A Word the Hebrew Does Not Have

• Anchor Bible Dictionary — “Jeremiah, Book of” (especially the section on the Greek text), “Prophecy, False,” “Prophetic Symbolic Actions,” “Hananiah,” “Shaphan,” “Jehoiachin,” “Exile, Babylonian,” “Seventy Years.”
• Dictionary of the Old Testament: Prophets (IVP) — “Prophecy and Society,” “Prophecy, False,” “Sign Acts,” “Intertextuality and Innerbiblical Exegesis,” “Text and Textual Criticism.”
• The New Interpreter’s Dictionary of the Bible — “Jeremiah, Book of,” §§ on chapters 26–36; “Yoke”; “Letter.”
• Where the dictionaries disagree, and the page says so rather than choosing: whether the quotation of Micah at 26:18 is Jeremiah’s or a later editor’s; whether chapters 26–29 come from Baruch, from a Deuteronomistic editor, or from Jeremiah; whether the “seventy years” is a round number, a Babylonian topos, or a calculation; whether the shorter Greek Jeremiah reflects an earlier Hebrew edition or the translator’s abridgement; and whether ψευδοπροφήτης is a translator’s coinage or an existing Greek word put to a new use. None of the five is decidable from any corpus held here.

Nothing in chapters 26–29 lets you identify a false prophet at the moment he is speaking. Read 28:1–11 slowly. Hananiah uses the standard messenger formula, speaks in the temple, performs a sign-act, and is called הַנָּבִיא by the narrator. Jeremiah’s first reply is Amen! May the Lord do so (28:6) — he does not have a rebuttal ready. He leaves. The word that answers Hananiah comes later (28:12), “after the prophet Hananiah had broken the yoke-bars.”

The only two tests the passage offers both work backwards. 28:9 says you will know when it happens; 29:9 says the mark of the false prophet is that God did not send him — which is true and unobservable. This is the problem the Greek column settles with a noun. Where the Hebrew leaves the reader in the room with two men holding the same title, the Greek labels one of them before he speaks. It is a real gain in clarity and a real loss of the situation, and this page does not pretend the two texts are doing the same thing.

Jeremiah 29:1–2 dates its letter after Jeconiah the king, the queen mother, the eunuchs, the officials and the craftsmen had gone from Jerusalem. Administrative tablets excavated from a vaulted building near the Ishtar Gate in Babylon list rations of oil issued to foreign dependants of the crown. Among them:

“10 (sila) to Ia-ku-u-ki-nu, the son of the king of Ia-ku-du (i.e. Judah); 2½ sila for the 5 sons of the king of Judah (Ia-ku-du) through Qana’a …”Babylon 28186, reverse ii 13–18 — Pritchard, ANET, p. 308

“… [Ia]-’-kin, king of Ia[…]; 2½ sila (oil) to [… so]ns of the king of Judah (Ia-a-hu-du); 4 sila to 8 men from Judah (a’ne Ia-a-hu-da-a-a) …”Babylon 28178, obverse ii 38–40 — Pritchard, ANET, p. 308

That is the addressee list of Jeremiah 29 as a Babylonian quartermaster saw it. A deposed king, his sons, and a group of men from Judah, drawing rations in a foreign capital. The letter tells them to build houses, plant gardens, marry off their children and pray for the place. What the page claims is only that the community existed and was administratively visible. The tablets say nothing about Jeremiah, nothing about a letter, and nothing about seventy years, and the page does not use them to argue anything about the biblical text.

The Lachish Ostraca are inscribed potsherds found in the burnt guardroom of the Lachish gate, from the last months before the Babylonian siege. Ostracon III is a field officer’s report to his commander, and two of its sentences touch this passage directly:

“And it hath been reported to thy servant, saying, ‘The commander of the host, Coniah son of Elnathan, hath come down in order to go into Egypt…’ And as for the letter of Tobiah, servant of the king, which came to Shallum son of Jaddua through the prophet, saying, ‘Beware!’, thy servant hath sent it to my lord.”Lachish Ostracon III — Pritchard, ANET, p. 322

Both mechanisms of these chapters are in that one potsherd. A senior officer travelling down to Egypt on the king’s business — which is what Jeremiah 26:22 describes, when Jehoiakim sends Elnathan the son of Achbor and others to Egypt to extradite the prophet Uriah. And a written warning passing through a prophet to a named official — which is the machinery of Jeremiah 29 in both directions.

