Whose Faith

• Anchor Bible Dictionary — “Habakkuk, Book of,” “Chaldeans,” “Theodicy,” “Pesher Habakkuk.”
• Dictionary of the OT: Prophets (IVP) — “Habakkuk: Book of,” “Faith,” “Lament,” “Theophany.”
• Dictionary of NT Use of the OT (Beale & Carson) — on Romans 1:17, Galatians 3:11, Hebrews 10:37–38 and Acts 13:41.
• NIDOTTE — אֱמוּנָה, צַדִּיק, חָזוֹן, חָמָס.

Habakkuk carries no Hebrew/English verse offset at any point. Masoretic, CATSS and the ESV all run 17 / 20 / 19 — 56 verses. Verified from content at each chapter seam rather than assumed. Of the twelve Minor Prophets, seven carry an offset somewhere and five do not; Habakkuk is one of the five, along with Amos, Obadiah, Zephaniah and Haggai.

The CATSS row reads: ἐὰν ὑποστείληται οὐκ εὐδοκεῖ ἡ ψυχή μου ἐν αὐτῷ · ὁ δὲ δίκαιος ἐκ πίστεώς μου ζήσεται

Hebrews 10:38 carries both halves; Romans and Galatians carry only the second. The deepest run is therefore on the material the other two do not use at all. And Hebrews prints the two halves in the opposite order to the row: it opens with the righteous living by faith and continues with the drawing back.

Checked from both ends: against all 7,927 New Testament verses, Habakkuk 2:4 reaches 9 with Hebrews 10:38 and with nothing else — tie band one, and the next depths are a single verse at 5 and a single verse at 4. Against all 30,348 CATSS verse units, Hebrews 10:38 reaches 9 with Habakkuk 2:4 and nothing else. The nine-word string stands in exactly one verse unit and one New Testament verse. ⚠️ Re-run on 2026-08-30 over the enlarged corpus and unchanged: when this paragraph was first written the scope was 23,061 units in 44 files; Exodus, Leviticus, Numbers, Joshua and the deuterocanon have since been added, taking the shelf to 30,637 units held in 64 files, of which 30,348 are screened — the Odes are set aside — and Hebrews 10:38 still finds Habakkuk 2:4 and nothing else. It found nothing else with the Odes in either.

One pronoun, and no two of these four put it in the same place.

The Hebrew attaches it to the faith and makes it the righteous man’s. The Greek row attaches it to the faith and makes it God’s. Hebrews attaches it to the righteous one. Romans and Galatians attach it to nothing. All four readings are printed here as their files have them, and the page does not order them, date them, or say what produced any of them.

All 56 CATSS rows of Habakkuk against all 7,927 verses of the Greek New Testament.

Three rows sit at the top and each has a tie band of one on the New Testament side: 2:4 with Hebrews 10:38 at nine, 1:5 with Acts 13:41 at seven, and 3:18 with Luke 1:47 at six. Nothing in the book reaches five or eight at all.

⚠️ Nine is deep for Habakkuk and not for the corpus. The same screen over all 30,348 screened verse units goes deeper elsewhere, and here is the whole top of the shelf as it now stands:

⚠️ A ROW STOOD IN THAT TABLE FOR ONE DAY AND HAS BEEN TAKEN OUT. IT IS NAMED HERE RATHER THAN DELETED.

Between the corpus fetch of 30 August 2026 and the exclusion of the Odes, this table carried an eleventh row — Odes 9:48 against Luke 1:48, a run of 18 at a tie band of 1, sitting between Joel and Amos — the twelfth-deepest Old-Testament-to-New-Testament pair on the whole shelf, with exactly eleven above it. It was Luke matching Luke. The Greek Odes are fourteen canticles appended to the Psalter in the great codices for liturgical use, and Ode 9 is the Magnificat and the Benedictus — Luke chapter 1 itself, excerpted for the daily office and copied back into a Greek Old Testament manuscript. Eighteen consecutive words is a very deep run on this shelf: screened from the New Testament side, exactly eleven pairs in the whole corpus go deeper, and every one of them is in the table above. A page that had reported it without opening the file would have printed a Gospel’s deepest Old Testament contact and been measuring the Gospel against itself.

