HistoricalChristian.Faith

Galatians 3:10–13 — the “curse of the law”: does Paul’s argument depend on the Septuagint’s text-form?

Type: supporting argument (off-timeline) · Assessment: sound — the catalog’s LXX-distinctive tag on Galatians 3:10 and its neither / free tag on 3:13 both hold: Paul’s “cursed is every one that continueth not in all things” carries the explicit “all” (πᾶς … ἐν πᾶσιν) that the Septuagint and the Samaritan Pentateuch read into Deuteronomy 27:26 against a Masoretic Hebrew that lacks it, and his “cursed is every one that hangeth on a tree” reshapes an LXX-based Deuteronomy 21:23 — swapping in the catchword ἐπικατάρατος from 3:10 and pointedly dropping the “of God.” But the strong framing that Paul’s justification argument hangs on a Greek plus the Hebrew does not supply overreaches on two counts, and the page fixes both.

See also: this row belongs with, but does not join, the Septuagint-gloss group (Isaiah 7:14, Psalm 40:6, Amos 9:12): those are the Greek reading a stable Hebrew, whereas here a Hebrew witness (the Samaritan Pentateuch) shares the divergent reading, so the fork the doctrine turns on — a translational gloss versus an older Hebrew the Masoretes lost — is genuinely live at this verse and is adjudicated below. The one place Paul entertains an outright corruption-charge here (“deleted by the Jews”) is the seam this doctrine shares with rabbinic corruption of the text.

The claim

“We consider it uncertain whether the seventy interpreters [of the Septuagint] have added ‘everyone’ and ‘in all’ or whether it was in the old Hebrew and deleted by the Jews. What makes me suspect this is that the apostle, a man skilled in Hebrew learning, would never have added these words everyone and all as if they were necessary to his meaning … unless they were in the Hebrew copies. Therefore, reading over the Hebrew copies of the Samaritans, I found the word kol written, which means ‘all’ or ‘in all’ and concurs with the Seventy.”

— Jerome, Commentary on Galatians 2.3 (c. 386), on Galatians 3:10. Sixteen centuries before the modern debate, Jerome states the exact fork this page must resolve: is the universalizing “all” a Greek plus (the Septuagint’s addition) or an older Hebrew the Masoretic tradition lost? He checks the Samaritan Pentateuch, finds the Hebrew word kol there, and leans toward the second answer. The interpretation this page weighs is the catalog’s classification and the reading it serves: that Paul’s curse-of-the-law proof is Septuagint-shaped at 3:10 (the explicit “all”) and a deliberate Greek-based adaptation at 3:13 (the dropped “of God”) — and, in its strong form, that the argument depends on that Greek text-form.¹

Who advances it

The text-form observation is old and undisputed: the ancient expositors who quote Galatians 3:10 reproduce Paul’s “all things … written in the book of the law” without comment — Justin Martyr already builds “the whole human race … under a curse” on the “all” (Dialogue with Trypho 95), and Chrysostom turns the two curses (Deut 27:26 and 21:23) into a single exchange. Jerome (above) is the first to name the divergence and to check a Hebrew witness against it. In the modern text-critical surveys the fact is a flat classification: Archer & Chirichigno and the standard quotation tables sort Galatians 3:10 among the citations that follow the Septuagint where it deviates from the Masoretic Hebrew, and note that the plus is shared by the Samaritan Pentateuch. The interpretive claim built on that fact — that Paul’s justification-by-faith argument leans on a Greek reading the Hebrew does not supply — is the load-bearing version this page tests; its strongest recent statement of the text-form’s role is Christopher D. Stanley’s study of Paul’s citation technique, which reads Paul’s wording here as an interpretive rendering rather than a mechanical copy of any one text.²

The evidence for

The “all” is really absent from the Masoretic Hebrew and really present in the Greek. The Masoretic Deuteronomy 27:26 reads “Cursed be he that confirmeth not the words of this law to do them” — no kol (“all”) stands in the consonantal text. The Septuagint reads “Cursed is every man (πᾶς ἄνθρωπος) who does not continue in all (ἐν πᾶσιν) the words of this law,” and Paul at Galatians 3:10 has “Cursed is every one (πᾶς) that continueth not in all things (πᾶσιν) which are written in the book of the law to do them.” Both universalizing words on which Paul’s “keep-all-or-be-cursed” point turns are the Greek reading, absent from the Hebrew as the Masoretes vocalized it. (The pull of the sense is such that even the King James translators supplied “all the words of this law” at Deuteronomy 27:26.)¹

