וַיַּ֤רְא הָֽאֱלֹהִים֙ אֶֽת־מַ֣עֲשֵׂיהֶ֔ם כִּי־שָׁ֖בוּ מִדַּרְכָּ֣ם הָרָעָ֑ה וַיִּנָּ֣חֶם הָאֱלֹהִ֗ים עַל־הָרָעָ֛ה אֲשֶׁר־דִּבֶּ֥ר לַעֲשׂוֹת־לָהֶ֖ם וְלֹ֥א עָשָֽׂה׃
wayyarʾ hāʾĕlōhîm ʾet-maʿăśêhem kî-šābû middarkām hārāʿâ wayyinnāḥem hāʾĕlōhîm ʿal-hārāʿâ ʾăšer-dibber laʿăśôt-lāhem wəlōʾ ʿāśâ.
וירא האלהימ את מעשיהמ כי שבו מדרכמ הרעה וינחמ האלהימ על הרעה אשר דבר לעשות להמ ולא עשהParsing · 18 words
| וַיַּ֤רְא | רָאָה | conj + qal wayyiqtol 3ms | and he saw — Apocopated wayyiqtol of a III-he root; the subject follows. |
| הָֽאֱלֹהִים֙ | אֱלֹהִים | art + noun mp abs | the God — Articular form; the narrator's designation through 3:9–10. |
| אֶֽת | אֵת | direct-object marker | [obj.] |
| מַ֣עֲשֵׂיהֶ֔ם | מַעֲשֶׂה | noun mp cstr + 3mp suffix | their deeds |
| כִּי | כִּי | conj | that — Epexegetical, specifying the content of what God saw. |
| שָׁ֖בוּ | שׁוּב | qal perfect 3cp | they turned back |
| מִדַּרְכָּ֣ם | דֶּרֶךְ | prep מִן + noun bs cstr + 3mp suffix | from their way — OSHB codes דֶּרֶךְ as of both genders (bs); the feminine adjective here fixes it as feminine. |
| הָרָעָ֑ה | רַע | art + adjective fs abs | the evil |
| וַיִּנָּ֣חֶם | נָחַם | conj + niphal wayyiqtol 3ms | and he relented |
| הָאֱלֹהִ֗ים | אֱלֹהִים | art + noun mp abs | the God |
| עַל | עַל | prep | concerning |
| הָרָעָ֛ה | רָעָה | art + noun fs abs | the evil |
| אֲשֶׁר | אֲשֶׁר | relative particle | which |
| דִּבֶּ֥ר | דָּבַר | piel perfect 3ms | he had spoken |
| לַעֲשׂוֹת | עָשָׂה | prep ל + qal infinitive construct | to do |
| לָהֶ֖ם | ל | prep ל + 3mp suffix | to them |
| וְלֹ֥א | לֹא | conj + negative particle | and not |
| עָשָֽׂה | עָשָׂה | qal perfect 3ms | he did — Same root as the infinitive and as מַעֲשֵׂיהֶם earlier in the verse. |
And the God saw their deeds, that they had turned back from their evil way; and the God relented concerning the evil which he had spoken to do to them, and he did not do [it].
And God saw what they did — that they turned back from their evil way — and God relented concerning the calamity he had spoken of doing to them, and he did not do it.
And God saw their deeds, that they turned from their evil way; and God relented of the harm he had said he would do to them, and he did it not.
Overview
The chapter’s resolution, and it is built out of repetitions rather than new vocabulary. Four things in the verse answer four things already said.
- God saw their deeds. Not their sackcloth, not their fasting, not their liturgy — מַעֲשֵׂיהֶם, what they did. The narrator chooses the word that the verse will close with (וְלֹא עָשָׂה, “and he did not do”), so that human doing and divine doing frame the sentence.
- They turned back from their evil way. This is a verbatim quotation of the king’s decree in v. 8 (וְיָשֻׁבוּ אִישׁ מִדַּרְכּוֹ הָרָעָה) — the narrator certifies that the command was obeyed.
