The text of Genesis 3:16, Genesis 4:7 and Song of Songs 7:11: a text-critical analysis

Part of the Genesis 3:16b reception study. Prepared 17 Oct 2026. Working method: Tov, Textual Criticism of the Hebrew Bible (3rd ed.), with the usual external criteria (date, independence and genealogy of witnesses) and internal criteria (scribal error types, translation technique, harmonisation, lectio difficilior). All primary readings below were read directly in the project's research library unless marked otherwise; the record references for every reading are kept in the project archive. Modern scholarship is quoted only briefly.

0. Verdicts in brief

Bottom line for the study. Nothing in the textual evidence threatens the Hebrew word תשוקה at Gen 3:16. The textual question that the study lists as a possible game-changer (STRONGEST_READING "What would change this conclusion", item 2) is close to closed, and the item should be reworded (§9). One consequence cuts the other way: if the "turning" renderings are interpretations of a word the translators may not have known, they are evidence of early reading, not of what the Hebrew word means. The study's section (e) and its table should weigh them accordingly.

1. Scope and sources

1.1 Witnesses read directly

Web checks (bibliographic and public facts only): the contents of 4QGen^b (Dead Sea Scrolls Bible Translations page for 4Q2; ORION bibliography for Davila, DJD XII 31–38); 4Q422; Ben Sira editions.

1.2 Not available (and what each would add)

2. The witnesses: date, language, and genealogy

External evidence is weighed by independence, not by counting. The Hebrew and versional witnesses fall into these groups:

The key point for 3:16b: the "turning" renderings (C, C', the Peshitta, and some Aramaic) are all translations. Every witness that gives us Hebrew letters (A, B, and the Hebrew implied by D, G, H) reads תשוקה.

3. Gen 3:16: inventory and variant tables

Base text (the three Masoretic editions read agree except for accent glyph conventions):

אֶֽל־הָאִשָּׁ֣ה אָמַ֗ר הַרְבָּ֤ה אַרְבֶּה֙ עִצְּבוֹנֵ֣ךְ וְהֵֽרֹנֵ֔ךְ בְּעֶ֖צֶב תֵּֽלְדִ֣י בָנִ֑ים וְאֶל־אִישֵׁךְ֙ תְּשׁ֣וּקָתֵ֔ךְ וְה֖וּא יִמְשָׁל־בָּֽךְ׃

Masoretic accents. The atnah on בָנִ֑ים divides the verse into 16a (childbearing) and 16b (husband). In 16b the accents are pashta (אִישֵׁךְ֙), munah + zaqef qaton (תְּשׁ֣וּקָתֵ֔ךְ), tiphha (וְה֖וּא), silluq (בָּֽךְ). The main disjunctive (zaqef) falls after תְּשׁוּקָתֵךְ, so the tradition reads "and toward your husband [is] your tᵉšûqāh | and he shall rule over you." Gen 4:7b carries exactly the same accent sequence (pashta, munah + zaqef, tiphha, silluq: וְאֵלֶ֙יךָ֙ תְּשׁ֣וּקָת֔וֹ וְאַתָּ֖ה תִּמְשָׁל־בּֽוֹ). The Masoretes parsed the two clauses identically. (A difference between the electronic editions at 4:7, ֙ vs ֨ on וְאֵלֶיךָ, is an encoding convention for the doubled pashta, not a textual difference.)

Witness texts at 3:16 (as read):

