HistoricalChristian.Faith

Does Josephus's twenty-two-book count (Against Apion 1.8) prove a closed, deuterocanon-excluding Hebrew canon?

Type: supporting argument (off-timeline) · Assessment: contested — for the strong inference the proponents draw, that the historian's tidy count fixes a settled, pan-Jewish, deuterocanon-excluding canon, centuries old and binding on the church. The bare datum is sound: writing c. AD 93, Josephus is the earliest datable witness to a Jewish reckoning that numbered the divine books at twenty-two — a Hebrew-sized table that does not count the deuterocanon among the inspired. But the proof is one witness, title-less, to one influential stream, and the conclusion the apologists hang on it overshoots the evidence on three counts (the count maps imperfectly onto the Protestant thirty-nine; Josephus himself drew on 1 Esdras and 1 Maccabees as histories; and "the Jews" of the apostolic age were a plurality, not yet agreed). Strong as testimony, short as proof — genuinely disputed

Related arguments: the closed Jewish canon — see also did the Jews reject the deuterocanon — was the first-century Jewish canon closed and Hebrew-only? (weak); was there a formal Pharisaic canon that Jesus and Paul accepted? (contested); does Philo's silence show a closed, Hebrew-only Alexandrian canon? (weak); does the threefold division prove a closed Jewish canon by Jesus' day? (weak); did the Council of Jamnia close or affirm the Hebrew canon? (weak); does the Bava Batra 14b baraita attest a pre-Christian closed Hebrew canon? (weak); were the Jews "entrusted with the oracles of God" (Romans 3:2)? (contested).

See also: the umbrella this prop serves — Did the Jews reject the deuterocanon — was the first-century Jewish canon closed and Hebrew-only? (weak) — and its major premise, Were the Jews "entrusted with the oracles of God" (Romans 3:2)? (contested); its weaker twin, the silence of the other Greek-writing Jew the proponents pair with Josephus — Does Philo's silence — never quoting the deuterocanon — show the Alexandrian Jews held a closed, Hebrew-only canon? (weak); the supposed canon-closing "Council of Jamnia" (a discarded hypothesis), Josephus's near neighbour in date; the cessation-of-prophecy reasoning on which Josephus rests his Artaxerxes cutoff, weighed on the witness page; and the Rabbinic Corruption of Scripture timeline, where the question of when — and against whom — the rabbinic table hardened is gathered.

The claim

Around the close of the first century, defending the antiquity of his nation against the grammarian Apion, the Jewish historian Josephus tells a Roman readership that the Jews hold not "an innumerable multitude of books … disagreeing from and contradicting one another," but

"…only twenty-two books, which contain the records of all the past times; which are justly believed to be divine … the time from the death of Moses till the reign of Artaxerxes … the prophets … wrote down what was done in their times in thirteen books. The remaining four books contain hymns to God, and precepts for the conduct of human life. It is true, our history hath been written since Artaxerxes very particularly, but hath not been esteemed of the like authority with the former by our forefathers, because there hath not been an exact succession of prophets since that time."

— Josephus, Against Apion I.8 (§§37–41).¹ Because the bound at Artaxerxes (mid-fifth century BC) leaves the Greek-only books (Tobit, Judith, Wisdom, Sirach, Baruch, 1–2 Maccabees) on the wrong side of the line, Protestant canon arguments lean on the passage as the single clearest Jewish witness for a settled, deuterocanon-excluding Hebrew canon already in force in the first century. The sub-claim weighed here is precisely that inference: that the count of twenty-two proves a closed, Hebrew-only canon from which these books had been deliberately excluded.

