it is quoted constantly and usually mangled:

The verdict of the language of Daniel is thus clear.  The Persian words presuppose a period after the Persian empire had been well established:  the Greek words demand, the Hebrew supports, and the Aramaic permits, a date after the conquest of Palestine by Alexander the Great (b.c. 332). — S. R. Driver, An Introduction to the Literature of the Old Testament, p. 508¹⁷

Note what Driver actually claims.  The Greek demands a post-332 date.  The Hebrew merely supports it.  The Aramaic only permits it.  Driver himself ranks his own evidence, and two-thirds of it is soft.

Greek.  Three words, all in the orchestra of Daniel 3:  qaytros (kithara), pesanterin (psalterion), sumponyah (symphonia).  Kitchen's reply is that Greek goods, mercenaries and artisans were in the Near East from the eighth century — Greek pottery at Samaria around 750, at Nineveh, and "even to Babylon of all places" — so a Greek instrument name in sixth-century Babylon is not impossible.¹³  Mitchell and Joyce, handling the words themselves, concede the weak point plainly:  psalterion is not attested before Aristotle, and if sumponyah is symphonia used of an instrument, "then its use in Daniel is consonant with a late date for the composition of the book."¹⁴  They propose instead that it transliterates a dialect form of tympanon, and stress that "some of them are extremely uncertain."¹⁴

Hebrew.  Driver's list of some thirty late expressions was answered word by word by W. J. Martin in the 1965 volume, who showed that several — malkut above all — occur in Numbers, Samuel, Jeremiah.¹⁸  What Martin does not overturn is the general impression Driver reports from Delitzsch:  Daniel's Hebrew "resembles the Hebrew of the Chronicler," and the Chronicler wrote around the start of the Greek period.¹⁷  This matters more than it looks, because Daniel 8–12 is the Hebrew part.  Whatever the Aramaic proves, it does not date the visions.

Aramaic.  Here the conservative case is at its strongest, and it is still not decisive.  Kitchen's fifty-page study concludes that Daniel's Aramaic is Imperial Aramaic — the chancery language of the Persian empire — and that its word order "places it squarely in full-blooded Imperial Aramaic, and in striking contrast with real Palestinian post-Imperial Aramaic of the second and first centuries BC as illustrated by the Dead Sea Scrolls."¹³  His summary is scrupulous:

If proper allowance be made for attested scribal usage in the Biblical Near East … then there is nothing to decide the date of composition of the Aramaic of Daniel on the grounds of Aramaic anywhere between the late sixth and the second century BC. … It is equally obscurantist to exclude dogmatically a sixth-fifth (or fourth) century date on the one hand, or to hold such a date as mechanically proven on the other. — K. A. Kitchen, "The Aramaic of Daniel," p. 79¹³

Anyone who cites Kitchen as having proved a sixth-century date has not read his last page.

Where Daniel sits in the Jewish canon

The Hebrew Bible has three divisions.  Daniel is not in the Prophets.  The Talmud's list is explicit:  the Prophets are Joshua, Judges, Samuel, Kings, Jeremiah, Ezekiel, Isaiah and the Twelve;  the Writings are "Ruth and the book of Psalms, and Job and Proverbs;  Ecclesiastes, Song of Songs, and Lamentations;  Daniel and the Scroll of Esther;  and Ezra and Chronicles."¹⁹

The late-date argument is that the prophetic collection closed before Daniel existed, so Daniel had to go in the third division.

This argument is weaker than it looks, and the reason is chronological.  Baba Batra is a rabbinic source of the fifth or sixth century AD.  Every earlier list that survives puts Daniel with the prophets — including two that explicitly claim to be reporting Jewish practice.

Melito, bishop of Sardis, writing about AD 170, says he travelled east specifically to settle the question — "when I went East and came to the place where these things were preached and done, I learned accurately the books of the Old Testament" — and his list runs:  "of Prophets, Isaiah, Jeremiah;  of the twelve prophets, one book;  Daniel, Ezekiel, Esdras."²⁰  Daniel sits inside the prophetic block, between the Twelve and Ezekiel.