• ⅏: ואל האשׁה אמר הרבה ארבה עצבוניך והריוניך בעצבון תלדי בנים ואל אישׁיך תשׁוקתיך והוא ימשׁל ביך
• 𝔊: καὶ τῇ γυναικὶ εἶπεν Πληθύνων πληθυνῶ τὰς λύπας σου καὶ τὸν στεναγμόν σου, ἐν λύπαις τέξῃ τέκνα· καὶ πρὸς τὸν ἄνδρα σου ἡ ἀποστροφή σου, καὶ αὐτός σου κυριεύσει.
• α´ σ´ θ´ (Field): 16a והרנך: α´ τὰς συλλήψεις σου; σ´ θ´ τὰς κυήσεις σου (OCR damaged; footnote sigla lost). 16b: α´ συνάφεια, σ´ ὁρμή (Cod. 127; Cod. X marg. "Α. Σ. συνάφεια, ὁρμή"); θ´ not preserved at 3:16. Jerome: Pro conversione Aq. societatem, Sym. appetitum vel impetum transtulerunt.
• 𝔏 (Sabatier): Et mulieri dixit: Multiplicans multiplicabo tristitias tuas et gemitum tuum: in tristitiis paries filios, et ad virum tuum conversio tua, et ipse tui dominabitur. Patristic forms: erit conversio tua ad virum tuum (Cyprian), erit tibi conversio ad virum tuum (Augustine, Gen. c. Man.), Sedulius Scottus' hybrid sub viri potestate eris, et erit reversio tua ad virum tuum.
• 𝔙: Mulieri quoque dixit: Multiplicabo aerumnas tuas, et conceptus tuos: in dolore paries filios, et sub viri potestate eris, et ipse dominabitur tui.
• 𝔖: ܘܠܐܢܬܬܐ ܐܡ̣ܪ. ܡܣܓܝܘ ܐܣܓܐ ܟܐ̈ܒܝܟܝ ܘܒܛܢܝ̈ܟܝ. ܘܒܟܐ̈ܒܐ ܬܐܠܕܝܢ ܒܢ̈ܝܐ. ܘܥܠ ܒܥܠܟܝ ܬܬܦܢܝܢ. ܘܗ̣ܘ ܢܫܬܠܛ ܒܟܝ. ("…your pains and your pregnancies … and to your husband you will turn, and he will rule over you")
• 𝔗O: לְאִתְּתָא אֲמַר אַסְגָּאָה אַסְגֵּי צַעֲרַיְכִי וְעִדּוּיַיְכִי בְּצְעַר תֵּלְדִי בְנִין וּלְוַת בַּעְלִךְ תְּהֵא תִּאוּבְתִּיךְ וְהוּא יִשְׁלַט בִּךְ
• 𝔗J: לְאִינְתְּתָא אָמַר אַסְגָא אַסְגֵי סְגוּפַיִךְ בַּאֲדַם בְּתוּלִין וְעִידוּיֵךְ בְּצַעַר תַּלְדִין בְּנִין וּלְוַת בַּעֲלִיךְ תְּהֵי מַתְוָךְ וְהוּא יְהֵי שַׁלִיט בִּיךְ לְמִזְכֵּי וּלְמֶחְטֵי
• 𝔗N (second-hand): ולות בעלך יהוי מתביך "to your husband will be your turning back" (Rösel, via Macintosh 2016, 375; Lohr 2011, 235–36); margin: "your safety" (רחצונך; McNamara via Lohr and Condren).
• Saadia: וקאל ללאמרה לאכ̇תרן מן משקתך וחמלתך במשקה תלדין אלאולאד ואלי בעלך יכון קיאדך והו אלמתסלט עליך באלאכ̇תיאר ("…your hardship and your pregnancy … and to your husband shall be your qiyād [leading, being led], and he is the one given authority over you, by choice")
• 𝔅o: ⲟⲩⲟϩ ⲡⲉϫⲁϥ ⲛ̀ϯⲥ̀ϩⲓⲙⲓ ϫⲉ ϧⲉⲛ ⲟⲩⲁ̀ϣⲁⲓ ⲉⲓⲉ̀ⲑ̀ⲣⲉ ⲛⲉⲙⲕⲁϩⲛ̀ϩⲏⲧ ⲁ̀ϣⲁⲓ ⲛⲉⲙ ⲡⲉϥⲓⲁϩⲟⲙ … ⲟⲩⲟϩ ⲉϥⲉ̀ϣⲱⲡⲓ ϧⲁ ⲡⲉϩⲁⲓ ⲛ̀ϫⲉ ⲡⲓϫⲓⲛⲧⲁⲥⲑⲟ ⲟⲩⲟϩ ⲛ̀ⲑⲟϥ ⲉϥⲉ̀ⲉⲣϭⲟⲓⲥ ⲉ̀ⲣⲟ ("…grief … groaning … and the returning shall be toward your husband, and he shall be lord over you"; ϫⲓⲛⲧⲁⲥⲑⲟ, nominal of ⲧⲁⲥⲑⲟ "return, turn back") — a rendering of 𝔊 (λύπας, στεναγμόν, ἀποστροφή).
• Philo, Leg. 3.220: „Καὶ πρὸς τὸν ἄνδρα σου“ … „ἡ ἀποστροφή σου“ (= 𝔊).
• Epiphanius, Pan. 49.3.2 and on 1 Cor 14:34; Severian; Chrysostom, Hom. Gen. 17: πρὸς τὸν ἄνδρα σου ἡ ἀποστροφή σου καὶ αὐτός σου κυριεύσει (= 𝔊; no Greek variant seen in any Father read).
• Josephus, Ant. 1.49: paraphrases only the childbearing pains; omits 16b entirely.
• Jubilees 3:24 (Ethiopic, Charles): "thy return shall be unto thy husband, and he will rule over thee"; Ethiopic megbāʾ "place of refuge/return" (VanderKam, via Lohr). The Hebrew Jubilees fragments from Qumran do not preserve 3:24 (Lohr n. 24).
• Apocalypse of Moses 25 (MSS A and B only, per Charles's note): "and thou shalt turn again to thy husband, and he shall be thy lord" (ANF English). Greek not legible in the Tischendorf OCR module.
• Genesis Rabbah 20:7: וְאֶל אִישֵׁךְ תְּשׁוּקָתֵךְ (= 𝔐), linked with Ps 65:10 וַתְּשֹׁקְקֶהָ (root שׁקק) and played on with תָּשׁוּבִי לִתְשׁוּקָתֵךְ "you will return to your desire".
• Dead Sea Scrolls: no biblical manuscript preserves Gen 3:16 (Lohr: the Qumran Genesis manuscripts "are missing 3:15–4:1"; 4QGen^k = 4Q10 has 3:1–2; 1QGen = 1Q1 has 3:11–14; 4QGen^b = 4Q2 has 4:2–11). A possible allusion to 3:16b in 4QInstruction (4Q416 2 iv 3–4, "toward you is her [desire]") depends on a reconstruction (Lohr n. 23; Macintosh 2016, 379).

