This is the hinge of James’s Abraham argument, and it does three things at once.

First, it supplies the proof-text: Gen 15:6, quoted verbatim from the LXX — the same verse Paul builds Rom 4 and Gal 3 on. Two writers, one sentence of Scripture, two very different arguments running out of it.

Second, it makes a claim about how that sentence relates to Gen 22. James says the text was filled up (ἐπληρώθη) when Abraham laid Isaac on the altar — an event some fifty-eight chapters and many years after the words were spoken. The word-declaration of Gen 15:6 found its full measure in the act of Gen 22.

Third, it closes with a clause that is not part of the quotation at all: “and he was called friend of God.” No single OT verse says this in those words; it is a title Abraham carries in Isa 41:8, 2 Chr 20:7, and thickly in Second Temple Jewish tradition. James appends it as the standing verdict that the filled-up Scripture produced.

The verse is grammatically simple and theologically dense. Every load-bearing decision is about how much English is allowed to resolve: πληρόω (fulfil or fill up?), λογίζομαι (impute or count?), εἰς δικαιοσύνην (as righteousness or with a view to it?), and whether ἐκλήθη names a divine act or a settled reputation. The policy throughout is the corpus policy — hold the Greek’s openness where English can hold it.

Built onbdag · na28

The corpus glossary fixes πληρόω as fill up / make full, with the vessel-image treated as live and the standing warning against abstracting it into complete or fulfil. That decision was made at Phil 2:2 and reaffirmed at Gal 5:14 (the perfect passive of the law) and Phil 4:19. It is retained here, and it is doing real work.

The issue: πληρόω with γραφή as subject is a fixed early-Christian formula for the realization of a prophetic word (Matt 1:22; 2:15; John 19:24, 36; Acts 1:16), and in that formula the standard English is “was fulfilled.” Gen 15:6, however, is not a prediction. It is a narrative report of a reckoning already made. So what does it mean for it to be filled up?

Three readings, all defended:

1. Consummative-vessel: the declaration of Gen 15:6 was a container that Gen 22 filled to the brim. The word was true when spoken and reached its full measure in the act. This coheres with ἐτελειώθη in v. 22 — the same movement described first of the faith, then of the text.
2. Prophetic-typological: Gen 15:6 functioned proleptically, anticipating an obedience not yet performed; the Aqedah is its realization in the standard fulfilment sense.
3. Confirmatory: the Scripture was shown to be true, vindicated by the event. Weakest lexically — πληρόω does not naturally mean verify.

The καί linking v. 22 to v. 23 is the strongest internal evidence: ἐτελειώθη and ἐπληρώθη are parallel passives with unexpressed divine agency, and the τελ-/πληρ- pairing runs right through this corpus (Phil 1:6; Gal 5:14). Rendering “was fulfilled” in the reading text quietly adopts reading 2 and severs the parallel. “Was filled up” is retained; “was fulfilled” is available in liturgical register, where the English tradition has fixed it.

Built onbdag · bdf

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

Two features.

The doubled article. Article–noun–article–participle is the second attributive position (BDF §270), which adds the modifier as a separate, mildly emphatic unit rather than embedding it. The corpus policy (fixed at Jas 1:9, and earlier at Phil 3:9-11, 4:17) is to retain the doubled article in the literal tier — “the Scripture, the one saying” — and resolve to an English relative clause in reading tiers. The construction is restrictive: this Scripture, the one with these words.

The personification. The glossary already fixes ἡ γραφή as the Scripture, capitalized and articular, with the note that where it governs a verb of speaking the personification is the point (Gal 3:8; 4:30; Rom 9:17). James has it speaking, and the present participle λέγουσα is durative — the text goes on saying this. Do not render “the passage which reads” or “as it says in Scripture”; both dissolve a speaking authority into a citation apparatus.

Note also what is absent: no γέγραπται, no καθὼς λέγει. James introduces the citation by asserting its fulfilment, not by flagging that he is quoting. The reader is expected to know the sentence.

Built onbdf · bdag

The citation is Gen 15:6 LXX verbatim, including the postpositive δέ, which belongs to the LXX sentence and not to James’s own syntax. Retaining it in English (“And Abraham believed God”) preserves the seam; deleting it as untranslatable filler (§10) hides that James lifted the clause whole.

