The tongue-paragraph (3:1-12) has run through three images of small things governing large ones — bit, rudder, spark — and then hardened into a verdict: no one among humans can tame it (v. 8). Verse 9 supplies the evidence, and it is not a catalogue of coarse speech. It is worship. The same organ that pronounces the blessing pronounces the curse, and James proves the tongue untamed not by pointing at its worst hour but at its best one.

Four things carry the argument.

The repeated phrase. ἐν αὐτῇ … καὶ ἐν αὐτῇ. James could have written the second clause without the prepositional phrase; the repetition is the whole point, and English must repeat it too.

The single article. τὸν κύριον καὶ πατέρα — one article over two titles, therefore one referent. This is not two objects of blessing but one, named twice, and the second name is what makes the curse monstrous: the God addressed as Father is the one whose likeness is on the people being cursed.

The Genesis allusion. καθʼ ὁμοίωσιν θεοῦ is Gen 1:26 LXX vocabulary, not Gen 1:27’s εἰκών. James is not making a general appeal to human dignity; he is quoting the creation mandate, and he applies it without qualification — τοὺς ἀνθρώπους, people as such, not fellow believers.

The perfect participle. γεγονότας, not an aorist. The having-been-made stands in force at the moment of the cursing. Whatever has happened to humanity since Genesis 1, the likeness has not lapsed, and that is why the curse lands on something that is still there.

The verb εὐλογοῦμεν is first plural. James is inside the indictment, not above it.

Proposed — pending reviewRalph P. Martin, James, Word Biblical Commentary 48 (Waco: Word Books, 1988)

ἐν + dative with an organ of the body as the object is instrumental: the tongue is what the blessing and the cursing are done with. Compare ἐν τῇ γλώσσῃ of speech elsewhere, and the parallel construction at v. 10 with the preposition ἐκ (ἐκ τοῦ αὐτοῦ στόματος) — the shift from ἐν (means) to ἐκ (source) across two verses is deliberate and both prepositions must be rendered distinctly.

The corpus policy on ἐν + dat keeps ‘in’ where sphere and means are both live and exegetically consequential. Here they are not: a tongue is not a sphere within which blessing occurs, and ‘in it we bless’ is unintelligible English. ‘With it’ is adopted, and the instrumental classification is recorded rather than hedged.

The antecedent of αὐτῇ is γλῶσσα (v. 8), feminine — which is why the pronoun is feminine, not because of anything about the noun’s referent.

The phrase is written twice. Greek permits the second clause to run on the first phrase (καὶ καταρώμεθα would have sufficed), and James declines the economy. Per §10 this is rhetorical data and the repetition is preserved in English at both points, in the same slot and with the same words; the standing corpus policy on repeated verbs and repeated pronouns (Phil 4:2, 4:17; 1 Pet 4:12) applies directly.

εὐλογέω in the LXX renders בָּרַךְ and covers both directions of the covenant relation: God blesses his people (Gen 12:2-3), and his people bless him (Ps 33:2 LXX; Ps 102:1 LXX εὐλόγει, ἡ ψυχή μου, τὸν κύριον). With God as object the sense is praise, ascribe blessedness to — not conferring a benefit on him, which would be nonsense.

The corpus already keeps εὐχαριστέω (‘give thanks’), ἐξομολογέομαι (‘confess / acclaim’), and εὐλογέω (‘bless’) lexically distinct, precisely because the LXX distributes Hebrew הודה and ברך across them and the reader should be able to see which Greek verb stands where. That policy governs here: render ‘bless’, not ‘praise’ (which would collapse it toward αἰνέω/ὑμνέω) and not ‘give thanks’.

The present tense is durative and the first person plural is inclusive — the standing practice of the assembly, James among them. This is liturgical vocabulary. The berakah formula (εὐλογητὸς ὁ θεός / בָּרוּךְ אַתָּה) was the ordinary shape of Jewish prayer, and the noun εὐλογία at v. 10 confirms that James has actual doxology in view, not vague well-wishing. The offense he names is thus located inside worship, not outside it.

One article governs both nouns. Per the standing corpus policy on article–noun–καί–noun (fixed at Phil 4:20, τῷ θεῷ καὶ πατρὶ ἡμῶν), a single article over two coordinated substantives normally marks one referent, and English must not repeat the article: ‘the Lord and the Father’ manufactures two addressees.