Origen, about seventy years later, gives the twenty-two books "as the Hebrews have handed them down," each with its Hebrew name, and the order is:  "Isaiah, Jessia;  Jeremiah, with Lamentations and the epistle in one, Jeremia;  Daniel, Daniel;  Ezekiel, Jezekiel;  Job, Job;  Esther, Esther."²¹  Same position.

Josephus, around AD 93, calls Daniel "one of the greatest of the prophets."¹⁵

The honest caveat is that Melito and Origen are Christians, and the Greek Bible they used groups Daniel with the prophets, so their arrangement may reflect Greek practice rather than Hebrew.  But both of them assert they are reporting the Jewish list, and neither shows any awareness of a three-part order with Daniel in the third part.  Whatever put Daniel in the Writings, the first evidence for it is centuries after the period at issue.

Placement in the Writings may also reflect a judgment about Daniel's role — a courtier and dream-interpreter, not a prophet sent to Israel — rather than about his date.  This argument should be demoted.  It is not evidence of a second-century composition;  it is evidence about how rabbinic Judaism eventually organised its Bible.

Ben Sira's silence

Around 180 BC Jesus ben Sira wrote a "Praise of the Fathers" running from Enoch to the high priest Simon.  He names Isaiah, Jeremiah, Ezekiel, "the twelve prophets," Zerubbabel, Joshua son of Jozadak, Nehemiah, Enoch, Joseph, Shem, Seth and Adam.³  He does not name Daniel.

If Daniel was a famous sixth-century figure with a book, an encomium of Israel's worthies written fifteen years before Antiochus should mention him.

The counter is that this is an argument from silence, and the same list omits Ezra — who was certainly real, certainly earlier, and has a book.  Neither Ezra nor Job is named anywhere in the Greek Sirach, and Mordecai is missing too.  The argument has some force, because Ben Sira reads like a man working through a collection.  But a praise-list that drops Ezra cannot be used to prove that what it drops did not exist.

The theology

Driver's third argument was that Daniel's theology "points to a later age than that of the exile":  bodily resurrection (12:2), an elaborate hierarchy of angels with individual names and national assignments — Gabriel (8:16, 9:21), Michael "your prince" (10:21, 12:1) — and a developed apocalyptic eschatology.  These features are common in second-temple literature and rare or absent in the pre-exilic prophets.

Driver himself flagged the weakness, noting that the argument "has sometimes been stated in an exaggerated form."¹⁷  He was right to.  Resurrection language appears in Isaiah 26:19 and Ezekiel 37 on any dating;  angels appear throughout the Torah;  and "this idea developed late, therefore this book is late" is circular whenever the book in question is one of the pieces of evidence for when the idea developed.

Treat the theological argument as corroborating at best.  It never decides anything by itself, and both sides know it.

The genre

Every surviving Jewish apocalypse that names an author names an ancient one.  1 Enoch is attributed to Enoch, Jubilees to Moses, the Testaments to the twelve patriarchs, 4 Ezra to Ezra, 2 Baruch to Baruch — all written centuries after the figures they name.  The rest, like the Qumran War Scroll, are anonymous.  Several contain reviews of history cast as prophecy.  If Daniel 7–12 belongs to that genre, then a second-century date and an attribution to a sixth-century sage are exactly what the genre produces, and no deception need be involved.  The critical statement of the point:  "'authorship' in the ancient Jewish world, unlike in the modern, was used more as a genre marker than as an author attribution."¹²

Mesopotamia had the same convention earlier.  Five Akkadian texts — the Marduk Prophecy, the Shulgi Prophecy, the Uruk Prophecy, the Dynastic Prophecy, and "Text A" — narrate real past events in the form of predictions about unnamed future kings.  Joyce Baldwin, arguing the conservative side in the 1978 Tyndale Old Testament Lecture, set out how critics read Daniel 11 in the light of the Shulgi Prophecy:

More significant than verbal similarities is the correspondence between the form of this work and the alleged vaticinium ex eventu in Daniel 11, according to which the general Babylonian setting gives the appearance of a work written in the sixth century B.C.  The author's true standpoint is to be discerned at the moment of transition, when history can no longer be demonstrated to correspond to the events prophesied. — Joyce Baldwin, "Some Literary Affinities of the Book of Daniel"²²

"According to which" marks this as the critics' reading, not Baldwin's own.  She rejects it for Daniel:  "I find no evidence in Daniel to suggest that any vaticinium ex eventu was intended."  What she grants is the likeness of form.  She agrees with the Assyriologist A. K. Grayson "that in form and style there is a striking resemblance," while holding that Daniel's purpose is different.²²  She also notes that the Akkadian texts originated in Babylon centuries before the exile.²²  That is a fair point about where the form was available, and no answer at all about the form itself.

