A Hidden Layer of Sifra beneath the Mishnah: A Hebrew Palimpsest in the Cairo Genizah, T-S E2.51
Cambridge University Library T-S E2.51 preserves an early Eastern version of the Mishnah, Tractate Pesaḥim. The upper part of the same parchment leaf, housed at the Bodleian Library in Oxford, Heb. C. 27/27, contains the conclusion of Tractate ʿEruvin. Both pieces have been catalogued and occasionally cited for textual variants.1 Yet one striking feature has gone unnoticed: this leaf is a Hebrew palimpsest.
T-S E2.51 verso
On the verso of the Cambridge fragment, the Mishnah was copied over an erased text. Beneath it – still partially visible – is a passage from the tannaitic midrash Sifra. Hebrew palimpsests are rare to begin with, and even rarer when both upper and lower texts are rabbinic compositions.2
A close inspection of the page allows us to reconstruct the scribal process that led to this layered text. The direction of writing in the two texts is reversed: the Mishnah runs from top to bottom, while the Sifra runs from bottom to top. The ruling of the parchment corresponds to the layout of the Sifra, not the Mishnah. The two top lines of the Mishnah (that is, the bottom edge of the original page) were written in the unruled margin, while the bottom lines of the Mishnah page – corresponding to the uppermost lines of the Sifra – contain the first lines of the Sifra text itself. These lines, written along the ruled lines and unimpeded by further overwriting, remain especially clear and easy to read. This coherence strongly suggests that the original plan was to copy the Sifra, and that the parchment was repurposed partway through. From this point on, I refer to the page according to the Sifra’s orientation.
The preserved Sifra text begins at the top of the page, but this is not the beginning of the work itself. The upper part of the leaf was cut, and the surviving text starts mid-passage. Since the preserved section comes from Nedavah, chapter 2, section 13, we must assume that several earlier lines – and likely at least one additional leaf – are missing. The verso bears no more than seven or eight lines of the lower text, after which no further writing is visible. The recto shows no trace of the Sifra at all, suggesting that it was never written there.3 It seems that the scribe began copying the Sifra on the verso, but for reasons unknown, abandoned the project after a few lines. The rest of the page was left blank, and the parchment was later reused for copying the Mishnah. The script of both layers is strikingly similar, and although certainty is elusive, it is highly likely that both texts were copied by the same hand. Based on the layout and physical evidence, we may conclude that the scribe initially intended to copy the Sifra but changed course midway and decided instead to copy the Mishnah.
Composite image of the upper and lower texts from T-S E2.51, showing the similarity of the hand
The surviving lines clearly contain material from chapter 2 and likely also from the start of parasha 2. Although the preserved text is brief, and witnesses to Sifra are not particularly rare, the palimpsest yields a distinctive textual variant.
Recently, Menahem Kahana has analyzed the aggadic passage that concludes Nedavah chapter 2.4 At this point, the witnesses differ regarding the sequence of the homilies. According to the version found in Vatican manuscript 66 (with my own corrections), the passage reads:5
1. לאמר – צא אמור להם דברי(ם֯) כבושים: בישבילכם נדבר עימי! שכן מצינו שכל שלֹשים ושמונה שנה שהיו ישראל [כמנוד]ים לא היה מדבר עם משה, שנ' "ויהי כאשר תמו כל אנשי המלחמה למות מקרב העם וידבר ייי אלי לאמר".
2. דבר אחר לאמר – צא ואמור להם והש[י֯]ביני.
3. מנין שהיה משה יוצא ומדבר עימהם? שנ' ״ויצא ודבר אל בני ישראל את אשר יצוה״.
4. מנין שהיה משה מישיב את הדברים לפני הגבורה? תל' לו' ״וישב משה את דברי העם אל ייי״.
5. אלעזר בן אחביי אומר: יכול לא היה מדבר עימו לצורך עצמו? תל' לו' לאמר. לאמר לישראל לא היה מדבר עימו, מדבר היה עימו לצורך עצמו.
1. (… and He spoke to him from the tent of meeting) lemor ('saying') – Say it to them in humble terms (so that they accept it gladly), viz.: For your sake does He speak with me. For we find that all of the thirty-eight years that Israel was out of grace (because of the sin of the spies) He did not speak with Moses, viz. (Deut. 2:16-17): "And it was, when all the men of war (those from the age of twenty) had finished dying from amidst the people that the L–rd spoke to me."
2. Another nuance: 'saying' – Go out and say it to them and return word to Me (if they accept it).
3. And whence is it derived that Moses went out and spoke with them? From (Exod. 34:34): "… and he would go forth and speak to the children of Israel what he had been commanded."
4. And whence is it derived that He returned word to the Almighty? From (Exod.19:8): "And Moses returned the words of the people to the L–rd"
5. Elazar b. Achvai says: I might think He didn't speak to him for his own needs; it is, therefore, written: 'to say' – for Israel's sake he didn't speak to him, but he spoke to him for his own needs.
