Paul breaks off his itinerary to swear. Having given a tightly controlled account of his movements (vv. 17-19) — no Jerusalem, then Arabia, then Damascus, then three years later fifteen days with Cephas and no other apostle but James — he stops and puts the whole of it under oath.

Three things carry the verse:

1. It is grammatically broken. There is no verb of swearing or affirming. Greek gives a fronted relative clause (“the things I write to you”), a presentative particle (“look!”), a prepositional phrase (“before God”), and then a ὅτι clause (“that I am not lying”). The governing verb is nowhere written. That is not a copyist’s loss; it is the shape of an oath blurted out. English has to supply something, and the apparatus here keeps the supplement visible.

2. ἐνώπιον τοῦ θεοῦ is courtroom language, not piety. In the LXX the phrase marks the presence in which a statement acquires legal weight. Paul is not saying his conscience is clear; he is naming the one before whom perjury would be committed.

3. The oath’s scope is contested. Does ἃ γράφω cover only v. 19 (the James claim), only the itinerary of vv. 15-19, or the whole autobiographical argument from v. 11? The relative has no stated antecedent, and the Greek does not settle it.

The pastoral logic is plain enough: in Galatia a competing account of Paul’s relation to Jerusalem is in circulation (cf. v. 1, οὐκ ἀπʼ ἀνθρώπων), and Paul’s own narrative cannot be verified by his readers. An oath is what a man offers when his testimony is all the evidence there is.

The construction is an anacoluthon: the sentence starts on one track and never completes it.

• ἃ δὲ γράφω ὑμῖν — a fronted relative clause. Standard Greek topicalization (casus pendens), and in the corpus we normally expect a resumptive pronoun to pick it up (as at Phil 3:7, ἅτινα … ταῦτα). Here nothing resumes it. The accusative ἅ is left governed by γράφω alone, hanging as an unattached topic.
• ἰδού — a frozen particle, not a verb that can take the ὅτι clause as its object.
• ἐνώπιον τοῦ θεοῦ — a prepositional phrase with nothing to modify.
• ὅτι οὐ ψεύδομαι — a content clause with no verb of speech, swearing, or testifying to depend on.

The usual grammatical descriptions are (i) ellipsis of a verb such as λέγω, μαρτυρῶ, or ὄμνυμι, or (ii) that ἰδού plus ἐνώπιον τοῦ θεοῦ has itself become a sufficient oath-formula and no verb was ever felt to be needed. The second is attractive: Rom 9:1 (ἀλήθειαν λέγω ἐν Χριστῷ, οὐ ψεύδομαι) and 2 Cor 11:31 (ὁ θεὸς … οἶδεν ὅτι οὐ ψεύδομαι) show Paul with a stated verb, which is precisely what makes its absence here look like compression under pressure rather than idiom.

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

The relative has no expressed antecedent, and the referent of the sworn material is genuinely open.

The apparatus does not resolve this. English must therefore leave the relative unspecified: “what I am writing to you” rather than “this that I have just told you” (narrow) or “everything I write” (broad). Both narrowings are available to a reader of the bare relative; neither may be written into the text.

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

ἐνώπιον is an improper preposition formed from ἐν + ὤψ (“face”), effectively absent from classical Attic prose and overwhelmingly a Septuagintalism: it is the standard LXX rendering of לִפְנֵי.

The decisive point is that in LXX usage the phrase is not locative decoration. It marks the presence in which an act acquires covenantal or legal weight — Gen 6:8 (Noah found favour ἐνώπιον κυρίου), 1 Kgdms 1:15 (Hannah pouring out her soul), 2 Chr 33:12. In oath and testimony contexts specifically it names the witness before whom a statement is made, which is exactly the function here.

Per the standing glossary entry: render “before” in reading tiers, “in the sight of” in declared liturgical register, and keep the face-etymology in the note. Do not substitute “in God’s presence,” which drifts back toward the locative, and do not render “as God is my witness,” which supplies the μάρτυς-vocabulary Paul uses elsewhere (Rom 1:9; Phil 1:8) but not here.

Note also the article: τοῦ θεοῦ is articular, against anarthrous θεοῦ πατρός in v. 1 and v. 3. Render “God” with the article’s definiteness carried by English usage.

Middle-only (deponent) and intransitive; the middle carries no reflexive force and must not be over-read.

Per the standing entry: ψεύδομαι is a verb of speech, not a character-state. It must stay lexically distinct from πλανάω (“be led astray”), ἀπατάω (“deceive”), φρεναπατάω (“deceive the mind of”), and παραλογίζομαι (“cheat in the reckoning”) — all four fixed elsewhere in this corpus. Excluded: “be false” (not a speech-verb) and “be a liar” (the direction of the Vulgate’s non mentior is right, but converting the verb into a predicate noun turns an act into a disposition).

Two features of the form:

• Present tense, durative. Not “I did not lie” about a past act of composition but “I am not lying” — the writing and the truth-telling are coextensive. Rom 9:1 and 1 Tim 2:7 use the identical present.
• οὐ, not μή. With the indicative, οὐ negates as asserted fact. The negation is flat and categorical; nothing hedged is available in the construction.

