This verse opens the letter’s most disputed paragraph, and almost every English rendering of it decides more than the Greek does.

Four things are worth seeing before the details:

1. The whole verse is two questions, not a statement. James asserts nothing here. He asks what benefit accrues (a nominal question with no verb written), and then asks whether this faith can save — with the particle μή, which telegraphs the expected answer ‘no’ without stating it. Rendering the second question as a declaration (‘such faith cannot save’) reaches James’s conclusion by a shortcut he deliberately declines.
2. The condition is hypothetical and the subject is indefinite. ἐάν + subjunctive presents a live possibility, and τις names nobody. James is building a test case, not indicting a party in the assembly.
3. The pivotal word is λέγῃ. The man says he has faith. James does not concede that he has it and then deny its efficacy; he reports a claim and asks what the claim is worth. Every later step in the paragraph (vv. 18-20) is an interrogation of the claim.
4. The article on ἡ πίστις is anaphoric. Greek marks the second πίστις as definite where the first was bare. James asks about the faith just described — workless, self-asserted — not about faith as such. This single article is the most load-bearing morpheme in the verse, and the traditional bare ‘can faith save him?’ obscures it.

Both πίστις and ἔργα stand anarthrous and fronted before their verbs, set in exact parallel; and both nouns are Pauline flashpoints. The apparatus below keeps the terms lexically constant with the rest of the corpus and flags rather than settles the relation to Paul.

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

ὄφελος occurs three times in the NT: here, at Jas 2:16 (where James repeats the question verbatim as the punchline of the illustration), and at 1 Cor 15:32 (τί μοι τὸ ὄφελος). It is a common enough word in classical and Hellenistic prose for advantage or profit accruing to someone, and the LXX has it at Job 15:3.

The construction is verbless: Τί ὄφελος with no ἐστίν. Per the corpus policy on elided copulas, the supplement is bracketed in the literal tier. The neuter interrogative is the predicate.

Register matters. ‘What good is it?’ is the natural modern equivalent and is used in the reading text. ‘What does it profit?’ preserves the commercial-advantage colouring the noun carries and is retained in the liturgical tier — the Vulgate’s quid proderit runs the same way. What must be avoided is any rendering that supplies a beneficiary the Greek withholds: 1 Cor 15:32 has the dative μοι and James does not. Whether the unstated benefit is to the claimant, to the poor of vv. 15-16, or to no one in particular is left open, and the repetition at v. 16 — where the destitute brother is plainly the one who gains nothing — suggests James wanted the question ambiguous on first hearing.

Note also that the question is not rhetorical in the same way as the second question in the verse. It has no μή and no οὐ; it genuinely asks, and vv. 15-17 answer it by illustration.

The plain form of James’s characteristic address (against the intensified ἀδελφοί μου ἀγαπητοί at 1:16, 1:19, and — decisively for this paragraph — 2:5). Per standing policy the possessive μου is rendered without exception: it places the writer inside the family he is pressing.

The vocative here is doing structural work. It marks the seam between the law-and-mercy argument of 2:8-13 and the faith-and-works argument of 2:14-26, exactly as the vocatives at 1:2, 1:16, 1:19, 2:1, and 2:5 mark earlier seams. Anyone reading James as a loose chain of maxims has to reckon with how regularly these addresses fall at argumentative joints.

Per §3 the masculine plural stands as written. That the referent is mixed is settled internally by ἀδελφὸς ἢ ἀδελφή at v. 15, one verse below — James marks the feminine when he chooses to, which is precisely why the masculine plural here may not be read as male-restrictive or expanded in the reading text.

Third-class condition: ἐάν + present subjunctive, presenting the case as a live possibility rather than as an accomplished fact. Per standing policy this is neither ‘when’ (which asserts the event) nor ‘though’ (which concedes it).

Two features of λέγῃ are load-bearing.

First, the present aspect. The saying is durative — a position the man holds and maintains, not a single utterance. ‘Claims’ in the reading text carries this; ‘says’ is retained in the literal tier.