Unit 1: "To the woman" (אֶל־הָאִשָּׁה)

BHS: "l c ⅏𝔊𝔖𝔘 וְאֶל" (read with them). Evaluation: The connective in SP and the versions smooths an asyndetic opening; 3:14 and 3:17 open with waw. Versions add "and" freely, so their support is weak. Asyndeton is the harder reading; SP harmonises often. Verdict: retain 𝔐 (moderate confidence). No effect on meaning.

Unit 2: "your pain(s)" (עִצְּבוֹנֵךְ)

All witnesses read a noun "pain/toil": ⅏ עצבוניך (Samaritan spelling of the 2fs suffix, as in אישׁיך, ביך); 𝔊 τὰς λύπας σου (plural); 𝔖 ܟܐ̈ܒܝܟܝ (pl.); 𝔗O צַעֲרַיְכִי; 𝔙 aerumnas tuas; Saadia משקתך. Skinner wonders whether 𝔊 read a plural (עצבוניך?); the plural is more likely Greek style. Verdict: no variant (orthographic only). Very high.

Unit 3: "and your pregnancy" (וְהֵרֹנֵךְ)

Evaluation. Every witness except 𝔊 and its daughters reads "pregnancy/conception". The Greek translator, faced with a unique contracted form, connected it with a sound-alike root (הגה "moan") or simply rendered by context (pain and groaning in labour). This is the same translator, in the same verse, who renders תשוקה by ἀποστροφή: a direct sign that he guessed at rare words. The BHS note "frt l c ⅏ וְהֶרָיוֹנֵךְ" is a vocalisation/morphology choice with no change of meaning. Hendiadys: "your pain and your pregnancy" is widely read as a hendiadys ("your pain in pregnancy", i.e. in childbearing), which the following בְּעֶצֶב תֵּלְדִי בָנִים makes explicit. That is a matter of syntax, not of text; an alternative reading (two burdens: toil, cf. 3:17 עִצָּבוֹן of the man, and frequent pregnancy) is interpretive and not decided by the textual evidence. Verdict: retain 𝔐 (high). 𝔊 = exegetical rendering.

Unit 4: "in pain" (בְּעֶצֶב)

⅏ בעצבון (harmonised with עצבונך and 3:17 בְּעִצָּבוֹן); 𝔊 ἐν λύπαις (repeats λύπ-, as Skinner notes; cannot distinguish); 𝔖 ܘܒܟܐ̈ܒܐ (with "and"); 𝔙 in dolore; 𝔗O בְּצְעַר. Verdict: 𝔐 בְּעֶצֶב retained (moderate-high). SP = harmonisation. The SP reading, if original, would tie the woman's "pain" more closely to the man's "toil" in 3:17; since SP's harmonising habit is well known, it should not be used that way.

Unit 5: "and to/toward your husband" (וְאֶל־אִישֵׁךְ)

Verdict: no variant. Very high.

Unit 6: "your tᵉšûqāh" (תְּשׁוּקָתֵךְ): the main unit

6a. The question: did the Old Greek read תשובתך?