The Hebrew of Gen 15:6 is וְהֶאֱמִן בַּיהוָה, with the hiphil of אמן — he trusted, held firm in YHWH. Per the divine-name policy (§1), the Greek τῷ θεῷ is rendered as written (God) and the Hebrew substrate is footnoted, never imported: the NT author’s own lexical choice is data to preserve.

The construction is πιστεύω + dative of person, not πιστεύω εἰς + accusative, which the glossary treats as the distinctively early-Christian committal idiom. The dative names the person whose word is credited. James’s own usage elsewhere in this chapter is instructive: at 2:19 he writes πιστεύεις ὅτι — propositional assent, which the demons share. Here the object is a person. Whether James intends the reader to feel that contrast is contested, but the syntactic difference is real and should not be flattened in English.

MT/LXX agreement here is close; the LXX supplies the subject Ἀβραάμ where the Hebrew has only the verb, and shifts to the passive ἐλογίσθη where MT has an active with an unnamed or divine subject (וַיַּחְשְׁבֶהָ). James follows the LXX in both.

Built onmuraoka-lxx · bdag · na28

Proposed — pending reviewHenry Barclay Swete, An Introduction to the Old Testament in Greek, rev. R. R. Ottley (Cambridge: Cambridge University Press, 1914)

λογίζομαι is fixed in the corpus glossary as reckon / count, with the accounting sense primary (BDAG 1) and the express requirement that an accounting word be kept in English so the Gen 15:6 imputation texts and the Pauline ledger argument remain visibly one vocabulary. That decision holds here without modification. Avoid think and consider (ἡγέομαι, which James does not use here) and avoid credit as base gloss only insofar as it obscures the passive.

The verb is a divine passive: the agent is unexpressed and must stay unexpressed. “God reckoned it to him” supplies what the Greek withholds — and, more importantly, the LXX itself chose the passive where the Hebrew has an active verb. That choice is data.

εἰς δικαιοσύνην. The preposition renders Hebrew לְ and is a Semitizing εἰς of resulting classification — the entry is posted under the heading “righteousness.” English as is the closest natural idiom; the literal tier keeps unto. Per the standing εἰς policy the goal/result senses need not be separated.

What is it that was reckoned? Greek αὐτῷ is masculine (“to him”) and the unexpressed subject of ἐλογίσθη is neuter-or-feminine by implication — most naturally the believing just named, matching MT’s feminine suffix on וַיַּחְשְׁבֶהָ. Do not supply “his faith” unbracketed.

δικαιοσύνη is fixed as righteousness across the forensic and ethical-conduct poles, with the instruction that where a passage forces one pole this is flagged in apparatus rather than varied in the gloss. That instruction applies with force here.

Built onbdag · muraoka-lxx · na28

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

This clause is outside the quotation. No OT text says of Abraham φίλος θεοῦ ἐκλήθη in those words. The nearest are:

• Isa 41:8 LXX — σπέρμα Ἀβρααμ ὃν ἠγάπησα (“Abraham whom I loved”); MT אַבְרָהָם אֹהֲבִי, Abraham my friend/lover.
• 2 Chr 20:7 LXX — τῷ σπέρματι Ἀβρααμ τῷ ἠγαπημένῳ σου; MT אַבְרָהָם אֹהַבְךָ.
• Dan 3:35 LXX (Prayer of Azariah) — Ἀβρααμ τὸν ἠγαπημένον ὑπὸ σοῦ.

The LXX consistently uses the ἀγαπ- root; the φίλος designation is a Jewish-tradition development, standard by the first century (Jub. 19:9; 30:20; Philo, De sobrietate 56, who explicitly discusses the LXX rendering). James is using a fixed honorific title, not making a fresh exegetical inference from a verse.

Lexical notes. φίλος names the friend as one admitted to confidence — Gen 18:17 (“shall I hide from Abraham what I am doing?”) is the narrative warrant the tradition drew on, and John 15:15 works the same logic. Do not weaken to “God’s loved one” (that would be ἀγαπητός) or strengthen to “the Friend of God” with a capital, which converts a title into a proper name.

The genitive θεοῦ is anarthrous and the whole phrase functions as a predicate title after a passive verb of naming — hence friend of God, not the friend of God.

ἐκλήθη. Aorist passive of καλέω, agent unexpressed. Two live construals: (a) divine naming — God called him this, which suits the LXX first-person texts; (b) the settled designation he came to bear, effectively “he got the name.” The Greek does not choose, and neither does the English. Note the deliberate chain of unexpressed-agent passives across vv. 22-23: ἐτελειώθη, ἐπληρώθη, ἐλογίσθη, ἐκλήθη. Four times James declines to name the actor. That reticence is the argument’s texture and must survive translation.

