Paul has been running an allegory (vv. 21-31) on Abraham’s two sons. Verse 29 observed that the son born κατὰ σάρκα persecuted the son born κατὰ πνεῦμα, and that the pattern holds now. Verse 30 does not answer that observation with strategy or with counsel; it answers it with a citation, introduced by the sharpest pivot available: ἀλλὰ τί λέγει ἡ γραφή; The Galatians’ situation is not to be adjudicated by experience but by the text.

Three things carry the verse. First, the personification: Scripture speaks, in the present tense, and it speaks an imperative. Second, Paul quotes Gen 21:10 LXX almost verbatim but changes the last phrase — Sarah’s “with my son, with Isaac” becomes “with the son of the free woman.” The change is not paraphrase; it is the hinge that binds the citation to the allegory Paul has just built and to the conclusion of v. 31. Third, the negation οὐ … μή with a future indicative is the strongest denial Koine can construct: not a prediction that it probably will not happen, but a categorical exclusion.

The pastoral edge is uncomfortable and should not be blunted. Sarah’s words in Genesis were spoken in jealousy, and Genesis reports Abraham’s distress (Gen 21:11). Paul nonetheless takes them as ἡ γραφή speaking. The command he thereby lays on the Galatian assemblies concerns the Torah-imposing mission, not the standing of Jewish people as such — but the verse is hard, and the translation should let it be hard.

Built onna28 · bdag · bdf

ἀλλά here is the strong clause-initial adversative already fixed in the glossary, not the ascensive ἀλλὰ καί of Phil 1:18. It cuts off the line of thought in v. 29 (this is how it is, and always has been) and replaces it with a different authority. Per §10 the connective is load-bearing and must be rendered; a bare “now” or an untranslated pivot loses the argumentative turn.

τί λέγει ἡ γραφή; is a Pauline formula (Rom 4:3; 11:2; Gal 4:30), and its rhetorical function is not to seek information but to submit the case to a higher court. The present λέγει is standard for citation in Paul (γέγραπται is the perfect alternative, used two verses earlier at v. 27) and should be rendered as a plain English present. The choice between the perfect “it stands written” and the present “it says” is not random: γέγραπται presents the text as a standing deposit, λέγει presents it as currently addressing the hearers. Here Paul wants the second.

Built onbdag · bdf · na28

ἡ γραφή in the singular with the article regularly denotes a specific passage of the sacred writings (Mark 12:10; John 19:37; Acts 8:32), while the plural αἱ γραφαί denotes the collection. But the construction here — Scripture as the subject of a verb of speaking, issuing a command — is the personification Paul also uses at Rom 9:17 and Gal 3:8, where ἡ γραφή does what God does. The identification is not casual: what Sarah said, Scripture says; what Scripture says, God says.

Rendering: keep Scripture capitalized and retain the article. “The passage” would be lexically defensible but would dissolve the personification that gives the verse its force.

Built onbdag · muraoka-lxx

παιδίσκη is formally a diminutive of παῖς but in Hellenistic usage denotes a female household slave without any necessary implication of youth. In the LXX it renders both אָמָה and שִׁפְחָה, the two Hebrew terms for a bondwoman, and at Gen 21:10 it stands for אָמָה. The papyri use it flatly of owned female labor.

Per §5 and the standing δοῦλος decision (‘slave,’ not ‘servant’), the ownership must survive into English. Slave woman is the rendering; “bondwoman” is acceptable in a declared liturgical register, where its archaism is a feature. “Maidservant,” “handmaid,” and “slave girl” are all excluded: the first two soften a legal institution the argument depends on, and the third imports an age the word does not carry.

Note that Paul never names Hagar in the citation, though he named her at vv. 24-25. Within the quotation she is only the slave woman — a status, not a person. That is Sarah’s rhetoric in Genesis, faithfully preserved, and English should preserve it too.

Built onbdag · muraoka-lxx

Ἔκβαλε is aorist imperative of ἐκβάλλω, ἐκ + βάλλω: throw out. The verb is used of expelling demons (Mark 1:34), driving traders from the temple (Mark 11:15), and ejecting a person from a community (John 9:34-35; 3 John 10). It is never gentle. LXX Gen 21:10 uses it to render Hebrew גָּרֵשׁ, the same verb used of expelling Adam from the garden (Gen 3:24) and the nations from Canaan.

Render with a verb of forcible expulsion: cast out (formal), drive out (modern). “Send away” and “get rid of” are both excluded — the first softens the violence, the second lowers the register to the colloquial. The aorist is a single decisive act, not an ongoing policy.

Built onbdag · muraoka-lxx

οὐ γὰρ μὴ κληρονομήσει — the doubled negative οὐ μή (here split by γάρ, which is postpositive and takes second position in its clause) is the emphatic negation, most commonly with the aorist subjunctive but also, as here, with the future indicative (BDF §365). The two negatives reinforce rather than cancel: Koine piles negatives for strength, as the glossary already records for μή + μηδείς.

The LXX uses οὐ μή heavily in solemn divine pronouncements, which is likely why it carries such weight in NT citation contexts. English must mark the emphasis: shall certainly not, will never, shall by no means. A bare “will not” under-translates a construction Greek reserves for its most categorical denials.

Note that the emphasis attaches to the legal exclusion, not to the expulsion. The son will not inherit — that is the ground (γάρ) on which the expulsion is commanded.