Arguments that the translator's Hebrew read תשובה (Ball, Nestle; considered by Skinner, Licht; "doubtful" in CATSS): 1. ἀποστροφή and ἐπιστροφή are, in the LXX, overwhelmingly equivalents of שׁוב and its derivatives. ἀποστροφή renders תְּשׁוּבָה at 1 Sam 7:17 (CATSS), מְשׁוּבָה at Jer 8:5, inf. abs. of שׁוב at Deut 22:1 and Ezek 33:11, שְׁבִית at Ezek 16:53 (verified in CATSS). Outside our three verses no LXX ἀποστροφή renders a word for "desire". 2. The Greek tradition is consistent: ἀποστροφή at Gen 3:16 and 4:7, ἐπιστροφή at Song 7:11 (a different translator). 3. The Peshitta, a translation from Hebrew, also uses "turn" (pny) in all three places; Onqelos (on one reading) and Neofiti use the Aramaic "return" root. 4. A sense "return" makes a plausible Hebrew wordplay with תָּשׁוּב in 3:19 (Nestle; Lohr's inclusio).

Arguments that the translator read תשוקה and interpreted it (Bergmeier, Wevers, Rösel, Macintosh, Lohr, Joüon): 1. Every Hebrew witness reads ק. MT; the Samaritan written text and, independently, the Samaritan oral reading tēšūqāttək; the Hebrew behind Aquila, Symmachus and the Vulgate (none of which renders "turning"); Genesis Rabbah; and at 4:7 the Qumran manuscript 4QGen^b. The Samaritan and Masoretic lines separated by the 2nd–1st c. BCE, so their agreement takes the reading back close to the date of the Septuagint. 2. The word existed in Second Temple Hebrew. תשוקה occurs in non-biblical Qumran texts (1QS 11:22; 1QM 13:12; 15:10; 17:4; 4Q418 168 3; 4Q446 3 1; 6Q18 2 4; Lohr n. 23), and in rabbinic Hebrew (Gen. Rab. 20:7 "four desires", with the root שׁקק). This removes Nestle's main premise (1904) that a Hebrew word of this root is attested only in these three verses. Macintosh (2016, 375–79) also rejects Yalon's and Licht's emendation of the Scroll passages to תשובה. 3. The same error three times is improbable (Joüon 1908). A ב/ק confusion would have to occur independently in Gen 3:16, Gen 4:7 and Song 7:11 in the translators' copies while never entering any surviving Hebrew copy. The simpler explanation of the Greek consistency is that the Genesis translator settled on an equivalent for an unknown word and kept it at 4:7 (Bergmeier: once found, "konnte an späteren Stellen … zurückgegriffen werden"), and that the later Song translator knew the Greek Pentateuch's equivalent. 4. ב and ק are not a common graphic pair in the Jewish square script of the 3rd–2nd c. BCE (contrast ב/כ, ד/ר, ה/ח, ו/י). CATSS itself labels the retroversion "doubtful". 5. Translation technique in the same verse. The same translator renders the unique הֵרֹנֵךְ by στεναγμόν (CATSS: "etymological exegesis"). At 4:7 he re-vocalises and re-divides much of the verse (διέλῃς, ἥμαρτες, ἡσύχασον; CATSS: "different vocalization"). He is demonstrably willing to construe rare or obscure Hebrew by sound and context. A rendering of תשוקה via the familiar שוב family is of a piece with that. 6. A Greek-internal motive is available. Three verses later he renders שׁוּבְךָ by ἀποστρέψαι (3:19). Using ἀποστροφή in 3:16 creates the woman→man, man→earth echo (Bergmeier relates it to the woman taken from the man, 2:21–24). Rösel (per Macintosh) thinks the translator substituted the familiar word deliberately to obtain a smooth sense ("refuge"); Wevers (per Macintosh and Condren) that he avoided a sexual sense. 7. ἀποστροφή can render the sense, not only the etymology. In Greek, ἀποστροφή πρός τινα is "recourse, resort to someone" (LSJ; NETS "recourse"; Macintosh: "attention to … a person"). That is a defensible sense-translation of a word meaning "devoted attention, focus toward", which would make the "variant" question moot. 8. The Peshitta does not settle it. Peshitta Genesis follows MT, not LXX, in the surrounding words (3:16a "pregnancies"; 4:7a "if you do well, I accept … at the door sin lies"), so it worked from Hebrew. Its "turn" for תשוקה may be a lexical borrowing from Greek for a hard word (a known Peshitta habit, debated by Weitzman and Lund), or a shared Jewish exegetical association (Maori 1995, as summarised by Lohr: such agreements "do not necessarily confirm variants"). Note too that the Peshitta's own 3:19 uses a different verb (ܬܗܦܘܟ) for שׁוב, so its pny at 3:16 is not an echo of its 3:19. Aramaic תְּיוּבְתָּא also means "reply, response", which (as Bergmeier argued for Onqelos) shows how easily an Aramaic-speaking exegete could hear "return" in a word near תְּאוּבְתָּא "longing". 9. Lectio difficilior. If two readings existed, the rare תשוקה would be the harder and the common תשובה the easier; scribes replace rare words with common ones far more often than the reverse.