The page does not identify anybody. Coniah son of Elnathan is not Elnathan son of Achbor; the names run in the same families and the identifications are contested and unprovable. What the ostracon establishes is the social fact: in Judah in that decade, prophets carried official correspondence, and men were sent to Egypt to fetch people back.

The image in chapters 27 and 28 is not agricultural sentiment. Submission to an overlord is described as a yoke across the Near East and across the Old Testament — and the closest biblical parallel is a political negotiation, not a farm. When Israel comes to Rehoboam at Shechem the demand is lighten the hard service of your father and his heavy yoke, and the answer is my father laid on you a heavy yoke; I will add to your yoke (1 Kings 12:4–14). Counted on this disk, nine of the 23 occurrences of κλοιός in the Greek Old Testament held here stand in that one chapter of 3 Kingdoms.

What Jeremiah adds is the physical object. He wears it. Hananiah takes it off him and breaks it in front of the crowd — a counter-sign-act, performed on the other man’s equipment — and the reply is a yoke that cannot be broken. There is no record here of anyone disputing that a sign-act was a legitimate way to argue. The dispute is entirely about which sign is true.

The seventy years of 29:10 became a number that later writers counted with. 2 Chronicles 36:21 names Jeremiah and uses it; Ezra 1:1 names Jeremiah and uses it; Daniel 9:2 names Jeremiah, says he was reading him, and uses it. All three are Scripture on the Pleasant Springs canon, not outside witnesses — and all three are printed in Stage 4 as counter-examples against this page’s own method, because the run screen finds almost nothing in any of them.

The Second-Temple memory of Jeremiah himself runs in the same direction as chapter 26: he is remembered for what was done to him. Ben Sira’s roll-call says ἐκάκωσαν γὰρ αὐτόν, “for they mistreated him” (Sirach 49:7), and 2 Maccabees 15:14 makes him an intercessor who prays much for the people and for the holy city. Both are read here as historical witnesses to what a later generation remembered, never levelled with Scripture.

A screen of every word form beginning ψευδοπροφητ- across the whole Greek Old Testament held here returns ten verses, and no more:

Eight of the ten are in the five chapters this page studies, and none at all is in the fifth of them. In every single case the Hebrew underneath reads the ordinary word for a prophet — the same word it uses of Jeremiah, of Micah and of Uriah. The judgement is carried by the Greek noun and by nothing in the Hebrew.

And the same noun stands eleven times in the Greek New Testament: Matthew 7:15, 24:11 and 24:24; Mark 13:22; Luke 6:26; Acts 13:6; 2 Peter 2:1; 1 John 4:1; and Revelation 16:13, 19:20 and 20:10. Eleven New Testament uses of a word whose Old Testament home, on this disk, is overwhelmingly Jeremiah 26–29.

⚠️ What this page claims and what it does not. It claims the two counts — ten and eleven — and the fact that no Hebrew noun corresponds. It does not claim that the Septuagint of Jeremiah coined the word, because the corpora held here cannot establish that: no classical Greek corpus, no papyri and no Second-Temple Greek outside the files on this disk were searched. It does not claim that the New Testament writers took the word from Jeremiah, because a shared noun is not verbatim contact and this page’s method cannot tell a borrowing from a common vocabulary. What it claims is the distribution, which is a count. Stage 5 names what would have to be checked to say more.

The Hebrew Jeremiah uses the singular עַבְדִּי of exactly three people, three verses each. The Greek held here uses δοῦλός μου or παῖς μου in exactly two verse units of the whole book, and both are Jacob. Verse by verse:

Two of the nine Hebrew verses are inside this passage, and neither survives as “my servant” in the Greek. The Hebrew of these chapters hands the title to the king of Babylon at 27:6 and takes it back for Jacob at 30:10; the Greek does neither. The page reports the arrangement of each text and asserts nothing about which came first, which of the two is an abridgement, or whether anybody found “Nebuchadnezzar my servant” difficult. Those are questions about history; this is a count.

Nebuchadnezzar. The Masoretic Jeremiah names him 35 times, in both of the book’s spellings, out of 76 occurrences in the 23,204-verse Masoretic corpus held here. Seven of the 35 are in these five chapters — 27:6, 27:8, 27:20, 28:3, 29:1, 29:3, 29:21. CATSS Jeremiah names him 14 times in the whole book and once inside LXX 33–37, at 34:6. Seven to one.