The collocation κύριος καὶ πατήρ is unusual. Elsewhere James pairs the Father-title with other elements (ὁ πατὴρ τῶν φώτων, 1:17; τῷ θεῷ καὶ πατρί, 1:27), and the NT norm is θεὸς καὶ πατήρ (Rom 15:6; 2 Cor 1:3; Eph 1:3; Col 1:3; 1 Pet 1:3). Sirach offers the nearest parallel to the wording here: κύριε πάτερ καὶ δέσποτα ζωῆς μου (Sir 23:1), κύριε πάτερ καὶ θεὲ ζωῆς μου (Sir 23:4).

Per §1, κύριος is rendered ‘Lord’ and never converted to YHWH, even where the LXX substrate is operative. Here the substrate is operative: the doxological formula being echoed is the Psalter’s εὐλογεῖν τὸν κύριον, where κύριος stands for the Tetragrammaton. That is footnoted, not restored — the Greek author’s lexical choice is data to preserve.

The referent is God the Father, not Christ. James does apply κύριος to Christ (1:1; 2:1), but the appended πατέρα under a single article fixes the reference here, and the following clause about God’s likeness in humanity confirms it.

καταράομαι is middle-only, and the middle is lexical rather than reflexive — no self-cursing is implied. In Koine it takes the accusative (as here and Mark 11:21; Luke 6:28) where classical Greek preferred the dative. The LXX field is dense and formal: Balaam is hired to curse Israel (Num 22-24 passim), Shimei curses David (2 Kgdms 16:5-13), and the Torah forbids cursing the deaf, a ruler, or one’s parents (Lev 19:14; Exod 22:27).

The register matters. A κατάρα in the ancient world is not a rude word but a speech-act invoking harm — a formal imprecation, the exact negative counterpart of a εὐλογία. That is why the two nouns pair naturally at v. 10 and why the LXX sets them together as the covenant alternatives (Deut 11:26-28; 30:1, 19). Render ‘curse’ and do not weaken to ‘insult’, ‘abuse’, ‘slander’, or ‘speak evil of’: each substitutes a general category of bad speech for a specific verbal act, and each destroys the symmetry with εὐλογέω on which the whole verse turns.

βλασφημέω (‘slander / blaspheme’, already fixed) and λοιδορέω were both available to James and are not used. The choice of the cultic-covenantal verb is what makes the clause an accusation of liturgical self-contradiction rather than a complaint about temper.

Present tense, durative, first person plural — the same standing practice as the blessing.

Proposed — pending reviewJ. H. Moulton and G. Milligan, The Vocabulary of the Greek Testament Illustrated from the Papyri and Other Non-Literary Sources (London: Hodder & Stoughton, 1930)

Articular masculine plural of ἄνθρωπος. Two questions are live and neither should be closed by the rendering.

Referent. Is the article generic (‘people as such’) or anaphoric to a group already in view (fellow members, or the rich of 2:6-7, or the teachers of 3:1)? The parallel with the second clause’s Genesis-wide allusion, the absence of any qualifier, and James’s use of ἀδελφοί when he means members all favor the generic reading: the curse falls on human beings, and the ground of the charge is their creation, not their membership. But James is writing to an assembly with named internal fractures, and an anaphoric reading is not excluded. The reading text keeps the bare ‘people’ and adds no qualifier — ‘our fellow men’, ‘others’, ‘fellow believers’ each adjudicate.

Gender. Per §3 the masculine plural is the Greek default for a referentially inclusive statement, and the referent here is unrestricted — the ὁμοίωσις of Gen 1:26-27 is explicitly conferred on ἄρσεν καὶ θῆλυ. The reading tiers render ‘people’; the literal tier retains ‘men’ to track the grammatical form. James himself shows he marks the feminine when he wants it (ἀδελφὸς ἢ ἀδελφή, 2:15), which is internal warrant for treating the masculine plural here as inclusive rather than restrictive.

The standing corpus entry on generic ἄνθρωπος (πᾶσιν ἀνθρώποις → ‘people’, literal tier ‘men’) governs directly.