- God relented concerning the evil. רָעָה stands twice in the verse: the Ninevites’ evil way, and the evil God had spoken of doing. One Hebrew word, two referents — their wickedness and his threatened calamity. The second is not moral evil in God; it is the disaster announced in 3:4.
- He did not do it. The blunt last clause, with no explanation and no softening. The threat of forty days (3:4) is not carried out.
The theological weight falls on וַיִּנָּחֶם, the niphal of נָחַם — God “relenting.” The verse makes no attempt to insulate the statement, and neither should the translation. The question of what divine relenting means, set against texts like Num 23:19 and 1 Sam 15:29, is the verse’s chief interpretive pressure point and is laid out below rather than resolved in the rendering.
Built onhalot · bdb · waltke-oconnor
Proposed — pending reviewHans Walter Wolff, Obadiah and Jonah: A Commentary, trans. Margaret Kohl, Continental Commentaries, Minneapolis: Augsburg, 1986.
Seeing deeds, not hearing words
מַעֲשֵׂיהֶם (“their deeds”) is the plural of מַעֲשֶׂה, the ordinary noun of action/work from עָשָׂה. The idiom רָאָה + מַעֲשֶׂה is not common; what is pointed here is the choice of this noun over any term for prayer, fast, or sackcloth, all of which vv. 5–8 had supplied in abundance. The narrator reports that God observed conduct.
כִּי after a verb of perception is epexegetical: it does not give a reason for the seeing but specifies its content — “he saw … namely, that they had turned back.” Joüon–Muraoka and Waltke–O’Connor both treat this as a standard object-clause use of כִּי after verbs of perception and speech.
The root עשׂה then recurs twice more (לַעֲשׂוֹת, עָשָׂה). Three occurrences of one root in a single verse, with the subject shifting from the Ninevites to God, is a deliberate structure: their doing is answered by his not-doing. English can preserve it with “deeds … to do … did not do,” and that is worth a slightly stiff line.
Built onhalot · bdb · jouon-muraoka · waltke-oconnor
שׁוּב: the hinge verb
שָׁבוּ (qal perfect 3cp) is the fourth occurrence of שׁוּב in nine verses — v. 8 (the king’s decree, וְיָשֻׁבוּ), v. 9 twice (יָשׁוּב of God, וְשָׁב from his burning anger), and here. The repetition is the chapter’s argument: the Ninevites turn, and God turns.
The Hebrew does not use a separate theological vocabulary for “repent.” שׁוּב is physical turning-back, applied to conduct by the prepositional phrase that follows it (מִדַּרְכָּם הָרָעָה, “from their evil way”). Rendering it “repented” would import a later religious category and, worse, would break the identity with v. 9, where the same verb has God as its subject and no one would render it “God repented of his anger.” “Turn back” at every occurrence keeps the chapter’s symmetry intact.
מִדַּרְכָּם — דֶּרֶךְ as a metaphor for conduct is thoroughly conventional in Hebrew (a dead metaphor by the standard of §8), but the quotation from v. 8 requires that the same English word be used, so the “way” image stands.
Built onhalot · bdb · dch
רָעָה twice over
The verse uses the root רעע twice, and the two uses are not the same thing.
- מִדַּרְכָּם הָרָעָה — here הָרָעָה is an attributive adjective (fs, agreeing with דֶּרֶךְ), “their evil way.” Moral wrongdoing.
- עַל־הָרָעָה אֲשֶׁר־דִּבֶּר לַעֲשׂוֹת — here הָרָעָה is a substantive, “the evil/calamity,” the announced disaster of 3:4. Harm, not wickedness.
Hebrew רָעָה covers both without strain, and Jonah exploits the span relentlessly: Nineveh’s evil (1:2), the sailors’ calamity (1:7–8), Nineveh’s evil way (3:8, 10), the threatened disaster (3:10), and Jonah’s own vexation (4:1, where the same noun describes his reaction to this very verse). Per the binding glossary decision, the literal line keeps “evil” at both occurrences so the wordplay survives; the modern line gives “evil way” and “calamity,” and this note records the gap.