Built onbdag · muraoka-lxx · na28

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)

James 2:23 and Rom 4:3 quote the same LXX sentence and reach conclusions that sound opposed. Honest handling requires stating the seam rather than hiding it (§4, §5).

What is genuinely different:

• The referent of ἔργα. Paul’s polemical target is ἔργα νόμου — circumcision, calendar, food laws, the boundary-markers of covenant membership (Gal 2:16; Rom 3:20). James’s ἔργα, defined by his own examples in this chapter, are feeding the naked and hungry (2:15-16) and the Aqedah (2:21). Neither writer is using the phrase to mean what the other means by it.
• The referent of πίστις. James defines the faith he is attacking at 2:19: assent to monotheism, which demons share. Paul’s πίστις is committal trust. The corpus glossary already flags πίστις as spanning both poles, and both writers are using one word at different points on its range.
• The moment in view. Paul takes Gen 15:6 as the moment of reckoning, prior to circumcision (Rom 4:10) — the chronology is his argument. James takes Gen 22 as the moment of filling up. They are looking at different points on the same story.

What is genuinely shared: neither writer thinks Abraham purchased anything. James’s own passives (ἐλογίσθη, ἐκλήθη) leave God as the unnamed actor throughout.

Built onbdag · na28

Proposed — pending reviewRichard Bauckham, James: Wisdom of James, Disciple of Jesus the Sage, New Testament Readings (London: Routledge, 1999)

The Fathers read the verse with less anxiety than the post-Reformation tradition, because they were not asking the question that makes it hard.

Clement of Rome (1 Clem. 10.1, 6; 17.2) already calls Abraham ὁ φίλος προσαγορευθεὶς τοῦ θεοῦ and cites Gen 15:6 — independent evidence that the friend-title and the Gen 15:6 citation travelled together in Jewish-Christian teaching before either was a controversial datum. This is the strongest single argument that James is drawing on catechesis rather than replying to Romans.

Origen and the Alexandrian tradition read the Aqedah as the demonstration of what Gen 15:6 had declared — closer to the consummative than the predictive sense of ἐπληρώθη.

Augustine (De fide et operibus) took up the Paul/James relation directly and resolved it by distinguishing the faith James denies (bare assent) from the faith Paul commends. That is a lexical judgment about πίστις, and it happens to align with what James himself says at 2:19 — which is why it has worn well even where its harmonizing motive is suspect.

Recorded as reception (§11): none of this is evidence for a rendering. But it does show that the φίλος θεοῦ title was a fixed formula in the earliest Christian generation, which bears on how the final clause of the verse should be construed.

Built onna28

The verse is stable. NA28 prints no variant of consequence in the main line.

Minor items in the tradition, none affecting sense:

• Some witnesses omit the postpositive δέ in Ἐπίστευσεν δέ, assimilating the citation to the more familiar anarthrous-opening form of the quotation as it appears in Rom 4:3 (which has δέ) and Gal 3:6 (which has καθώς). Retaining δέ is the harder and better-attested reading and preserves the LXX seam.
• A scattering of witnesses supply ὁ θεός as an expressed subject with ἐκλήθη or read ἐκάλεσεν (active). This is scribal clarification in the direction the apparatus warns against — supplying an agent the divine passive withholds. Reject.
• Orthographic variation in Ἀβραάμ (single/double alpha) is spelling, not text.

The editorial punctuation — where the quotation opens and closes — is not a manuscript matter. The earliest majuscules are unpunctuated, and NA28’s raised dot after λέγουσα and its comma placement are the editors’ judgment. Recorded per corpus policy rather than treated as evidence.

Built onna28 · metzger-tc

James 2 is paraenetic argument with diatribe features — rhetorical questions (vv. 4, 5, 6, 7, 14, 20, 21, 25), an imagined objector (v. 18), a vocative of rebuke (ὦ ἄνθρωπε κενέ, v. 20). Verse 23 is the argument’s evidential high point, and its register lifts accordingly: a scriptural citation, a formulaic honorific title, and four solemn passives in two verses.

The English should not flatten this into explanatory prose. Per §7, match the source’s register: the citation is set off, the title stands as a title, and the chain of passives is kept passive throughout. The temptation in every clause here is to supply an actor and smooth the syntax — and every time that happens, something the Greek deliberately withheld is put back in.

Built onbdf