Bergmeier's Aramaic hypothesis (1967): the translator relied on an Aramaic (proto-Onqelos) gloss תיאובתיך ("your longing"), misheard or misread it as the near-homograph תיובתיך ("your return"). The mechanism works for Aramaic readers, and it explains why "return" appears in the Targum tradition. As an account of a 3rd-century-BCE Alexandrian translation it requires a written Aramaic Targum of Genesis at that date, which is not attested. It should be treated as possible, not demonstrated. (Bergmeier's n. 3, that the Vulgate renders תשוקה "korrekt mit desiderium" at 3:16 and 4:7, is wrong: the Vulgate has sub viri potestate eris and appetitus.)

Verdict on the LXX question. The weight of evidence favours an interpretive rendering by a translator who read תשוקה (or, at least, that no Hebrew תשובה ever stood in the transmitted text of Genesis). I would put the odds at roughly 3:1 against a תשובה Vorlage; confidence moderate to high. What remains genuinely open is why he chose ἀποστροφή: an etymological association with שוב, a contextual guess shaped by 3:19 and 2:21–24, an existing Jewish exegetical association (later visible in the Aramaic "return" renderings), or a sense-translation "recourse". These are not mutually exclusive, and they cannot be decided without the Göttingen apparatus and Wevers's notes.

Verdict on the Hebrew of 3:16b (unit 6). תְּשׁוּקָתֵךְ is secure. Even if the Septuagint's exemplar had read תשובתך, it would be a secondary (easier) reading in one copy, outweighed by two independent Hebrew lines (proto-MT and Samaritan), the Hebrew implied by every Jewish reviser, and the Second Temple currency of the word. Confidence very high.

What the versions do and do not show about meaning. No ancient version renders תשוקה at 3:16 as hostility or control. They render it as turning/recourse (𝔊 and daughters; 𝔖; 𝔗N), joining (α´), eager impulse (σ´), longing (𝔗O as vocalised in the edition read; probably 𝔗J), or they restate the clause as subjection (𝔙; Saadia). Because the most widespread rendering may rest on a guess at an unfamiliar word (Bergmeier, Rösel), its value for the lexical meaning of תשוקה is limited. It is good evidence for how the verse was read from the 3rd c. BCE on.

Unit 7: "and he will rule over you" (וְהוּא יִמְשָׁל־בָּךְ)

Verdict: no variant. Very high. The 𝔗J and Saadia expansions are early exegesis (both qualify the rule morally and both attach the same qualification to 4:7) and belong in the reception data, not the apparatus.

4. Gen 4:7

Base text: הֲל֤וֹא אִם־תֵּיטִיב֙ שְׂאֵ֔ת וְאִם֙ לֹ֣א תֵיטִ֔יב לַפֶּ֖תַח חַטָּ֣את רֹבֵ֑ץ וְאֵלֶ֙יךָ֙ תְּשׁ֣וּקָת֔וֹ וְאַתָּ֖ה תִּמְשָׁל־בּֽוֹ׃

