Paul has just argued that the inheritance comes by promise, not by law (vv. 15-18), and that a covenant ratified 430 years earlier cannot be annulled by a later enactment. The obvious objection follows at once, and he puts it himself: Τί οὖν ὁ νόμος; — what about the law, then? Why have it at all?

The answer is four compressed clauses, and each one subordinates the law to something else. It was added (a later supplement, not the foundation). It was added for the sake of the transgressions — a phrase that has been read in at least three incompatible ways since antiquity. It had a terminus: until the seed should come. And it arrived through angels and by the hand of a mediator — two degrees of remove, against a promise God made to Abraham directly (v. 18, κεχάρισται ὁ θεός).

Nothing here says the law is bad. Paul explicitly denies that inference two verses later (v. 21, μὴ γένοιτο). What the verse says is that the law is secondary, provisional, and indirectly delivered. That is enough for his argument, and it is all he claims.

The translator’s main obligations in this verse are negative: do not resolve χάριν, do not supply an agent for the divine passive, do not name the mediator, and do not render διά as though it were ὑπό.

A verbless question of four words. Two construals are live and both are ancient.

τί as predicate nominative/accusative: “What, then, is the law?” — a question about the law’s nature or status, answered by what follows.

τί as accusative of respect: “Why then the law?” — the idiomatic Greek way of asking after purpose (cf. τί γάρ; Phil 1:18, already fixed in this corpus as a frozen idiom). On this reading the question is not what the law is but what it is for, which is exactly what προσετέθη then answers.

The second is favored by the answer: Paul does not define the law, he states why it was given and for how long. But the Greek supplies no verb, so the copula and its force are supplied by the reader. Both the copula and the question mark are editorial — the earliest majuscules are unpunctuated — and the apparatus must say so rather than citing the printed text as evidence.

The inferential οὖν must be rendered (§10). It draws the objection out of vv. 15-18: given that the inheritance is by promise, what is the law doing here?

προστίθημι = πρός + τίθημι: to put to, lay alongside, add to what is already there. The LXX uses it constantly for adding to a stock, a number, a span of days (Gen 30:24; Deut 4:2; 1 Kgdms 12:19).

Two things are load-bearing here.

The prefix. Something added presupposes something already standing. That is the whole point of vv. 15-18: the διαθήκη was in force, and the νόμος came 430 years later as a supplement. Any rendering that drops the πρός- (“the law was given,” “the law came”) deletes the word on which the argument turns. Note that Deut 4:2 and 13:1 LXX use this very verb to forbid adding to what God commanded — a resonance worth recording but not pressing.

The passive. The agent is unexpressed. Paul could have written ὁ θεὸς προσέθηκεν and did not; he had just named God as the direct giver of the promise (v. 18, κεχάρισται ὁ θεός) and here declines to name the giver of the law. Verse 19b then supplies two intermediaries instead. Supplying “God added it” in the reading text undoes a reticence that the clause’s own continuation explains, and it also pre-empts a contested question about the law’s provenance. Keep the passive passive.

The variant ἐτέθη (D* F G and some Latin) is a simplification toward the plain verb and must not be followed; it is a smoothing of exactly the kind this prefix invites.

This is the crux of the verse, and the Greek underdetermines it.

χάριν is the frozen accusative of χάρις functioning as an improper preposition with the genitive, always postpositive (Luke 7:47; Eph 3:1, 14; 1 Tim 5:14; Titus 1:5, 11; 1 John 3:12; Jude 16). It means “for the sake of, on account of,” and it is genuinely open between purpose (in order to produce or deal with X) and cause (because of X). English “for the sake of” and “on account of” hold the same latitude, which is why both are retained.

παράβασις (already fixed in this corpus as “transgression,” παρά + βαίνω, “step across”) names an overstepping of a boundary that has been drawn. Paul makes the dependency explicit at Rom 4:15: οὗ δὲ οὐκ ἔστιν νόμος, οὐδὲ παράβασις — where there is no law, neither is there transgression. So the word itself already implies the standard it violates.

The articular plural τῶν is anaphoric or generic; it does not settle anything by itself, but it tells against rendering “because of sin” in the abstract, which would be ἁμαρτίας.

Proposed — pending reviewJ. Louis Martyn, Galatians: A New Translation with Introduction and Commentary, Anchor Bible 33A (New York: Doubleday, 1997)

ἄχρις οὗ is a fixed temporal conjunction (ἄχρι + genitive relative, with χρόνου understood). The final ς before a vowel is orthographic and carries no weight; the ἄχρι reading in part of the tradition is not a different word.

