How many male Levites more than one month old did Moses and Aaron count?

What follows is one scholar’s way of reading this disagreement — a critical-historical view that often concludes the texts genuinely disagree. It draws on the peer-reviewed sources listed below and stands as one position among the answers above, presented for transparency, not as GEBible’s position.

How critical scholarship weighs this: Genuinely disputed among scholars

This summarizes where critical scholarship tends to land on this tension — a report on the field’s center of gravity, not GEBible’s own ruling.

Numbers 3:39 gives the total of all Levite males a month old and upward as 22,000. The clan subtotals in the same chapter — 7,500 Gershonites (3:22), 8,600 Kohathites (3:28), and 6,200 Merarites (3:34) — sum to 22,300 in the Masoretic Text. The 300-Levite discrepancy is internal to a single chapter and turns on simple addition.

The long-running apologetic harmonization is the firstborn-deduction theory, traceable to Keil and Delitzsch and carried forward by Gleason Archer and (cautiously) Timothy Ashley in NICOT *Numbers*. Its premise is the substitution scheme of 3:40-51: the Levites are taken in place of the firstborn of Israel, with 22,273 Israelite firstborn counted (3:43) and a 273-firstborn surplus redeemed at five shekels each (3:46-48). The argument runs that the 300 'missing' Levites are themselves *firstborn*, and a firstborn Levite cannot stand in for another firstborn — he already bears his own consecrated status — so the 22,000 figure in 3:39 is not the gross total but the count of Levites *available for substitution*, with 300 firstborn Levites set aside. This has the genuine virtue of solving the problem from the chapter's own logic, and the substitution arithmetic of 3:43-48 really does require careful bookkeeping of firstborn.

But the harmonization carries two weaknesses. First, the text does not say it. Numbers 3:39 reads, 'all that were numbered of the Levites... were twenty and two thousand' — *kol-pequdei ha-leviyyim*, an unqualified total. To read it as a substitution-eligible subset is to import a restriction the verse does not state. Second, the figure of 300 firstborn Levites is reverse-engineered to balance the sum; there is no independent attestation of that number anywhere in the chapter, and the redemption transaction in 3:40-51 concerns the surplus of *Israelite* firstborn over *all* Levites, not over a reduced count. The harmonization rescues the arithmetic at the cost of inventing a datum.

The text-critical reading is more economical. A commonly proposed scholarly explanation is a numeral corruption in the Kohathite total: where the MT of 3:28 reads 8,600, a number of commentators conjecture an original 8,300. With 8,300 Kohathites, the three subtotals (7,500 + 8,300 + 6,200) sum to exactly 22,000 — matching the grand total in 3:39. It must be stated precisely what kind of evidence this is. This 8,300 figure is a conjectural emendation proposed to harmonize the sum, not a securely attested reading of the Septuagint; the standard LXX (Rahlfs and the major uncials) reads the Kohathite figure in agreement with the MT, and the 8,300 emendation is a scholarly proposal sometimes noted in critical apparatus rather than a manuscript variant. (The Septuagint and the Samaritan Pentateuch broadly preserve the same census numbers as the MT, so the discrepancy is not resolved by an existing version; the 8,300 reading is a reconstruction.) The conjecture is nonetheless attractive because a hundreds-digit slip (300 vs. 600) is exactly the sort of error that arises easily in numeral transmission, where digits and digit-signs are notoriously unstable. George Buchanan Gray's ICC *Numbers* sets out the arithmetic difficulty and the proposed emendation in detail.

This sharpens rather than dissolves the critical picture, and it is not the only critical option. A second view treats the round 22,000 as an independent, possibly schematic total drawn from a different stratum than the itemized clan figures, with the two never fully reconciled — in which case there is nothing to 'emend' because the numbers were never a single arithmetical system. Others simply leave the discrepancy unresolved. Baruch Levine (Anchor Bible) and Jacob Milgrom (JPS) both note that the Priestly census figures do not consistently reconcile and bear the marks of complex transmission; Milgrom discusses the firstborn-substitution scheme and the arithmetic of 3:39-51 in detail. Emanuel Tov's *Textual Criticism of the Hebrew Bible* documents precisely this class of phenomenon — mismatches between a rounded grand total and the sum of its components — as a recurrent feature of numerical transmission in the Hebrew Bible, where digits were especially prone to corruption. Whether one judges the MT 8,600 a copyist's slip (with a conjectured original of 8,300) or treats the whole census apparatus as a layered Priestly composition whose round total and itemized subtotals never fully cohered, the result is the same: the discrepancy is a transmission/composition artifact, not a substantive claim about firstborn bookkeeping.

Where the critical reading lands: most plausibly a scribal/numeral corruption or a composite-total seam. The firstborn-Levite theory is contextually motivated and internally coherent, and one cannot strictly rule it out — but it depends on an unstated qualification of 3:39 and a stipulated figure of 300. A single hundreds-digit corruption in the Kohathite total (8,600 for an original 8,300) yields a consistent 22,000 and is the simpler account, even though it rests on conjecture rather than an attested version.