At 3:13 Paul reshapes a Greek-based Deuteronomy 21:23. The Hebrew of Deuteronomy 21:23 reads “for he that is hanged is qilelat ʾəlōhîm” — “a curse of God” / “accursed of God” — and the Septuagint renders it “cursed by God (κεκατηραμένος ἥπὸ θεοῦ) is everyone hanging on a tree (ἐπὶ ξύλου).” Paul’s Galatians 3:13 keeps the LXX-shaped “every one … on a tree” but makes two changes: he substitutes ἐπικατάρατος (“accursed,” the very word that opens the Deuteronomy 27:26 citation two verses earlier) for the Septuagint’s κεκατηραμένος, welding the two proof-texts by catchword; and he drops the “of God” / “by God” that both the Hebrew and the Greek carry. The omission is pointed: where Tertullian, quoting the fuller passage against the Jews, still writes “cursed by God is every one who shall have been suspended on a tree,” Paul’s form says only “cursed is every one that hangeth on a tree” — Christ becomes “a curse for us” without the sentence making God the direct agent of the cursing.⁵

Reading Deuteronomy 21:23 as a crucifixion-curse was a live pre-Christian move. Paul is not inventing the link between the “hanged … accursed” law and the cross. The Qumran Temple Scroll (11QTemple 64:6–13) already applies Deuteronomy 21:22–23 to hanging a living man for treason — execution by suspension — and even expands the curse to “cursed by God and men”; the Nahum pesher (4QpNahum 3–4 i.7–8), the “Lion of Wrath” who “hangs men alive,” is widely read of crucifixion. The same “tree” idiom carries the Deuteronomy-21 frame straight into the earliest Christian preaching (“whom ye slew and hanged on a tree,” Acts 5:30, 10:39; “bare our sins in his own body on the tree,” 1 Peter 2:24). So the exegetical move at 3:13 rests on a shared Jewish reading, not a Christian novelty.⁶

Adversarial assessment

What holds. The catalog’s two tags are correct. At 3:10 Paul’s explicit “all” is the Septuagint’s reading, absent from the Masoretic consonantal text — a genuine LXX-distinctive divergence — and at 3:13 his citation is neither pure Septuagint (he swaps the verb and drops “of God”) nor pure Hebrew, a real free-adaptation shaped by the Greek. Paul’s Bible here is a Greek Bible, and the wording shows it. The two sub-claims are one reading, not rival ones: the “all” and the dropped “of God” both serve the single thesis that the proof-texts are text-form–shaped, so they fold onto one page rather than splitting.

What must be refined, so the tag is not misread.

1. A Hebrew witness shares the “all,” so this is not a gloss over a stable Hebrew — but neither is it a demonstrable older Hebrew the Masoretes trimmed. The plus is not uniquely Greek: the Samaritan Pentateuch reads kol (“all”), as Jerome verified by hand and as later collations confirmed (the word also appears in a handful of medieval Hebrew manuscripts and in Targum copies). That already separates this row from Isaiah 7:14 and Psalm 40:6, where no Hebrew witness carries the Greek reading. Jerome pressed the point further — Paul, a Hebraist, would not have leaned on words he did not find in Hebrew copies, so (he reasons) “all” was original and “deleted by the Jews.” That is a real possibility, and it is exactly the older-Hebrew fork the doctrine turns on — but here it cannot be carried, for three reasons. First, the decisive witness is missing: Deuteronomy 27:26 is unattested at Qumran — the extant Deuteronomy scrolls cover chapter 27 only through v. 10 (4QDeut^f), so the Dead Sea Scrolls can neither confirm nor deny the “all,” unlike the christological clause of Deuteronomy 32:43 that 4QDeut^q does preserve. Second, the Samaritan Pentateuch is the textbook expansionist text, prone to filling a phrase out toward its parallels. Third, “all the words of this law” is a stock Deuteronomic formula standing twice in the immediate context (Deuteronomy 27:3, 27:8, of the words written on the plastered stones), precisely the gravitational pull that generates an assimilating plus. So the mainstream judgment — that the Masoretic short reading is likely original and the shared Samaritan/Greek “all” a secondary harmonization — is the better bet; the honest verdict is that this is an LXX-and-Samaritan plus, most probably harmonizing, which is why the row stays tagged LXX-distinctive rather than being re-tagged older-Hebrew.³

