The Eagles, the Soul That Sins, and the Man in the Gap

The lexica give מָשָׁל a very wide range — proverb, byword, taunt, allegory, oracle, parable — unified by the idea of a comparison or a likeness. חִידָה is narrower: an enigma, a puzzle set to be solved, the thing Samson proposes at his wedding feast and the thing the Queen of Sheba brings to Solomon. The dictionaries agree that the two overlap and that 17:2 pairs them deliberately.

What the files show, rather than the dictionaries. The bare noun חִידָה, unprefixed and without the article, stands in exactly two verses of the 23,204-unit Masoretic corpus: Judges 14:12 and Ezekiel 17:2. The verb-plus-noun sequence חוד חידה stands in one: this verse. And the Greek column does not carry a riddle at all — CATSS reads διήγημα, a tale, which stands in six verses of the whole 30,348-unit screened Greek Old Testament and in only this one prophetic verse.

A funeral dirge, sung over a dead man by professional women, with a characteristic limping metre (the so-called qinah line, three beats followed by two) that the handbooks describe and dispute in equal measure. Amos 5:1–2 sings one over Israel while Israel is still standing, and that is the move Ezekiel 19 makes.

What the files show. קִינָה stands in fifteen verses of the Hebrew Bible. Eight are in Ezekiel — 19:1, 19:14, 26:17, 27:2, 27:32, 28:12, 32:2 and 32:16 — and the rest are in 2 Samuel, Jeremiah, Amos and one place-name in Joshua. Inside chapter 19 it stands in the first verse and the last and nowhere between. The chapter is a labelled box, and the label is on both ends.

Ezekiel’s standing word for idols, and the reference works are unanimous that it is an insult. The most widely repeated proposal connects it with גֵּל, a lump or pellet, and specifically with dung pellets, vocalised on the pattern of שִׁקּוּצִים, detestable things. Others read it as rolled things, blocks or logs. The dictionaries flag the etymology as uncertain and the contempt as certain.

What the files show. The word stands in thirty-nine verses of the Hebrew Bible. Thirty-six of those thirty-nine are in Ezekiel, and seven of them are in chapter 20 alone (20:7, 8, 16, 18, 24, 31, 39). The CATSS Greek does not reproduce the insult: it renders the word with ἐπιτηδεύματα, practices, or εἴδωλα, images — at 20:39 it uses ἐπιτηδεύματα for both occurrences in the verse.

Repeatedly in chapter 20 God says he lifted his hand — וָאֶשָּׂא יָדִי twice in 20:5, נָשָׂאתִי יָדִי at 20:6 and 20:15, and נָשָׂאתִי אֶת־יָדִי at 20:23, 20:28 and 20:42 — and every English version renders it I swore. The reference works treat it as the standard gesture of oath-taking, comparable to the raised hand in Deuteronomy 32:40 and Daniel 12:7.

What the files show. The idiom is the hinge of this chapter’s deepest external run. Ezekiel 20:28 and 20:42 each share seven consecutive Hebrew words with Exodus 6:8, tie band one apiece — אל הארץ אשר נשאתי את ידי לתת, to the land which I lifted my hand to give. Exodus 6:8 is the verse in which the promise to the patriarchs is restated to Moses. The exact sequence נשאתי את ידי stands in seven verses of the Hebrew Bible: Exodus 6:8, Numbers 14:30, and five in Ezekiel — 20:23, 20:28, 20:42, 36:7 and 47:14. And the Greek loses it: Greek Ezekiel 20:28 against Greek Exodus 6:8 runs THREE, because the two Greek verses render the raised hand differently.

The dictionaries take פֶּרֶץ as a gap torn in a city wall and גָּדֵר as the dry-stone field wall or terrace wall. The combined image at 22:30 is a man physically filling a hole in a defence, and it is usually read alongside Psalm 106:23, where Moses stands in the breach and turns the wrath away.

What the files show. The two-word sequence גֹּדֵר־גָּדֵר stands in exactly one verse of the Hebrew Bible and this is it. The exact form בַּפֶּרֶץ stands in two: Ezekiel 22:30 and Psalm 106:23. A third verse, Ezekiel 13:5, has the plural בַּפְּרָצוֹת with the same verb — and it is an accusation that the prophets did not go up into the breaches. 22:30 is the answer to 13:5, nine chapters later: the man who was told to go up was never found. Neither image survives into the Greek column at either verse.

