Verse 9 is not a new argument but the same curse laid down a second time, deliberately and in a changed shape. Paul has just put the most extreme case he can construct — himself, or an angel out of heaven (v. 8, with a hypothetical subjunctive, ἐὰν … εὐαγγελίζηται). Now he drops the hypothetical and states the case as live: εἴ τις … εὐαγγελίζεται, first-class condition with the indicative. Somebody is doing this.

Four features carry the verse.

1. The doubling is announced, not accidental. ὡς προειρήκαμεν … καὶ ἄρτι πάλιν λέγω is a formal notice of repetition, the same move as 2 Cor 13:2 (προείρηκα καὶ προλέγω) and 1 Thess 4:6. Paul tells the reader he is repeating himself and then repeats himself.
2. The verbs shift person. Plural προειρήκαμεν, singular λέγω. Whatever the earlier saying was, it was said by more than one voice; the present restatement is Paul’s alone, and v. 10 goes on to ask about Paul alone.
3. The standard shifts from Paul’s preaching to the Galatians’ reception. Verse 8 measured the rival message against ὃ εὐηγγελισάμεθα ὑμῖν — what we proclaimed. Verse 9 measures it against ὃ παρελάβετε — what you received. The norm is now in their hands, which is why the question of v. 6 (that they are shifting) is a question about them and not only about the agitators.
4. ἀνάθεμα ἔστω is repeated verbatim. Two words, unchanged, closing both verses. Paul does not soften the second statement; the repetition is the point (§6: a hard text stays hard).

What the verse does not do is name anyone, specify the content of the rival message, or say who executes the curse. All three silences are preserved in the rendering.

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

ὡς heads a comparative clause and καί in the second limb is apodotic — it answers ὡς rather than simply joining a further clause (BDF §442 on καί in the apodosis after a comparative or temporal protasis; cf. Matt 6:10, ὡς ἐν οὐρανῷ καὶ ἐπὶ γῆς).

This matters for English. Rendering “As we have said before, and now I say again” reads as two coordinated reports. Rendering “As we have said before, so now I say again” shows the second clause completing the first. The corpus requires that adjunctive and correlative καί always be rendered and never dropped (§10), so the particle stands in every tier; the literal tier keeps ‘and’ to display the form and the reading text uses ‘so’ to show the correlation.

Neither limb may be reduced to a discourse marker. ὡς προειρήκαμεν is an assertion about something actually said on an earlier occasion, not a stock phrase meaning ‘as I was saying.’

Built onbdf

The glossary already fixes the lemma: προλέγω / προεῖπον → ‘say beforehand, tell in advance / told before,’ with the rule that the πρό- prefix is genuinely open between temporal forewarning and prior utterance, and that ‘before’ must stay audible in English.

Two further points belong to this verse.

The tense is perfect, not aorist. προειρήκαμεν asserts a completed act of speaking whose result stands: the earlier word is still in force at the moment of writing. English present perfect carries this exactly. A simple past (‘as we said before’) severs the standing result and makes v. 9 a reminiscence rather than a re-enactment.

The reference is contested. Three readings are live:

The rendering must be compatible with all three, which ‘as we have said before’ is. Excluded: ‘as I have just said’ (adopts the first and deletes the plural) and ‘as we warned you when we were with you’ (adopts the second and supplies a setting).

Built onbdag · bdf

The plural προειρήκαμεν stands against the singular λέγω in the same sentence. Paul uses the first person plural freely in Galatians (1:8 ἡμεῖς … εὐηγγελισάμεθα; 2:5, 16), so the plural here cannot be pressed as decisive evidence for a named group; but neither may it be levelled.

The distinction is load-bearing in one direction at least: v. 10 immediately turns to Paul singly (ἀνθρώπους πείθω … Χριστοῦ δοῦλος), so the move from ‘we’ to ‘I’ in v. 9 is the hinge into the self-defence of vv. 10-24. English must preserve both persons as written.

ἄρτι is the sharper of the two Greek ‘now’ words: punctiliar, the present instant, as against the broader νῦν (Matt 26:53; John 13:37; 1 Cor 13:12). The glossary already fixes it as ‘just now / at this moment,’ with the requirement that it stay in its Greek word-order slot. Its recurrence at the head of v. 10 (Ἄρτι γὰρ ἀνθρώπους πείθω) binds the two verses.

