This is the hardest verse in James, and the difficulty is not lexical. Every word is ordinary; the Greek is clear. What is unclear is who is talking, and where the quotation stops — and since Greek manuscripts carry no quotation marks, that is a question the text itself does not answer.

The shape of the paragraph is plain enough. James has just said that faith without works is dead in itself (v. 17). He now raises a voice — ἀλλʼ ἐρεῖ τις, “but someone will say” — a standard diatribe move, the same one Paul uses at Rom 9:19 and 11:19. What that voice says is: Σὺ πίστιν ἔχεις κἀγὼ ἔργα ἔχω, “You have faith and I have works.” Then comes a challenge: show me faith without works, and I will show you faith out of works.

The trouble is that the objection, as it stands, does not object to anything James has said. James has just argued that faith and works belong together; an interlocutor who separates them into two possessions (“you have one, I have the other”) is saying something James would want to attack — yet the challenge that follows is Jamesʼ own argument, put in the first person. So either the objector says only the first clause and James answers from “show me,” or the objector says both and James is quoting a position he then demolishes in vv. 19-26, or the second-person σύ does not point at James at all.

No solution is free of strain. The apparatus below sets out the readings without pretending one is settled. What is not in doubt is the argumentative payload: faith is not demonstrable in the abstract. The only evidence admissible for its existence is works. That claim survives every reconstruction of the speakers.

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

ἀλλʼ ἐρεῖ τις is diatribe formula. The future indicative of λέγω with an indefinite subject conjures an interlocutor who is not present and whose objection the writer has composed himself. Compare Rom 9:19 (ἐρεῖς οὖν μοι), Rom 11:19 (ἐρεῖς οὖν), 1 Cor 15:35 (ἀλλὰ ἐρεῖ τις) — the last being verbally identical to James here. The device belongs to the Hellenistic moral-philosophical tradition and is one of the marks of Jamesʼ cultivated register (cf. the literary ἔοικα at 1:6, 23).

Two points bear on the rendering. First, ἀλλά is the strong clause-initial adversative and per the corpus policy is rendered plainly “but,” kept clause-initial: the abruptness of the pivot is the rhetoric. It is not the ascensive ἀλλὰ καί of Phil 1:18. Second, the future ἐρεῖ is anticipatory, not predictive of a specific person — the objection is entertained before it is made. English must keep the future; a present (“someone says”) turns an anticipated objection into a report of one already circulating, which is more than the Greek claims.

τις is left studiedly indefinite. James does not name the speaker and the rendering must not either — “an objector will say,” “one of you will say,” and “a critic will say” all supply identification the pronoun withholds.

Proposed — pending reviewStanley K. Stowers, The Diatribe and Paul's Letter to the Romans, SBLDS 57 (Chico, CA: Scholars Press, 1981)

The whole difficulty can be stated as two questions: where does the quotation end, and who is σύ?

The reading adopted in the smooth text follows NA28ʼs punctuation (option 1) because the printed critical text is the base of this project (§4), and because δεῖξόν μοι … κἀγώ σοι δείξω reads most naturally as the challenge of the one who has just been talking about dead faith. But the punctuation of the editions is editorial judgment, not manuscript evidence — the earliest majuscules are unpunctuated here, and the reader should know that the choice was made by editors, not scribes.

Note too that James switches to the singular σύ for the whole stretch vv. 18-22, having addressed the assembly in the plural throughout vv. 1-17 and returning to the plural at v. 24 (ὁρᾶτε). Whoever the singular is, the argument narrows to one imagined man and then widens again.

Proposed — pending reviewMartin Dibelius, A Commentary on the Epistle of James, rev. Heinrich Greeven, trans. Michael A. Williams, Hermeneia (Philadelphia: Fortress, 1976) · Peter H. Davids, The Epistle of James: A Commentary on the Greek Text, NIGTC (Grand Rapids: Eerdmans, 1982)

πίστις occurs four times in this verse and ἔργον four times; the two nouns are the whole vocabulary of the argument. Neither may be varied in English.