Witness texts (as read): ⅏ הלוא אם תטיב שׂאת ואם לא תטיב לפתח חטאת רבץ ואליך תשׁוקתו ואתה תמשׁל בו (= 𝔐, defective תטיב). 4QGen^b: preserves parts of 4:7 including "its desire is for you" ([ו]תשוקתו per Lohr 2011, 232), = 𝔐; it also lacks the 4:8 plus "let us go out to the field" with 𝔐 against ⅏ and 𝔊 (DSS Bible Translations). 𝔊: οὐκ, ἐὰν ὀρθῶς προσενέγκῃς, ὀρθῶς δὲ μὴ διέλῃς, ἥμαρτες; ἡσύχασον· πρὸς σὲ ἡ ἀποστροφὴ αὐτοῦ, καὶ σὺ ἄρξεις αὐτοῦ (= 1 Clem 4:4–5 verbatim). σ´: … παρὰ θύραν ἁμαρτία ἐγκάθηται· καὶ πρὸς σὲ ἡ ὁρμὴ αὐτῆς, ἀλλ’ ἐξουσιάσεις αὐτῆς. θ´: … ἐπὶ θύρᾳ ἁμαρτία ἐγκάθηται· καὶ πρὸς σὲ ἡ ἀποστροφὴ αὐτοῦ, καὶ σὺ ἄρξεις αὐτοῦ (Codd. 127, 135) but πρὸς σὲ ὁρμὴ αὐτοῦ, καὶ ἄρξεις αὐτοῦ (Montfaucon's schedae). α´: only "(οὐκ) ἐὰν ἀγαθύνῃς, ἀρέσεις" preserved. Jerome QHG: et ad te societas eius: sed tu magis dominare eius (imperative), and the false claim that ḥaṭṭāʾt is masculine in Hebrew. 𝔏: Quiesce; ad te enim conversio eius, et tu dominaberis illius (Augustine, Civ. 15.7; Irenaeus' Latin). 𝔙: sed sub te erit appetitus eius, et tu dominaberis illius. 𝔖: ܥܠ ܬܪܥܐ ܚܛܗܐ ܪܒܝܥ. ܐܢܬ ܬܬܦܢܐ ܠܘܬܗ ܘܗ̣ܘ ܢܫܬܠܛ ܒܟ ("at the door sin lies; you will turn to it, and it will rule over you"). 𝔗O: repentance paraphrase (… חֶטְאָךְ נְטִיר … אִם לָא תְתוּב וְאִם תְּתוּב יִשְׁתְּבֵק לָךְ) with no equivalent of 4:7b except the root תוב. 𝔗J: … וּלְוָותָךְ יְהֵי מַתְוַיֵהּ וְאַנְתְּ תְּהֵי שַׁלִיט בֵּיהּ בֵּין לְמִזְכֵּי בֵּין לְמֶחְטֵי (antecedent: the evil inclination). Saadia: … ראבץ̇ ואליך קיאדה ואנת אלמסלט עליה באלאכ̇תיאר. 𝔅o: ⲡⲉϥϫⲓⲛⲧⲁⲥⲑⲟ ⲉϥⲉ̀ϣⲱⲡⲓ ϩⲁⲣⲟⲕ ⲟⲩⲟϩ ⲛ̀ⲑⲟⲕ ⲉⲕⲉ̀ϣⲱⲡⲓ ϩⲓϫⲱϥ ("his returning will be toward you, and you will be over him"; from 𝔊). GenR 20:7: וְאֵלֶיךָ תְּשׁוּקָתוֹ, antecedent the evil inclination.

Variation units

The grammar problem and the conjectures

חַטָּאת is feminine; רֹבֵץ and the suffixes of תְּשׁוּקָתוֹ and בּוֹ are masculine. No witness has a consonantal text that removes the mismatch. The proposed solutions, all conjectural or interpretive:

• רֹבֵץ as a substantive, "a croucher" (often compared with Akkadian rābiṣu, a lurking demon), with the masculine suffixes referring to it (GKC §145u; Castellino 1960 treats "sin" as effectively masculine by personification).
• Schlimm (2012): repoint לְפֶתַח חַטָּאת as a construct, "at the entrance of sin", so that the masculine רֹבֵץ is the subject. No consonantal change.
• Crouch (2011), after H. Duhm (1904): חַטָּאת is an interpolative gloss; remove it and רֹבֵץ is the subject.
• Ehrlich (1908): redivide חַטָּאת רֹבֵץ → חַטָּא תִּרְבַּץ; suffixes refer to Abel.
• The suffixes refer to Abel (𝔊, where "sin" is a verb; Calvin; Ehrlich; Ben Yashar; Wöller).
• BHS (Eissfeldt) 4:7 b–b: "orig prb ואלַיִךְ תשׁוקתו וְאַתְּ תמשְׁלִי בוֹ, gl ad 3,16", i.e. the clause was probably originally a gloss on 3:16 addressed to a woman ("toward you [f.] is his tᵉšûqāh, and you [f.] shall rule over him"), later displaced into 4:7. At 3:16 BHS cross-refers "d–d cf 4,7b–b". This conjecture has no manuscript support, but it records the long-standing view (also Westermann's "mechanical citation") that 4:7b is secondary and modelled on 3:16b.
• Westermann (1974): the whole of 4:6–7 is an expansion; "erhebliche Textverderbnis".

Each ancient version resolved the mismatch differently (sin as subject: σ´, Jerome, Augustine; the evil inclination: 𝔗J, GenR; Abel: 𝔊, Chrysostom; Cain turning to sin: 𝔖). Verdict: the consonantal text of 4:7b is secure; its syntax and referent are not. Any reading of 3:16 that depends on a specific construal of 4:7 depends on a conjecture or an interpretation, not on a settled text. Confidence that the text is as 𝔐: high. Confidence in any single solution of the grammar: low.