Who advances it

In its careful scholarly form it is F. F. Bruce's — "Even Jews who wrote in Greek at the beginning of our era, like Philo and Josephus, recognized only the Canon of the Hebrew Bible, although they used the 'Septuagint'" — and John Piper's, that Josephus and Philo "used the LXX but did not accord the Apocrypha authority." Popularly: James White — "Josephus … in Against Apion 1:7 gives the number of books as 22. He specifically rejects those books written after Malachi, that is, the Apocryphal books … There is no reason to believe that Josephus' canon is recent … a canon that had been in place for 300 years"; Mike Winger — Josephus "refers to the books that are laid up in the temple as being 22 in number, which means the apocrypha was not part of them"; Allen Parr — the Jews "themselves rejected the Apocrypha," tying it to the silent centuries from Malachi to Christ; CARM (Matt Slick) — "Josephus explicitly rejected the Apocrypha and listed the Hebrew Canon to be 22 books"; Steve Christie — "Josephus affirmed the same thirty-nine Old Testament book canon of later Protestants, just in a condensed twenty-two-book format"; and Bible.org — "Josephus, the Jewish historian (A.D. 95), indicated that the 39 books were recognized as authoritative."² The Josephus prop is the minor premise's strongest plank; the umbrella it shores up — that "the Jews" held a closed, Hebrew-only canon the church should follow — is weighed on the deuterocanon-excluded page, the Romans 3:2 major premise on the oracles-of-God page. The adversarial side — Gary Michuta and Trent Horn for the Catholic case, and the mainstream canon historians (Sundberg, Lim, McDonald, Barton) — answers that the passage attests one stream's reckoning, not a pan-Jewish settlement.³

The evidence for

Josephus is a real, early, datable witness — and the proof's true strength. The argument is not built on nothing. Unlike the much later rabbinic enumerations, Josephus writes within living memory of the Second Temple, c. AD 93, and states plainly that the reverenced books are a bounded, inviolable set — "no one has been so bold as either to add any thing to them, to take any thing from them" — which the Jews would "willingly … die for." His twenty-two is a Hebrew-sized count (the same total Origen and Jerome report, folding Ruth into Judges and Lamentations into Jeremiah to match the letters of the Hebrew alphabet), and the post-Artaxerxes writings — the very window in which the deuterocanonical histories and wisdom books fall — he expressly relegates to a lower order. So a genuine, ancient, Hebrew-measured Jewish reckoning of Scripture existed by the late first century, it did not number the deuterocanon among the divinely-credited books, and it is the lineal ancestor of the rabbinic table the Reformers later invoked. That much is sound, and it is the real datum the argument trades on — the same Hebrew-measure reading this timeline treats as a serious, live position from Jerome onward.⁴

Adversarial assessment

What holds. That Josephus attests a Hebrew-measured twenty-two-book reckoning, c. AD 93, that did not count the deuterocanon among the inspired books, is sound and undisputed — he is the earliest and clearest Jewish witness on the point, far firmer than the flat popular slogan "the Jews rejected the Apocrypha." The deuterocanon really is absent from the Hebrew Bible that the Pharisaic-rabbinic tradition transmitted, and Josephus stands at the head of that stream.

Where the strong inference gives way. The proof has to carry more than "a Hebrew-measured reckoning existed by AD 93." It is enlisted to show a closed, pan-Jewish, deuterocanon-excluding canon already settled centuries earlier and binding on the church. On that it falls short, on three counts.

1. One witness to one stream — not "the Jews." Josephus was a priest of Pharisaic sympathies writing apologetic for a Roman audience; his "twenty-two" is the reckoning of the priestly–proto-rabbinic mainstream he belonged to, projected as the settled mind of his whole nation. But the leading canon scholarship denies that any single Jewish canon was yet fixed in the first century: the Law and the Prophets were closed, the Writings still fluid, and "the Jews" were a plurality — Pharisees, Sadducees, Essenes, the Qumran community, Samaritans, and the Greek-reading Diaspora — with overlapping but not identical Scriptures. The Hebrew- and Aramaic-reading sectarians of Qumran kept Tobit, Ben Sira, and the Epistle of Jeremiah; the Diaspora read the longer Greek Bible. To read off Josephus a verdict the whole of Judaism had jointly rendered is to mistake one influential reckoning for a national act of canon-closing that the period's historians say had not happened. This is the umbrella weakness argued in full on the deuterocanon-excluded page; here it is enough that Josephus, however strong as a single witness, cannot by himself supply the uniformity the conclusion needs.⁵