Second, and more importantly, the verb is λέγῃ and not ἔχῃ. James does not write ἐὰν πίστιν ἔχῃ τις, ἔργα δὲ μὴ ἔχῃ. He writes that the man says he has faith. The construction is an infinitive of indirect discourse (ἔχειν) with the subject accusative unexpressed and recoverable from τις.

This is the hinge of the whole paragraph and it is very often lost. Read with λέγῃ in view, James is not conceding a real faith and then denying its power; he is examining a claim. That is exactly how the argument proceeds: v. 18 demands a demonstration (δεῖξόν μοι), v. 19 tests the claim’s content against what demons also manage, v. 20 calls the claim ἀργή.

τις is studiedly indefinite. No party in the assembly is named, in contrast with the concrete ὑμεῖς of vv. 6-7. English must keep ‘someone’ and must not specify (‘a man who professes to be a believer’, ‘one of you’).

Proposed — pending reviewPeter H. Davids, The Epistle of James: A Commentary on the Greek Text, New International Greek Testament Commentary (Grand Rapids: Eerdmans, 1982)

The second limb of the same protasis: James does not repeat ἐάν, and the subjunctive ἔχῃ continues under the single conditional particle. English must not split this into two sentences or introduce a second ‘if’.

δέ is adversative here — the corpus distinguishes this from the ascensive use (Phil 2:8), the continuative use (Jas 1:15), and the correlative μέν … δέ. Render ‘but’. It is never dropped (§10).

μή is the grammatically obligatory negative in a subjunctive protasis and carries no independent semantic weight, exactly as with μή + participle. Do not read emphasis out of its presence.

The word order is the point. Both πίστιν and ἔργα stand anarthrous and before their verbs:

πίστιν λέγῃ τις ἔχειν — ἔργα δὲ μὴ ἔχῃ

Two fronted, articleless nouns in exact symmetry. The fronting is topicalizing; the absence of articles makes both qualitative — faith of a certain kind, works of a certain kind, not two defined entities. Preserving the early position of both nouns in English is possible and should be attempted, since the parallelism is what makes the paragraph’s question a single question rather than two.

πίστις is rendered faith throughout, per the corpus policy fixing one gloss for the noun and its cognates. What varies is not the English word but what James is talking about, and that is carried by the article, the context, and the notes — never by shifting the gloss.

The diachronic check (§9) matters here. First-century πίστις spans trust reposed in a person, the trustworthiness that makes one reliable, and — by a metonymy already active in the NT — the content believed. James uses at least the first and (at v. 19) something close to the third: the man of v. 19 believes a proposition (ὅτι εἷς ἐστιν ὁ θεός) and James grants it, then points out that δαιμόνια manage the same. That is a datum about the kind of πίστις under examination, not a lexical shift.

One thing this verse does settle: whatever James denies saving power to, he specifies with an article in the second clause. See the next section.

Proposed — pending reviewLuke Timothy Johnson, The Letter of James, Anchor Bible 37A (New York: Doubleday, 1995)

This is the most consequential single decision in the verse.

The first πίστις (v. 14a) is anarthrous. The second (v. 14b) is articular: ἡ πίστις. The switch is not stylistic. Greek marks with the article what has just been introduced, and the natural reading is anaphoric — the faith just described, the one claimed without works.

The consequence for English is direct. The traditional ‘can faith save him?’ reads as a question about faith as such, which is not what James asks and would set him against Rom 3:28 and Eph 2:8 in a way the Greek does not require. ‘That faith’ or ‘this faith’ renders the article as it stands.

The reading is not merely inferred from the article. James supplies the referent himself three verses later: οὕτως καὶ ἡ πίστις, ἐὰν μὴ ἔχῃ ἔργα, νεκρά ἐστιν (v. 17), and again at v. 20 (ἡ πίστις χωρὶς τῶν ἔργων ἀργή ἐστιν) and v. 26. Each time the articular πίστις is immediately qualified by a workless condition. The article at v. 14 points at the same thing.

The corpus rule established by this verse: where a second, articular occurrence of a noun follows an anarthoruous first occurrence in the same sentence, the article is rendered demonstratively in English if the anaphora is exegetically load-bearing.

