Paul as a Jew? The Hebraisms Attributed to Paul

The Hebraisms Attributed to Paul: Nine Protestant Objections and Their Philological Answer

Whoever maintains that Paul’s Greek is not that of a talmid chakham exposes himself to a predictable rejoinder: “he has Hebraisms.” The rejoinder usually comes from Protestants with good grammars in hand (Blass–Debrunner–Funk, Moulton–Turner, Zerwick) and with a roll of authors regularly cited in their support (F. F. Bruce, E. Earle Ellis, Richard Longenecker, N. T. Wright and, from within Lutheranism, Martin Hengel). It is not answered by denying the obvious, because there are Semitisms in the letters, and whoever denies it loses the debate in the first exchange. This article sets out the terrain before the discussion: it defines what counts as a Hebraism, fixes the burden of proof, passes each objection through four sieves, and marks in advance what must be conceded. The thesis is that almost every Semitism cited against our position is of another kind (ordinary koine, Septuagintism, received formula or Qumranic parallel) and that none of those kinds proves formation in Yerushalayim.

I. Before arguing: what counts as a Hebraism

The word “Hebraism” hides five different things, and the attacker usually passes from one to another without warning. They should be fixed from the outset.

Type	What it is	What it proves about the author	Example
Koineism	A construction that seems Hebrew but exists in untranslated Greek (papyri, Polybius)	Nothing	ἐν τῷ + infinitive (Gal 4:18)
Septuagintism	A calque that the Greek translation had already created	That he read or heard the Greek Bible	δικαιοσύνη with the field of צדקה; “flesh and blood”
Received formula	A Semitism in a text that the author says he received	The tradition, not his language	1 Cor 15:3–5; ἀββά
Qumranic parallel	An idiom with a parallel in first-century Hebrew, but not rabbinic	Contact with that milieu	υἱοὶ φωτός; ἔργα νόμου
Direct interference	A feature that only the Semitic language explains	Competence in that language	None secure in Paul (Fitzmyer)

Only the last row proves anything about the author’s linguistic formation, and it is the one that criticism has not been able to fill: W. C. van Unnik searched for Aramaisms in Paul (Sparsa Collecta I, 1973, pp. 129–143) and Joseph Fitzmyer (A Wandering Aramean, 1979) judges that he did not succeed in establishing a single secure one. The other four rows serve another purpose: to say that Paul read the Bible in Greek, received traditions from Yerushalayim and frequented a Judaism whose pious vocabulary circulated in Greek. Nobody disputes that.

II. The burden of proof: four sieves

Whoever affirms that an expression is a Hebraism of Paul must show that it passes four sieves. Sieve 1, the koine: does it exist in contemporary untranslated Greek (documentary papyri, inscriptions, Polybius, Epictetus)? If it does, it is not a calque. Sieve 2, the LXX: was it fixed by the Greek translation? If so, it is a Septuagintism, and it proves reading, not formation. Sieve 3, the received tradition: does it live in material that Paul declares he received (παρέλαβον, 1 Cor 15:3; 11:23)? If so, it belongs to the tradition and not to his prose. Sieve 4, contemporaneity: does it have a parallel in first-century Hebrew or Aramaic (Qumran, Murabbaʿat, Bar Kokhba, Nabataean) and not only in the Mishnah, the Talmud or the late targumim? If it has it only in later sources, it is an anachronism. What survives all four deserves discussion, and it is very little.

A control of frequency should be added. The most recognisable Semitisms of biblical Greek (the initial καὶ ἐγένετο, the cognate dative that renders the infinitive absolute, the resumptive pronoun) are rare or non-existent in Paul’s own prose and abound in Luke–Acts 1–15 and in the Apocalypse. An author who thinks in Hebrew and writes in Greek does not choose which interferences he lets through.

III. The objections

Objection 1. “Sons of light” and the Semitic genitives

The objection. υἱοὶ φωτός (1 Thess 5:5) and σῶμα τῆς ταπεινώσεως (Phil 3:21) are cited as genitives of quality calqued on the Hebrew construct state. The testimony of 1 Thessalonians as transcribed by the ECM reads: πάντες γὰρ ὑμεῖς υἱοὶ φωτός ἐστε καὶ υἱοὶ ἡμέρας (“for you are all sons of light and sons of day”).