How it was caught, since that is the part worth having. The run was flagged because it was implausibly clean: eighteen words unbroken, at a tie band of one, in a book nobody had checked, in a verse of Luke with no known Old Testament source. Opening the CATSS file settled it in a line — Ode 9 is headed with the Magnificat. The instrument was working perfectly. It found a real, exact, eighteen-word agreement between two files on this disk. What it could not know is that the two files are the same text. That is the standing hazard of a corpus screen and the reason every deep run on this shelf is opened by eye before it is printed.

The Odes are now excluded from anything this shelf calls “the Greek Old Testament”, for that reason and for a second: ten of the fourteen canticles are second copies of passages already in the corpus — Ode 1 is Exodus 15, Ode 2 Deuteronomy 32, Ode 3 1 Samuel 2, Ode 4 is Habakkuk 3, Ode 5 Isaiah 26, Ode 6 Jonah 2, Odes 7 and 8 Daniel (Th) 3, Ode 10 Isaiah 5, Ode 11 Isaiah 38 — so a verse standing once appears to stand twice and any tie band touching a canticle is doubled. They stay on the disk, they are still read, and Ode 4 is discussed at length in stage 3 as what it is: a second copy of this book’s third chapter. They are not counted as an Old Testament witness against the New. The shelf holds 30,637 verse units in 64 files and screens 30,348 of them; the difference is exactly the 289 units of the Odes.

⚠️ This table has been corrected three times, and the page prints every correction rather than swapping numbers in silence. When this page was first written the Greek Old Testament held here was 35 files and 18,183 verse units, and Deuteronomy 6:5 at 21 was the deepest run on the shelf. 2026-08-19: Isaiah, Jeremiah, Lamentations, Ezekiel and both Greek Daniels were added — 44 files, 23,061 verse units — and Deuteronomy 6:5 stopped being the deepest: Isaiah 6:10 reached 24 against Matthew 13:15 and LXX Jeremiah 38:34 reached 23 against Hebrews 8:11. 2026-08-30: Exodus, Leviticus, Numbers, Joshua (A and B) and the whole deuterocanon were added — 64 files, 30,637 verse units held. The top of the table did not change hands again: Isaiah 6:10 is still first at 24. What the enlargement added was three new rows above Amos — Isaiah 54:1 / Galatians 4:27 and Jeremiah 38:32 / Hebrews 8:9, both at 23, which were already on disk and are here because the table was re-derived rather than patched; and Odes 9:48 / Luke 1:48 at 18, which is the artefact described above and has since been removed. 2026-08-31: the Odes were taken out of the screening corpus — 64 files still, 30,637 units held, 30,348 screened — and the Odes row left the table. No other row moved. Nothing measured about Habakkuk changed at either enlargement: every Habakkuk figure on this page is measured against the Greek New Testament, which did not grow, and each was re-run both times and held.

What the measurement is, and its scope. Contiguous runs of word forms, Greek against Greek: the CATSS FORM column (cols 0–24, beta-code transliterated) across 30,348 verse units screened — 30,637 held in 64 .mlxx files, less the 289 units of the Odes, against the MorphGNT surface-word column across 7,927 verses in 27 files. Both sides lowercased, accents and breathings dropped, final sigma folded, then filtered to Greek letters — the same filter on both corpora. Word forms, never lemmas. Every CATSS file is now held here and there is no Greek Old Testament gap left on this disk — the Major Prophets arrived 2026-08-19, and Exodus, Leviticus, Numbers, Joshua (A and B), Judith, Tobit (BA and S), Wisdom, Baruch, the Letter of Jeremiah, 2–4 Maccabees, Susanna, Bel, Odes and Psalms of Solomon on 2026-08-30. Every figure in this section was re-run against the enlarged corpus. It says nothing about which text is prior and nothing about citation.

The opening word of God’s first answer is addressed to different hearers in the two columns.

The seven-word run between the row and Acts 13:41 comes later in the verse — ὃ οὐ μὴ πιστεύσητε ἐάν τις ἐκδιηγῆται — but the address at the front is the more visible difference, and it is a mutual maximum: Acts 13:41’s deepest partner among all 30,348 screened verse units is this row, and this row’s deepest among all 7,927 New Testament verses is that verse — re-run 2026-08-30 over the enlarged corpus, both bands still one.

The same Greek root turns up again at 1:13, where the row has ἐπὶ καταφρονοῦντας for the Hebrew בּוֹגְדִים, traitors.