This order of the sentences appears in most manuscript witnesses. However, in Yalkut Shim'oni and in a thirteenth–fourteenth century Sephardic Genizah fragment (T-S C5.3), the fifth sentence appears before the second. Several other witnesses reflect knowledge of both versions and seem to allude to them.6
Our fragment contributes two notable insights. First, its sequence matches that of Yalkut Shim'oni and the Sephardic Genizah fragment. Thus, we now have an early Eastern textual witness that supports the sequence reflected in European sources. This is significant, because that sequence has often been suspected of being a later revision. Logically, it makes far more sense for the statement of Eleazar ben Aḥbai to follow directly after the first homily, since it disputes its content. According to this logic, the original order might be that found in most manuscripts, and Eleazar’s statement was originally appended at the end – out of place. A later revisor might then have moved it to its more appropriate position, as reflected in the European witnesses. On this view, the European version is textually superior but secondary. The discovery of an early Eastern witness preserving the same order suggests that this process may have occurred quite early.
Second, the version preserved in the palimpsest differs from all other known witnesses (including the European ones) in the first sentence. In the palimpsest, the words attributed to Eleazar ben Aḥbai appear identical to those of the anonymous tanna in the first homily. That homily ends with the line: לאמר <לי>ש<ר>א֯ל֯ לא היה מדב֯ר֯ ע[֯מ֯]?? מי֯דבר היה עימו לצורך עצמו – which closely parallels Eleazar’s statement. This could, of course, be a scribal error: the scribe may have accidentally incorporated Eleazar’s wording into the previous homily. If that were the case, however, we would expect the scribe to omit Eleazar’s statement later on, and yet it appears in full. We may therefore consider the possibility that this is not a mistake, but an alternative version of the text in which the anonymous tanna and Eleazar ben Aḥbai express the same view. Perhaps Eleazar’s teaching served as the basis for the homily in Sifra.
Even if we cannot fully reconstruct the textual development here, the scribe who began copying Sifra and then changed his mind has nonetheless offered us a meaningful contribution to the textual history of this midrashic passage.
Transcription of Sifra text from the palimpsest
[...] ע֯ם֯ מ֯ש֯ה֯ שנ' ויהי כאשר תמו כל אנשי המלחמה ל[מו]ת מק֯רב הע֯ם וידבר י֯['י אלי לאמר]
לאמר [לי]ש[ר]א֯ל֯ לא היה מדב֯ר֯ ע[֯מ֯]?? מי֯דבר היה עימו לצורך עצמו O֯ אלע֯ז֯ר֯ [בן מהבאי]
אומר יכו֯ל֯ לא היה מידבר עמו לצורך עצמו תל' לומ֯[ר] לאמר לאמר ל֯י֯שר֯א֯ל֯ לא היה מידב֯[ר
עמו מידבר] ה֯י֯ה֯ ע֯[י]מ֯[ו] ל֯צ֯ו֯ר֯ך֯ עצמו O֯ ד֯ב֯ר֯ [אחר] ל[אמר] צ֯[א ו]א֯מ֯[ר] ל֯ה֯ם֯ ו֯[השיבני]
א֯ – .Uncertain reading
[א] – .Tentative reconstruction
? – .Illegible or unclear letter in the manuscript
Footnotes
1 See for example: Y. Sussmann, Y. Rosenthal and A. Shweka, Thesaurus of Talmudic Manuscripts, vol. 1: Austria – England (Jerusalem: Ben-Zvi Institute, 2012; Hebrew), p. 114, no. 954; A. Goldberg, Commentary on the Mishnah: Tractate ʿEruvin (Jerusalem: Magnes Press, 1986; Hebrew), Introduction, p. 77, no. 31; G. Kochavi, The Impact of the Destruction of the Second Temple on the Halakhah and Customs of the Jewish Festivals (PhD diss., The Hebrew University of Jerusalem, 2003; Hebrew), p. 91 (Section G7).
2 For some examples of Hebrew over Hebrew palimpsests: Judith Olszowy-Schlanger, “An Early Palimpsest Scroll of the Book of Kings from the Cairo Genizah,” in B. Outhwaite and S. Bhayro (eds.), From a Sacred Source (Leiden & Boston: Brill, 2011), especially p. 240; Ronny Vollandt, “Palimpsests from the Cairo Genizah and the Qubbat al-Khazna in Damascus,” in New Light on Old Manuscripts, ed. Claudia Rapp, Giulia Rossetto, Jana Grusková, and Grigory Kessel (Vienna: Austrian Academy of Sciences, 2023), pp. 231–253. Note that some of the examples cited there are not true palimpsests but rather marginal annotations. Another example: Manchester, Gaster heb. frag. 139 (1v), where the upper text is the Bible (Esther), and the lower is likely a piyyut for Yom Kippur.
3 If we assume that the entire beginning of Sifra was copied on this page, one might suggest that the recto served as a title page. However, this is unlikely, as it would require an unusually large sheet. Alternatively, one might propose that the fragment reflects a version of Sifra that began near the surviving lines, but without further evidence this too seems implausible.
4 Menahem Kahana, “The Tosafot (‘Marginal Annotations’) in Sifre on Deuteronomy,” Tarbiz 87 (2020), pp. 215–217 (Hebrew).
5 The translation is based on Rabbi Sheraga Silverstein translation found in Sefaria, with some corrections. The manuscript, Vat.ebr.66, can be viewed here: https://digi.vatlib.it/view/MSS_Vat.ebr.66. The relevant passage is on f. 5v.
6 For a detailed description, see Kahana pp. 215–217.
If you enjoyed this Fragment of the Month, you can find others here.
Contact us: genizah@lib.cam.ac.uk
The manuscripts in this article are part of the Cairo Genizah Collection in Cambridge University Library. To see more items from this collection visit: https://cudl.lib.cam.ac.uk/collections/genizah/1