The honest statement is that the genre argument establishes possibility, not fact.  "Apocalypses are usually pseudonymous" does not entail "this apocalypse is."  But it does dissolve the common apologetic move that a pseudonymous Daniel would have to be a forgery designed to deceive.  On the evidence of the genre, it would not.

Part 2 — The case for the sixth century

Belshazzar

For most of Christian history Daniel 5 was the book's worst embarrassment.  No king named Belshazzar appears in Herodotus, in Berossus, in Ctesias, or anywhere else;  every ancient list makes Nabonidus the last king of Babylon.  Commentators identified Daniel's Belshazzar with one or another known king and moved on.⁵

Cuneiform recovered him, in three stages.

A foundation inscription, c. 540 BC.  Nabonidus buried clay cylinders in the ziggurat at Ur recording his repairs to the temple of Sin.  The text ends in a prayer, and the prayer has two names in it:

As for me, Nabonidus, king of Babylon, save me from sinning against your great godhead and grant me as a present a life long of days, and as for Belshazzar, the eldest son — my offspring — instill reverence for your great godhead in his heart and may he not commit any cultic mistake, may he be sated with a life of plenitude. — Nabonidus Cylinder from Ur, iii.3–31⁶

This is a contemporary royal document, written a year or two before Babylon fell, praying for the king and for one named son alongside him.  That is what a king does for a co-regent, and it is not what he does for an ordinary prince.

The Nabonidus Chronicle, published by Pinches, then showed the crown prince holding Babylonia with the army while the king was away at Tema.⁵

The Verse Account, published by Sidney Smith in 1924, said what that arrangement was:

when the third year was about to begin — he entrusted the army to his oldest son, his first born, the troops in the country he ordered under his command.  He let everything go, entrusted the kingship to him, and, himself, he started out for a long journey. — Verse Account of Nabonidus, ii.5–6⁸

Popular treatments routinely merge the first and third of these into one document.  They are separate texts, found seventy years apart, and the case is stronger for their being independent.

Taken together this is a genuinely impressive result for the book, and it does three things at once.

It vindicates the name.  It vindicates the setting — a king in Babylon while the actual monarch was elsewhere.  And it explains a detail that had made no sense for two thousand years:  Belshazzar offers to make the interpreter of the writing "third ruler in the kingdom" (5:7, 5:16, 5:29).¹  Why third?  Because first was Nabonidus and second was Belshazzar himself.  Third was the highest thing he had to give.

The force of this is easy to understate.  In 165 BC, in Judea, the information required to write Daniel 5 did not exist in any source we know of.  Greek historiography did not have it.  Jewish tradition did not have it.  It was in Akkadian, in Babylon, in documents nobody read again until the nineteenth century.

Two honest qualifications.

The "third ruler" reading is not unanimous.  Montgomery argued the Aramaic means "one of three" or a šalšu-officer, and Shea's defense of "third ruler" is an argument, not a demonstration.⁵

And no cuneiform text calls Belshazzar king.  Contracts continued to be dated by Nabonidus's regnal years;  Belshazzar never performed the New Year ceremony;  oaths were sworn in his name jointly with his father's, which is not the same thing.  Shea, defending the chapter, concedes it:  "Entrusting the kingship to Belshazzar, as mentioned in the Verse Account, is not the same as making him king."⁵  He then proposes that Belshazzar may have assumed the title late, after Nabonidus's capture — a reasonable conjecture with no text behind it.

So Daniel 5 knows things no second-century Judean could know, and calls Belshazzar something the records do not.  That is the actual shape of the evidence, and both halves are real.

