The Scripture That Cannot Be Found

• Anchor Bible Dictionary — “Booths, Feast of,” “Siloam,” “Synagogue,” “John, Gospel of,” “Blindness.”
• Dictionary of Jesus and the Gospels (IVP) — “Feasts,” “Light,” “Signs and Wonders,” “Healing,” “Synagogue.”
• NIDNTTE — σκηνοπηγία, φῶς, τυφλός, ἀποσυνάγωγος.
• Dictionary of NT Use of the OT (Beale & Carson) — the John 7:38 entry, which is where the disagreement about the citation is laid out.

John is a New Testament book. There is no Hebrew of it and no Septuagint of it, so the MT/LXX offset that governs this shelf’s Old Testament pages does not apply. The Greek New Testament and the ESV agree on every chapter and verse number in John 7–9. What they do not agree on is how many verses there are.

The gap is exactly twelve verses and it is exactly John 7:53–8:11. The file runs 7:52 and then 8:12. This is not a chapter-division difference, and it is not an offset: no verse in the file carries a number that the ESV gives to a different sentence. Twelve verses are simply not there.

What this page does with that. It prints no Greek for 7:53–8:11, offers no measurement of them, and takes no position on them. Every count on the page is a count over 141 rows, and where a count would change if the twelve verses were included — the Moses distribution in the structure diagram is the clearest case — the page says so at the point where the count is made. The ESV prints the passage with a bracketed note of its own; readers who want the question argued should start there and with a critical commentary, not here.

The word σκηνοπηγία, Tabernacles, occurs once in the Greek New Testament — John 7:2 — and nine times in the Greek Old Testament files held here.

Zechariah 14 is the only Old Testament chapter that holds both of the things John 7 holds. It has three of the nine occurrences of the festival name, and at 14:8 it has ὕδωρ ζῶν — living water — going out from Jerusalem. John 7 is set at Tabernacles and speaks of living water. That is a thematic convergence and this page reports it as one.

⚠️ And now the counter-measurement, which is the point. John 7:38 against Greek Zechariah 14:8 returns a contiguous run of one: the word εἰς. The two-word sequence ὕδωρ ζῶν that Zechariah 14:8 uses does not appear in John 7:38 at all — John has ὕδατος ζῶντος, a different case, and the spelling fold this shelf applies does not join two different case forms. The convergence is of subject matter. It is not verbal contact, and a page that printed only the first half of that would be misleading its readers.

Direction is never asserted. Nothing here says John is using Zechariah, and nothing says he is not. The measurement can say what words the two passages share, which is almost none, and the thematic reading is left where it belongs — with the reader and the commentaries.

⚠️ This is the largest hole in the page and it is being put in the middle of the page rather than at the end of it.

Nearly every sermon and commentary on John 7:37 and 8:12 explains them from two ceremonies described in the Mishnah — the daily water-drawing from Siloam and the lighting of the great lamps in the Court of the Women during Tabernacles (Mishnah Sukkah 4–5). The connection is old, it is widely held, and it is the reason “if anyone thirsts” and “I am the light of the world” are read the way they are.

No rabbinic text was consulted for this page. The Mishnah is not on this shelf. It is not held in any corpus this page measures, it was not read, and nothing above or below rests on it. It is named here as the kind of evidence the standard reading depends on, not as something checked — and a reader should notice that the strongest cultural argument for how to hear John 7:37 and 8:12 is precisely the argument this page cannot make.

What the page can say from its own corpora is narrower and duller. Neither ceremony is described anywhere in the Old Testament’s own instructions for the feast: Leviticus 23:33–43, Numbers 29:12–38 and Deuteronomy 16:13–15 prescribe booths, offerings, rejoicing and pilgrimage, and say nothing about drawing water or lighting lamps. That was checked rather than assumed: across those 41 Greek verses, not one carries the root ὑδ- (water), λυχν- (lamp), λαμπ- or φω- (light) at all. The ceremonies are later practice, and later practice is exactly what a biblical-text corpus cannot see.

9:7 stops to gloss the name: Siloam (which means Sent). The Gospel does this twice in the passage — it glosses Tabernacles as the feast of the Jews at 7:2, and it glosses Siloam at 9:7. Both glosses are the narrator’s, not a character’s.