πάλιν is governed by the standing entry: render ‘again,’ and where scope is open, leave it open. Here the three live senses of πάλιν — repetition of an act, resumption of earlier teaching, and additive ‘furthermore’ (1 Cor 12:21; Heb 1:5) — line up with the three construals of προειρήκαμεν above, and English ‘again’ is equally hospitable to all of them.

The accumulation ἄρτι πάλιν is Koine reinforcement, not redundancy to be trimmed (cf. πάλιν ἄνωθεν, Gal 4:9, already fixed on exactly this ground). Both adverbs are rendered.

Built onbdag

The decisive change from v. 8. There Paul wrote ἐὰν … εὐαγγελίζηται — third-class condition, aorist/present subjunctive, a case entertained as possible. Here he writes εἴ τις … εὐαγγελίζεται — first-class condition with the present indicative.

Per the standing first-class-condition policy, the protasis is entertained as real for the sake of the argument, and ‘since’ is excluded because it closes the rhetorical gap the conditional deliberately leaves. But the shift of construction between two adjacent verses is itself the evidence that Paul has moved from the constructed extreme case (himself, an angel) to the actual one.

The present indicative εὐαγγελίζεται is durative: not a single act of preaching but a campaign under way, matching the present participles of v. 7 (οἱ ταράσσοντες … θέλοντες) and the present μετατίθεσθε of v. 6.

The pronoun τις is unqualified. Paul names nobody, assigns no office, and specifies no number. Supplying ‘these agitators,’ ‘the Judaizers,’ or ‘anyone, whoever he may be’ writes in what the indefinite withholds (§4, §5). Note also that the accent distinguishing indefinite τις from interrogative τίς is editorial — the earliest majuscules are unaccented — though no one has seriously proposed the interrogative here.

Built onbdf · silva-becnt

The glossary fixes the gloss: proclaim the good news (literal tier), preach the gospel (smooth). The εὐαγγελ- root must stay visible so the link to εὐαγγέλιον (1:6, 7, 11) holds.

Galatians 1 uses the verb six times in six verses (1:8 twice, 1:9, 1:11, 1:16, 1:23), and the density is deliberate: Paul is contesting ownership of an activity, not merely of a doctrine. The verb is construed here with a double accusative (ὑμᾶς, the persons; the content carried by the παρʼ ὅ clause), a construction attested at Luke 3:18; Acts 8:25; 1 Cor 15:1-2.

Note what the verb does not do: it does not evaluate. Paul applies the same word to the rival message as to his own. The condemnation comes from παρʼ ὅ and from ἀνάθεμα, not from the verb — so the English must stay neutral. ‘Peddles a gospel,’ ‘promotes,’ ‘touts’ import a verdict the lexeme lacks (compare the policy already fixed for καταγγέλλω at Phil 1:17-18, where Paul applies one proclamation-verb to sincere and self-seeking preachers alike).

The deponent middle carries no reflexive force and must not be over-read.

Built onbdag · silva-becnt

παρά + accusative is the crux of both v. 8 and v. 9, and the two principal senses yield materially different charges.

The two are not far apart in effect — Paul’s point in vv. 6-7 is precisely that an addition constitutes a contradiction — but they are distinguishable, and no single English preposition covers both cleanly. ‘Beyond’ is retained in all tiers because it leaves the addition-reading primary while permitting the stronger one; ‘contrary to’ is recorded as the flagged alternative. ‘Other than’ is excluded: it adopts the contrary-to reading outright and, in English, suggests an alternative rather than a supplement.

The relative ὅ is neuter and its antecedent is unexpressed. It is not τὸ εὐαγγέλιον grammatically attracted (that would be ὅ by agreement anyway, but no antecedent stands in the clause); it is the content received, left as a bare neuter. The literal tier keeps ‘that which’ and supplies no noun.

Built onbdag · silva-becnt

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 glossary fixes παραλαμβάνω → ‘receive (as something handed on),’ the technical receiving-end counterpart of παραδίδωμι, standard in Paul for the transmission of a fixed deposit (1 Cor 11:23; 15:1, 3; 1 Thess 2:13; 4:1; 2 Thess 3:6; Col 2:6). ‘Were taught’ and ‘learned’ are excluded as collapsing it into μανθάνω and hiding the tradition-vocabulary.