Seventy years. The Hebrew phrase שִׁבְעִים שָׁנָה stands 13 times in 12 verses of the Hebrew Bible; three of the twelve are Jeremiah’s (25:11, 25:12, 29:10) and the rest are Genesis 5:12, Genesis 11:26, Isaiah 23:15 (twice in the one verse), Isaiah 23:17, Zechariah 1:12, Zechariah 7:5, Psalm 90:10, Daniel 9:2 and 2 Chronicles 36:21. All three of Jeremiah’s survive into the Greek — LXX 25:11, 25:12 and 36:10 — which matters, because the seventy years is the one thing in these chapters that later Scripture quotes back.

Across the 105 CATSS rows against 7,927 New Testament verses: all 105 reach one, 103 reach two, 74 reach three, 27 reach four, 2 reach five, and nothing reaches six. The whole top of the table is two rows long:

Before anything is made of the two fives, here is the baseline. Screened chapter by chapter against the New Testament, 40 of the 52 chapters of Greek Jeremiah top out at five or lower. These five chapters are not unusually silent; they are ordinary. What is not ordinary in this book is the other end: Jeremiah 38 (Hebrew 31) reaches 23 twice, against Hebrews 8:9 and 8:11, because Hebrews is quoting it at length. A page that printed ‘the deepest run in Jeremiah 26–30 is only five’ without that comparison would be selling the reader a number.

Acts 4:10. Peter, on trial before the high-priestly family, says let it be known to all of you and to all the people of Israel. Jeremiah 27:16 is a prophet addressing the priests and all the people. The situations rhyme almost too well — and the shared material is ὑμῖν καὶ παντὶ τῷ λαῷ, an ordinary way of addressing a crowd. It stands in one New Testament verse and one Greek Old Testament verse, so the tie band on both sides is 1. A tie band of 1 means the string is rare. It does not mean the string is a citation.

Mark 9:39, and this one is sharper. Greek Jeremiah 36:9 says the prophets prophesy unjust things to you ἐπὶ τῷ ὀνόματί μου and I did not send them. Mark 9:39 has Jesus say do not stop him, for no one who does a mighty work ἐπὶ τῷ ὀνόματί μου and will be able soon afterward to speak evil of me. The same five Greek words are doing opposite work: one is a reason to stop people speaking in God’s name, the other a reason not to.

⚠️ And here is the deflation, printed at full strength. The tie band of 1 is produced by a conjunction. Measured across the whole Greek New Testament:

Four of the five words are a stock phrase; the fifth is and. Two of the seven verses that carry the four-word phrase — Matthew 24:5 and Mark 13:6 — are warnings about people who will come in Jesus’ name and lead many astray, which is Jeremiah’s subject exactly. That connection is real and it is thematic. The run screen cannot see it, and the run screen’s answer — Mark 9:39 — is not it.

Jeremiah 26:18 is the passage’s hinge and, so far as the standard literature goes, the clearest case anywhere in the Hebrew Bible of one writing prophet being quoted by name inside another prophetic book. The page cannot prove that negative by measurement — no screen held here can establish that nothing else in the canon does it — and says so. What it can do is measure the quotation. The result is that it is not verbatim, and that it breaks in different places in the two languages.

Words 4 and 5 are printed unpointed, because the difference is in the consonants and the vowel-points would hide it. Both are read as the qere; neither is a ketiv/qere.

Eight of the ten words are identical; two differ by a single letter each; the longest unbroken run is five. On the Hebrew screen Jeremiah 26:18 therefore comes in at 5 with a tie band of 7 — and six of those seven partners are not Micah at all but verses carrying the messenger formula לֵאמֹר כֹּה אָמַר יְהֹוָה צְבָאוֹת (Jeremiah 27:4, Jeremiah 39:16, Zechariah 1:4, 1:14, 1:17, 6:12). The one explicit quotation in the passage does not even stand alone at the top of its own row.

And in Greek the quotation is deeper than in Hebrew. Seventeen words are quoted on each side; fourteen are identical; three differ, and none of the three is a spelling — they are lexical choices:

Five in Hebrew, six twice in Greek. That is a real reversal and it is worth stating carefully. It does not mean the Greek is a better witness, and it does not mean anything about priority. It means that the two Hebrew texts differ by two letters and the two Greek texts differ by two words, and the run engine measures letters, not sense. A reader who wants to know whether Jeremiah is quoting Micah should look at the sense, where the answer is obvious; the measurement is here to show how little a measurement can settle.