The name is measurable, and it is rarer than readers expect. Σιλωαμ occurs once across all 30,348 screened verse units of the Greek Old Testament held here — Isaiah 8:6, the waters of Siloam that go softly. It is not carried over into the Greek of Nehemiah 3:15, where an English Bible reads the Pool of Shelah. That one occurrence was not reached by any run screen on this page; it is named because the place is the same place, which is a geographical statement and not a textual one.

No archaeological report was consulted. The excavation of the Second-Temple pool south of the City of David is well known and is not evidence this page handled.

The two words ἐγώ εἰμι stand in six verses of John 7–9.

A word-form screen cannot tell those apart, and the page will not pretend otherwise. Across the whole Greek New Testament the two-word sequence stands in 47 verses, 24 of them in John. Across the Greek Old Testament it stands in 182 (183 until 30 August 2026 — the 183rd was Odes 2:39, which is Deuteronomy 32:39 copied out again as a canticle, and the Odes are no longer screened). Most of those 182 are somebody ordinary saying “it is I”. Which occurrences carry the weight of the divine name is a judgement made from context, from the reaction of the hearers, and from the shape of the sentence — not from a count.

The clearest internal evidence in these chapters is not lexical but narrative: after 8:58 they pick up stones, and after 9:9 they ask the man how his eyes were opened. The same two words, and two entirely different rooms.

He says I am the light of the world twice, and the Greek is not the same twice.

Because the verb sits between the noun and its genitive at 9:5, a contiguous-word screen does not join these two verses past two words. They are the same claim and they are not the same string, and this is the clearest illustration on the page of what a sequence screen is blind to. The ESV renders both identically, which is a translator’s judgement about sense and is the right judgement — but it hides the difference the Greek shows.

All 141 Greek rows of John 7:1–9:41 were screened three times: against the Greek Old Testament, against the New Testament outside John, and against the rest of the Greek New Testament including John’s own Gospel. The results are not close.

Nothing in these three chapters reaches six shared words with anything in the Greek Bible — except with John.

What this does not mean. It does not mean John 7–9 is unconnected to the Old Testament. Chapter 7 argues about circumcision and the law of Moses; chapter 8 argues about Abraham; chapter 9 argues about the Sabbath. The passage is soaked in the Old Testament at the level of subject. What the measurement shows is that it is not quoting: it argues about Scripture without reproducing its sentences.

The Odes came off this shelf’s Greek Old Testament. They are still on disk; they are no longer screened. Two of the numbers this page published were counting the same verse twice, and a third was a tie band inflated the same way.

Nothing else on the page moved. The deepest run against the Greek Old Testament is still five, still two rows, still John 7:13 against Esther 8:17 and John 9:40 against Genesis 24:54, both tie band one. The distribution table is unchanged at every threshold. John 7:38 still measures four with a band of two, and the band is still Judges 3:22 in the A text and the B text.

Why the Odes were removed, in the shelf’s words rather than a summary. The Odes are the fourteen liturgical canticles bound after the Greek Psalter, and they are not an independent witness to anything. Ten of the fourteen are second copies of passages the corpus already holds — Ode 1 is Exodus 15, Ode 2 Deuteronomy 32, Ode 3 1 Samuel 2, Ode 4 Habakkuk 3, Ode 5 Isaiah 26, Ode 6 Jonah 2, Odes 7 and 8 Daniel (Theodotion) 3, Ode 10 Isaiah 5, Ode 11 Isaiah 38. A verse standing once appeared to stand twice, so every tie band touching a canticle was doubled, which is precisely what happened at John 8:28.

⚠️ And the other reason is worse, though it never touched this page. Two of the Odes are New Testament text. Ode 9 is Luke 1 — the Magnificat and the Benedictus — and Ode 13 is Luke 2, the Nunc Dimittis. Screened as Greek Old Testament, Ode 9:48 matches Luke 1:48 at eighteen consecutive words. A page could report that as a Gospel’s deepest Old Testament contact while measuring nothing but its own text. No hit removed from this page was in Ode 9 or Ode 13 — John is not Luke, and the two Odes rows this page ever touched were Ode 2 and Ode 6, both plain duplicates. The trap is named here anyway, because the instrument can do that and a reader is entitled to know it was checked.