The answer. First sieve: the genitive of quality exists in Hellenistic Greek without interference, and σῶμα τῆς ταπεινώσεως (“body of humiliation”, “lowly body”) has a more economical explanation within Greek; the grammars catalogue it as a possible Semitism but document it in other texts. Second sieve: the construction “son of X” as membership of a class was fixed by the Septuagint (υἱὸς ἀδικίας, 2 Sam 7:10 LXX, “son of injustice”; υἱὸς ἀνομίας, Ps 88:23 LXX, “son of lawlessness”). Fourth sieve: for υἱοὶ φωτός there is a first-century Hebrew parallel, בני אור, in 1QS and 1QM, and it is the only point at which the objection survives. But Qumran is not the Mishnah, and the formula appears also in Luke 16:8, John 12:36 and Eph 5:8, that is, it circulated in Greek-speaking Christian communities. Verdict: partial concession. It is a Jewish idiom of piety, of Palestinian origin, that circulated in Greek. It shows contact with a milieu, not a formation.

Objection 2. “Flesh and blood” and “all flesh”

The objection. σαρκὶ καὶ αἵματι (Gal 1:16; 1 Cor 15:50) would be the rabbinic idiom בשר ודם, and οὐ δικαιωθήσεται πᾶσα σάρξ (Gal 2:16; Rom 3:20) the Semitism οὐ πᾶς for “none” (כל… לא).

The answer. Gal 1:16 says οὐ προσανεθέμην σαρκὶ καὶ αἵματι (“I did not consult with flesh and blood”). The idiom is authentically Semitic, but it was already in Greek well over a century earlier: Sir 14:18 (γενεὰ σαρκὸς καὶ αἵματος, “a generation of flesh and blood”), where Hebrew manuscript A from the Genizah has בשר ודם, and Sir 17:31. Second sieve, then. What is characteristic of rabbinic literature is something else: the mashal of the “king of flesh and blood” (מלך בשר ודם) as against the Holy One, blessed be He. Paul never uses it; his “flesh and blood” is the human or mortal condition (1 Cor 15:50). For Gal 2:16, ὅτι ἐξ ἔργων νόμου οὐ δικαιωθήσεται πᾶσα σάρξ (“because by works of the law no flesh will be justified”), the source is Ps 142:2 LXX, οὐ δικαιωθήσεται ἐνώπιόν σου πᾶς ζῶν (“no living person will be justified before you”), a literal translation of כִּי לֹא יִצְדַּק לְפָנֶיךָ כָּל חָי (“for no living being is righteous before you”), with σάρξ in place of ζῶν; and “all flesh” (כל בשר) is already Hebrew Bible (Gen 6:12; Isa 40:5). A scriptural echo is not a spontaneous calque. Verdict: full concession on the idiom, none on the inference. They are Hebraisms of biblical Greek, not of the speaker.

Objection 3. “Works of the law” and 4QMMT

The objection. ἐξ ἔργων νόμου (Gal 2:16; 3:2, 5, 10; Rom 3:20, 28) calques מעשי התורה, and the parallel is first-century Hebrew and not a conjecture: מקצת מעשי התורה (4QMMT C 27, “some of the works of the Torah”) and מעשי תורה (4QFlor 1–2 i 7). It is the opponent’s best card, and it should be played before he plays it. Fitzmyer and Dunn have discussed it in the wake of 4QMMT.

The answer. I concede the parallel: it is probable that the Pauline expression reflects a Hebrew formula of first-century Judaea. But that concession does not lead where the objector wishes, for three reasons. First, the formula is Qumranic, not rabbinic: as far as I can tell, no equivalent technical phrase exists in the Mishnah. Second, Paul never uses it positively: οὐ δικαιοῦται ἄνθρωπος ἐξ ἔργων νόμου (Gal 2:16, “a man is not justified by works of the law”), ὅσοι γὰρ ἐξ ἔργων νόμου εἰσὶν ὑπὸ κατάραν (Gal 3:10, “for as many as are of works of the law are under a curse”), διότι ἐξ ἔργων νόμου οὐ δικαιωθήσεται πᾶσα σάρξ (Rom 3:20, “because by works of the law no flesh will be justified”). He always negates it, and always in the context of the dispute with those who come from Yerushalayim (Gal 2:11–16). Third, and most important: a formula that Paul inherits in a controversy, as the slogan of the other side which he takes up in order to negate it (sieve 3), proves that Paul and his adversaries were arguing in a vocabulary that came from Yerushalayim, not that Paul was formed there. Verdict: concede it, and use it. It is the only Semitism of Paul with a Hebrew parallel so specific that it excludes chance, and its context assigns it to the interlocutor.