ὁμοίωσις is the -σις action-noun of ὁμοιόω, rare in the NT (only here) and drawn straight from Gen 1:26 LXX: ποιήσωμεν ἄνθρωπον κατʼ εἰκόνα ἡμετέραν καὶ καθʼ ὁμοίωσιν. The LXX pair renders Hebrew צֶלֶם (εἰκών) and דְּמוּת (ὁμοίωσις), and Gen 5:1 LXX has κατʼ εἰκόνα θεοῦ ἐποίησεν αὐτόν while Gen 5:3 has κατὰ τὴν ἰδέαν αὐτοῦ καὶ κατὰ τὴν εἰκόνα αὐτοῦ.

That James chose ὁμοίωσις rather than εἰκών is worth recording rather than smoothing. εἰκών is the word Paul and the wider NT use for the image (1 Cor 11:7; 2 Cor 4:4; Col 1:15; 3:10), and it is available to any writer echoing Genesis. James’s noun is the second member of the Genesis pair, and the -σις formation is formally open between the act of likening and the resulting likeness.

Render ‘likeness’. The corpus keeps ὁμοίωμα as ‘likeness’ (Phil 2:7), and no English distinction between ὁμοίωμα and ὁμοίωσις is worth manufacturing — the two are recorded in the parsing table and differentiated in note, not in the rendering. Excluded: ‘image’ (that is εἰκών, and using it here erases James’s choice), ‘in God’s own image’ (imports the other Genesis noun and a possessive the Greek lacks), and ‘as a copy of God’ (misconstrues the noun as a product-noun).

Proposed — pending reviewDale C. Allison Jr., A Critical and Exegetical Commentary on the Epistle of James, International Critical Commentary (London: Bloomsbury T&T Clark, 2013)

κατά + accusative in the norm/pattern sense, per the standing corpus policy (which already covers disposition-nouns, credential-lists, concrete targets, quantity-nouns, and personal names). Here the accusative names the pattern according to which the coming-to-be occurred. Render ‘according to’ in literal and study tiers; ‘in the likeness of’ is permitted in smooth registers, though it drifts toward a locative that κατά does not carry.

θεοῦ is anarthrous. Per Apollonius’ canon the genitive of a proper-noun-like θεός is regularly anarthrous without loss of definiteness, and English supplies the article or a possessive (‘God’s likeness’). No indefiniteness may be read in — ‘a god’s likeness’ is excluded, and the capital is warranted by the referent.

The wording is not a verbatim citation: Gen 1:26 LXX has καθʼ ὁμοίωσιν without θεοῦ (the genitive would be redundant in first-person divine speech), and Gen 1:27 / 5:1 supply κατʼ εἰκόνα θεοῦ. James has fused the Genesis vocabulary rather than quoted a verse, which is why no citation formula stands. Per §7 the allusion is identified and footnoted; per §4 the absence of a citation formula is itself recorded rather than repaired.

The second article converts the whole prepositional phrase plus participle into an attributive modifier of τοὺς ἀνθρώπους — the fuller of the two attributive constructions, adding the modifier as a separate weighted unit rather than embedding it. The standing corpus policy on article + prep phrase in attributive position (Phil 4:21; 1 Pet 5:2; Jas 1:9) applies: retain the doubled article in the literal tier so the construction is visible (‘the men, the ones according to God’s likeness having come to be’), and resolve to an English relative clause in reading tiers.

The modifier is restrictive in form but not in effect: it does not select a subset of humans who happen to bear the likeness, since on the Genesis premise there is no other kind. Its function is to make the ground of the indictment explicit at the very end of the sentence, in the emphatic final slot. English should therefore preserve the late arrival of the participle — the sentence must land on γεγονότας, as the Greek does. A dash before the relative clause in reading tiers reproduces the effect without converting a restrictive modifier into a non-restrictive aside.

The word order is also worth noting: the prepositional phrase is fronted inside the attributive unit, before the participle it modifies. Greek permits this freely and no emphasis need be read into it beyond the general topicalization of the pattern-phrase.

Perfect active participle of γίνομαι. Per the standing corpus entry, γίνομαι is rendered ‘come to be / become’, preserving the contrast with εἰμί and ὑπάρχω — a verb of entry into a state, not of bare being.

The tense is the theological hinge of the verse and must not be levelled.