Proposed — pending reviewDouglas J. Moo, The Letter of James, Pillar New Testament Commentary (Grand Rapids: Eerdmans, 2000)

ἔργα is rendered works without variation, in keeping with the corpus policy that keeps ἔργον on one English root across gospel labor (Phil 1:6; 2:30), the singular of an individual’s own conduct (Gal 6:4), and the Pauline ἔργα νόμου. Varying the English here — ‘deeds’, ‘actions’, ‘good works’, ‘what you do’ — would hide from the reader that James and Paul use the same word.

Anarthrous and plural: works of a kind, itemizable acts, not a defined body of requirements. James defines them by example rather than by definition: giving τὰ ἐπιτήδεια τοῦ σώματος to a naked and hungry brother or sister (vv. 15-16), Abraham’s offering of Isaac (v. 21), Rahab’s reception of the messengers (v. 25). Two of the three are acts of hospitality and relief; none is a cultic or calendrical observance.

Do not supply a qualifying adjective. ‘Good works’ reads a verdict into a bare plural noun, and — in English — carries a freight of merit-theology (§9, §11) that the Greek plural does not. The Vulgate’s opera is the right level of restraint.

μή heading an independent clause with the indicative marks a question expecting the answer ‘no’ (BDF §427; cf. μὴ γὰρ οἰκίας οὐκ ἔχετε, 1 Cor 11:22; μὴ πάντες ἀπόστολοι, 1 Cor 12:29). It is not a prohibition, which would require a subjunctive or imperative, and it is not the ordinary negative.

English has a close equivalent in the negative tag question: ‘That faith cannot save him, can it?’ This is what the reading text uses. It carries the expectation of ‘no’ without asserting it, exactly as the Greek does.

What is excluded from the reading text is a flat declarative — ‘such faith cannot save him’ — which is the answer rather than the question, and reaches James’s conclusion four verses ahead of him. The Vulgate’s numquid fides poterit salvare eum? keeps the interrogative, and so should English. A bare ‘Can that faith save him?’ is acceptable in the study tier, where the expected answer can be recorded separately, but it loses the μή.

δύναται is kept lexically distinct from ἰσχύω (‘have strength for’, already fixed at Phil 4:13 and Gal 5:6). δύναμαι is plain capability with a complementary infinitive; James is asking whether the thing is capable of accomplishing the act, and the aorist infinitive σῶσαι names the act as a single whole.

σῴζω is rendered save per the standing σωτηρία/σῴζω policy, which fixes one gloss and footnotes narrower contextual senses rather than varying the English.

The aorist infinitive is complexive: the saving considered as one completed act, not a process. Nothing in the tense fixes when.

James’s own usage is the best guide to the referent. At 1:21 he speaks of the implanted word able σῶσαι τὰς ψυχὰς ὑμῶν; at 4:12 of the lawgiver who is δυνάμενος σῶσαι καὶ ἀπολέσαι; at 5:20 of turning a wanderer and so σώσει ψυχὴν αὐτοῦ ἐκ θανάτου. The 5:20 formula, with ἐκ θανάτου, and the judgment-language of 2:12-13 immediately preceding, place the reference in the domain of final rescue at judgment rather than of deliverance from a present peril.

That said, the word itself does not say ‘at the last judgment’, and no adjective may be supplied. ‘Save him eternally’, ‘save him in the end’, and ‘give him final salvation’ each specify what James leaves to the surrounding argument (§4). The immediately preceding ἡ γὰρ κρίσις ἀνέλεος τῷ μὴ ποιήσαντι ἔλεος (v. 13) does the specifying, and the reader should be able to see it doing so.

αὐτόν refers to τις and is masculine by grammatical concord with the indefinite pronoun; per §3 the referent is not thereby male-restricted.

NA28 prints the verse without a bracketed reading, and the sense is not in doubt at any point. Three minor divergences in the tradition are worth recording, all of them showing the direction in which scribes and versions smoothed.

Word order in the first protasis. NA28 reads πίστιν λέγῃ τις ἔχειν with 01, 02, 03, 44, and the bulk of the early witnesses. Part of the Byzantine tradition transposes to λέγῃ τις πίστιν ἔχειν, normalizing the fronted object into the ordinary verb-object sequence. The harder reading, with πίστιν first, is original, and it is the reading that preserves the symmetry with the fronted ἔργα in the second limb.