Across the 119 Masoretic rows: 1 is a zero row, 17 stop at two, 37 at three, 25 at four, 18 at five, 9 at six, 7 at seven, 4 at eight, and one reaches twelve. The rows that reach seven or more:

Every deep partner but one is a verse of Jeremiah. The exception is Ezekiel 36:28, which shares seven consecutive Hebrew words with Jeremiah 30:22 — the covenant formula, and you shall be my people and I will be your God. The two prophets were contemporaries; the page reports the contact and declines to say which way it runs, whether either man knew the other, or whether both were using a formula already current. It notes only what the measurement adds: the Greek held here does not have Jeremiah 30:22 at all, so a reader of the Greek Jeremiah cannot make the connection from this verse.

The one zero row is Jeremiah 29:24 — “To Shemaiah of Nehelam you shall say” — which shares no two consecutive words with any other verse in the Hebrew Bible. (On this screen a 0 means exactly that; it may still share individual words.) And one zero row in 119 is below Jeremiah’s own rate, not above it: screened chapter by chapter across the whole book, Jeremiah has 41 zero rows in 1,364 verses, 3.0%, against 0.8% here. These are not unusually original chapters. They are unusually repetitive ones, and the page prints the baseline so that its single zero row cannot be mistaken for a finding.

1. Daniel 9:2 says outright that it is reading Jeremiah, and the run screen sends it to the wrong verses. “I, Daniel, perceived in the books the number of years that, according to the word of the Lord to Jeremiah the prophet, must pass before the end of the desolations of Jerusalem, namely, seventy years.” It names the prophet. It names the source. It is doing the thing every intertextual argument wants to find. Screened against every verse of Masoretic Jeremiah:

Daniel’s deepest contact with the whole book of Jeremiah is a chapter heading he is not quoting, in four verses he is not talking about. The two words that carry what he actually says he is doing come in at two.

2. 2 Chronicles 36:21 and Ezra 1:1 name Jeremiah too, and fare no better. Chronicles reaches two and its band is sixty-six. Ezra reaches four, band 1, on a verse about Media.

Three explicit, self-declared, named uses of Jeremiah, and this page’s method scores all three at or near the level of noise — and points one of them somewhere else entirely. That is the most important paragraph in Stage 4. Nothing on this page’s tables should be read as measuring use; they measure wording, which is a different thing.

3. Deuteronomy 18:22 is not sitting under Jeremiah 28:9. The test of a prophet — wait and see whether the word comes to pass — is unmistakably the same rule in both places. Measured: one consecutive Hebrew word (הַנָּבִיא) and two consecutive Greek words (ὁ προφήτης). The same goes for Jeremiah 29:13 and Deuteronomy 4:29, you will seek me and find me when you seek me with all your heart: one word in Hebrew, two in Greek.

4. The yoke of Acts 15:10 is not this yoke, measurably. At the Jerusalem council Peter asks why they should place a yoke on the neck of the disciples — and Greek Jeremiah 35:14 says God has placed an iron yoke upon the neck of all these nations. The run is three (ἐπὶ τὸν τράχηλον). Against Matthew 11:29, “take my yoke upon you,” it is one; against Matthew 11:30, “my yoke is easy,” it is zero — not one word in common. (On the Greek screen a 0 means no shared word at all, which is the strong sense.) Two texts can share an image, a body part and a piece of farm equipment and share almost no words. The page prints that here so that its own runs of five are not read as more than they are.

• Commentaries. William McKane, Jeremiah II (XXVI–LII) (ICC, 1996) — the most cautious on the Greek. Jack Lundbom, Jeremiah 21–36 (Anchor Yale, 2004). William Holladay, Jeremiah 2 (Hermeneia, 1989). Leslie Allen, Jeremiah (OTL, 2008). Georg Fischer, Jeremia 26–52 (HThKAT, 2005), which argues the opposite of the usual line on the Greek.
• On the two editions of Jeremiah. Emanuel Tov, The Septuagint Translation of Jeremiah and Baruch (1976) and Textual Criticism of the Hebrew Bible; Hermann-Josef Stipp, Das masoretische Sondergut des Jeremiabuches; Albert Pietersma and Marc Saunders in NETS; Andrew Shead on the two forms of chapter 32. The counts in Stage 3 are offered as data these works interpret, not as a verdict on them.
• On prophetic conflict. The standard monographs on true and false prophecy — among them James Crenshaw, Prophetic Conflict, and Thomas Overholt, The Threat of Falsehood — and the specialist literature on Jeremiah 27–29 and its Greek text. None was consulted at first hand for this page.
• Ancient Near Eastern comparisons. James Pritchard, ANET (3rd ed.) — the Babylonian ration tablets at p. 308 and the Lachish Ostraca at pp. 321–322 were read at first hand from that volume for this page.