• Perfect, not aorist: a completed act with a standing result. The making is done; the likeness holds. Some witnesses read the aorist γενομένους, which converts a state in force into a past event and thereby weakens the charge — the curse would then fall on people who were once made in God’s likeness, and the question of whether they still bear it would be left open. English present perfect (‘who have come to be’) tracks the Greek and must be used in every tier; ‘who were made’ is excluded from the reading text.
• Active, not passive: γίνομαι has no passive, so no divine-passive construction is available and none should be imported. The rendering must not supply an agent (‘whom God made’), even though God is plainly the maker. The corpus policy on unexpressed agents applies: what Greek withholds, English withholds. Note that the Vulgate’s facti sunt and much of the English tradition (‘who are made in the similitude of God’) supply exactly this, and the direction of that smoothing is to be avoided.

The perfect also answers the perfect δεδάμασται two verses earlier (v. 7): every kind of creature has been tamed, and the human being has come to be in God’s likeness. Both perfects name standing states, and the tongue stands over against both.

The verse is two clauses of identical shape joined by bare καί:

• ἐν αὐτῇ | εὐλογοῦμεν | τὸν κύριον καὶ πατέρα
• ἐν αὐτῇ | καταρώμεθα | τοὺς ἀνθρώπους τοὺς καθʼ ὁμοίωσιν θεοῦ γεγονότας

The connective is καί, not ἀλλά or δέ, and per the standing corpus policy on coordinating καί it must be rendered ‘and’. Rendering ‘but’ or ‘yet’ manufactures an adversative the particle withholds; the horror of the sentence is precisely that the two acts are laid side by side as equally routine practice, and the verdict on them is postponed to v. 10 (οὐ χρή). §6 forbids shading toward comfort, and it equally forbids shading toward a rebuke the syntax has not yet delivered.

The asymmetry is in the objects, not the structure. Clause one has a two-word title under one article; clause two has a noun plus a substantial attributive modifier. The second clause is deliberately heavier, and it is the modifier that closes the circuit: the God blessed is the Father, and the people cursed bear his likeness, so the two objects are not independent. Cursing the second is an act against the first.

This is what proves the thesis of v. 8. An untamed thing is not one that only misbehaves; it is one that cannot be relied on to be consistent even in its best exercise.

Two variants, one of substance.

τὸν κύριον καὶ πατέρα (𝔓20 ℵ A B C P Ψ 33 81 and the bulk of the early evidence) against τὸν θεὸν καὶ πατέρα (the Byzantine tradition, and reflected in the Latin Deum et Patrem). NA28 prints κύριον without brackets. The internal evidence runs with the external: θεὸν καὶ πατήρ is the overwhelmingly common NT formula (Rom 15:6; 2 Cor 1:3; Eph 5:20; Col 1:3; 3:17; 1 Pet 1:3), and James himself has τῷ θεῷ καὶ πατρί at 1:27, so a scribe copying κύριον καὶ πατέρα had every reason to normalize and none to innovate. κύριον is the harder reading and is original.

The substitution is not doctrinally trivial. With κύριον the clause echoes the Psalter’s εὐλογεῖν τὸν κύριον, where κύριος renders the Tetragrammaton, and the sentence sits inside the LXX doxological idiom. With θεόν it assimilates to the Pauline-Petrine benediction formula. The reading adopted preserves the LXX register — which is exactly what makes the second half of the verse a Genesis argument rather than a general appeal.

γεγονότας (perfect) against γενομένους (aorist), read by a scattering of witnesses. The perfect is both better attested and the harder reading, and the aorist is the predictable simplification. See γεγονότας above for why the tense matters.

Vulgate: In ipsa benedicimus Dominum et Patrem, et in ipsa maledicimus homines, qui ad similitudinem Dei facti sunt. Three things to note. Dominum attests the harder Greek reading against the Byzantine θεόν. In ipsa renders ἐν αὐτῇ locatively, which is a defensible Latin rendering of a Greek instrumental but has pulled the English tradition toward ‘therewith / by it’ only unevenly. And facti sunt converts an active perfect participle into a passive, supplying the agent-implication that the Greek withholds — the direction the corpus policy on unexpressed agents rules out.

Peshitta: renders the paired objects with the standard Syriac terms for Lord and Father and keeps the two clauses parallel; the Genesis allusion is preserved with the Syriac counterpart of דְּמוּת, which brings the Hebrew substrate of ὁμοίωσις back into view. This is a useful check: the Syriac translator plainly heard Gen 1:26 behind the clause.

Neither version differentiates ὁμοίωσις from εἰκών lexically — similitudo would serve for either — which is one reason the later patristic distinction between image and likeness had freer play in Latin and Greek theological usage than the translation evidence warrants.

The early reading of the verse consistently treats the second clause as the argument, not as ornament.