The second הָרָעָה is the object of דִּבֶּר לַעֲשׂוֹת — what God had said he would do. In Hebrew idiom a judicial disaster brought by God is routinely called רָעָה (cf. Exod 32:14; Jer 18:8; 26:3). The word describes the effect on its recipients, not the moral character of the agent. An English rendering that reads “the evil that God intended” invites exactly the inference the Hebrew does not make.
Built onhalot · bdb · dch
נָחַם (niphal)
וַיִּנָּחֶם is the niphal of נָחַם, the third occurrence of the root in the chapter after וְנִחַם in v. 9. The niphal’s attested range in Biblical Hebrew runs along two lines:
- Be sorry, regret, change one’s course of action — of God in Exod 32:14; 2 Sam 24:16; Jer 18:8; 26:3, 13, 19; Amos 7:3, 6; Joel 2:13–14; and here. Where the complement is עַל + a named action, as here, the sense is the retraction of an announced course.
- Be comforted, console oneself — Gen 24:67; 2 Sam 13:39; Ps 119:52.
The construction here (עַל־הָרָעָה אֲשֶׁר־דִּבֶּר) fixes the first sense beyond reasonable doubt, and the following clause spells it out: וְלֹא עָשָׂה. The lexicons agree on the semantics of the construction; HALOT and BDB differ in how they subdivide the niphal’s senses, not in what this verse means.
On the English word: “relent” is chosen because it names a change in an announced course of action without either (a) asserting an emotional state that the Hebrew leaves unspecified, as “was grieved” would, or (b) implying prior moral fault, as “repented” unavoidably does in modern English. “Change his mind” is defensible for the sense but imports a cognitive idiom the Hebrew does not use. The rendering is applied identically at v. 9 and v. 10 so the Ninevites’ hope and its fulfilment match.
Built onhalot · bdb · dch · jouon-muraoka
הָאֱלֹהִים with the article
Both occurrences of the divine designation in this verse are הָאֱלֹהִים, articular, as in v. 9. This matters for Jonah’s larger pattern. In ch. 3 the Ninevites believe בֵּאלֹהִים (v. 5, anarthrous), call out אֶל־אֱלֹהִים (v. 8, anarthrous), and then the narrator speaks of הָאֱלֹהִים (vv. 9, 10). YHWH, by contrast, is the name used when Israel’s prophet and Israel’s God are in view (1:1–3; 3:1–3; 4:2–4).
Per the binding glossary decision, אֱלֹהִים is rendered “God” (of Israel’s God) and never collapsed into YHWH. The article is not represented in the modern line — “the God” is not English — but it is kept in the literal line so the reader can see what stands there. The force of the articular form in a Gentile setting is debated: some read it as “the true God,” others as the narrator’s ordinary designation in this section with no special emphasis. The distribution is reported here rather than adjudicated.
Built onwaltke-oconnor · jouon-muraoka · bdb
“Which he had spoken to do”
אֲשֶׁר־דִּבֶּר לַעֲשׂוֹת־לָהֶם — the relative clause has הָרָעָה as its head, with דִּבֶּר (piel perfect) governing an infinitive of purpose/content, לַעֲשׂוֹת. The construction דִּבֶּר + ל + infinitive construct for “said that he would do / spoke of doing” is well attested of divine announcements (cf. Exod 32:14, where the same idiom appears with the same two roots, דִּבֶּר … לַעֲשׂוֹת, of the רָעָה threatened against Israel).
The referent of הָרָעָה is the oracle of 3:4: forty days, and Nineveh overturned. The verse does not call that oracle false; it says God had spoken it and did not do it, and leaves the two facts side by side.
The final וְלֹא עָשָׂה has no object in Hebrew. The literal line supplies “[it]” in brackets; the modern line supplies “it” plainly. Nothing else should be supplied — the clause is as abrupt in Hebrew as it looks.
Built onwaltke-oconnor · jouon-muraoka · bdb
Divine relenting and divine constancy
This verse is the Hebrew Bible’s most compact statement that God announced a disaster and then did not bring it. It stands in the canon alongside texts that deny God relents at all.