Ezekiel 17 is an allegory of a real, datable political event: Nebuchadnezzar installed Zedekiah on the throne of Judah under oath in 597 BC, and Zedekiah then sent envoys to Egypt for horses. 2 Kings 24:17–20 and 2 Chronicles 36:13 report the same sequence, and 2 Chronicles specifies the oath was sworn by God.

Neo-Assyrian and Neo-Babylonian vassal treaties — the Esarhaddon succession treaties are the best preserved — follow a shape the chapter reproduces: a suzerain, an installed local king, an oath sworn by the gods of both parties, and a long list of curses for breaking it. What the chapter does with the form is the unusual part: 17:19 makes the broken oath an offence against God’s covenant, not Babylon’s.

What that licenses. It makes the allegory datable. It says nothing about when the poem was composed, and this page does not use it that way. Cited from Pritchard’s Ancient Near Eastern Texts and the standard published treaty editions; no cuneiform was read for this page.

Hebrew 21:26 (English 21:21) describes the king of Babylon halting where the road forks and using three techniques: shaking arrows, consulting תְּרָפִים, and inspecting a liver. All three are attested in Mesopotamian practice, and extispicy in particular is one of the best-documented activities in the cuneiform record — there are clay liver models with inscribed omens, and whole libraries of bārû reports.

What the two columns do with it. The Greek keeps all three: τοῦ ἀναβράσαι ῥάβδον (to shake up a rod), ἐπερωτῆσαι ἐν τοῖς γλυπτοῖς (to enquire by the carved images), and the single compound verb ἡπατοσκοπήσασθαι, to inspect a liver. Neither column offers any comment on whether the divination worked. The oracle simply says the lot fell on Jerusalem.

Cited from Pritchard and the standard published Akkadian omen literature at second hand; no Akkadian was read for this page.

Kings of the ancient Near East cut Lebanese cedar and said so in their inscriptions — Egyptian, Assyrian and Babylonian royal texts all record cedar expeditions, and the tree functions as a standing image of unreachable height. The Bible uses it the same way at Isaiah 2:13, Ezekiel 31 (Assyria as a cedar) and Daniel 4 (Nebuchadnezzar as a tree).

What the files show. Ezekiel 17 uses the image twice and reverses it in the middle: at 17:3 a foreign eagle takes the top of the cedar, and at 17:22 God says I myself will take from the top of the cedar. The verb and the object are the same in both verses — לקח plus צַמֶּרֶת — and צַמֶּרֶת occurs in five verses of the Hebrew Bible and all five are in Ezekiel — 17:3, 17:22, and 31:3, 31:10 and 31:14, where Assyria is the cedar.

Chapter 20 tells the exodus and omits almost everything the Pentateuch narrates. There is no plague cycle, no sea, no Sinai, no calf, no spies. What there is instead is a repeated legal formula (I gave them my statutes… they did not walk in them) and a repeated motive clause (I acted for the sake of my name). The commentaries treat the chapter as a rewriting of the tradition rather than a summary of it, and disagree sharply about which written sources, if any, the writer had.

What the measurement can and cannot say about that. It can say that the two verses in the chapter that reach deepest outside Ezekiel land on Exodus 6:8 (seven words, twice) and Leviticus 18:5 (six words). It cannot say whether the writer had those verses in front of him, whether all three draw on a common formula, or which is earlier. The shelf never asserts direction of influence.

The letters משל stand in eighteen places across eleven verses of Ezekiel — but one of the eleven, 42:6 (מְשֻׁלָּשׁוֹת, in three storeys), is a different root altogether and is excluded from everything that follows. Inside this passage: 17:2 (וּמְשֹׁל מָשָׁל, the commission), 18:2 and 18:3 (the proverb taken away), 19:11 (מֹשְׁלִים — a homograph, rulers, not proverbs), and 21:5 (מְמַשֵּׁל מְשָׁלִים, the dismissal). Outside it: 12:22, 12:23, 14:8, 16:44 and 24:3.

The homograph at 19:11 is worth naming because it is a trap. The consonants are identical to the word for proverb and the meaning is unrelated: the vine had strong stems for the sceptres of rulers. A screen that counted roots rather than forms would fold the two together and produce a thread that is not there. Here even the form does not separate them — 18:2 and 19:11 carry the identical form מֹשְׁלִים — so the two are told apart above by reading, not by the screen, and the separation is editorial and is labelled as such.

נֶפֶשׁ in 18:4 is rendered soul by the ESV, ψυχή by the Greek, and person or life by most modern versions. The word covers throat, appetite, life, self and person; whether 18:4 is asserting anything about an immaterial soul is a question about theology and translation history, not about this text.