The article at ἡ πίστις. A handful of late witnesses omit the article, producing a general question about faith. This is the reading behind much of the traditional English (‘can faith save him?’), and it is secondary — the omission is easily explained as assimilation to the anarthrous πίστιν of the first clause, whereas an addition of the article serves no obvious scribal motive. The critical text’s article stands.

δέ. Isolated witnesses drop the δέ before μή, and the Latin tradition’s et (as against a stronger adversative) flattens it. Per §10 the particle is rendered.

None of this is doctrinally consequential, but the second item is exegetically consequential, which is precisely the reason to say so rather than to leave a reader wondering why one English tradition reads ‘faith’ and the source text reads ‘the faith’.

The verse is the header of a tightly built unit, and the rendering should serve the structure rather than obscure it.

• v. 14 — two questions: what benefit? can that faith save?
• vv. 15-16 — the test case: a brother or sister naked and short of daily food, dismissed with a blessing and nothing else. The paragraph’s opening question is repeated verbatim as its punchline: τί ὄφελος.
• v. 17 — the inference: οὕτως καί — so also faith, if it has no works, is dead by itself.
• vv. 18-19 — the interlocutor, and the demand for demonstration (δεῖξόν μοι).
• vv. 20-25 — Abraham and Rahab.
• v. 26 — the closing analogy, returning to νεκρά.

Two consequences for translation. First, τί ὄφελος must be rendered with identical English words at v. 14 and v. 16, since the repetition is the frame of the illustration and any variation for style destroys it. Second, the ἐάν … δέ construction of v. 14 recurs in the same shape at vv. 15-16 (ἐὰν … ὑπάρχωσιν … εἴπῃ δέ … μὴ δῶτε δέ) and at v. 17 (ἐὰν μὴ ἔχῃ ἔργα): James is running one syntactic pattern through the whole paragraph and English should let it be seen.

Note also the catchword linkage backward: λειπόμενοι at v. 15 picks up λείπεται at 1:5 and λειπόμενοι at 1:4, and κρίσις/κριτής at 2:12-13 picks up κριταί at 2:4. This paragraph is not a fresh start but the next turn of an argument about hearing-versus-doing that began at 1:22.

Proposed — pending reviewMartin Dibelius, James: A Commentary on the Epistle of James, rev. Heinrich Greeven, Hermeneia (Philadelphia: Fortress, 1976)

The early readers of this paragraph divide instructively, and their divisions track the exegetical questions rather than obscuring them.

The most striking early witness is not a commentator but a user. 1 Clement, writing from Rome in the 90s AD, holds both together without embarrassment: Abraham is justified διὰ πίστιν καὶ φιλοξενίαν (10.7), and Rahab is saved διὰ πίστιν καὶ φιλοξενίαν (12.1) — the same two figures James pairs at 2:21-25, in the same order, with the same coupling of faith and hospitable action. Whatever the literary relation, the pairing was evidently already traditional, which weighs against reading James 2 as an ad hoc anti-Pauline construction.

Greek commentators on the Catholic Epistles regularly turn on the very point the article makes: what James denies is not faith but a faith that stands alone, and the question of v. 14 is a question about the claim (λέγῃ), not about the possession. The distinction is available in the Greek and the Fathers reading Greek generally saw it.

Augustine takes the paragraph as directed against a misreading of Paul rather than against Paul, and treats the two apostles as answering different errors — a reading that has shaped Western handling of the passage ever since. It should be recorded as reception (§11) and not used as evidence for a rendering.

What the patristic material does not support is either of the two flattenings modern English has produced: the reading that makes James deny the saving power of faith as such, and the reading that quietly turns his ἔργα into ἔργα νόμου so the collision with Paul can be dissolved by lexical means.

Proposed — pending reviewBo Reicke, The Epistles of James, Peter, and Jude, Anchor Bible 37 (Garden City, NY: Doubleday, 1964) · Gerald Bray, ed., James, 1-2 Peter, 1-3 John, Jude, Ancient Christian Commentary on Scripture: New Testament XI (Downers Grove, IL: InterVarsity Press, 2000)