- Num 23:19 — “God is not a man, that he should lie, nor a son of man, that he should relent (וְיִתְנֶחָם).”
- 1 Sam 15:29 — “the Glory of Israel does not lie or relent (יִנָּחֵם), for he is not a man, that he should relent.”
- Yet Exod 32:14; Jer 18:7–10; 26:3, 13; Amos 7:3, 6; Joel 2:13–14 all use the same niphal of God doing exactly what Jonah 3:10 reports — and Jer 18 states the principle as a policy: an announced judgment is conditional on the response of those addressed.
The conditional-oracle reading. Jeremiah 18:7–10 is taken as the governing hermeneutic: prophetic threats of judgment are implicitly conditional, so the non-fulfilment at Jonah 3:10 is the oracle working as designed, not a reversal in God. Num 23:19 and 1 Sam 15:29 then deny something different — that God lies or is capricious like a man — and the two sets of texts are about different questions.
The anthropomorphic-accommodation reading. נָחַם applied to God is figurative language for a change in God’s dealings consequent on a change in his creatures, not a change in his purpose; what alters is the human situation and therefore the appropriate divine response, and Scripture describes this from the human side. This reading is the classical one and is close to what Jer 18 and Joel 2:13 presuppose.
The plain-reading / divine-responsiveness position. The narrator says God saw, God relented, God did not do it, in sequence, as a real response to a real event. On this view the harmonizing moves above explain away the force of the narrative, which turns on God actually being movable by repentance — a point Jonah himself makes bitterly in 4:2 by quoting Exod 34:6–7 plus נִחָם עַל־הָרָעָה.
The Hebrew of 3:10 supports none of these against the others; it reports the sequence and stops. The rendering must do the same.
Two further observations the Hebrew does make. First, the verse withholds any statement that Nineveh’s turning merited the reprieve; God sees, and God relents, and the causal link is asserted by sequence (and by the epexegetical כִּי) rather than by any language of desert. Second, the king’s own hope in v. 9 was framed by מִי־יוֹדֵעַ (“who knows”) — the narrative grants no claim on God, and v. 10 is given as God’s freedom exercised, not as a transaction completed.
Built onhalot · bdb · dch
Proposed — pending reviewJack M. Sasson, Jonah: A New Translation with Introduction, Commentary, and Interpretation, Anchor Bible 24B, New York: Doubleday, 1990.
Echoes inside the Hebrew Bible
Three links are made by shared wording, not merely shared theme.
Exod 32:14. וַיִּנָּחֶם יְהוָה עַל־הָרָעָה אֲשֶׁר דִּבֶּר לַעֲשׂוֹת לְעַמּוֹ — God relenting concerning the רָעָה he had spoken of doing to his people, after Moses’ intercession at Sinai. The clause in Jonah 3:10 is all but identical, with the divine designation changed from יְהוָה to הָאֱלֹהִים and “his people” replaced by “them.” The parallel is close enough that Jonah’s narrator is working with a known formula; what Nineveh receives is what Israel received at Sinai.
Jer 18:7–10. The potter oracle states the mechanism: if a nation about which God has spoken disaster turns (שׁוּב) from its evil (רָעָה), God will relent (נִחַם) of the disaster he planned. Jonah 3:10 reads as that principle in narrative form, using the same three roots.
Joel 2:13–14. וְנִחָם עַל־הָרָעָה … מִי יוֹדֵעַ יָשׁוּב וְנִחָם. Joel supplies both the “who knows” of Jonah 3:9 and the relenting formula of 3:10, in a passage calling for fasting and turning. The direction of dependence between Jonah and Joel is disputed and should not be asserted; that the two share a formulaic cluster is clear.
Jonah 4:2 then quotes the same cluster in Jonah’s own mouth (וְנִחָם עַל־הָרָעָה), which is why 3:10 must be rendered in wording that 4:2 can match.
Built onbdb · halot · dch
Proposed — pending reviewPhyllis Trible, Rhetorical Criticism: Context, Method, and the Book of Jonah, Minneapolis: Fortress, 1994.