Other things the court tales get right

The smaller details cluster the same way.  Nebuchadnezzar boasting "Is this not the great Babylon that I have built for a royal residence by my own mighty strength?" matches a king whose building inscriptions are among the most extensive in Mesopotamian history.¹  Babylonian execution by fire and Persian execution by animals fit their respective settings.  "The law of the Medes and Persians which cannot be revoked" fits Persian legal ideology.  Daniel 2:37's "king of kings" is often called a Persian title and is not:  Wiseman traces it in Assyrian and Babylonian royal inscriptions back to Tiglath-pileser I.⁹

One more belongs on the list, and it is better than the rest because it is a negative detail.  Daniel 5 has Babylon change hands in a single night, during a banquet, with no siege and no battle described.  The Nabonidus Chronicle says the same thing in the flattest possible terms:  Sippar "was captured without a battle," and on the sixteenth of Tashritu Ugbaru and the army of Cyrus "without battle they entered Babylon."⁷  A writer inventing the fall of a great empire would give it a siege.  This one gives it a dinner party, and the Babylonian record agrees.

Individually these prove nothing.  Together they describe a writer who knows the sixth-century Babylonian court from inside, which is at least a fact in need of explanation.

The "Chaldeans" objection cuts less than it is made to.  Daniel uses kasdim both for the ethnic Babylonians and for a class of diviners, and critics call the second use an anachronism.  Two Greek historians use it the same way, and both are describing periods long before the second century.  Herodotus, in the fifth century, speaks of "the Chaldeans who are the priests of this god" in his description of the ziggurat at Babylon.²³  Diodorus Siculus is more explicit still:  the Chaldeans "have about the same position among the divisions of the state as that occupied by the priests of Egypt;  for being assigned to the service of the gods they spend their entire life in study," and the training "is passed down in the family, and son takes it over from father, being relieved of all other services in the state."²⁴  That is exactly Daniel's usage.  This objection should be dropped from the late-date brief.

Eight copies at Qumran

Fragments of eight Daniel scrolls were recovered from Caves 1, 4 and 6, covering substantial parts of chapters 1–11:  1Q71, 1Q72, 4Q112–4Q116, and 6Q7.¹¹  That is "more copies than most books of the Hebrew Bible," in the words of three scholars who hold the late date.¹²

More important than the count is the status.  4Q174, the Florilegium, introduces a quotation from Daniel 12:10 with the formula normally reserved at Qumran for scripture — "which is written in the book of Daniel the prophet."²⁵  Daniel is being read as a prophetic book, by a community whose own foundation is usually dated to the mid-second century.

The argument built on this is the "canonization speed" argument:  if Daniel was written in 164 BC, then within a few decades a sectarian community was copying it eight times and citing it as prophetic scripture.  That is fast.

It is not impossible.  The reply from the critical side is direct:  "If Daniel was written in the 160s BCE there were obviously manuscripts of the book in existence at that time and in the following years.  There is no evidence one way or another that this copy of Daniel was viewed as 'canonical.'"¹²  That answer is fair as far as it goes;  it does not address 4Q174's citation formula, which is about a different manuscript.

The wider diffusion points the same way without settling anything.  1 Maccabees, written around 100 BC, has the dying Mattathias hold up Daniel's friends and Daniel himself as examples — "Hananias, Azarias and Misael, because of their faith, were saved from fire.  Daniel, by his simplicity, was rescued from the mouth of lions."³  The Old Greek translation was made in the later second century BC, and a second Greek translation followed in the next century.¹²  Additions accumulated — Susanna, Bel and the Dragon, the Prayer of Azariah and the Song of the Three — which shows the book being expanded, not that it was old.  All of this is consistent with a book of 164 BC that spread fast, and equally consistent with a book of 530 BC.  It cannot bear the weight conservatives put on it.

The radiocarbon date, done properly

This is the newest evidence and the most mishandled, so it is worth working through carefully.