2. Paul’s argument does not depend on the plus; it survives on the Hebrew. The strong framing — “the whole point leans on the Septuagintal ‘all’” — overstates it. Deuteronomy 27:26 is the concluding, summarizing curse of the series (the twelfth of the Shechem dodecalogue, 27:15–26): after eleven specific curses it curses whoever fails to uphold “the words of this law” as a whole — the totality is carried by its function as the summary curse, with or without the word. And the covenant-curse logic is spelled out with the very word “all” a few lines on: the curses fall on Israel if it will not “observe to do all his commandments” (Deuteronomy 28:15; cf. “Keep all the commandments,” 27:1). So the Hebrew Deuteronomy already says what Paul needs; the Septuagint makes explicit the comprehensiveness the context implies rather than supplying a premise the Hebrew withholds. This is the standard commentators’ reading — Silva takes the LXX “all” to underline a totality already present, and Bruce reads Paul’s Deuteronomy 27:26 as invoking the whole pentateuchal law. (What is genuinely contested at 3:10 is a different premise entirely — whether Paul assumes the impossibility of perfect obedience, or means Israel’s corporate/exilic curse — the Sanders–Dunn–Wright debate; but that turns on covenant theology, not on the text-form, and it is untouched by the “all.”) So the row is like Psalm 40:6, not like Isaiah 7:14: the Greek wording is exploited and real, but the argument would stand on the Hebrew.⁴

3. The dropped “of God” is best read as theological disambiguation, exploiting a real Hebrew ambiguity — not a distortion, and not certainly a text-critical datum. Three accounts compete, and the reading survives on any of them. (a) Theological motive (the dominant view): Paul declines to say Christ was cursed by God directly — he has Christ “become a curse” (a status shouldered for us), keeping the Torah’s legal ἐπικατάρατος while dropping the agent-phrase. This is not sleight of hand, because the Hebrew qilelat ʾəlōhîm is genitivally ambiguous: a subjective genitive (“cursed by God”), which the Septuagint’s ἥπὸ θεοῦ chose, or an objective one (“an affront to God,” a thing God’s law abhors) — Paul’s omission simply declines to press the reading that makes God the direct curser. (b) Catchword abbreviation (Fitzmyer): Paul is chaining ἐπικατάρατος across 3:10 and 3:13 to bind the two texts, and “of God” is trimmed as excess to that verbal join. (c) A known variant / testimonia tradition (Wilcox): Paul’s form may reflect an early Jewish-Christian “tree”-testimony rather than a fresh cut from the Septuagint. The lead’s framing — a “theologically motivated adaptation” — is therefore well within reach, but it should be held as the most probable of three live explanations, not as a proven authorial intention.⁵

Assessment

Sound. The interpretation this argument serves — that Paul’s curse-of-the-law proof is Septuagint-shaped at both hinges (the explicit “all” of Deuteronomy 27:26 at Galatians 3:10, and the reshaped, LXX-based Deuteronomy 21:23 at 3:13) — holds: the “all” is genuinely absent from the Masoretic Hebrew and present in the Greek, and 3:13 is a real free-adaptation (catchword ἐπικατάρατος in, “of God” out). Both catalog tags stand. What the tags must be trusted to mean, and no more, is fixed by three refinements. First, a Hebrew witness (the Samaritan Pentateuch) shares the “all,” so this is not the Greek glossing a stable Hebrew as at Isaiah 7:14 — yet because Deuteronomy 27:26 is unattested at Qumran, the Samaritan text is expansionist, and “all the words of this law” is a stock formula standing twice in the immediate context, the plus is most probably a harmonization, not a demonstrable older Hebrew the Masoretes trimmed; Jerome’s “deleted by the Jews” remains a live minority possibility the evidence cannot carry, so the row stays LXX-distinctive. Second, the argument does not depend on the plus: Deuteronomy 27:26 is the summary curse of the whole law, and 28:15’s “to do all his commandments” supplies the comprehensiveness in the Hebrew, so the Greek makes explicit what the context already implies — the wording is exploited, as at Psalm 40:6, not load-bearing as at Isaiah 7:14. Third, the dropped “of God” at 3:13 is best read as theological disambiguation that trades on the real ambiguity of the Hebrew qilelat ʾəlōhîm, one of three live explanations, none of which distorts the text. Kept to that, Galatians 3:10–13 is a clean instance of the apostle arguing from a Greek Old Testament; stretched into “the Septuagint preserves the older Hebrew here” or “the doctrine of justification hangs on a Greek plus,” it overreaches.

Sources