Objection 4. The imperatival participle of Romans 12

The objection. In Rom 12:9–13 a chain of participles serves as imperatives: ἀποστυγοῦντες τὸ πονηρόν, κολλώμενοι τῷ ἀγαθῷ (“abhorring what is evil, clinging to what is good”), τῷ πνεύματι ζέοντες (“fervent in spirit”), τῷ κυρίῳ δουλεύοντες (“serving the Lord”). David Daube, in an appendix to Selwyn’s commentary on 1 Peter (1946), saw in this usage the model of the participle of Hebrew legal rules. It is a brilliant hypothesis, and it comes from a Jewish scholar, which makes it harder to dismiss.

The answer. First sieve: the participle with imperatival force exists in untranslated Greek; C. F. D. Moule (An Idiom Book of New Testament Greek, 1953) collects examples from the documentary papyri. Third sieve: Rom 12:9–21 is traditional paraenetic material, with parallels in 1 Pet 3:8–12, Matt 5:44 and Luke 6:27–28, and the chain of participles belongs to that catechetical block and not to Paul’s own syntax. Fourth sieve: the Hebrew parallel that Daube adduces comes in good part from later rabbinic texts, so that it collides with the requirement of contemporaneity. Verdict: minimal concession. It is possible, it is not demonstrable and it does not discriminate.

Objection 5. “Paul thinks in Hebrew”: the anthropology

The objection. Paul thinks in Hebrew because his anthropology is Hebrew: σάρξ, ψυχή, καρδία, πνεῦμα and δικαιοσύνη loaded with the field of בשר, נפש, לב, רוח and צדקה. It is the most popular objection: “Hebrew thought” against “Greek thought”.

The answer. The lexicon is conceded, and it is level 1: it is the same inventory that Josephus shows, the Jew of Yerushalayim who wrote in Greek in the same century, and it was loaded by the Septuagint; whoever acquired it did not need to know Hebrew. The refutation lies in the layer that is not Hebrew. In Rom 7:22–23 Paul writes:

συνήδομαι γὰρ τῷ νόμῳ τοῦ θεοῦ κατὰ τὸν ἔσω ἄνθρωπον, βλέπω δὲ ἕτερον νόμον ἐν τοῖς μέλεσίν μου ἀντιστρατευόμενον τῷ νόμῳ τοῦ νοός μου

(“For I delight in the law of God according to the inner man, but I see another law in my members waging war against the law of my mind”).

Three pieces have no Hebrew correlate. First, the νοῦς as an instance of moral discernment set against the σάρξ (except in the citation of Isa 40:13, where the LXX renders רוח as νοῦς). Second, the “inner man”, ὁ ἔσω ἄνθρωπος, which is Plato’s ὁ ἐντὸς ἄνθρωπος (Rep. 589a). And third, the very structure of the passage, the psychomachia:

ὃ γὰρ κατεργάζομαι οὐ γινώσκω· οὐ γὰρ ὃ θέλω τοῦτο πράσσω, ἀλλ᾽ ὃ μισῶ τοῦτο ποιῶ (Rom 7:15, “for what I work out I do not understand; for what I want, this I do not practise, but what I hate, this I do”),

which Rudolf Bultmann placed beside Epictetus (Diss. 2.26.4, ὃ θέλω οὐ ποιῶ, καὶ ὃ μὴ θέλω ποιῶ, “what I want I do not do, and what I do not want I do”) and the Medea of Euripides (1078–80) and of Ovid (Met. 7.20–21, video meliora proboque, deteriora sequor, “I see the better and approve it; I follow the worse”). To this are added συνείδησις (Rom 2:15; 9:1), of which the Septuagint records only a couple of uses (Eccl 10:20; Wis 17) and which Paul employs as a moral instance of Stoic popularisation; and the distinction ψυχικός/πνευματικός of 1 Cor 15:44–46, which cannot derive from נֶפֶשׁ חַיָּה, the name that the Torah already gives, in Bereshit 1:20 and 1:24, to fish and beasts. The conclusion is the inverse of the pulpit’s: Paul uses words of the Greek Bible to articulate a conflict of faculties that is Greek. Verdict: concession on the lexicon, refutation on the structure.