2. The count is title-less, and Josephus himself used the disputed histories. Josephus names no books; exactly which writings fill his thirteen "prophets" and four "hymns and precepts" is inferred, not stated — John Barton's caution stands, "it is impossible to know exactly which books he included, or how he counted them." The bare "twenty-two" does not by itself pin the Protestant list: it can be made to include or exclude different books, and it differs from the rabbinic "twenty-four" of Bava Batra 14b and 4 Ezra. Worse for the tidy equation, Josephus did not treat the post-Artaxerxes Greek books as worthless: he followed 1 Esdras for the post-exilic narrative (the unique tale of the three bodyguards, Antiquities XI.iii) and drew on 1 Maccabees as his main source for the Hasmonean history (Antiquities XII–XIII). He bounds the inspired, prophetic books at Artaxerxes, but he reads and trusts the very histories the argument treats him as having rejected — so "Josephus affirmed the same thirty-nine, condensed" claims more than his own practice will bear.⁶

3. The "since Artaxerxes" cutoff rests on the cessation-of-prophecy premise — and proves too much when pressed. Josephus's reason for the lower rank of the later books is that "there hath not been an exact succession of prophets since that time." Note the precise claim: it is the exact succession of prophets that he says lapsed, not prophecy as such — which Rebecca Gray, in the standard study of prophecy in Josephus, reads as his marker of the continuous historical record from Genesis to Esther, a criterion of the very antiquity his apologetic is defending, not a dogma that inspiration stopped — and Josephus elsewhere claims a prophetic gift for himself and reports later prophetic figures (indeed, retelling the Maccabean rededication he drops 1 Maccabees' own "until there should come a prophet," 1 Macc. 4:46), so the passage will not bear the flat "all prophecy ceased after Malachi" that proponents like Allen Parr load onto it (weighed on the witness page). And the same late-rabbinic move that excludes Sirach and "everything written after" it from Scripture would, in the mouths of the rabbis who made it, exclude the Christian Gospels too — an argument that, pressed to prove a binding canon for the church, proves far too much. White's gloss that Josephus "specifically rejects those books written after Malachi, that is, the Apocryphal books," and that this was "a canon … in place for 300 years," reads a deliberate, itemized rejection of named books into a passage that names none and dates none.⁷

Proponent overreach to flag. Bible.org pairs Josephus with the claim that the "Council of Jamnia (A.D. 90) officially recognized our 39 OT books" — but no such council closed any canon (a discarded hypothesis), so the prop is propped on a myth. White's "300 years" and "specifically rejects those books written after Malachi" and Christie's "the same thirty-nine … condensed" each convert a title-less count into an itemized, dated, deliberate exclusion the text does not supply. What every one of these framings is entitled to is the narrower, true point: Josephus attests an early, Hebrew-measured Jewish reckoning that did not number the deuterocanon among the inspired books — a real contribution to those books' graded standing, not a proof of a settled, church-binding canon.⁸

A reliability caveat — and its limit. The Catholic side adds a credibility point: Josephus is a known embellisher with a pro-Roman agenda, so his sweeping apologetic flourishes here — no disagreement among the books, nothing ever added or removed, every Jew ready to die for them — are better read as patriotic rhetoric than as a sober census. The exhibit is his own most famous act of scriptural reading: the "ambiguous oracle … in their sacred [writings]" that "one from their country should become governor of the habitable earth" he refers not to a Jewish Messiah but to Vespasian, "appointed emperor in Judea" (Jewish War VI.5.4) — scripture bent to a Roman point. The honest form of this is "exaggerating, not lying," and it rightly discounts the rhetorical absolutes. But it has a limit the critics should grant in turn: a writer can be tendentious in interpreting an oracle and still reliable on a plain bibliographic fact — how many books his tradition reverenced — so the Vespasian episode tempers Josephus's flourishes without falsifying his twenty-two-count. The bare datum stands; what the embellishment denies it is the affidavit-grade certainty the strong inference wants.⁹