Built onbdf · bdag · muraoka-lxx

κληρονομέω (κλῆρος, ‘allotted portion’ + νέμομαι) is the standard LXX verb for entering into an inheritance, and in Deuteronomy and Joshua it is near-technical for Israel’s possession of the land. Paul has been working the κληρο- root through this chapter: κληρονόμος at 4:1 and 4:7 (‘heir’), and now the verb here. Galatians 3:29 has already declared the Galatians κατʼ ἐπαγγελίαν κληρονόμοι.

One English root across all of them — inherit / heir / inheritance — or the argument’s spine becomes invisible. Do not vary to “receive a share” or “obtain the estate.”

The future tense is legal-declarative, not merely predictive: it states what the law of the household will and will not permit. Compare the codified prohibition in the household law of the ancient Near East on the one hand and, on the other, Deut 21:15-17, which protects a disfavored wife’s firstborn — the tension is real and Genesis itself registers it in Abraham’s distress (Gen 21:11).

Built onbdag · muraoka-lxx

μετὰ τοῦ υἱοῦ τῆς ἐλευθέρας — μετά with the genitive marks association, and that is precisely what is excluded. The Greek does not say the slave woman’s son gets nothing; it says he does not inherit in company with the free woman’s son. Joint heirship is the impossibility.

This matters for the argument. Paul’s target is not the existence of Torah or of Israel but the attempt to place law-observance alongside promise as a co-equal basis of inheritance. Render with or alongside and keep the associative force visible; “as well as” drifts toward addition, and dropping the preposition into “instead of” inverts the sense entirely.

Built onbdag · bdf

LXX Gen 21:10 reads: Ἔκβαλε τὴν παιδίσκην ταύτην καὶ τὸν υἱὸν αὐτῆς· οὐ γὰρ μὴ κληρονομήσει ὁ υἱὸς τῆς παιδίσκης ταύτης μετὰ τοῦ υἱοῦ μου Ισαακ.

Paul makes three changes:

1. He drops both instances of ταύτην / ταύτης (‘this’). Sarah pointed at a particular woman; Paul is stating a principle, and the demonstrative would tie the sentence back to one household.
2. He replaces τοῦ υἱοῦ μου Ισαακ (‘my son Isaac’) with τοῦ υἱοῦ τῆς ἐλευθέρας (‘the son of the free woman’). The first-person possessive belonged to Sarah; Paul is not Sarah. More importantly, the substitution converts a name into a category, and the category — ἐλευθέρα — is the one the whole allegory has been building toward (vv. 22, 23, 26, and the conclusion at v. 31).
3. He omits the introductory καὶ εἶπεν Σαρρα τῷ Αβρααμ. The words are no longer Sarah’s; they are ἡ γραφή’s.

The Hebrew of MT Gen 21:10 reads גָּרֵשׁ הָאָמָה הַזֹּאת וְאֶת־בְּנָהּ, and closes עִם־בְּנִי עִם־יִצְחָק, ‘with my son, with Isaac.’ The LXX is a close rendering; Paul’s departures are his own, not inherited from a variant Greek text.

This is canonical-intertextual reinterpretation by a later biblical author (§9), not post-biblical reception, and it should be footnoted as biblical theology. The substitution is the interpretive act of the verse, and any translation that quietly restores “Isaac” — as a few later witnesses were tempted to do — erases Paul’s argument.

The text of the verse is stable in its essentials. Two points are worth recording.

τῆς ἐλευθέρας. A small group of witnesses, under the pull of the Genesis wording, read μου Ἰσαάκ or add the name. The reading τῆς ἐλευθέρας is supported by the great uncials and by the internal argument — no scribe would replace “Isaac” with an abstraction, whereas assimilation to the LXX source text is the commonest of scribal reflexes. NA28 prints τῆς ἐλευθέρας without brackets.

ταύτης. Some witnesses insert the demonstrative after παιδίσκης in the second half, again harmonizing to Gen 21:10 LXX. The shorter text is original on the same reasoning.

Neither variant is doctrinally consequential; both illustrate the standard direction of scribal drift toward a known source text. The apparatus is noted here rather than resolved silently (§4).

Built onna28 · metzger-tc

What the verse does not license, on any reading, is a verdict on the Jewish people as such. Paul is arguing about the terms of inheritance among those who claim Abraham, and he places himself and his Jewish-Christian colleagues on the side of the free woman. Reading the verse as an ethnic pronouncement retrojects a category Paul is not using and inverts his argument at Rom 9:1-5 and 11:1 (§5).

Built onbdag · bauckham-god-crucified

The patristic tradition is nearly unanimous that the expulsion concerns the law-covenant rather than persons.

Chrysostom (Homilies on Galatians, ad loc.) reads the casting out as the abolition of the old covenant’s claim: the son of the bondwoman is cast out not because he is despised but because slavery cannot share an inheritance with freedom. He is careful to note that Paul’s alteration — “the son of the free woman” rather than “Isaac” — is deliberate and makes the verse a statement about categories.

Jerome (Commentary on Galatians) takes ἐκβάλλω as directed against the Judaizing teaching, and observes that Paul substitutes the general term for the proper name precisely so that the reader will not restrict the sentence to one household.

Augustine, in his commentary on Galatians and again in City of God 15.2, uses the passage for the two-cities scheme: the earthly Jerusalem in bondage, the heavenly free. He is the source of much later allegorizing, some of it far beyond what Paul’s text supports.

The Vulgate renders eice ancillam et filium eius; non enim heres erit filius ancillae cum filio liberae — note eice, a strong expulsion verb, and heres erit (‘will be heir’) which converts the verb κληρονομήσει into a predicate noun. The Latin loses the κληρο- verbal link to Gal 4:1, 7. This is reception, not evidence (§11).

Built onmuraoka-lxx · bdag