Held and screened are now different numbers, and this page uses the right one in each place. The shelf HOLDS 30,637 CATSS verse units in 64 files. It SCREENS 30,348 in 63; the difference is exactly the 289 units of the Odes. Every “screened against” sentence on this page reads 30,348.

Genesis 24:54 is Abraham’s servant and his men eating and drinking at Bethuel’s house. The five words are those who were with him, and. There is no argument to be made from that and the page makes none.

Esther 8:17 is the better warning. The five words — because of the fear of the Jews — are a real phrase with real content, and they sit in a verse about Gentiles submitting out of dread. It would be easy to build something on that. Do not. The same five words stand twice more in John itself — 19:38, where Joseph of Arimathea is a secret disciple, and 20:19, where the doors are locked — and nowhere else in the New Testament. It is John’s own stock phrase, used three times, and its appearance in Esther is a coincidence of common vocabulary.

⚠️ The deepest run on the page is built on a house formula and is therefore the weakest kind of eight. Four of John 8:34’s eight words are ἀμὴν ἀμὴν λέγω ὑμῖν — a formula that occurs twenty times in the Greek New Testament and all twenty in John. A writer with a catchphrase will score high against himself. An eight-word run inside one author’s own book is evidence of idiom, not of quotation, and reporting it as anything more would be exactly the error these pages exist to avoid.

Which is why the finding is stated in the negative and not the positive. The claim is not John 8:34 quotes John 10:7. The claim is that the passage has no six-word partner outside its own Gospel at all — not in the Law, not in the Prophets, not in Paul, not in the Synoptics.

A note on the spelling fold, checked rather than assumed. This shelf folds οὐ / οὐκ / οὐχ and ἐκ / ἐξ together before matching, so a run is not cut at a word two editions merely spell differently — which means a fold can lengthen a run that the printed texts do not actually share. The four rows in the table above were checked against the unfolded text and none of them depends on it: John 8:23, 17:14 and 17:16 all read οὐκ in the SBLGNT, so the six-word run stands without the fold. The fold is applied to every screen on every page and is documented in the footer.

John 7:38 says καθὼς εἶπεν ἡ γραφή — as the Scripture said. It is the only explicit citation formula in these three chapters. Here is every candidate, measured.

The last column is the longest contiguous run, not a list of every word the two verses share. A row measured at one can still share several isolated words — John 7:38 and Jeremiah 2:13 share ὁ, ἐμέ and ὕδατος; John 7:38 and Zechariah 14:8 share εἰς, ἐκ and αὐτοῦ — but in none of them do two shared words stand next to each other. A zero means no word at all is shared.

The deepest Old Testament contact of the only verse in John 7–9 that claims a Scripture is Ehud’s sword in Eglon’s belly. Judges 3:22: the fat closed over the blade, for he did not draw the sword out of his belly. Four words, tie band of two — and the two are the same verse in the A text and the B text, so the band is really one verse counted twice. It is not a source. It is what a four-word run looks like when there is nothing better in the corpus.

Two supporting counts. The two-word sequence ὕδατος ζῶντος occurs in exactly three Greek Old Testament verses — Genesis 21:19, Genesis 26:19 and Song of Songs 4:15 — and every one of the three is a literal well or spring, none of them a person. And ποταμοὶ ὕδατος, rivers of water, occurs nowhere — zero in the Greek Old Testament and zero in the Greek New Testament, John 7:38 included, because the two words stand five words apart here rather than side by side.

Two honest limits on this result. First, a contiguous-run screen measures reproduction, not paraphrase; a writer summarising several passages at once would score exactly this low, and many commentators think that is what is happening here. The measurement rules out copying; it does not rule out composite allusion, and nothing on this page should be read as saying the verse has no Old Testament background. Second, where the sentence divides is editorial. The oldest manuscripts carry no punctuation, and whether out of his heart belongs to the believer or to Christ depends entirely on where a modern editor puts the stop. This page prints the Greek as the SBLGNT punctuates it and takes no position on the division.

All 41 rows of John 9 were screened against the Synoptic blind-man pericopes — Mark 8:22–26, Mark 10:46–52, Matthew 9:27–31, Matthew 20:29–34 and Luke 18:35–43, 32 verses in all.