Objection 6. The exegetical method: qal vachomer, pesher and targum

The objection. E. Earle Ellis (Paul’s Use of the Old Testament, 1957) and Richard Longenecker (Biblical Exegesis in the Apostolic Period, 1975) maintain that Paul uses the methods of Pharisaic exegesis: qal vachomer (πολλῷ μᾶλλον, “much more”, Rom 5:9–10, 15, 17; 11:12, 24), gezerah shavah (Rom 4:3–8, the linking of Gen 15:6 and Ps 32:2 by λογίζομαι) and pesher (τοῦτ᾽ ἔστιν, “that is”, Rom 10:6–8). The strongest version brings in the targumim: the Palestinian targumim of Deut 30:12–14 turn “beyond the sea” into the abyss and make Jonah the one who descends, which would explain the unusual τίς καταβήσεται εἰς τὴν ἄβυσσον (“who will descend into the abyss?”, Rom 10:7); the tradition of the well that followed Israel (Tos. Sukkah 3:11) would explain ἀκολουθούσης πέτρας (“the following rock”) in 1 Cor 10:4; and Eph 4:8 (Ps 68:19) would read “gave” with the targum and the Peshitta and not “took” with the Masoretic Text.

The answer. Qal vachomer. Reasoning a minori ad maius is in Aristotle (Rhet. 2.23.4, ἐκ τοῦ μᾶλλον καὶ ἧττον, “from the more and the less”) and in Cicero, and David Daube (“Rabbinic Methods of Interpretation and Hellenistic Rhetoric”, 1949) and Saul Lieberman (Hellenism in Jewish Palestine, 1950) showed that Hillel’s own middot have Hellenistic rhetorical parallels: the argument is doubly neutral. Gezerah shavah. It works in Greek (λογίζομαι) and in Hebrew (חשב) and does not distinguish. Romans 10:6–8. Paul cites Deut 30:12–14 with the change of “sea” for “abyss” and glosses it with τοῦτ᾽ ἔστιν Χριστόν. First, the pair “heaven/abyss” is in biblical Greek: Ps 106:26 LXX, ἀναβαίνουσιν ἕως τῶν οὐρανῶν καὶ καταβαίνουσιν ἕως τῶν ἀβύσσων (“they go up to the heavens and go down to the depths”, for the Hebrew עלו שמים ירדו תהומות, “they mount up to heaven, they go down to the depths”), and Baruch 3:29–30, which had already read Deut 30 with Wisdom, without need of any targum. Second, the Palestinian targumim are late in their form (the Neofiti manuscript is of 1504), and the date of their traditions is a matter of debate (McNamara against Kaufman). Third, the gloss τοῦτ᾽ ἔστιν is that of Hellenistic commentary: Philo uses it to gloss allegories. 1 Cor 10:4. The tradition of the well exists in Palestinian sources, but Paul does not speak of a well but of a spiritual πέτρα identified with Xristos, and that identification is Philo’s, for whom the rock is divine wisdom (Leg. all. 2.86). The very verse that is cited in favour of the Palestinian haggadah contains its Alexandrian gloss. Ephesians 4:8. It is the soundest case, and it falls outside: the authorship of the letter is disputed, and to the extent that the objector uses it, we are dealing with an isolated reading. Verdict: concession on the technique, rejection on the origin. That Paul argues by analogy and with a pesher gloss shows that he received a Jewish exegetical culture, not that he passed through the Beit HaMidrash.