Chrysostom, commenting on the honor due to what bears God’s likeness, presses the point that reverence for the Creator and contempt for the creature made after him cannot coexist in the same mouth — the offense is against God, not merely against the neighbor. This is the reading the syntax supports: the attributive modifier exists to make the curse an act against the one blessed.

Bede, in his commentary on the Catholic Epistles, takes the verse as an indictment of the assembly’s actual speech-practice rather than a hypothetical, and reads τοὺς ἀνθρώπους without restriction — the likeness is the ground, so the scope is human, not congregational. He also links the verse to the Matt 5 prohibition of contemptuous speech.

The image/likeness distinction (εἰκών retained, ὁμοίωσις lost and restored) becomes a major Greek theological instrument from Irenaeus onward. It is worth recording as reception (§11) and worth naming as a genuine question in the history of interpretation — but per §9 it supplies no lexical evidence for rendering ὁμοίωσις here, since the LXX translators’ own usage at Gen 1:26 and 5:1-3 does not sustain a stable distinction. The rendering stays neutral.

What the patristic reading gets right, and what the modern English tradition sometimes loses, is that this verse is about worship. It is not a general complaint about bad language; it is the observation that the blessing and the curse issue from the same organ in the same assembly.

Proposed — pending reviewBede the Venerable, Commentary on the Seven Catholic Epistles, trans. David Hurst, Cistercian Studies 82 (Kalamazoo: Cistercian Publications, 1985)

Two Hebrew backgrounds converge.

Gen 1:26 — נַעֲשֶׂה אָדָם בְּצַלְמֵנוּ כִּדְמוּתֵנוּ, LXX κατʼ εἰκόνα ἡμετέραν καὶ καθʼ ὁμοίωσιν. James’s ὁμοίωσις is the Greek of דְּמוּת. Note that Gen 9:6 grounds the prohibition of bloodshed on the same premise (בְּצֶלֶם אֱלֹהִים עָשָׂה אֶת־הָאָדָם), which is the nearest OT analogue to James’s move: an act against a human being is answerable because of what the human being is.

The berakah. בָּרוּךְ אַתָּה יְהוָה is the standing shape of Jewish prayer, rendered in LXX with εὐλογητὸς κύριος / εὐλογέω τὸν κύριον (Ps 33:2; 102:1, 22; 144:1 LXX). Per §1 the κύριος of the Greek text is rendered ‘Lord’ and the underlying Tetragrammaton is footnoted here rather than restored — but the substrate is what gives the clause its liturgical weight. James is describing the assembly at prayer.

The blessing/curse pair is covenant vocabulary (Deut 11:26-28; 27-28; 30:1, 19; Josh 8:34), where εὐλογία and κατάρα are the two alternatives set before Israel. That pairing is why the two nouns can stand together at Jas 3:10 without further explanation, and why a tongue that produces both is not merely inconsistent but is doing something covenantally incoherent.

Proposed — pending reviewPatrick J. Hartin, James, Sacra Pagina 14 (Collegeville: Liturgical Press, 2003)

Verse 8 asserts that no human can tame the tongue. Verse 9 is the evidence, and the choice of evidence is the argument.

James does not prove the tongue untamed by pointing to obscenity, or gossip, or the boasting of v. 5. He points to the moment when the tongue is doing what it was made for. The blessing is real — εὐλογοῦμεν is not sarcastic, and no ancient reader would have heard it so. The problem is that the same organ, in the same community, in the same week, does the other thing too, and both are habitual (present tense, both verbs).

The attributive modifier is what makes the two acts one problem rather than two. Cursing a human being is not an offense in a separate category from blessing God; it is directed at something that stands in a determinate relation to the God being blessed, and the perfect participle insists that the relation holds now. If the likeness had lapsed, there would be an escape from the charge. It has not.

Note also the first person plural. James, who has just warned against many becoming teachers because teachers face greater judgment (3:1), and who has just said no one can tame this thing, puts himself in the indictment. That is consistent: the paragraph’s logic is not that some members have a speech problem but that the tongue as such is beyond human mastery — which is why the paragraph ends not with a technique but with the flat οὐ χρή of v. 10 and the natural-impossibility images of vv. 11-12.

Proposed — pending reviewSophie Laws, A Commentary on the Epistle of James, Black's New Testament Commentaries (London: A. & C. Black, 1980)