For the 3:16 question: 4:7 provides no independent Hebrew evidence about תשוקה. The Greek ἀποστροφή at 4:7 is the same translator reusing his 3:16 equivalent, not a second witness. Symmachus is the one ancient witness that makes "sin" the subject with a feminine pronoun and renders ὁρμή, an eager impulse or onset.

5. Song of Songs 7:11 (Eng. 7:10)

Base: אֲנִ֣י לְדוֹדִ֔י וְעָלַ֖י תְּשׁוּקָתֽוֹ׃ ס

Evaluation. The conjecture וְאֵלַי harmonises the Song with Gen 3:16 and 4:7. No witness supports it: 𝔊's ἐπί and 𝔖's ܥܠ both reflect עַל. The Greek Song translator is extremely literal (NETS introduction), which is the strongest single argument that his copy might have read תשובתו; but a literal translator who did not know a rare word could as easily take over the equivalent already established in the Greek Pentateuch, adjusting the prefix (ἐπι- "toward" rather than ἀπο-, which can suggest "away") to suit a lover. Verdict: 𝔐 secure (high).

6. Early rewritings and quotations (reception, not text)

• Jubilees 3:24 ("thy return shall be unto thy husband") survives only in Ethiopic, which was translated from Greek. Its "return" therefore most probably entered at the Greek stage under 𝔊's influence (Bergmeier) and cannot be cited as a 2nd-century-BCE Hebrew witness. The Qumran Hebrew Jubilees fragments do not preserve 3:24.
• Apocalypse of Moses 25 (MSS A, B): "thou shalt turn again to thy husband": a Greek retelling dependent on 𝔊.
• Philo Leg. 3.220 quotes 𝔊 exactly; QG 1.49 not checked (Armenian).
• Josephus, Ant. 1.49, omits 3:16b.
• 1 Clement 4:4–6 (c. 96) quotes Gen 4:6–8 in the exact LXX form, including ἀποστροφή: the earliest dated Christian witness to the Greek of 4:7.
• New Testament: no quotation of 3:16b. Possible allusions (1 Cor 14:34 "as the law also says"; 1 Tim 2:11–14; Eph 5:22–24) carry no wording of the clause and have no text-critical value here.
• 4QInstruction (4Q416 2 iv 3–4): possibly the earliest reuse of 3:16b ("toward you is her [desire]") in a marriage context, but the key word is restored (from 1Q26); García Martínez restored לבבה instead (Macintosh n. 41). Not usable as textual evidence.
• Genesis Rabbah 20:7 quotes 𝔐's wording at all three places and links תשוקה with שׁקק (Ps 65:10), while playing on תשובה ("you will return to your desire"). So the rabbis read ק and knew the two words as distinct.

7. What is established, what is judgment, what could not be checked

Firmly established (from texts read directly)

1. MT, the Samaritan Pentateuch and Genesis Rabbah read תשוקה at 3:16; MT, SP and (per Lohr) 4QGen^b read it at 4:7; MT and GenR at Song 7:11. SP differs only in spelling (אישׁיך, תשׁוקתיך, ביך) at 3:16b.
2. BHS records no Hebrew manuscript variant for any word of 3:16b, 4:7b or Song 7:11b. Its only proposals there are conjectures (4:7b as a displaced gloss; Song 7:11 וְאֵלַי).
3. The LXX has ἀποστροφή at Gen 3:16 and 4:7 and ἐπιστροφή at Song 7:11; every Greek Father read quotes ἀποστροφή at 3:16. Field (who reports hexaplaric evidence, not full manuscript collations) records no Greek variant for ἀποστροφή at 3:16.
4. Aquila (συνάφεια) and Symmachus (ὁρμή) do not render "turning" at 3:16; Symmachus renders ὁρμή at 4:7 with a feminine pronoun. Theodotion's 4:7 reading is contested (ἀποστροφή in Codd. 127, 135; ὁρμή in Montfaucon's schedae).
5. The Old Latin (conversio) and the Bohairic Coptic (ϫⲓⲛⲧⲁⲥⲑⲟ "returning") render the LXX; both also follow the LXX in 3:16a ("groaning").
6. The Peshitta uses the root pny "turn" at all three places; at 4:7 it makes Cain the subject of the turning and copies 3:16b's rule clause ("and it will rule over you").
7. The LXX's "groaning" at 3:16a stands alone against MT, SP, the Three, Peshitta, Targums, Vulgate and Saadia ("pregnancy/conception").
8. No Qumran biblical manuscript preserves Gen 3:16 or Song 7:11 (Lohr 2011 n. 22, consistent with the published contents of 4Q1–4Q10 and 1Q1).