In June 2025 a team led by Mladen Popović published radiocarbon dates for 24 Dead Sea Scroll samples, using a cleaning protocol designed to remove the castor oil that contaminated earlier attempts, plus a machine-learning model trained on the resulting dates.¹⁰  One of the samples was 4Q114 — 4QDaniel^c — which preserves Daniel 10:5–9, 10:11–16, 10:21, 11:1–2, 11:13–17 and 11:25–29.²⁶  That last block is Antiochus IV's first Egyptian campaign.  This is a manuscript of the disputed material, not of the safe part of the book.

The measurement:  2168 ± 15 BP, calibrated in OxCal 4.4.4 against IntCal20.^{10  10}

The paper's own text says only that "the accepted 2σ calibrated range for 4Q114, 230–160 BCE, overlaps with the period in which the final part of the biblical book of Daniel was presumably authored."¹⁰  The older band is set aside "based on comparative typological evidence from date-bearing Aramaic manuscripts from that period" — that is, on palaeography.¹⁰

Then the two popular readings arrived, and both are wrong.

Associates for Biblical Research reported "230–160 (mean 195) BC," concluding it is "highly improbable that Daniel was written after the death of Antiochus IV."¹¹  A bimodal calibrated posterior does not have a usable mean;  averaging the endpoints of one hump is not a statistic.

Three critical scholars replied in February 2026 that the range "represents a span within which there is an equal probability throughout, rather than evidence favoring any specific year or decade within that range."¹²  The published intervals say otherwise.  A uniform distribution over 230–160 BC would not produce a 1σ highest-density interval of 205–175 BC holding 34.5% of the total probability.  Thirty years out of seventy carrying 34.5 of 45.9 percentage points is a density of about 1.15% per year against a band average of 0.66%.

Stated properly:  if the palaeographic rejection of the fourth-century band is accepted, then about three-quarters of the remaining probability — 34.5 of 45.9 — falls between 205 and 175 BC.  That is before Antiochus IV became king.  It is a copy, not an autograph;  the paper notes its scribal errors make it unlikely the scribe was the author.¹⁰  So on the study's own numbers, the most likely reading is that Daniel 11 was already being copied, imperfectly, a generation before the crisis it is said to describe.

Now the caveats, and they are not small.

This is one sample.  Parchment dates the animal, not the writing, and skins could be stored.  The study found systematically older dates than palaeography across many samples, which is either a real correction of a contaminated baseline or an uncorrected offset — and until it is replicated by another laboratory nobody can say which.  And the rejection of the 355–285 BC band is a palaeographic judgment made inside a paper whose central claim is that palaeographic judgments have been unreliable.  If that rejection is wrong, the date is fourth-century and the Maccabean thesis for Daniel 10–11 collapses;  if the whole measurement is offset, the date could be exactly where Cross put it.

What can be said without overreach:  4Q114 is now the single most important physical datum in the debate, it sits awkwardly for the standard critical model, and it is not yet strong enough to settle anything by itself.  A replication on a second Daniel fragment would be worth more than another decade of argument about musical instruments.

Three Greek words and no Greek government

The loanword profile is genuinely odd for a Hellenistic composition.

Daniel's Aramaic has around twenty Persian loanwords and three Greek ones.  All three Greek words are musical instruments in one list in one chapter.  There is no Greek word in Daniel for a governor, a garrison, a tax, a coin, a magistrate, a soldier, or an assembly — the vocabulary a Judean writing in 165 BC, after a century and a half of Seleucid and Ptolemaic administration, would use without thinking.

Kitchen puts the point sharply:  "Many Old Persian words alongside hardly any Greek words in our text suggest a date in the Persian age;  a document of Hellenistic date with a penchant for loan-words should have taken them from Greek (or Middle Persian)."¹³

The Persian words compound it.  They are Old Persian, not Middle Persian, which indicates no fresh borrowing after about 300 BC.  And four of them are so badly handled by the Septuagint translators that their meanings had already been lost — which argues, Kitchen says, "for a date before the second century BC."¹³

The obvious counter is that a skilled author writing historical fiction set in Babylon would avoid anachronisms.  It is a real counter, and it has a cost:  an author careful enough to suppress every Greek administrative term did not notice three Greek instrument names, and did not notice that Darius the Mede never existed.  Selective archaizing of that quality is hard to credit.