Textual note
The consonantal text is stable. records no variant of substance in this verse, and I am aware of no divergent Hebrew witness at any point in it; the Jonah material among the Dead Sea Scrolls is fragmentary and the state of this verse should be checked against the apparatus before any claim is made about it.
Two points of pointing and accentuation are worth noting rather than emending:
- מִדַּרְכָּם — OSHB parses דֶּרֶךְ as common gender (
Ncbsc). The adjective הָרָעָה is feminine singular, so the noun is feminine here; דֶּרֶךְ is attested with both genders, and the agreement in this verse settles it for this occurrence. - The ʾatnaḥ falls on הָרָעָ֑ה at the end of the first half, dividing the verse exactly between what God saw and what God did. The Masoretic phrasing supports the rendering’s em-dash structure, placing the כִּי-clause inside the first half.
Built onbhs · bhq
Draft — witness readings here are an AI’s recollection and have not been checked against the apparatus (BHQ, the Göttingen Septuagint, DJD). Verify before relying on them.
Septuagint
Jonah 3:10 renders closely: καὶ εἶδεν ὁ θεὸς τὰ ἔργα αὐτῶν, ὅτι ἀπέστρεψαν ἀπὸ τῶν ὁδῶν αὐτῶν τῶν πονηρῶν, καὶ μετενόησεν ὁ θεὸς ἐπὶ τῇ κακίᾳ ᾗ ἐλάλησεν τοῦ ποιῆσαι αὐτοῖς, καὶ οὐκ ἐποίησεν.
Three features deserve comment.
- ὁδῶν … πονηρῶν, plural against Hebrew singular מִדַּרְכָּם. Most likely translator’s idiom (Greek prefers the plural for habitual conduct) rather than a different Hebrew text; the of the Twelve pluralizes such collectives elsewhere. No Hebrew witness for a plural is known to me here.
- μετενόησεν for וַיִּנָּחֶם. The Greek verb is the ordinary equivalent for the niphal of נָחַם in this sense, and in Hellenistic usage μετανοέω means “change one’s mind, think again.” It does not yet carry the fully moralized sense of “repent” that later Christian usage gives it — a diachronic point that matters, because the same Greek root drives the New Testament’s reading of Jonah.
- κακία for the second הָרָעָה. Here the Greek is less helpful than the Hebrew: κακία leans toward moral badness, where Hebrew רָעָה in this clause means the announced calamity. The Greek choice preserves the Hebrew’s single-word repetition at the cost of sharpening the moral overtone. This is one place where translating from the Greek rather than the Hebrew would mislead.
The has misertus est (“he had mercy”) for the relenting, an interpretive softening that replaces a change of course with compassion; and non fecit for the last clause, matching the Hebrew’s bluntness. The Targum to the Prophets characteristically avoids predicating נָחַם of God directly and speaks instead of the decreed evil not being brought about; the precise wording should be checked against an edition before it is quoted.
Built onmuraoka-lxx · goettingen-lxx-xii
Proposed — pending reviewAlexander Sperber, The Bible in Aramaic, vol. 3: The Latter Prophets according to Targum Jonathan, Leiden: Brill, 1962.
Draft — witness readings here are an AI’s recollection and have not been checked against the apparatus (BHQ, the Göttingen Septuagint, DJD). Verify before relying on them.
Reception
What follows is how later readers used the verse, not evidence for what the Hebrew means.
New Testament. Jesus appeals to Nineveh’s response in Matt 12:41 / Luke 11:32: the men of Nineveh μετενόησαν at Jonah’s preaching and will rise in judgment against his hearers. The Greek verb is the ’s word from this verse’s own vocabulary field, though the Gospels apply it to the Ninevites’ turning (Jonah 3:5–8) rather than to God’s relenting. The argument is comparative — Gentiles turned at a lesser preaching — and does not comment on divine relenting.
Rabbinic tradition. Jonah 3 is read publicly as the haftarah for the afternoon service of Yom Kippur, and this verse is the reason: it is the canonical narrative of repentance accepted. The Mishnah’s discussion of fast-day liturgy (Taʿanit) cites the Ninevites as the model, with the point pressed that God saw their deeds — not their sackcloth and fasting. That emphasis is a reading of מַעֲשֵׂיהֶם in this verse and is widely echoed in later Jewish commentary.