The corpus has already fixed ἄχρι as marking a terminus reached, not merely approached (Phil 1:5-6). That force is the point here: the law has an end-date written into its own charter. Paul is not saying the law was weak or that people grew out of it; he is saying it was given with a boundary. The παιδαγωγός of v. 24-25 makes the same point with a different image — ἐλθούσης δὲ τῆς πίστεως οὐκέτι ὑπὸ παιδαγωγόν ἐσμεν.

The aorist subjunctive without ἄν is ordinary Koine (cf. the ἕως + subjunctive entry already fixed at Jas 5:7): the indefiniteness attaches to the timing, never to whether the event occurs. No hedging may be introduced into English — “until such time as the seed might come” is excluded. Witnesses supplying ἄν are normalizing toward literary Greek and are secondary.

τὸ σπέρμα is articular and anaphoric: the seed of v. 16, where Paul has already pressed the collective singular of Gen 22:18 LXX into a reading ὅς ἐστιν Χριστός. Render “the seed” and do not substitute the identification into the text; the article does the pointing.

ᾧ ἐπήγγελται is syntactically interesting. The relative is dative — the party to whom the promise was made — and grammatically neuter in agreement with σπέρμα, though the referent is personal. English “to whom” carries the same slide.

ἐπήγγελται is a perfect of a middle-deponent verb used here with passive force: the promise stands made. Two features must survive: the perfect’s completed-act-with-standing-result (a simple past, “to whom it was promised,” severs the result from the present, which is precisely what Paul needs — the promise is still in force when the seed arrives), and the unexpressed agent. Paul names God as giver of the promise at v. 18 and withholds him here, consistent with the reticence of προσετέθη in the same sentence.

A small body of witnesses reads ἐπήγγελτο (pluperfect) or ἐπηγγείλατο (aorist middle active, “he promised”); the second supplies the agent and is clearly secondary, the first levels the tense.

διατάσσω (διά + τάσσω) belongs to the ordering family already represented in this corpus by ὑποτάσσω, ἀντιτάσσομαι, and στοιχέω. The sense is to set in order, arrange, promulgate as an ordinance — the verb of a commander disposing troops or a magistrate issuing a standing instruction (Matt 11:1; Luke 8:55; Acts 7:44; 1 Cor 7:17; 11:34; 16:1; Titus 1:5). Keep the τάσσ- root’s ordering-force audible: “ordained” is permitted in declared liturgical register but has drifted in modern English toward clerical appointment, and “put into effect” drains the τάσσω.

The aorist passive participle agrees with ὁ νόμος and is coincident with προσετέθη: the adding and the ordering are one transaction described twice.

διά + genitive is intermediate agency, and the corpus keeps it rigorously distinct from ἀπό/ἐκ (source) and ὑπό (personal agent of a passive). This is the verse’s sharpest structural claim. Paul does not write ὑπὸ ἀγγέλων. The angels are the channel, not the originators; and the whole weight of vv. 19b-20 is that a channel was used at all, over against the promise, which God made to Abraham directly.

On ἄγγελος the corpus policy (Gal 4:14) is to flag the messenger/angel ambiguity where the referent is open. Here it is not open. Deut 33:2 LXX has ἐκ δεξιῶν αὐτοῦ ἄγγελοι μετʼ αὐτοῦ at Sinai, and the tradition of angelic mediation at Sinai is well established in Second Temple Judaism (Jubilees 1:27-29; Josephus, Ant. 15.136) and taken up in the NT at Acts 7:38, 53 and Heb 2:2. Render “angels.”

Proposed — pending reviewRichard N. Longenecker, Galatians, Word Biblical Commentary 41 (Dallas: Word Books, 1990)

ἐν χειρί is a Semitism. It calques Hebrew בְּיַד, the standard idiom for agency exercised by a named intermediary — repeatedly used in the LXX precisely of the law given through Moses: ἐν χειρὶ Μωυσῆ (Lev 26:46; Num 4:37, 41, 45; 15:23; 36:13; 3 Kgdms 8:53, 56). That the phrase is idiomatic Greek-of-translation rather than natural Attic is itself the datum: Paul is using the Pentateuch’s own formula for the Mosaic promulgation without naming Moses.

Per §8 the hand stays in the rendering where English permits. “By the hand of a mediator” is idiomatic enough in English to carry the construction; “by a mediator” and “through an intermediary” delete the organ and the LXX echo with it.