Scholarly judgment (argued above; others may weigh differently)

1. ἀποστροφή is an interpretation of תשוקה, not a witness to תשובה (moderate–high).
2. The Peshitta's and the Aramaic "return" renderings are exegesis, not variants (moderate–high).
3. 3:16a: retain MT וְהֵרֹנֵךְ and בְּעֶצֶב; the opening asyndeton is original (moderate–high).
4. 4:7b is textually secure, but its grammar is unresolved; BHS's displaced-gloss conjecture is possible but unprovable.

Could not be checked

Göttingen Genesis and Wevers's Notes; BHQ Genesis; Kennicott and de Rossi; Sperber's Onqelos; Neofiti and its margins; Fragment Targums; Samaritan Targum (Nestle's claim of a תוב form is unverified); Leiden Peshitta; Beuron Vetus Latina; Ethiopic, Armenian, Sahidic and Syro-Hexaplaric texts (known only second-hand, partly through Bushnell 1921); Field vol. 2 (Song); DJD XII, XIII, XXXIV; Hendel 1998; Philo QG 1.49; Masorah of L; the Greek of ApMos 25; Targum Song 7:11.

8. Open questions (in order of value)

1. Göttingen LXX and Wevers's Notes (1993, pp. 45, 54–56). Any Greek manuscript with ἐπιστροφή or another noun at 3:16/4:7; and Wevers's own account of the translator's choice.
2. Sperber's Onqelos. Which spelling (תיובתיך / תאובתיך) the best Yemenite and Babylonian-vocalised manuscripts have. This decides whether Onqelos is a "return" or a "longing" witness.
3. Neofiti 1 and the Samaritan Targum (Tal). To confirm the Palestinian Aramaic "return" renderings first-hand.
4. Hendel, The Text of Genesis 1–11, for his decisions at 3:16 and 4:7.
5. BHQ Genesis (Tal 2016), p. 85 commentary, to quote its judgment exactly rather than through Condren.
6. DJD XII (4QGen^b) to record the exact letters preserved at 4:7.
7. Sirach: an earlier draft of the study listed Ben Sira among the occurrences of תשוקה. None of the sources read here gives one (Lohr's list of 7 + 4 Qumran cases; Macintosh's list; HALOT), so it has been dropped. Condren 2024 and a Ben Sira concordance should confirm.

9. Sources consulted

Primary texts: Biblia Hebraica Stuttgartensia with its apparatus (Genesis ed. O. Eissfeldt; Song ed. F. Horst); the Westminster and ETCBC Hebrew texts; the Samaritan Pentateuch; the Septuagint (Rahlfs; Swete) and NETS with its notes; the CATSS MT–LXX parallel alignment (Tov and Polak); F. Field, Origenis Hexaplorum quae supersunt I (1875); P. Sabatier, Bibliorum sacrorum latinae versiones antiquae I (1743); the Clementine Vulgate; the Peshitta (ETCBC); Targums Onqelos and Pseudo-Jonathan; Saadia Gaon, Tafsīr; the Bohairic Coptic; Philo, Legum allegoriae 3.220; Josephus, Antiquities 1.49; 1 Clement 4:4–6; R. H. Charles, Apocrypha and Pseudepigrapha II (Jubilees 3:24; Apocalypse of Moses 25); Genesis Rabbah 20:7; Epiphanius, Severian of Gabala and Chrysostom (Hom. Gen. 17).

Lexica: HALOT; Brown–Driver–Briggs.

Studies: R. Bergmeier, ZAW 79 (1967) 77–79; J. N. Lohr, JBL 130 (2011) 227–46 and CBQ 71 (2009); A. A. Macintosh, JSS 61 (2016) 365–87; E. Nestle, ZAW 24 (1904) 312–14; P. Joüon, MFOB 3 (1908) 330–31; A. B. Ehrlich, Randglossen I (1908); C. L. Crouch, ZAW 123 (2011) 250–58; M. R. Schlimm, ZAW 124 (2012) 409–15; G. R. Castellino, VT 10 (1960); J. Skinner, Genesis (ICC, 1910); J. C. Condren, JETS 60 (2017) 227–45.

Public reference pages (bibliographic facts only): Dead Sea Scrolls Bible Translations, 4Q2 (https://www.dssenglishbible.com/scroll4Q2.htm); ORION bibliography entries for DJD XII (4QGen^b), 4Q422 and Ben Sira editions.

← Back to the interactive site