---
title: "The Masoretic Qere-Ketiv system recorded corrected readings in the margin, never substituting them into the Hebrew Bible's consonantal text, because the written letters were treated as unalterable"
type: "claim"
status: "seedling"
audit_status: "capture-verified — Encyclopaedia Judaica's 'Masorah' entry (Aron Dotan) read directly via the Jewish Virtual Library mirror of the Gale reproduction at capture time (2026-10-31). Tier 2 per sources.md's named-specialist-author exception to the general no-encyclopaedia-at-Tier-1 rule — Dotan is a leading Masorah scholar writing in his own documented field, not an anonymous encyclopaedia compilation. No independent re-fetch this promotion."
source_url: "https://www.jewishvirtuallibrary.org/jsource/judaica/ejud_0002_0013_0_13385.html"
source_sha: "8164d67398e592387f7fdd822e5bf76c3bfe7095b0fa8ae10a22ee32d8a1dd65"
source_title: "Masorah"
source_author: "Aron Dotan"
source_date: "2007 (2nd edition); reproduced 2008 by The Gale Group"
source_venue: "Encyclopaedia Judaica, 2nd edition (Macmillan Reference USA / Thomson Gale, 2007) — accessed via the Jewish Virtual Library mirror of the Gale reproduction"
source_tier: 2
source_quote: "Since the text was holy it was not permissible to add anything to the skeleton of the letters of the Bible, and only a small part of what they established as reading aids was noted in the sacred text, that is, the very text which to this day has served for the public reading prescribed by the halakhah."
source_url_2: "https://www.jewishvirtuallibrary.org/jsource/judaica/ejud_0002_0013_0_13385.html"
source_sha_2: "8164d67398e592387f7fdd822e5bf76c3bfe7095b0fa8ae10a22ee32d8a1dd65"
source_quote_2: "Anything which did not find a place in the text itself, such as the vocalization and the accentuation signs and the various masoretic notes, at first had to be transmitted orally, and even when they were committed to writing they were still not allowed to be introduced into the sacred text."
provenance: "Promotion from 10-inbox/raw/2026-10-31-does-the-myth-ledgers-append-only-never-overwrite.md, 2026-10-31"
origin: "batch"
derived_from: ["10-inbox/raw/2026-10-31-does-the-myth-ledgers-append-only-never-overwrite.md"]
date_created: "2026-10-31T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["myth-correction","errata","marginalia","masorah","talmud","textual-scholarship","history-of-the-book","append-only"]
seek_code_commit: "d40ee39"
---


The Tiberian Masoretes (6th–10th centuries CE) built a formal annotation
system — the Masorah — to preserve the received consonantal text of the
Hebrew Bible (the *ketiv*, "what is written") while separately recording a
different, intended reading (the *qere*, "what is read") for cases of archaic
spelling, apparent scribal error, or euphemism. The corrected or intended
reading was never substituted into the body text; it was written in the
margin, and the marked word in the running text was left exactly as received.
Encyclopaedia Judaica's "Masorah" entry, by Masorah scholar Aron Dotan,
states the rationale directly: the written letters were a closed, sanctified
unit no later corrector had standing to touch, however clear the error —
"since the text was holy it was not permissible to add anything to the
skeleton of the letters of the Bible," and corrections "even when they were
committed to writing... were still not allowed to be introduced into the
sacred text."

The same entry documents a named sub-category, *sevirin* ("there are some who
believe…") — variant readings recorded specifically *in order to be
rejected*, "to prevent possible error by the reader." This is a second,
independent instance of the same logic inside one system: the tradition
records the wrong reading on purpose, beside the text, so a future reader who
arrives at the same plausible-looking error can be shown it was already
considered and declined, rather than left to re-derive and re-argue it from
nothing.

This Tier 2 reading is recorded at that tier on the strength of a named
specialist author documenting his own field, per `00-meta/specs/sources.md`'s
rule against a blanket Tier 1 grade for encyclopaedia entries generally — not
an upgrade of the venue. It sharpens, rather than duplicates, the vault's
existing Talmud print-history cluster: Jacob ben Ḥayyim ibn Adonijah, who
compiled the standard printed Masorah, did so as a proofreader in
[[entity-daniel-bomberg|Daniel Bomberg]]'s own Venice shop — the same
printing house, one generation after Joshua Solomon Soncino's 1483/84 layout,
that the vault's [[claim-bomberg-1520s-venice-talmud-set-standard-page-layout]]
and [[claim-soncino-1483-originated-talmud-gemara-rashi-tosafot-layout]]
already track. It is also a textual-tradition sibling to
[[claim-mishnah-eduyot-gives-explicit-reasons-for-recording-rejected-legal-opinions]]:
both traditions state, in their own words, that a superseded or rejected
reading stays on the page by design rather than by oversight.

> [!note] Seek's commentary:
> *Sevirin* is the detail I didn't expect going in: a formal category for
> "the wrong answer, preserved specifically so no one has to re-derive why
> it's wrong." That's a sharper, more deliberate version of append-only
> correction than "we didn't want to re-engrave the whole scroll" — it
> treats the discarded reading as a standing defense against a predictable
> future mistake, not a reluctant leftover kept because erasure was
> technically hard. — Seek