The standing corpus decision permits “faith” or “faithfulness” for πίστις and requires that the choice be flagged in virtue-lists. Here neither pole is a virtue-list item: the noun is the thing claimed to be possessed and the thing demanded as evidence, and James himself glosses it at v. 19 as assent to a proposition (πιστεύεις ὅτι εἷς ἐστιν ὁ θεός). “Faith” is therefore correct throughout, and the reader should notice that James is arguing about the sort of πίστις that can be stated — which is why he can call it ἀργή (v. 20) and νεκρά (vv. 17, 26).

ἔργα is anarthrous in the objectorʼs mouth (ἔργα ἔχω, qualitative — works of some kind) and articular thereafter (χωρὶς τῶν ἔργων, ἐκ τῶν ἔργων μου). The articles are anaphoric: the works, the ones this whole paragraph has been about since v. 14 (ἔργα δὲ μὴ ἔχῃ) and specified concretely at vv. 15-16 as clothing and feeding a destitute brother or sister. English idiom permits dropping the article in smooth registers, but the literal tier retains it, because the definiteness is what keeps James from making an abstract claim about works-in-general.

κἀγώ stands twice, and both elements have to survive into English. The crasis is καί + ἐγώ: an adjunctive/correlative καί plus an expressed nominative pronoun, and Greek verb endings already carry the person, so the pronoun is marked (§4; cf. the standing entries on ἐγώ and καὶ αὐτός).

In the first instance (κἀγὼ ἔργα ἔχω) the καί correlates the speakerʼs possession with the addresseeʼs: “you have X, and I for my part have Y.” The pairing with the equally emphatic Σύ at the head of the clause makes the sentence a two-column comparison, which is precisely what makes it hard to hear as an objection to James.

In the second (κἀγώ σοι δείξω) the καί answers the imperative: do this, and I will do the corresponding thing. Rendering “and I will show you” is adequate for smooth English, but the study tier should carry “and I for my part,” since the contrast between the two demonstrations is the point — one is impossible, the other is producible on demand.

Dropping either καί or reducing the pronoun to a bare first-person verb (“I will show you”) loses the symmetry on which the challenge rests.

δείκνυμι is the verb of pointing something out so that it can be seen — exhibiting, producing, displaying to view. It is not “prove” in the sense of demonstrating a proposition by argument (that is ἀποδείκνυμι or συνίστημι), and it is not “explain.” BDAG gives “point out, show, make known”; the concrete pole is dominant in NT usage (Matt 4:8, showing the kingdoms; Luke 4:5; Luke 22:12, showing a room; John 20:20, showing hands and side; Acts 7:3).

That concreteness is exegetically load-bearing here. The demand is not argue for your faith but produce it, put it where I can look at it. And the retort is not I will reason from my works to my faith but I will exhibit the faith, and the works are what I exhibit it out of. Rendering “prove to me your faith” shifts the whole exchange from display to argumentation and quietly concedes that faith might be established some other way.

The aorist imperative δεῖξόν is a single decisive act: one demonstration, now. The future δείξω is a genuine future undertaking, matched to it. Keep both tenses; a present (“I show you”) collapses the challenge-and-answer structure.

χωρίς + genitive marks absolute separation, not degree — the corpus fixes this at Phil 2:14 and it holds here. “Without the works” means with them wholly removed from view, not “with few works” or “with works left aside for the moment.” The word recurs at v. 20 (χωρὶς τῶν ἔργων ἀργή) and v. 26 (χωρὶς ἔργων … χωρὶς πνεύματος), where the body/spirit comparison makes the absoluteness unmistakable: a corpse is not a body with reduced animation.

Note the article in vv. 18 and 20 (τῶν ἔργων) against its absence in v. 26 (ἔργων). The articular form is anaphoric to the works of vv. 15-16; the anarthrous form at the end generalizes. The literal tier preserves both.

ἐκ + genitive marks source, kept distinct in this corpus from ἀπό (origin), ὑπό (personal agency), and διά (means). The works are what the demonstration proceeds out of — the evidentiary material from which the faith becomes visible.