Objection 7. “Paul cites the Hebrew”: citations that depart from the LXX

The objection. In several citations Paul departs from the Septuagint towards the Masoretic Text: ἐξήγειρά σε in Rom 9:17 (“I raised you up”, active, like the hiphil הֶעֱמַדְתִּיךָ of Exod 9:16, “I have made you stand”, as against the passive διετηρήθης, “you were preserved”, of the LXX); πέτραν σκανδάλου in Rom 9:33 (“a rock of offence”, like צוּר מִכְשׁוֹל of Isa 8:14, as against the πτώματι, “a fall”, of the LXX); and ὁ δρασσόμενος τοὺς σοφοὺς ἐν τῇ πανουργίᾳ αὐτῶν in 1 Cor 3:19 (“who catches the wise in their craftiness”, as against ὁ καταλαμβάνων σοφοὺς ἐν τῇ φρονήσει, “who seizes the wise in their prudence”, of Job 5:13 LXX). Therefore, it is said, he consulted the Hebrew.

The answer. I concede the three facts and dispute the inference. First, there existed a biblical Greek revised towards the Hebrew before and during Paul’s lifetime: the Greek scroll of the Minor Prophets from Naḥal Ḥever (8ḤevXIIgr), of between the first century BCE and the first century CE, shows the kaige tradition that Dominique Barthélemy described in Les devanciers d’Aquila (1963; edition by Emanuel Tov, DJD VIII, 1990). Second, the reading of 1 Cor 3:19 coincides with that of Theodotion in Job 5:13 (ὁ δρασσόμενος σοφοὺς ἐν πανουργίᾳ αὐτῶν): Paul cites a Greek already revised by another. Third, the contrary pattern exists and weighs more: Paul stays with the LXX against the Hebrew when the LXX serves him. In Rom 10:18 he cites Ps 18:5 LXX, εἰς πᾶσαν τὴν γῆν ἐξῆλθεν ὁ φθόγγος αὐτῶν (“their sound has gone out into all the earth”), while the Hebrew of Ps 19:5 says בְּכָל הָאָרֶץ יָצָא קַוָּם, “their line has gone out through all the earth”: the “sound” is the Greek translator’s, and Paul gives it as Scripture. In Rom 15:12 he cites Isa 11:10 as ἔσται ἡ ῥίζα τοῦ Ἰεσσαί, καὶ ὁ ἀνιστάμενος ἄρχειν ἐθνῶν, ἐπ᾽ αὐτῷ ἔθνη ἐλπιοῦσιν (“there shall be the root of Jesse, even he who rises to rule the nations; in him shall the nations hope”), when the Masoretic Text says that he will stand “as a banner of peoples” (לְנֵס עַמִּים) and that “the nations will seek” him (גּוֹיִם יִדְרֹשׁוּ): the “ruling” and the “hoping”, which sustain the argument of Romans 15, are the LXX’s. Fourth, what is missing remains missing: not a disputed reading, not a qeri/ketiv, not a “the Hebrew says”. Verdict: concession on the datum, rejection of the conclusion. An intermittent access to a revised Greek is not equivalent to consulting the Hebrew, and where the Hebrew and the Greek diverge on a point that matters, Paul stays with the Greek.

Objection 8. “Hebrew of Hebrews”, Pharisee and “Hebrew dialect”

The objection. Paul calls himself Ἑβραῖος ἐξ Ἑβραίων (Phil 3:5, “a Hebrew of Hebrews”), κατὰ νόμον Φαρισαῖος (“a Pharisee according to the law”) and ζηλωτὴς τῶν πατρικῶν παραδόσεων (Gal 1:14, “zealous for the traditions of the fathers”), and Luke shows him speaking Ἑβραΐδι διαλέκτῳ (“in the Hebrew dialect”) before the crowd (Acts 21:40; 22:2). N. T. Wright (Paul: A Biography, 2018) and Martin Hengel (The Pre-Christian Paul, 1991) maintain that these declarations describe a Pharisaic formation in Yerushalayim, and that “Hebrew” indicates competence in Hebrew or Aramaic.