Four of the six threes are narrative furniture — and he said, Jesus said to him, Jesus said to them, he said to him, go. The other two are worth naming: John 9:6 and John 9:15 both share with Mark 8:25 ἐπὶ τοὺς ὀφθαλμούς, on the eyes — and Mark 8 is the one Synoptic healing in which Jesus uses spittle and touches the eyes, as here.

Three shared words, and one of them the preposition. That is not a literary relationship and this page does not claim one. It is reported because the two episodes are constantly compared, and a reader deserves to know that the comparison is a comparison of method, made by readers, and not of wording, which the texts do not share.

1. The Mishnah was not consulted, and it is the backbone of the standard reading. Mishnah Sukkah 4–5 on the water-drawing and the lamps in the Court of the Women is where nearly every commentator grounds John 7:37 and 8:12. It is not on this shelf, it was not read, and no claim on this page rests on it. That means the page’s cultural stage is the thinnest stage it has, and stage 2 says so in its own body rather than burying it here.

2. John 7:53–8:11 was not measured, because the Greek file does not have it. No manuscript, no critical apparatus beyond the SBLGNT’s own sigla, and no patristic citation was examined. The page cannot say anything about those twelve verses and does not try. The one structural claim they touch — that Moses is not named in chapter 8 — is flagged at the diagram itself.

3. Every number here is a number about two specific editions. “Nothing reaches six” is a statement about the SBLGNT as distributed by MorphGNT, screened against the CATSS Greek Old Testament files held on this shelf. A different Greek New Testament, or a fuller Septuagint, could move it. The CATSS collection here does not include every Greek witness, and where a book exists in two Greek forms — Judges A and B, Daniel OG and Theodotion — both are counted, which inflates tie bands. The Judges 3:22 band of two in stage 4 is exactly that: one verse in two recensions.

No Greek critical apparatus beyond the SBLGNT’s own marks. For scale: the SBLGNT flags 126 apparatus notes across these three chapters, in 83 of the 141 verses — nearly three verses in five carry at least one point where the edition records a difference from another printed text. None of those was examined; the page uses the SBLGNT’s text as it stands and strips the marks for display.

No Mishnah, Tosefta, Talmud or Targum. No Josephus. No Dead Sea Scrolls. No patristic commentary — and John 9 has a very large one in Augustine’s tractates, which was not read. No Syriac, Latin or Coptic version. No archaeological report on the Pool of Siloam.

And one methodological limit that applies to every screen above. These screens find verbatim contact — contiguous runs of the same word forms. They do not find allusion, echo, paraphrase, or argument from a passage. A zero on the Greek side means the two verses share no word at all, which is a strong result about wording and a weak one about meaning. John 7–9 argues with Scripture continuously and reproduces it almost never, and only the first half of that sentence is something this page measured.

• Raymond E. Brown, The Gospel According to John I–XII (Anchor Bible 29).
• C. K. Barrett, The Gospel According to St John, 2nd ed.
• D. A. Carson, The Gospel According to John (Pillar).
• Craig S. Keener, The Gospel of John: A Commentary — the fullest treatment of the Tabernacles background.
• Chris Keith, The Pericope Adulterae, the Gospel of John, and the Literacy of Jesus.
• None of these were consulted for the measurements on this page, which were run on the text files named in the footer.

Chapter 9 is an interrogation. The neighbours question him, the Pharisees question him, the parents are questioned and refuse to answer, and then the Pharisees question him again. He is asked the same thing five different ways by people who already know what answer they want.

And in the middle of it he says the sentence that has outlived every argument in the chapter: whether he is a sinner I do not know; one thing I do know, that though I was blind, now I see. He does not win the theological exchange. He is not equipped to. What he has is one fact he cannot be argued out of, and he holds it until they throw him out.

Then — and this is the part the interrogation scene is built to set up — Jesus goes and finds him (9:35). The man is not restored to the synagogue. He is found by the one who healed him, outside it.

What the numbers on this page are for. Not much of chapter 9 could be traced to anywhere else; 18 of its 41 verses share no more than a single isolated word with the Synoptic healings, and the whole passage has no six-word partner in the Old Testament. The measurement’s honest conclusion is that this story is told in John’s own words and nobody else’s. Which is worth knowing before anyone builds a sermon on a parallel that the wording does not support.