The central claim about ψευδοπροφήτης is a claim about one corpus, and the corpus is small. The page counts ten occurrences in the 30,348 Greek Old Testament verse units held here. It did not search classical Greek, the documentary papyri, Philo, Josephus, or any Second-Temple Greek outside the CATSS files on this disk. It therefore cannot say that the translator coined the word, that the word was new, or that the New Testament writers got it from Jeremiah. A reader who leaves this page believing the Septuagint invented the word has read more than is written. That is the sentence a critic should press first.

No Qumran text is held here. The decisive witnesses for the shorter Greek Jeremiah are the Hebrew scrolls 4QJer^b and 4QJer^d, reported to preserve a short Hebrew text agreeing with the Greek against the Masoretic Text. The page prints Jeremiah 27:1, 30:10 and 30:22 with an empty Greek column and cannot say whether a Hebrew manuscript ever lacked them.

The Greek here is Rahlfs, not Ziegler. Every Greek figure on this page is a figure about the CATSS digitisation of Rahlfs’s Septuaginta. No Göttingen apparatus is held, so no variant is known to this page and nothing is claimed from one. In particular the fourteen absent verses, and the reading ψευδοπροφήτης in each of its eight places here, are Rahlfs’s text.

The Greek New Testament here is the SBLGNT via MorphGNT, and no New Testament apparatus was consulted.

“The only prophet quoted by name” is not measured. Stage 4 says that Jeremiah 26:18 is, in the standard literature, the clearest such case in the Hebrew Bible. No screen held here can prove that negative, and the page does not claim to have proved it. What is measured is the quotation itself.

The word-counts are counts of forms, not of meanings. The Hebrew figure of 134 for the root נבא counts every word form containing that consonant sequence, verb and noun together; the Greek figure of 98 counts every form containing the stem προφητ-, including augmented aorists and the compound with ψευδο-. They are comparable to each other only loosely, and the load-bearing part of the finding is the zero at the end of both, which no definition of the stem changes.

Direction of influence is nowhere asserted. Not for Micah and Jeremiah, not for Jeremiah and Ezekiel 36:28, not for the Septuagint and the New Testament, and not between the two editions of Jeremiah. The page reports arrangements and counts.

No Akkadian or Hebrew epigraphic corpus is held. The Babylonian ration tablets and the Lachish ostracon in Stage 2 were read in Pritchard’s English translations, not in the original languages, and were not screened against anything. The ostracon does not name Jeremiah, and Coniah son of Elnathan is not identified here with Elnathan son of Achbor.

No commentary was consulted at first hand. The works listed above are named as where a reader should go. No claim on this page rests on any of them.

The ESV column was typed from the printed 2016 text edition, because no ESV API key has ever been issued to this site. Any transcription error is ours. Where the ESV itself departs from the Masoretic Text — at 27:1, where it prints “Zedekiah” with a footnote at that word naming Jehoiakim — the page says so in the row rather than quietly printing three columns that disagree.

Where the page is weakest. Its headline is a count of a noun, and a count of a noun in one digitised edition of one corpus is a fragile thing to build on. If a wider Greek corpus turned up ψευδοπροφήτης in quantity outside Jeremiah, the distribution would still hold — eight of ten is eight of ten — but the sense that something is being done in these chapters would weaken considerably. The page would rather write that sentence itself than have it written about it.

• The text does not give you a way to tell them apart in the moment. Hananiah used the formula, stood in the temple, and performed a sign. Jeremiah’s first answer was “Amen!” and then he walked away with nothing to say (28:6–11). The word that answered Hananiah came afterwards. Anyone who thinks discernment is a skill you deploy on the spot should read Jeremiah 28 twice.
• The Greek reader was told and the Hebrew reader was not. This page’s central measurement is that the Greek column settles the question with a noun the Hebrew does not have. That is a real gain in clarity and a real loss of the difficulty. Be careful of any teaching that hands you the labels already attached — and be equally careful of pretending the labels never turn out to be right. Hananiah did die in the seventh month.
• The letter is the strangest thing in these chapters. To people who have just lost everything, God says: build houses, plant gardens, have children, and pray for Babylon. Not endure Babylon — seek its welfare, because in its welfare you will have welfare (29:7). Jeremiah 29:11 is addressed to those people, in that situation, about seventy years, and it is not a promise that this year will go well.
• One concrete response. Name one place you live, work or worship that you have been treating as somewhere to survive rather than somewhere to seek the welfare of. This week, do one specific thing for its good — and pray for it by name. That is Jeremiah 29:7, and it was written to people with far better reasons than yours to refuse.