The answer. I distinguish between what Paul declares and what he does. First, Ἑβραῖος in Hellenistic Greek is an ethnic and cultural label, not a linguistic certificate. In 2 Cor 11:22, Ἑβραῖοί εἰσιν; κἀγώ· Ἰσραηλῖταί εἰσιν; κἀγώ· σπέρμα Ἀβραάμ εἰσιν; κἀγώ (“Are they Hebrews? So am I. Are they Israelites? So am I. Are they the seed of Avraham? So am I”), the triad is one of belonging, and it answers rivals who claimed it beforehand. Second, κατὰ νόμον Φαρισαῖος defines an adscription to a school with respect to the Torah; it does not say where or with whom. Josephus (Vita 10–12) declares himself a Pharisee after trying the three schools and leaves, moreover, all the traces of the language in which he studied (glosses, etymologies, a Hebrew text of his own); Paul declares and leaves none. Third, the ζηλωτής of Gal 1:14 and the διώκων of Phil 3:6 describe a zeal, not a curriculum; the model is Pinchas, not Hillel. Fourth, the “Hebrew dialect” is Luke’s, not the letters’, and according to most specialists the Ἑβραΐς διάλεκτος of Acts designates Aramaic. Fifth, if Paul had been a disciple of Gamliel, the proof would have to be in his work, and what is offered, the list of Semitisms, is what we have just subjected to the four sieves. Verdict: the declaration is a datum, and it is answered with performance.

Objection 9. Paul’s Aramaic: the creed, ἀββά and μαρανα θά

The objection. There is Aramaic in Paul: ἀββά (Rom 8:15; Gal 4:6), μαρανα θά (1 Cor 16:22), Κηφᾶς, and the creed of 1 Cor 15:3–5 has Semitic features (καὶ ὅτι… καὶ ὅτι… καὶ ὅτι…, “and that… and that… and that…”; τῇ τρίτῃ ἡμέρᾳ, “on the third day”; ὤφθη, “he appeared”). Joachim Jeremias went further and proposed Aramaic originals for hymns such as Phil 2:6–11. If Paul preserves Aramaic, he spoke it; and if he spoke it, his formation was Palestinian.

The answer. First, ἀββά and μαρανα θά are liturgical formulae, and the first comes glossed in Greek in the same verse (ἀββὰ ὁ πατήρ, Rom 8:15). An Aramaic preserved as a sonorous fossil and with a gloss is that of a communal bilingualism, not that of one who thinks in that language; μαρανα θά is not even segmented with certainty (maran atha or marana tha), which suggests that Paul transmits it as sound. Second, 1 Cor 15:3–5 is what Paul says he received (παρέλαβον) and transmitted (παρέδωκα): the Semitisms of that text belong to the formula and to those who formulated it, and that is precisely what Paul declares (Gal 1:18, fifteen days with Kefa). It is the proof that he received tradition, not that he was formed in it. Third, the hypothesis of an Aramaic original is disputed both for the creed and for the hymns, and several “Semitic” features are Septuagintisms (ἐγήγερται κατὰ τὰς γραφάς, “he has been raised according to the Scriptures”, is biblical Greek). Fourth, the specialist verdict stands: van Unnik searched for Aramaisms in the letters, and Fitzmyer judges that he did not succeed in establishing a single secure one. Verdict: concession on the existence, rejection of the inference. Paul preserves communal Aramaic; there is no school Aramaic.

IV. Minor objections, in one line

“He has anacolutha, therefore he translates mentally.” The anacolutha (Rom 5:12–14; Gal 2:6–9) are normal in dictated Greek and in rhetoric; Fitzmyer concludes that Paul’s Greek is neither deficient nor a literal translation of Aramaic.

“Parallelism and chiasmus are Hebrew poetry.” They are also figures of Greek rhetoric (ἀντίθεσις, ἀντιμεταβολή); they do not discriminate.

“He uses ἐν with instrumental value, like the Hebrew בְּ.” The instrumental ἐν is attested in untranslated koine, and its frequency in the LXX is explained by the LXX (sieves 1 and 2).

“He writes in parataxis, like Hebrew.” His syntax is hypotactic: Rom 1:1–7 is a single sentence, and the typical syntactic Semitisms (καὶ ἐγένετο, the cognate dative) almost never appear in his prose.

“His name Σαῦλος is Hebrew.” The name appears only in Acts; in the seven undisputed letters it is always Παῦλος, a Latin cognomen, and his companion is Σιλουανός, not Σίλας.

V. Debate protocol: four questions

When someone cites a “Hebraism”, do not discuss it: pass it through the four questions, in this order.