What can be said from the files is only this: the clause הנפש החטאת is unique to two verses; the Greek renders it the same way in both; and neither column offers any comment on what a נֶפֶשׁ is. Anyone wanting the doctrinal discussion should read it somewhere it is actually being argued, not off a run screen.

Screened against the other 23,022 Masoretic verse units, twenty-one of the 182 Hebrew rows reach six consecutive words or more. Between them they carry fifty-three partner verses at their reported maxima. Forty-nine of the fifty-three are inside Ezekiel. Exactly four are not, and they name three verses:

Exodus 6:8, twice — from 20:28 and from 20:42, seven consecutive words each, tie band one each: אל הארץ אשר נשאתי את ידי לתת.

Leviticus 18:5, once — from 20:13, six consecutive words, tie band one: אשר יעשה אתם האדם וחי בהם.

Isaiah 49:26, once — from Hebrew 21:10 (English 21:5), six consecutive words, tie band one: וידעו כל בשר כי אני יהוה. Hebrew 21:4 reaches five with the same Isaiah verse.

Three of the four are in chapter 20, the chapter that retells the Torah. The single deepest row in the whole passage — nine consecutive words at 20:30, tie band one — does not leave the book: its partner is Ezekiel 14:6, and the nine words are a stack of two messenger formulae, לכן אמר אל בית ישראל כה אמר אדני יהוה. The deepest measured thing in these six chapters is boilerplate. Say that plainly, because it is the sort of number a page could quietly let a reader misread.

Ezekiel 18:2 and Jeremiah 31:29: TWO consecutive words. The same proverb, quoted by two prophets, and the shared run is בסר ושני and nothing else. Jeremiah has a perfect verb and no article on sons; Ezekiel has an imperfect and an article. Two morphemes cut the run from six to two.

Ezekiel 22:30 and Psalm 106:23: ONE consecutive word — בפרץ. A prefixed vav on the participle and a suffix on before do the rest.

Ezekiel 21:27 and Genesis 49:10: ONE consecutive word, and the word is לא. That row’s absolute maximum against the entire Hebrew Bible is two, with a tie band of 114.

Ezekiel 17:23 and Mark 4:32: FOUR Greek words, tie band one. This is the closest thing on the page to a real allusion — a great tree, birds sheltering under its shade — and four is what an allusion measures. Compare the six-word run at 20:39, which is almost certainly not an allusion at all.

If a page prints only the numbers that flatter its thesis it is not a measurement, it is an argument with decoration. These four are printed near the top on purpose.

Screened against all 7,927 verses of the Greek New Testament, the 182 CATSS rows give 182 at two, 134 at three, 32 at four, 3 at five, 1 at six, nothing at seven. The single six is Ezekiel 20:39 into Revelation 3:12, tie band one: μου καὶ τὸ ὄνομά μου τὸ, breaking on ἅγιον against καινόν.

Here is the reason not to read it as a quotation. Revelation 3:12 is also the sole maximum partner of Greek Ezekiel 20:14 and of Greek Ezekiel 20:22, both at four, both tie band one. Three rows of one chapter, one partner verse. Ezekiel 20 says my name over and over; Revelation 3:12 says τὸ ὄνομα three times in one sentence. That is two formulae colliding.

For scale: across all 1,267 rows of the whole Greek book of Ezekiel only THREE reach six or more — 24:21 at seven (Acts 2:17), 3:3 at six (Revelation 10:10) and this one. So the row is genuinely near the top of its own book, and still not a citation.

The runners-up on the Greek side are two rows at five, tie band one apiece: Ezekiel 20:6 and Ezekiel 20:9 into Hebrews 8:9, sharing ἐξαγαγεῖν αὐτοὺς ἐκ γῆς Αἰγύπτου. A five with a band of one looks like contact.

It is not. Hebrews 8:9 shares TWENTY-THREE consecutive Greek words with Greek Jeremiah 38:32 (Masoretic and English Jeremiah 31:32) — τοῖς πατράσιν αὐτῶν ἐν ἡμέρᾳ ἐπιλαβομένου μου τῆς χειρὸς αὐτῶν ἐξαγαγεῖν αὐτοὺς ἐκ γῆς Αἰγύπτου ὅτι αὐτοὶ οὐκ ἐνέμειναν ἐν τῇ διαθήκῃ μου. Hebrews is quoting the new-covenant passage of Jeremiah at length, and Ezekiel happens to use the same five-word exodus idiom.