The limit on this argument must be stated, because it is the limit almost nobody states.  The loanwords are in the Aramaic.  Daniel 11 is in Hebrew.  The loanword evidence is excellent evidence about chapters 2–7 and says nothing whatever about chapters 8–12.

Ezekiel's Daniel

Ezekiel, writing in the same exile, names a proverbially righteous "Daniel" beside Noah and Job (14:14, 14:20) and a proverbially wise one (28:3).¹  Conservatives read this as the book's Daniel.  Critics read it as the Ugaritic Dan'el of the Aqhat legend, a righteous judge from a text a thousand years older, which fits the company of Noah and Job — two non-Israelites from the distant past — better than a contemporary civil servant does.

The critical reading is the stronger one here, and conservatives should stop leaning on Ezekiel 14.  Ezekiel 28:3 addressing the king of Tyre with "you are wiser than Daniel" makes more sense as a reference to a figure the Tyrian court would recognize than to a young Judean official in Babylon.

Weak arguments that should be retired

Three conservative staples do not survive inspection, and an article that keeps them is not worth reading.

Susa "in the province of Elam" (8:2).  Archer argues that only an early author would place Susa in Elam, since the Persians reorganized the region as Susiana.  But "Elam" remained standard Hebrew usage for the region long afterwards, and Ezra kills the argument outright:  its list of deported peoples reads "the people of Susa (that is, the Elamites)" — a Persian-period text equating exactly the two terms Archer says only a sixth-century writer could have equated.¹  The argument proves nothing.

Josephus's story that Alexander was shown Daniel's prophecy (Antiquities 11.337).  A late, legendary, and self-serving anecdote.  It is evidence that Daniel existed by Josephus's day, which nobody disputes.

"The book was already canonical by Qumran, so it must be old."  The premise smuggles in what it needs to prove.  Being copied and cited is not the same as being canonized, and the critical reply on this point is correct.

Part 3 — Weighing it

What each side has to explain away

The late date has to explain:  Belshazzar, and the third place in the kingdom, and the fact that this knowledge was unavailable in second-century Judea.  Around twenty Old Persian loanwords with no Greek administrative vocabulary.  An Imperial Aramaic word order unlike anything at Qumran.  Four Persian words already unintelligible to the Septuagint translators.  And 4Q114's radiocarbon date, whose likeliest value predates Antiochus IV's accession.

The early date has to explain:  a vision whose accuracy runs out in 164 BC and then fails.  Four ascending numbers for one crisis, two of them in consecutive verses.  A Median king who has no room in the Babylonian record and no inscription anywhere.  Hebrew that reads like the Chronicler's.  Ben Sira's silence.  And a genre in which every other surviving example is pseudonymous.

Two items that usually appear on the second list have been struck from it here.  The Writings-not-Prophets argument fails on chronology:  every list older than Baba Batra — Melito's, Origen's, and Josephus's usage — puts Daniel with the prophets.  The "Chaldeans" anachronism fails on Herodotus and Diodorus.  Neither is evidence for anything.

Neither list is answerable by the other side's best argument, because the two lists are about different halves of the book.

The strongest argument on each side

For the late date:  Daniel 11:40–45.  Not the accuracy of 11:2–39 — accuracy is what a prophecy is supposed to have — but the failure at the end, located exactly where the theory says the author's present must be.  No other explanation accounts for both the precision and the specific placement of the error.

For the early date:  Belshazzar, with 4Q114 second and rising.  Belshazzar is not a lucky guess.  It is a cluster of correct, obscure, mutually reinforcing facts about a court that had been forgotten for two millennia.

Where the uncertainty actually is

Strip out the polemics and the residue is this.

Chapters 1–6 look old.  The Babylonian knowledge is real, the Aramaic is Persian-period, and the loanword profile is wrong for a Hellenistic author.  Even scholars who date the visions to 165 BC generally allow that the tales are older.  On the evidence, the court tales are probably pre-Hellenistic, and the conservative case for them is better than the academy usually admits.

Chapters 8–12 are the contested ground, and there the evidence is bad for the early date in every respect except one — and that one, the radiocarbon date of 4Q114, is a single measurement on a single scrap that has not been replicated.