“Where does it appear in untranslated Greek?” Ask for a papyrus, an inscription or a passage of Polybius or Epictetus, and supply one yourself if you have it. If it exists, it is not a calque.

“Is it in the Septuagint?” If it is, the Hebraism belongs to the book and not to the author: it proves reading, not formation.

“Did he receive it or write it?” If it lives in material that Paul says he received (παρέλαβον), it belongs to the tradition.

“What first-century parallel does it have in Hebrew or Aramaic?” The Mishnah, the Talmud and the late targumim do not count without an argument from dating; Qumran, Murabbaʿat and Bar Kokhba do.

And a fifth, that of frequency: “How many times, and where?” A Semitism that appears once, in a citation or in a formula, is not a habit of thought.

VI. What you must concede

Paul has Semitic colouring at the level of the LXX (lexicon and formulae) and handles it with ease.

There are at least two first-century Hebrew parallels so specific that they exclude chance: ἔργα νόμου / מעשי התורה and υἱοὶ φωτός / בני אור, both Qumranic.

His citations sometimes depart from the LXX towards the Hebrew, probably by way of an already revised Greek.

He preserves communal Aramaic (ἀββά, μαρανα θά, Κηφᾶς, Σατανᾶς, ἀρραβών).

He practises arguments by analogy that rabbinic exegesis also practises.

VII. What you must not assert

“Paul has no Semitisms at all.” It is false, and it loses the debate in the first exchange.

“Paul did not know Hebrew.” It is not demonstrable; what is shown is that there is no trace of that competence.

“Paul was not a Jew.” It is a separate thesis (Maccoby’s), and it ruins the credibility of the philological argument.

“The targumim are all late.” There are targumim of the first century BCE at Qumran (4QtgLev, 4QtgJob, 11QtgJob); what is late is the form of Onkelos and of Yonatan.

“Mishnaic Hebrew did not exist.” It existed in proto-Mishnaic form (4QMMT, the Copper Scroll, the Bar Kokhba letters).

Conclusion

The defensible position is narrower than the one usually enunciated, and for that reason it holds. It does not say that Paul lacks Semitisms. It says that those he has belong to kinds that do not prove formation in Yerushalayim (the Greek Bible, the received tradition, a pious vocabulary that circulated in Greek) and that, where Josephus, the natural control, leaves the unmistakable traces of one who studied in Hebrew and Aramaic (glosses such as σάββατα and κόρβαν, etymologies, a Hebrew text of his own: Ant. 1.33–34 and 6.171; Against Apion 1.167), Paul leaves none. Whoever wishes to refute it must show a feature that survives the four sieves. So far, criticism has not done so.

Minimal bibliography:

F. Blass, A. Debrunner and R. Funk, A Greek Grammar of the New Testament (1961); J. H. Moulton and N. Turner, A Grammar of New Testament Greek IV: Style (1976); M. Zerwick, Biblical Greek (1963); W. C. van Unnik, “Aramaisms in Paul”, Sparsa Collecta I (1973), 129–143; J. A. Fitzmyer, A Wandering Aramean (1979); D. Daube, appendix to E. G. Selwyn, The First Epistle of St. Peter (1946); C. F. D. Moule, An Idiom Book of New Testament Greek (1953); D. Daube, “Rabbinic Methods of Interpretation and Hellenistic Rhetoric”, HUCA 22 (1949); S. Lieberman, Hellenism in Jewish Palestine (1950); E. E. Ellis, Paul’s Use of the Old Testament (1957); R. Longenecker, Biblical Exegesis in the Apostolic Period (1975); D. Barthélemy, Les devanciers d’Aquila (1963); E. Tov, The Greek Minor Prophets Scroll from Naḥal Ḥever, DJD VIII (1990); D.-A. Koch, Die Schrift als Zeuge des Evangeliums (1986); C. D. Stanley, Paul and the Language of Scripture (1992); M. McNamara, Targum and Testament (1972); S. A. Kaufman, “Dating the Language of the Palestinian Targums” (1994); R. Bultmann, “Römer 7 und die Anthropologie des Paulus” (1932); M. Hengel, The Pre-Christian Paul (1991); N. T. Wright, Paul: A Biography (2018); Flavius Josephus (ed. Niese; Loeb).