Six words of Habakkuk 3:18 stand in Luke 1:47 — and so do six words of Micah 7:7.

ἐπὶ τῷ θεῷ τῷ σωτῆρί μου. Screened against the 23,061 CATSS verse units this page was written against, Luke 1:47 reached six with two of them, tied: Habakkuk 3:18 and Micah 7:7. The screen did not separate them and this page did not either. In the other direction, Habakkuk 3:18’s deepest New Testament partner was Luke 1:47 alone — one-to-one from the Habakkuk side, two-to-one from the Luke side, which is exactly the kind of asymmetry a maximum reported from one direction hides.

★ On 2026-08-30 the shelf grew to 64 files and 30,637 verse units, and for one day the screen came back with a different answer — a wrong one. The page prints what it said, on what corpus, and what it says now.

⚠️ And Odes 9:47 is Luke 1:47. The Greek Odes are fourteen canticles appended to the Psalter in the great codices for liturgical use, and Ode 9 is the Magnificat followed by the Benedictus — Luke chapter 1 itself, lifted out of the Gospel and copied into a Greek Old Testament manuscript. The eleven-word “match” is Luke against Luke. It is not a parallel, it is not a source, and it tells you nothing whatever about Habakkuk. The Odes are held on this shelf as a measurement WITNESS and not as Scripture — the same standing Psalm 151 and Jubilees have here — and this page names them every time they appear.

So the finding survives, and it survives on a stated exclusion rather than on an accident of what the disk happened to hold. With Ode 9 set aside as what it is, the tie is exactly where it was: Habakkuk 3:18 and Micah 7:7, six words each, and the screen still does not separate them. That is a better position than the page was in before, because before, the tie held only because the corpus was short of a book nobody had checked.

⚠️ The exclusion is now the shelf’s default and not this page’s special pleading, which matters. For one day — between the fetch of 30 August 2026 and the exclusion — the top line of that table was the live answer, and the live answer was that Luke 1:47’s deepest Old Testament partner was Luke 1:47. The Odes have since been removed from everything this shelf calls the Greek Old Testament: 30,637 units held in 64 files, 30,348 screened, the 289-unit difference being the canticles. The page no longer has to argue for setting Ode 9 aside in the middle of a paragraph; the screen no longer sees it. Both states are printed above because a reader is entitled to know that the instrument returned Luke against Luke and how long it took to notice.

And with the canticles out, Luke 1:48 has an answer worth printing too. The verse the Odes were matching at eighteen — ὅτι ἐπέβλεψεν ἐπὶ τὴν ταπείνωσιν τῆς δούλης αὐτοῦ — reaches 5 against the screened corpus, tie band 1, and its partner is CATSS 1 Kingdoms 1:11 (= 1 Samuel 1:11): Hannah’s vow at Shiloh, if you will indeed look on the humiliation of your servant. That is a real Old Testament contact and the eighteen-word one was not. The page reports the five and does not build anything on it: five words at a band of one is a signal, not a citation, and Luke does not flag it as a quotation.

Ode 4 of the Greek Odes is headed προσευχὴ Αμβακουμ — the prayer of Habakkuk — and it is this chapter.

Odes 4:18 and CATSS Habakkuk 3:18 are the same nineteen Greek words in the same order: ἐγὼ δὲ ἐν τῷ κυρίῳ ἀγαλλιάσομαι χαρήσομαι ἐπὶ τῷ θεῷ τῷ σωτῆρί μου, a run of thirteen after normalisation with nothing broken anywhere in it. Habakkuk 3 was taken into the church’s daily prayer as a canticle, and the copy in the canticles is on this disk now alongside the copy in the book.

Which is why the four-place count for the phrase reads as it does. ἐπὶ τῷ θεῷ τῷ σωτῆρί μου stood in four of the 30,637 verse units held — and the four were two verses. Habakkuk 3:18 and Micah 7:7 are the two; Odes 4:18 is Habakkuk 3:18 again and Odes 9:47 is Luke 1:47. A count that did not say so would have been a false four. Screened now, over the 30,348 units that exclude the canticles, the count is two — Habakkuk 3:18 and Micah 7:7, which is what the phrase was always worth. ⚠️ This is exactly the doubling the exclusion exists to prevent, and it happened to be visible here because the duplicated canticle is this book’s own third chapter.