Two things follow for this verse.

The subject is the Galatians. Verse 8 measured the rival message against Paul’s act of proclamation (ὃ εὐηγγελισάμεθα ὑμῖν); verse 9 measures it against their act of reception. The standard is not Paul’s authority in the abstract but a deposit now in their possession — which is why the preceding verses can hold them responsible for shifting (μετατίθεσθε, 1:6) and why vv. 11-12 then have to establish where the deposit came from.

The vocabulary concedes something to the opponents’ frame and then turns it. παράδοσις is the word Paul uses of his own ancestral traditions at 1:14 (τῶν πατρικῶν μου παραδόσεων), which he has just renounced as a credential. The same transmission-language is now claimed for the gospel, with the crucial difference that its origin is δι’ ἀποκαλύψεως Ἰησοῦ Χριστοῦ (1:12) rather than παρὰ ἀνθρώπου.

The aorist is complexive: the receiving taken as a single completed event, the founding mission as a whole.

Built onbdag · silva-becnt

ἀνάθεμα is the later-Greek by-form of ἀνάθημα (‘a thing set up, a votive offering’), and in the LXX it becomes the standard equivalent of Hebrew חֵרֶם — the thing devoted irrevocably to God and therefore removed from common use, which in the conquest narratives means destroyed (Lev 27:28-29; Deut 7:26; 13:18; Josh 6:17-18; 7:12; Zech 14:11). The object is not merely cursed in the sense of being spoken ill of; it is handed over.

NT occurrences: Rom 9:3 (Paul’s wish to be ἀνάθεμα for Israel’s sake); 1 Cor 12:3 (ἀνάθεμα Ἰησοῦς); 1 Cor 16:22; Acts 23:14 (ἀναθέματι ἀνεθεματίσαμεν, the cognate figure); and Gal 1:8-9. In every case the devotion is to God and the outcome is destruction or exclusion.

Rendering policy. The transliteration ‘anathema’ is retained in the literal tier, per §4: a cultic-legal term with no adequate English equivalent keeps its Greek form plus a glossary entry. ‘Accursed’ stands in the reading text, where English idiom requires an adjective and where the older tradition has fixed the word. Excluded from reading tiers:

• ‘eternally condemned’ — supplies a duration and a verdict the noun does not carry, and converts a devotion-formula into a doctrinal statement about final destiny.
• ‘damned’ — imports post-biblical infernal content (§9).
• ‘excommunicated’ / ‘expelled from the church’ — retrojects a later ecclesiastical procedure onto a first-century devotion-formula, and supplies an executing body (§9).
• ‘under God’s curse’ — supplies the agent the Greek withholds (see below).

Built onbdag · muraoka-lxx · silva-becnt

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

Third-person present imperative of εἰμί. Per the standing third-person-imperative policy (γνωσθήτω, Phil 4:5; αἰσχυνέσθω, 1 Pet 4:16; ἤτω, Jas 5:12), English has no third-person imperative and a ‘let’ periphrasis is unavoidable — but three constraints follow.

The third person must not become second. Paul does not address the agitators; he rules about them, to the Galatians. ‘You must treat such a person as accursed’ converts an impersonal pronouncement into an instruction to the congregation and thereby adopts the communal-exclusion reading above.

No agent may be supplied. The clause has no subject beyond the implied ‘he,’ and no executor. ‘Let God curse him’ and ‘God’s curse be on him’ both write in what the imperative withholds.

‘Let him be’ must not be heard as permission. Where modern English would read the periphrasis as withheld consent, reading tiers may use a stronger construction, with the shift recorded. The literal tier retains ‘let him be’ so the construction stays visible.

The present aspect is durative/stative — a standing condition, not an act performed at a moment.

Built onbdf

Two features of the verse exist only because it is a repetition, and both must survive translation.

ἀνάθεμα ἔστω is verbatim. The same two words close v. 8 and v. 9, unaltered. Per the corpus anti-variation discipline (already applied to παρακαλῶ … παρακαλῶ, Phil 4:2; ἐπιζητῶ … ἐπιζητῶ, Phil 4:17; ἀλλήλων across Gal 5), identical Greek receives identical English at both occurrences. Varying the second for elegance — ‘let him be accursed’ / ‘let such a one be condemned’ — destroys the only feature that makes v. 9 a repetition rather than a further point.