Parallel Pauline formulations are worth setting side by side, because they show how compressed Gal 1:20 is:

A frozen aorist middle imperative of εἶδον functioning as a presentative particle. Per the standing entry (fixed at Jas 3:4): it does not agree with the number of its addressee and frequently governs a nominative rather than an accusative, which is why it cannot be treated as a live verb governing the ὅτι clause.

It must always be rendered — never dropped as filler — because it marks a discourse hinge: the point at which Paul stops narrating and turns to the reader. Here it is doing precisely that, breaking into the middle of a relative clause.

“Look!” in modern registers; “behold” retained in declared liturgical register only. Keep distinct in the parsing table from the deictic ἴδε (Gal 5:2) and from the genuine plural imperative Ἴδετε (Gal 6:11), so that the reader can see which of the three forms stands where — Galatians uses all three.

Postpositive and here transitional rather than adversative. Paul is not contradicting v. 19; he is stepping out of the narrative to certify it. Per §10 the particle is rendered and never dropped: “now” in reading tiers.

Rendering “but” would manufacture an antithesis between the James claim and the oath, which is the opposite of their relation — the oath underwrites the claim. The corpus already forbids manufacturing antitheses from ascensive δέ (Phil 2:8) and continuative δέ (Jas 1:15, 2:25); the same discipline applies here.

Present indicative, and the tense is load-bearing in two ways.

First, it is self-referential: the verb names the act in progress, as at Gal 6:11 (ἔγραψα, where the aorist raises a different question) and 1 Cor 4:14; 14:37; 2 Cor 1:13. The reader is being pointed at the document in his hands.

Second, it is durative, which bears on the scope question above: “what I am writing” sits more naturally with the material presently under the pen than with the completed letter. This is evidence, not proof — the epistolary present can look back over a whole composition.

Do not render with a perfect (“what I have written”), which closes the act and tips the scope question backward, and do not render with a future.

The verse is textually stable in the critical tradition; NA28 records no variant requiring decision, and the Greek is uniform across the principal uncials.

Two matters worth recording, neither of which is a textual variant:

1. The punctuation is editorial. The earliest majuscules are unpunctuated. Whether a comma, a dash, or a colon stands after ὑμῖν, and whether ὅτι opens direct or indirect discourse, is the judgment of modern editors and not manuscript evidence. This is the same discipline applied in this corpus at Gal 5:11 (ἄρα/ἆρα), Gal 6:15 (τι/τί), Jas 1:19 (Ἴστε), and 1 Pet 5:8 (τινα/τίνα). The apparatus follows NA28’s punctuation in the reading text and states that it is editorial.

2. The absence of a governing verb is original. No witness supplies λέγω, μαρτυρῶ, or ὄμνυμι. Had the ellipsis been felt as a defect by copyists, we would expect exactly such a supplement — the scribal tradition fills gaps readily elsewhere in this corpus (τοῦ νόμου at Jas 4:11; αὐτό at Jas 4:17; ἐκκλησία at 1 Pet 5:13). Its absence here is good evidence that the compressed oath-formula was intelligible as it stood.

Paul elsewhere uses a cluster of asseveration-formulae: μάρτυς μου ὁ θεός (Rom 1:9; Phil 1:8), ὁ θεὸς οἶδεν (2 Cor 11:11), μαρτύρομαι (Gal 5:3), and the compressed form here.

The apparatus does not adjudicate. What the rendering must not do is settle it: “I swear before God” adopts the first reading and imports the lexeme Paul declined; “I can honestly say” adopts the second and drains the verse.

Per §6, the governing conviction of this project is declared, and the honest report is that this is one of the places where a declared tradition does not relieve the exegetical pressure. The text is hard here, and the translation should be hard too.

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

The Greek commentators read the verse as Paul’s response to a specific accusation rather than as general self-defence.

Chrysostom, on Galatians, takes the oath to be occasioned by the charge that Paul was a dependent of the Jerusalem apostles — that his gospel was received and could be audited. Chrysostom’s observation is that Paul swears precisely where his account cannot be checked: the Galatians were not in Damascus or Arabia and could not verify fifteen days in Jerusalem. He also notes the delicacy of the position — that a man forced to describe his own career risks appearing to boast, and that the oath functions to transfer the burden from Paul’s credibility to God’s witness.

Jerome, in his commentary on Galatians, treats the construction as elliptical and supplies a verb of affirmation, and he raises the swearing question explicitly, resolving it by distinguishing a solemn appeal to God from the oath-taking prohibited to the perfect.

Both readings are reception and are recorded as such (§11): they show how the verse was heard, not what the Greek requires.

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; reprint Grand Rapids: Eerdmans) · Jerome, Commentary on Galatians, trans. Andrew Cain, Fathers of the Church 121 (Washington, DC: Catholic University of America Press, 2010)