Ode 4 is still on the disk and is still worth reading. Habakkuk 3 was taken into the church’s daily prayer, and the fact that the copy in the canticles is word for word the copy in the book is a fact about transmission that this page is glad to have. It is simply not a second witness, and a rarity count that treats it as one is counting the same verse twice.

Each is printed as its file has it. Why any of them stands as it does is not something these columns can settle. The third has had a long afterlife in Christian art and none of that is measured here.

The two deepest Habakkuk rows in chapter 2 land on two consecutive verses of Hebrews, in the same sequence.

Habakkuk 2:3 ends ὅτι ἐρχόμενος ἥξει καὶ οὐ μὴ χρονίσῃ and Hebrews 10:37 carries four of those words. The page reports the sequence and does not explain it — that two adjacent rows match two adjacent verses is what the columns show.

Neither is a short quotation; both are short runs, and for different reasons.

Romans 1:17 reads ὁ δὲ δίκαιος ἐκ πίστεως ζήσεται — six words, of which five run unbroken against the row. The break is the row’s μου between πίστεως and ζήσεται.

Galatians 3:11 reads ὁ δίκαιος ἐκ πίστεως ζήσεται — without δέ, so the run starts one word later and reaches three.

One particle and one pronoun account for the whole difference between five and three — a pattern this shelf has met before, most sharply at Joel 3:5, where a single ἄν against ἐάν moved a run from seven to five.

Run length is not importance. That Hebrews 10:38 measures nine and Galatians 3:11 measures three says which text reproduces more consecutive word forms of this particular Greek column. It says nothing about which use of the verse mattered more to anyone, and the page makes no such claim.

The scope is 64 Greek Old Testament files — and that is now the whole of the Greek Old Testament as CATSS transmits it, with no file missing. Isaiah, Jeremiah, Lamentations, Ezekiel and both Greek Daniels were added on 2026-08-19; Exodus, Leviticus, Numbers, Joshua (A and B), Judith, Tobit (BA and S), Wisdom, Baruch, the Letter of Jeremiah, 2–4 Maccabees, Susanna, Bel, Odes and Psalms of Solomon on 2026-08-30 — 44 files and 23,061 units → 64 files and 30,637 held, of which 30,348 are screened: the Odes were removed from the screening corpus on 2026-08-31 because two of the canticles are New Testament text and ten are second copies of passages already counted. Every “only” on this page was re-run against the enlarged corpus and again after the exclusion. All held. The one that wobbled is set out in stage 3: for a single day Luke 1:47’s two-way tie was displaced by Ode 9 — which is Luke 1 itself — and with the canticles out it is a two-way tie again. Every “only” here means on the CATSS FORM column across those files.

And 2:4 is one of the most contested verses in the Hebrew Bible. Whether אֱמוּנָה is best rendered faith or faithfulness, and how the Qumran pesher on Habakkuk reads it, are questions with large literatures. None of that was consulted. This page reports four columns and stops.

No Hebrew or Greek critical apparatus. No Qumran material — which is a particularly sharp omission here, since 1QpHab is a running commentary on Habakkuk 1–2 and quotes 2:4 directly, and the Minor Prophets scroll from Nahal Hever bears on the Greek of these books. Neither was consulted. No Targum, Syriac, Vulgate, Aquila, Symmachus or Theodotion, and no text-critical work on the Hebrews 10:38 manuscript tradition, which varies at exactly the possessive discussed above.

And nothing about the date or the Chaldeans. A word count cannot touch them.

Habakkuk 2:4 is two halves. Romans and Galatians take the second. Hebrews takes both, and the nine words it shares with the CATSS row are all from the first — the half about drawing back, which is the part of the verse least often heard.

Nine words, and each verse is the other’s deepest match: no other of the 7,927 New Testament verses reaches nine against that row, and no other of the 30,348 verse units reaches nine against that verse — re-run on 2026-08-30 against the enlarged corpus and unchanged.

And then the pronoun. The Hebrew says the righteous will live by his faith. The Greek row says by my faith. Hebrews says my righteous one will live by faith. Romans and Galatians say the righteous will live by faith, and leave the pronoun out. Four texts, one small word, four different homes for it — and a great deal of what has been built on this verse depends on which of the four you are reading.

Eighth study on the Minor Prophets shelf, and the whole of Habakkuk in one page. Five books complete: Hosea, Joel, Amos, Jonah, Habakkuk.