μεσίτης (from μέσος) is the one who stands in the middle: the guarantor of an oath, the arbitrator, the go-between of a transaction. The word is attested in the papyri in exactly this legal-commercial sense. In the NT it occurs at Gal 3:19, 20 and at 1 Tim 2:5 and Heb 8:6; 9:15; 12:24.

The noun is anarthrous and qualitative and the mediator is unnamed. Supplying “Moses” in the reading text is excluded: Paul withholds the name here and at v. 20, and the withholding is what allows the argument of v. 20 (ὁ δὲ μεσίτης ἑνὸς οὐκ ἔστιν) to run as a general proposition about mediators rather than a point about one man. Record the identification in a note.

Proposed — pending reviewF. F. Bruce, The Epistle to the Galatians: A Commentary on the Greek Text, New International Greek Testament Commentary (Grand Rapids: Eerdmans, 1982)

The Fathers felt the force of the objection Paul is answering and were careful not to let the verse become an attack on the law.

Chrysostom, commenting on Galatians, reads τῶν παραβάσεων χάριν restrainingly: the law was added to check wrongdoing, as a bridle on a people inclined to it — and he presses that Paul says προσετέθη, added, precisely to show the law is a supplement to something prior and not the ground of the inheritance.

Jerome, in his Galatians commentary, works through both the restraining and the revelatory senses of the phrase and notes the two degrees of remove (angels, mediator) as the point of the second half of the verse.

Augustine, in his exposition of Galatians, leans toward the revelatory-provocative sense, reading it alongside Rom 5:20 and 7:7-13: the law makes sin visible as transgression and thereby drives the sinner to grace.

That the Greek and Latin Fathers divide along the same lines modern commentators do is a reasonable sign that the ambiguity is in the text and not in our distance from it.

Proposed — pending reviewJohn Chrysostom, Commentary on the Epistle to the Galatians, in Nicene and Post-Nicene Fathers, First Series, vol. 13 (Edinburgh: T&T Clark, 1889) · Jerome, Commentary on Galatians, trans. Andrew Cain, Fathers of the Church 121 (Washington, DC: Catholic University of America Press, 2010) · Augustine, Commentary on Galatians (Epistolae ad Galatas expositionis liber unus), ed. and trans. Eric Plumer (Oxford: Oxford University Press, 2003)

The verse is relatively stable, but three places deserve recording.

προσετέθη. A few Western witnesses (D* F G, with Latin support) read ἐτέθη, “was laid down.” This drops the πρός- and with it the argument that the law is a later supplement — a simplification in the direction the surrounding argument makes tempting. The compound is clearly original.

ἄχρις οὗ. The ἄχρι / ἄχρις alternation is orthographic (ς before a vowel). Some witnesses add ἄν with the subjunctive, which is the literary normalization; the Koine omission is the harder and original reading.

ἐπήγγελται. Read as a perfect middle/passive by NA28. Variants include ἐπήγγελτο (pluperfect) and the active ἐπηγγείλατο, which supplies God as subject. Both are secondary: the first levels the standing-result force, the second fills in the agent the divine passive withholds.

Some later witnesses also read ὃ ἐπήγγελται (neuter relative, agreeing formally with σπέρμα) — a grammatical tidying that loses the personal reference the dative masculine carries.

The verse is the hinge of Galatians 3. Having established that the inheritance comes δι᾽ ἐπαγγελίας (v. 18), Paul must account for the law’s existence, and he does so with four subordinating moves in a single sentence.

1. Added — temporally and logically secondary to the promise.
2. For the sake of transgressions — serving a function defined by sin, not by inheritance.
3. Until the seed should come — terminable, with its end written into its commission.
4. Through angels, by a mediator’s hand — delivered at two removes, against the promise given to Abraham directly.

Each of these is a limitation, and none is a condemnation. Verse 21 makes that explicit: Ὁ οὖν νόμος κατὰ τῶν ἐπαγγελιῶν τοῦ θεοῦ; μὴ γένοιτο. The law is not opposed to the promises; it simply does something else, for a while, through others.

The translator’s discipline here is restraint. Paul’s own sentence is reticent at four points — the adder is unnamed, the purpose is stated with an ambiguous preposition, the promise’s giver is unnamed, and the mediator is unnamed — and every one of those silences does work in the argument that follows. A rendering that fills them in produces a clearer verse and a worse one.

Proposed — pending reviewHans Dieter Betz, Galatians: A Commentary on Paul's Letter to the Churches in Galatia, Hermeneia (Philadelphia: Fortress, 1979)