Five against twenty-three. That is what the tie band cannot tell you and a second screen can. A run figure is only ever a floor on how much work a reader still has to do.

Eight of the 182 Hebrew rows come out at zero: 17:6, 17:7, 19:3, 19:6, 21:21, 22:5, 22:10 and 22:15 (all in Hebrew numbering; 21:21 is English 21:16). The Hebrew-side screen runs with the unigram fallback OFF, so zero means “shares no two consecutive words with any other verse in the Hebrew Bible”. Every one of the eight shares individual words.

The distribution of the zeros is the finding, not the zeros themselves. Four of the eight are pictures: the vine and the second eagle in chapter 17, the two lion cubs in chapter 19. Three are in the indictment of chapter 22. Chapters 18 and 20 — the two chapters that argue a case — have no zeros at all, and between them hold twelve of the twenty-one rows that reach six or more.

An honest reading of that: when this book reasons it uses the shared vocabulary of Israelite prophecy, and when it draws it stops resembling anything. Whether that is one writer changing register or two hands is a question the screen cannot reach.

Three of this passage’s measured connections behave differently in Hebrew and in Greek, and all three go the same way — deep in Hebrew, shallow in Greek:

20:13 to Leviticus 18:5: six in Hebrew, THREE in Greek.
20:28 to Exodus 6:8: seven in Hebrew, THREE in Greek.
22:30 to Psalm 106:23: one in Hebrew, and in Greek the shared image is not there to measure, because the CATSS Ezekiel has no breach.

And one goes the other way: 18:31 to 36:26 measures five in Hebrew and five in Greek — καρδίαν καινὴν καὶ πνεῦμα καινόν — the only cross-reference on this page that survives translation at full depth.

No conclusion is drawn from this about which text is older or what the translator was doing. It is reported because a page that measured only one column would have produced a different set of headlines and not known it.

No critical apparatus was consulted for the Hebrew. The Masoretic text used is Sefaria’s digitization of Miqra according to the Masorah. No variant readings were collated from BHS or BHQ, and no Dead Sea Scroll was examined — the Masada and Qumran Ezekiel fragments included.

The Greek is Rahlfs as digitized by CATSS, and that is one edition. Ziegler’s Göttingen Ezechiel was not consulted, and neither was Papyrus 967, which is the manuscript question in Greek Ezekiel: it carries a different chapter order in the second half of the book and is missing 36:23c–38. Every statement above of the form “the Greek reads…” is a statement about one digitized edition on one disk. The claim that most concerns 967 — the missing wall and breach at 22:30 — was not checked against it and could look different there.

No ESV file exists on this disk for Ezekiel and no Crossway API key exists in this project. The English column above was typed. Its verse alignment rests on verse counts and chapter openings, not on a held text. The Hebrew and Greek columns were pulled programmatically from the corpus files and can be re-derived by anyone with the repository.

Greek Exodus, Leviticus and Numbers ARE held here and were searched — the whole-corpus screens run over 30,348 Greek Old Testament verse units including all three — but they sit outside the named book map used by the per-book helper, so the Greek Torah comparisons above were run by explicit file reference rather than by book name. If that mapping is wrong, the two figures at risk are the Greek 20:13–Leviticus 18:5 run of three and the Greek 20:28–Exodus 6:8 run of three. Both were spot-checked by printing the two Greek verses in full and reading them.

The Odes are excluded from every Greek screen on this page. The CATSS collection held here is 64 files and 30,637 verse units; the screened corpus is 30,348, because Ode 9 is Luke 1 and Ode 13 is Luke 2 copied into a Greek Old Testament manuscript, and ten more Odes are second copies of Exodus 15, Deuteronomy 32, 1 Samuel 2, Habakkuk 3, Isaiah 26, Jonah 2, Daniel 3, Isaiah 5 and Isaiah 38. Leaving them in would inflate any tie band that touched one, and two of them would have put New Testament text into an Old Testament screen.

Chapter 20 rewrites the exodus with the salvation removed and adds a sentence — I gave them statutes that were not good (20:25) — that has generated a very large literature. The main proposals are that it refers to child sacrifice (reading 20:26 as its explanation), that it refers to a body of law the writer regarded as illegitimate, that it is a judicial hardening comparable to Isaiah 6:9–10, or that it is a rhetorical concession rather than a claim about the Sinai legislation.

This page has nothing to contribute to that and says so. What it can report is that 20:25 is a deliberate negative of 20:11 and 20:13 in both columns — the same two nouns, the same live by them — and that 20:11 and 20:13 are the verses standing closest to Leviticus 18:5. The measurement establishes that the negation is aimed at a specific sentence. It cannot tell you what the negation means.