That is an uncomfortable place to stop.  It is where the evidence stops.

What would settle it

• A second radiocarbon date on another Daniel fragment, ideally from chapters 8–12, run by an independent laboratory. If 4Q112 or 4Q113 returns a pre-175 BC date, the Maccabean thesis is in serious trouble. If they cluster after 160, 4Q114 is an outlier and the conservative case loses its best physical evidence.
• Any cuneiform or Persian-period reference to Darius the Mede. One inscription would end the oldest objection to the book.
• A pre-Maccabean citation of Daniel. Ben Sira's silence would be broken by any second- or third-century text quoting or naming the book.
• Independent attestation of Nebuchadnezzar's illness in a Babylonian source, which would decouple Daniel 4 from the Nabonidus tradition.

Part 4 — What actually survives at 165 BC

Suppose the critical date is right.  What breaks?

Daniel 11 as prophecy — completely.  There is no version of the argument that survives.  Anyone using Daniel 11 as evidence for the supernatural is betting the whole argument on the sixth-century date.

The four kingdoms ending in Rome — probably.  Daniel 11:30's "ships of Kittim" is Rome, and the Roman embassy that turned Antiochus back from Egypt happened in 168 BC.  A second-century author knew about Rome.  The identification of the fourth kingdom as Rome, which Jerome treated as "the traditional interpretation of all the commentators of the Christian Church,"⁴ then becomes a later re-reading — which is exactly what the Jewish apocalypse 4 Ezra says it is doing, around AD 100:  "The eagle that you saw coming up from the sea is the fourth kingdom that appeared in a vision to your brother Daniel.  But it was not explained to him as I now explain to you."²⁷

The seventy weeks — no.  Whatever Daniel 9:24–27 means, a document from 165 BC that speaks of an anointed one being cut off, and of a coming people destroying the city and the sanctuary, still predates the crucifixion by about two centuries and the destruction of the temple by more than two.  The most-used apologetic argument from Daniel does not depend on the date at all.  This deserves emphasis, because both sides argue as though it does.

The son of man — no.  Daniel 7:13–14's figure receiving everlasting dominion is pre-Christian on any dating, and it is the title Jesus used of himself more than any other.

Resurrection — no.  Daniel 12:2 is pre-Christian on any dating.

The book's own voice — yes, and this is the real issue.  Daniel 7:2, 8:1, 9:2, 10:2 and 12:5 speak in the first person.  If the visions were written in 165 BC, those sentences were written by someone who was not Daniel, about visions he did not have, dated to reigns he did not live under.  Whether that is compatible with inspiration is a theological question with a long literature and no empirical test.  It should be argued on those terms, and not smuggled into a historical debate as though it were evidence.

And Jesus' reference in Matthew 24:15 — "the abomination of desolation, spoken about by Daniel the prophet" — is a real datum for anyone who holds a high view of Christ's teaching.¹  It is not a knockdown historical argument, since naming a book by its traditional author was ordinary usage.  But it is not nothing, and pretending otherwise is dishonest in the other direction.

Objections and responses

"The consensus is settled;  only fundamentalists question it."  The consensus is real and it is not stupid, but it has an inconvenient feature:  the best conservative technical work — Kitchen on the Aramaic, Wiseman on the Babylonian records — has never been refuted so much as bypassed.  And the consensus itself has moved:  the composite view now concedes that chapters 1–6 are older than 165 BC, which is a substantial retreat from Driver.

"Only inerrantists question it — no serious scholar holds an early date."  Both halves of this are false.  Evangelical scholars are on both sides:  Lucas, Goldingay, Buster and Walton hold the mainstream view;  Steinmann, Sprinkle and Tanner hold the traditional one.¹²  The dividing line runs through the evangelical academy, not around it.

"Daniel 11 is too detailed to be prophecy — no real prophecy works like that."  This is an argument from what prophecy ought to look like, which is not evidence.  The evidential argument is the different one made above:  the accuracy is lopsided in favor of the second century and it fails at 11:40.