This is the exact preposition James will use three times in vv. 21-25 (ἐξ ἔργων ἐδικαιώθη, ἐκ τῶν ἔργων ἡ πίστις ἐτελειώθη, ἐξ ἔργων δικαιοῦται), and the concordance link should be visible in English: keep “out of” in the literal tier at every point rather than varying to “by” or “through,” which would convert source into agency or means and pre-empt vv. 21-24.

The fronting of ἐκ τῶν ἔργων μου before its object τὴν πίστιν is emphatic and worth preserving in English word order where possible — “I will show you, out of my works, the faith.” The sentence ends on the noun it is about.

The final words are printed by NA28/UBS5 as τὴν πίστιν — articular, with no possessive pronoun, in deliberate asymmetry with τὴν πίστιν σου earlier in the verse.

A substantial body of witnesses adds μου, giving “I will show you my faith out of my works,” and this is the reading behind most English versions. The addition is the easier text in every respect: it balances the two halves of the challenge, it supplies the possessive that English idiom expects, and it removes the slight oddity of a bare articular noun at the clause end. The direction of scribal activity in such cases runs toward supplying the pronoun, not deleting it, so the shorter reading is to be preferred as the harder one.

The difference is not trivial for the sense. Without μου, the object of the demonstration is the faith — the thing under discussion, faith as such — and the claim is that works are the only medium in which faith of any kind becomes visible. With μου, the claim narrows to a comparison of two personal possessions. The smooth rendering here keeps “my faith” because English will not tolerate a bare “the faith” in that slot without sounding like a reference to the Christian religion; but the literal and liturgical tiers retain the article without the possessive, and the reader should know that the tidier familiar wording rests on the weaker text.

A further variant tradition reads χωρὶς τῶν ἔργων σου (adding the pronoun to the first phrase as well), again smoothing toward symmetry.

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

The patristic tradition read the verse without much of the modern anxiety about speaker-identification, largely because it did not feel the pressure of a Paul-versus-James problem in the form later exegesis gave it.

The surviving Greek commentary tradition on James (the catena material transmitted under the names of Didymus and others, and later Oecumenius and Theophylact) generally takes the challenge from δεῖξόν μοι as Jamesʼ own and reads the whole as a demand for evidence: faith that cannot be exhibited is not faith of the kind that saves. That is the reading the printed editions punctuate.

The Latin tradition is instructive on a different point. The Vulgate renders ἀλλʼ ἐρεῖ τις as sed dicet quis, keeping the future and the indefinite subject, and translates δεῖξόν μοι with ostende mihi — the display-verb, not a verb of proof. Jerome, that is, did not read the exchange as an argument about demonstrating propositions, which supports keeping the concrete force of δείκνυμι in English. The Latin also has fidem meam at the end, reflecting the fuller Greek reading with μου.

Augustine, in De fide et operibus, treats the passage as directed against those who thought bare confession sufficient — which is exactly the target James names at v. 19 with the demons who believe and shudder. That the demons believe is the pointed thing: whatever this πίστις is, it is not nothing, and it is not enough. Reception of this kind is recorded here as reception (§11); it does not decide the punctuation.

Proposed — pending reviewAugustine, De fide et operibus, in Corpus Scriptorum Ecclesiasticorum Latinorum 41 (Vienna, 1900)

1. Punctuation follows NA28, with the objection limited to the first sentence — but the apparatus records that the punctuation is editorial and the long-quotation reading is live. The reading text does not use a footnote-free construction that hides the seam; the quotation marks are where the editors put them and are flagged as such.

2. ἀλλά rendered “but,” clause-initial. Not “yes, but” (which would make it ascensive), not dropped (§10).

3. Both κἀγώ preserved in full. The adjunctive καί and the expressed pronoun are separately rendered.

4. δείκνυμι kept as “show,” not “prove.” The display sense is what the argument runs on.

5. χωρίς kept as absolute separation (“apart from” / “without”), matching vv. 20 and 26.

6. ἐκ kept as “out of” in the literal tier so that the concordance link to vv. 21-25 is visible.

7. The final τὴν πίστιν is rendered “my faith” in smooth English only, with the text-critical situation flagged; the literal tier keeps the bare article.

8. σύ and the singular verbs are retained as singular through the verse, against the pull toward a plural congregational address, because the number-shift from vv. 1-17 is discourse structure.