The standard treatments read Ezekiel 18 and Jeremiah 31:29–30 as two prophets taking up the same current proverb, and disagree about whether either knew the other, and about whether both are reacting to Exodus 20:5 and Deuteronomy 5:9.

The screen’s answer is two consecutive words, and that answer is compatible with every one of those positions. A shared proverb quoted from memory in two different grammatical shapes is exactly what a low run looks like; so is no relationship at all. The number rules out one thing only: that either verse was copied from the other in writing. It cannot rule out dependence, and this page does not claim it does.

Is the Greek of 22:30 translating a different Hebrew text, or paraphrasing? Unanswerable from one digitized edition. The page reports that the same file does the same thing at 13:5, which is consistent with either.

Is chapter 19 poetry in a strict qinah metre, and does that metre exist? Metrical analysis is not something a word-form screen can do, and the handbooks disagree about the metre itself.

Why does the Greek add εἰς παραβολὴν θρήνου at 19:14? The page counts the word and stops.

Does the fivefold repetition of the recognition formula in chapter 20 structure the chapter? Probably; the section markers do not agree with it, and this page built its diagram from the markers because the markers are in the file.

Why does Greek Ezekiel reach the New Testament so shallowly for a book of 1,267 verses? Only three of its rows reach six or more — 24:21, 3:3 and 20:39 — against 258 that reach four. Whether that is a fact about Ezekiel’s Greek, about how the New Testament uses Ezekiel, or about the screen, this page does not know.

One. The ESV column is typed, not held. It is the only column on this page that was not machine-read from a file, and the two rows carrying the chapter-20/21 offset are the ones where a typing error would be hardest to see.

Two. The eight-zero finding rests on eight rows out of 182, and the claim that four of them cluster in the “picture” chapters is an observation about a small number. It is reported as an observation and nothing is built on it.

Three. Stage 2 is entirely second-hand and none of it can be checked here.

Four. The comparison passage is one comparison. Ezekiel 33–37 was measured on the same day with the same settings; no systematic survey of the whole book was run, and the claim that this passage is unusually deep on the Hebrew side is supported against that one neighbour and against the general run of this shelf.

Five. The literary-structure diagram is built from one edition’s section markers. The Greek file has no markers, so the internal shape shown is a Hebrew-side shape and is labelled as one.

The three best-known cross-references in these chapters measure two, one and one. The two deepest connections outside the book are to Exodus 6:8 and Leviticus 18:5, and almost nobody cites either. What everyone quotes and what actually repeats are two different lists, and only one of them can be counted.

That is not an argument that the famous cross-references are wrong. Ezekiel 18 and Jeremiah 31 really are answering the same proverb; Ezekiel 22:30 and Psalm 106:23 really are the same picture. It is an argument that agreement of meaning and agreement of wording are two different measurements, and a page that reports one should never let a reader take it for the other.

Six consecutive Greek words, tie band one, and they break on the adjective. Ezekiel: my name, the holy one, you shall not profane again. Revelation: and my name, the new one. Only three rows in the whole Greek book of Ezekiel reach six or more, and this is one of them.

The page has already told you not to call it a citation, and the reason is printed above: the same Revelation verse is the top partner of two other verses in the same Ezekiel chapter. So the honest question is not “did John quote Ezekiel here”. It is: what is it about the phrase my name that makes two books three centuries apart build the same six words without either one reaching for the other?

22:30 is the verse that has launched the most sermons in these six chapters, and it is the verse where the two ancient columns are furthest apart. The Hebrew wants a builder standing in a hole in a wall. The Greek wants a man whose conduct is upright, standing wholly before God. Neither column has anybody who qualifies.

Read 13:5 first — the accusation nine chapters earlier that the prophets did not go up into the breaches — and then read 22:30 as its verdict. And notice what the chapter does next: it does not find a man. It pours out the wrath. The passage that begins with an eagle carrying off a king ends with God looking for one person and not finding one.

Anyone with Ziegler’s Ezechiel, or with a transcription of Papyrus 967, can check three things this page explicitly could not: whether another Greek edition has a wall and a breach at 22:30, whether 19:14 carries εἰς παραβολὴν θρήνου outside CATSS, and whether the chapter numbering of Ezekiel 17–22 is stable across the Greek witnesses as it is in this file.

Send corrections. A page that names its weak points is asking to be checked at them.