"The failure at 11:40–45 refers to a future Antichrist."  Possible, and Jerome's church held it.  But it requires an unmarked change of subject in mid-paragraph, and Daniel already uses "time of the end" for a vision about Antiochus (8:17).  A reading that has to be supplied to save a text is weaker than a reading the text supplies.

"The Dead Sea Scrolls prove an early date."  They do not prove it.  Eight copies and a scriptural citation formula make a 164 BC composition tight, not impossible.  The radiocarbon date on 4Q114 is stronger than the manuscript count, and it is one measurement.

"A pseudonymous book cannot be Scripture."  That is a theological premise, not a historical finding, and it is not obviously true — Hebrews is anonymous, the Psalm headings are debated, and ancient attribution conventions differed from modern ones.  Whatever one concludes, it should not be allowed to decide a question about Babylonian regnal records.

"Belshazzar proves the whole book."  It does not.  It proves that chapter 5 rests on real sixth-century Babylonian information.  It says nothing about who wrote the Hebrew visions three chapters later, in a different language, about a different century.

"The popular defenses of Daniel already answer all this."  They do not, and the gap is worth naming, because it is the reason this article exists.  Take the best-known free treatment:  Mike Winger's "Defending Daniel," a fifty-one-minute lecture from 2016 that has served as many people's introduction to the question.  It is careful in places — it is right that the Persian loanwords are Old Persian, and it supplies the canon-list point used above, which most academic treatments omit.²⁸  But it lists seven critic arguments and Daniel 11 is not one of them.  The failure at 11:40–45 is never mentioned.  Neither is Ben Sira, nor the 1,290-and-1,335-day revisions.²⁸  Its first and longest "critic argument" is that "prophecy is impossible," presented as Porphyry's actual reason.²⁸  Jerome, our only source for Porphyry, reports something quite different — that the book "contained authentic history" up to Antiochus and then went "false."⁴  Answering "prophecy is impossible" is easy.  It is also answering the wrong argument.

Smaller slips in the same talk illustrate how the popular version drifts:  it counts four Greek loanwords where there are three, and it presents the Nabonidus Cylinder and the Verse Account as quotations from "the same cylinders," when they are separate documents published seventy years apart.²⁸  None of that is dishonest.  It is what happens when a case is transmitted rather than checked.

Verdict

The date of Daniel cannot presently be settled, and the reason is that the book is not one thing.

The court tales of chapters 1–6 carry information about the Babylonian court — Belshazzar's existence, his position, and why third place was the best on offer — that was not available to anyone in second-century Judea.  Their Aramaic is Imperial, their Persian loanwords are Old Persian and partly already unintelligible by the time the Greek translation was made, and they contain no Greek administrative vocabulary at all.  On the evidence, these chapters are pre-Hellenistic, and probably considerably older than the visions.

The visions of chapters 8–12 are a different matter.  They are in late Hebrew.  Their historical detail is dense exactly where a second-century author's knowledge would be dense and thin exactly where it would be thin.  They give four different, ascending durations for one crisis.  And they end with a specific prediction about the death of Antiochus IV that did not happen — he died at Tabae in Persia, not between the seas and the holy mountain.

Against all of that stands one hard measurement:  4Q114, a copy of Daniel 10–11 including the Antiochus material, whose radiocarbon date most probably falls between 205 and 175 BC.  If that date holds up under replication, the Maccabean thesis for these chapters is finished.  If it does not, the case for the late date is very strong.

So the position taken here is:  chapters 1–6, early;  chapters 8–12, unresolved, with the balance of the literary evidence favoring the second century and the balance of the physical evidence favoring the third or early second.  That is not a satisfying answer.  It is the answer the evidence currently supports, and the alternative — picking the half of the evidence that suits one's theology and calling the question closed — is what both sides have been doing for a hundred and fifty years.

One last observation, for readers on either side.  The argument that most Christians actually need from Daniel — the seventy weeks, the son of man, the resurrection — survives the late date intact.  The argument that dies is the one from Daniel 11, which is the one nobody in the pews has read.  It would be a strange trade to stake the credibility of the book on the chapter fewest people can quote.

Sources

Bible quotations are from the NET Bible unless otherwise noted.¹