Early Church. The Fathers read the verse chiefly through the problem of the unfulfilled oracle. Jerome, in his commentary on Jonah, treats the threat as conditional — the mercy was latent in the warning, since a warning given is already an opening — and this becomes the standard Latin position. Augustine takes up the same case in City of God (bk. 21), arguing that the prophecy was fulfilled in the sense that matters: the Nineveh that was threatened, the wicked city, was in fact overturned, though in its morals rather than its walls. Both readings are attempts to hold the oracle’s truthfulness together with v. 10, and both acknowledge that the verse creates the difficulty rather than solving it.
Later tradition. The verse is a standing proof text in discussions of divine immutability, cited on both sides: by those arguing that Scripture presents God as genuinely responsive, and by those arguing that such language is accommodation to human perspective. The debate is reception, not exegesis of the Hebrew, and belongs here rather than above.
Built onbdb
Proposed — pending reviewJerome, Commentarii in Ionam prophetam, in S. Hieronymi Presbyteri Opera, Pars I, Opera Exegetica 6, Corpus Christianorum Series Latina 76, Turnhout: Brepols, 1969. · Augustine, De civitate Dei, bk. 21, ed. B. Dombart and A. Kalb, Corpus Christianorum Series Latina 47–48, Turnhout: Brepols, 1955.
Draft — the later uses cited here (New Testament, rabbinic, patristic) are an AI’s recollection and have not been checked against the sources. Verify before relying on them.
Elliger, K., and W. Rudolph, eds. Biblia Hebraica Stuttgartensia. 5th rev. ed., ed. A. Schenker. Stuttgart, 1997.
Biblia Hebraica Quinta. Stuttgart: Deutsche Bibelgesellschaft, 2004–. Vol. 13: The Twelve Minor Prophets, ed. A. Gelston. 2010.
Ziegler, J., ed. Duodecim prophetae. Septuaginta: Vetus Testamentum Graecum, vol. 13. Göttingen, 1943.
Koehler, L., and W. Baumgartner. The Hebrew and Aramaic Lexicon of the Old Testament. Revised by W. Baumgartner and J. J. Stamm; translated and edited under the supervision of M. E. J. Richardson. 5 vols. Leiden, 1994–2000.
Brown, F., S. R. Driver, and C. A. Briggs. A Hebrew and English Lexicon of the Old Testament. Oxford, 1906.
Clines, D. J. A., ed. The Dictionary of Classical Hebrew. 9 vols. Sheffield, 1993–2016.
Muraoka, T. A Greek–English Lexicon of the Septuagint. Leuven, 2009.
Waltke, B. K., and M. O'Connor. An Introduction to Biblical Hebrew Syntax. Winona Lake, IN, 1990.
Joüon, P., and T. Muraoka. A Grammar of Biblical Hebrew. Translated and revised by T. Muraoka. 2nd ed. Subsidia Biblica 27. Rome, 2006.
Jerome, Commentarii in Ionam prophetam, in S. Hieronymi Presbyteri Opera, Pars I, Opera Exegetica 6, Corpus Christianorum Series Latina 76, Turnhout: Brepols, 1969.
Augustine, De civitate Dei, bk. 21, ed. B. Dombart and A. Kalb, Corpus Christianorum Series Latina 47–48, Turnhout: Brepols, 1955.
Hans Walter Wolff, Obadiah and Jonah: A Commentary, trans. Margaret Kohl, Continental Commentaries, Minneapolis: Augsburg, 1986.
Jack M. Sasson, Jonah: A New Translation with Introduction, Commentary, and Interpretation, Anchor Bible 24B, New York: Doubleday, 1990.
Phyllis Trible, Rhetorical Criticism: Context, Method, and the Book of Jonah, Minneapolis: Fortress, 1994.
Alexander Sperber, The Bible in Aramaic, vol. 3: The Latter Prophets according to Targum Jonathan, Leiden: Brill, 1962.