The Leningrad Codex (LC) contains errors, omissions and ambiguities as in any ancient, hand written document. What to do with these problems challenges each publisher of the text. For example, the LC-diplomatic Biblia Hebraica Leningradensia (BHL) of Aron Dotan (z-l) lists 779 specific places where the body text differs from the LC in Appendix A. Many other divergences from the LC in qere/ketiv forms are described in the Forward, p. xx, xxi. Similarly, the UXLC 2.1 text contains divergences from the LC in qere/ketiv forms because its source, WLC 4.20, was originally a transcription of the eclectic Biblia Hebraica Stuttgartensia, BHS. The Unicode/XML Leningrad Codex (UXLC) intends to faithfully transcribe the LC text without editing. Consequently, the current changes are directed at improving the fidelity of the UXLC to the actual LC qere/ketiv forms; the number of changes is small and the changes will not cause difficulties for most readers. This effort was motivated by Ben Denckla's post "creating a pointed qere isn't always easy" (28 Feb 2024) to the bhebrew.biblicalhumanities.org list.

A source of the problem is the Masoretic representation of vowelized qeres. Because the qere markings are applied to the ketiv letters in the body text while the unmarked qere letters appear in the margins and the number ketiv and qere letters may differ, the placement of the marking on the qere letters may be ambiguous. In addition, the Masoretes were scrupulous with dageshes in the ketiv body text. For example, if a body text letter could not take a dagesh, they would not place a dagesh in the body text even when the corresponding letter in the qere can take a dagesh and needs it.

Two studies of qere/ketiv words noted in BHL and in WLC 4.20 are made here. The term unsupported indicates a mark in a qere is not actually present in the LC. Unsupported marks are added to qere forms by publishers to make the text conform to convention; they should not be present in the UXLC. A UXLC transcription note 'q' has been added for instances in which a UXLC qere word has been changed to match the LC rather than convention.

• q - "UXLC differs from the conventional word for the qere."

The bold numbers in open parentheses (n) indicate the change number n in each study in the 2024.10.19 - Changes.xml file.

• 2024.04.12 - 48 words mentioned in BHL Forword "Qere and Ketiv", p. xx-xxi.

  Four cases are described: Cases 1 and 2 describe usual qere and yatir-as-qere practice. Four instances of unsupported characters were restored (1, 2, 3, 9) and were given a transcription note q. In case 3 BHL "obliged to add" 19 qeres (10-28) all but seven are not marked as qeres in the LC. The UXLC 2.1 has a qere/ketiv pairs on Job 39:30.2 (9), 1 Sam 19:23.5 (25) and 2 Sam 20:8.15 (26); they will be removed. UXLC 2.1 has a questionable qere/ketiv pair on 1 Chr 12:19.7 (28); a 't' transcription note will be added and the qere will remain. In case 4 BHL removed 13 qeres (29-41) "[f]or the sake of convenience and for aesthetic reasons". None of these qeres will be removed from UXLC. A yatir in the LC for "Artaxerxes" in Ezra 4:11.8 will be added as a qere (43) to the UXLC following usual practice.

  An additional unsupported dagesh in a qere word will be removed and a 'q' note applied in 2 Sam 8:3.13 (47).

  Of 48 BHL change items only 5 change items(1, 2, 3, 9, 47), all noted 'q', will be made for fidelity of qeres to the LC text. One routine correction to an existing normal word Dan 6:1.2 (22) will be made and one 't' transcription note added (28). Three qere/ketiv pairs (9, 25, 26) will be removed and replaced with a single normal word (tag: w) and one qere/ketiv (43) will be added (tags: k, q) .

  Items 46, 47 provide examples of qere-without-ketiv and ketiv-without-qere words. These are the mechanisms by which the Masoretes added and removed entire words to create an improved qere text while still respecting the unmarked letter text.

• 2024.04.15 - 39 words having WLC 4.20 'a' bracket notes: "Adaptations to a Qere which L and BHS, by their design, do not indicate."

  Of 39 WLC 4.20 change items, 30 change items, all noted 'q', will be made to match UXLC qeres to the LC. The majority (26) of the note change items remove unsupported dageshes ( 1, 2, 4, 6, 7, 8, 9, 11, 12, 14, 17, 18, 19, 21, 23, 25, 26, 30, 32, 33, 34, 35, 36, 37, 38, 39). (Item 15 was done previously in 2024.04.12 - 47 and is a NoAction.) Some (9) of these unsupported dageshes were indicated by a qubuts in the LC; these qubuts will be restored to the text ( 1, 2, 7, 11, 12, 21, 26, 37, 39). The remaining unsupported characters were 3 maqafs ( 3, 28, 29) and a single sheva (22).

  Four of the changes (5, 27, 31) will be ordinary corrections. (Item 10 was done previously in 2024.04.01 : 2024.01.14 - 1 and is not counted in the total.) Six change items are marked as NoAction (13, 10, 15, 16, 20, 24).

These examples demonstrate that LC-diplomatic texts, such as the BHL and WLC 4.20, differ substantially from the LC in their qere/ketiv forms. Other differences between UXLC and LC qere/ketiv words remain to be discovered.

UXLC 2.1 has 305,492 unmarked words (words with w and k tags) with 1,278 qere words (words with q tags) representing 0.42% of the unmarked words. The total number of 'q' transcription notes added is 35; representing 3% of all qere words and 0.01% of all unmarked words. Thus the actual number of qere words in the UXLC corrected by these changes will be small.

2 May 2024