The escalation runs from the hypothetical to the actual. Verse 8: we, or an angel from heaven, subjunctive. Verse 9: anyone, indicative. Paul’s method is to establish the principle on the hardest imaginable case and then apply it to the case at hand, and the grammatical shift is how he does it. English must preserve ‘if anyone … is proclaiming’ as an indicative against v. 8’s ‘even if we … should proclaim.’

Chrysostom read the doubling as deliberate pedagogy: Paul repeats so that no one may suppose the first statement was spoken in heat. The observation is reception, not lexical evidence, but it registers that the repetition was felt as rhetorical from early on.

The verse is textually stable in its load-bearing elements. NA28 prints no variant in the apparatus requiring a decision, and the UBS committee did not treat the verse. Four minor points of variation in the manuscript tradition are worth recording:

• εὐαγγελίζεται / εὐαγγελίζηται. Part of the tradition reads the subjunctive, assimilating v. 9 to the ἐάν + subjunctive of v. 8. The indicative is the harder reading and is required by εἰ; the variant shows scribes levelling the two verses, which is precisely the direction to be avoided, since the mood-shift is the argument.
• παρʼ ὅ / παρʼ ὧν / παρὰ τοῦτο ὅ. Scattered minuscule variation around the relative, all of it simplification of a bare neuter with no expressed antecedent.
• Addition of a verb in the ὡς clause. A few witnesses expand προειρήκαμεν (e.g. with an object), filling the ellipsis — the same impulse visible in the versional tradition.
• Punctuation. The earliest majuscules are unpunctuated, so the placement of the colon before εἴ τις, and the question of whether ὡς προειρήκαμεν … λέγω forms one sentence with the conditional or a separate preface, is editorial judgment in every printed edition. This must be reported as editorial and never as manuscript evidence.

None of these affects the sense. The verse’s difficulty is exegetical, not text-critical.

Built onna28 · metzger-tc

Three early readings bear on the verse’s register and on the scope of the curse.

Chrysostom (Hom. in Gal. 1) treats the repetition as proof of deliberation: Paul says it twice so that the Galatians cannot dismiss the first as passion. He also presses the self-inclusion of v. 8 — Paul puts himself under the same sentence — as the thing that makes the curse credible rather than tyrannical.

Jerome (Comm. in Gal.) renders ἀνάθεμα with anathema transliterated rather than translated, which is a useful witness that the Latin tradition also found no adequate equivalent; the Vulgate’s anathema sit keeps the Greek word. He takes παρά as praeter, ‘beyond,’ which weighs for the addition-reading above.

Augustine (Expos. Gal.) emphasizes that the standard is the gospel as received and not the status of the preacher, drawing the conclusion that no apostolic authority can revise it — the same inference Paul draws in vv. 10-12.

None of this is lexical evidence (§11). It is recorded because it shows that the doubling, the self-inclusion, and the ‘beyond’ sense of παρά were all felt as the verse’s load-bearing features by readers much nearer the Greek than we are.

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; repr. Grand Rapids: Eerdmans)

Vulgate: sicut praediximus et nunc iterum dico: si quis vobis evangelizaverit praeter id quod accepistis, anathema sit. Three points. (i) praeter adopts the ‘beyond, in addition to’ sense of παρά, the reading argued above. (ii) anathema sit leaves the Greek noun untranslated — Latin had no equivalent either, which supports the §4 transliteration policy. (iii) evangelizaverit is a perfect subjunctive, which obscures the indicative/subjunctive contrast with v. 8 that the Greek maintains; Latin conditional syntax forced this, so it is not evidence about the Greek.

Peshitta renders the curse with a term of exclusion/separation, which leans toward the communal reading of ἀνάθεμα. This is one witness to how the formula was heard in a Syriac-speaking church, not evidence about Paul’s sense, and it should be recorded on the exclusion side of the contested question above rather than used to settle it.

LXX background is the decisive versional datum for ἀνάθεμα: the word’s content in biblical Greek is supplied by its use for חֵרֶם, not by classical ἀνάθημα ‘votive offering.’ See the lexical section.

Proposed — pending reviewJerome, Commentarii in epistulam Pauli apostoli ad Galatas, Corpus Christianorum Series Latina 77A (Turnhout: Brepols, 2006)