A second overreach in the same clause — and its limit. Josephus's inviolability boast runs further than "nothing added or removed": "during so many ages as have already passed, no one has been so bold as either to add any thing to them, to take any thing from them, or to make any change in them." Taken as textual history that flat "no change through so many ages" does not hold, and Gary Michuta presses the gap with Frank Moore Cross and the text-critical consensus the Dead Sea Scrolls forced: down to the first century Judaism transmitted no single stabilized Hebrew text but several coexisting text-types — proto-Masoretic, proto-Samaritan, and the Hebrew Vorlage behind the Septuagint — and a standardized consonantal text was not fixed until between the two Jewish revolts, c. AD 70–135. The Qumran copies of the very books Josephus reverenced vary across those families, so his "no change" states an ideal of reverence, not the manuscript record. Yet the defeater has the same limit as the Vespasian one: it discounts a rhetorical absolute, not the bibliographic count. White's reply is fair as far as it goes — Josephus is reporting what the Jews believed about their sacred books (that they were untouchable), not writing a history of textual transmission — and the sub-claim weighed here is which books were counted, not the wording within them, on which Cross's finding does not bear. So the stabilization evidence rightly deflates one more of Josephus's sweeping absolutes without loosening the twenty-two-count — the same lesson as the Vespasian oracle, reached from the side of the text rather than its interpretation. (When and against whom the rabbinic consonantal text finally hardened is gathered on the Rabbinic Corruption of Scripture timeline.)¹⁰

“Laid up in the temple” — which books? — and Beckwith's fair rejoinder. A popular form of the argument (Mike Winger's above) runs the twenty-two straight through the temple: Josephus “refers to the books that are laid up in the temple as being 22 in number.” But Gary Michuta presses a precise point in cross-examining White. The only three places Josephus explicitly says a writing was laid up in the temple name just the Pentateuch and Joshua — the scripture “laid up in the temple” that foretold water from the rock (Antiquities III.1.7), Moses's hexameter song “left behind in a book that was kept in the temple” (Antiquities IV.8.44), and the long day “made clear in the scripture laid up in the temple” (Antiquities V.1.17). So the temple-deposit phrase, taken on its own words, attests three books, not twenty-two; it cannot by itself carry a Hebrew-sized canon. White's reply, via Roger Beckwith (The Old Testament Canon of the New Testament Church, pp. 85–86), is fair and should be granted: it is improper to limit the temple's contents to the handful of writings three narrative asides happen to mention — Numbers and the Hagiographa were there too, and the fuller content is given by the twenty-two-count of Against Apion, not by these incidental references. Just so — but the concession cuts both ways: it means the “laid up in the temple” motif is not itself the proof, and the popular slogan that fuses phrase and count (“the books laid up in the temple, 22 in number”) claims a directness the texts do not supply. The weight rests where it always did — on the bare twenty-two-count and everything already said about it — not on the temple deposit, which by name reaches only three books.¹¹

Assessment

Contested. The witness is strong by the historian's lights and short by the canon-historian's. What Josephus soundly establishes — a datable, late-first-century Jewish reckoning of twenty-two divinely-credited books, a Hebrew-sized table that did not count the deuterocanon among the inspired — is genuine and is the firmest single datum the Protestant case has on the Jewish side; it is real evidence for the Hebrew-measure reading this timeline traces from Jerome. But the inference the proponents draw — that the count proves a closed, pan-Jewish, deuterocanon-excluding canon centuries old and binding on the church — outruns it: a single apologist's tidy, title-less figure cannot supply the uniformity the conclusion needs (the leading scholarship holds no Jewish canon was yet fixed); the "twenty-two" does not map cleanly onto the thirty-nine; and Josephus himself read 1 Esdras and 1 Maccabees as history. The strong reading is the one weighed weak in full on the umbrella page; the bare attestation is sound. Trust Josephus as a witness to a Hebrew-measured stream; flag — on the proponents' accounts — the slide from his count to a settled, church-binding canon, and Bible.org's lean on the discredited Jamnia council.

Sources