The MS Munich, Cod. hebr. 95: The cover spine and a sample page from the manuscript. Munich Digitalization Center

Surviving Manuscripts of the Talmud: An Overview

What we know about the surviving manuscripts, and how they contribute to Talmud study 1
Dr. Menachem Katz

 From Oral to Hand-Written to Printed Editions

For centuries, rabbinic literature, including the Babylonian Talmud, was produced, transmitted, and studied orally.2 Only in approximately the 9th century did scribes commit the text of the Babylonian Talmud to writing. In 1483, half a millennium later, the first edition of the Talmud was printed in Soncino, Italy.3

With the exception of some sparse documentation in the writings of the Geonim, there is virtually no remaining evidence for the Talmudic text as it existed during the era of oral transmission.4

The Surviving Manuscripts

The principal surviving evidence of the medieval Talmudic text is comprised of manuscripts written in various parts of the Jewish world from the 9th century until the rise of the printing press, at the end of the 15th century and the beginning of the 16th (and in the case of Yemen, the 17th century). These manuscripts (as well as the early printed editions, discussed below) are called “direct witnesses” to the Talmudic text.

Various genres of Geonic and Rishonic literature, such as commentaries, responsa, and halakhic treatises, also contain much important evidence for the medieval text of the Talmud; these witnesses, known as “indirect” or “secondary witnesses,” will not be discussed here.

The Scope of the Manuscripts

The only surviving manuscript that covers the entirety of the Talmud is the famed MS Munich, Cod. hebr. 95, which formed the basis of the widely respected, Eastern European Talmud scholar, Rabbi Raphael Nathan Nata Rabbinovicz’, Dikudei Sofrim (Munich, 1877) – the first major work of Talmudic manuscript variants.

The bulk of the manuscripts that have come down to us contain just one or two tractates, although some manuscripts cover an entire order (e. g. Seder Mo‘ed in MS Oxford, Opp. Add. fol. 23; Seder Kodashim in Vat. ebr. 120–121).  Several surviving manuscripts contain only a few folios, while in still others, especially from the Cairo Genizah and fragments discovered in the book-bindings of other books (the so-called “European Genizah”), only part of a single folio is left.

The number of surviving manuscripts thus varies from one tractate to another. For example, two manuscripts each of tractates Makkot and Horayot have survived, while nine of Sukkah and ten of Pesachim are known. The ratio of surviving manuscripts for specific tractates may reflect whether they were included in yeshiva curricula. Arguably, tractates that were commonly studied survive in a higher number of manuscripts, while more obscure tractates survived in fewer tractates.

Variation in Manuscripts

The manuscripts of complete orders in some cases contain tractates copied from different manuscripts and are thus uneven in quality. In some instances, different chapters of the same tractate were copied from different manuscripts. Care must be thus be taken to limit any conclusions about the character of a given section of manuscript based on its other sections.

Most of the manuscripts contain only the text of the Talmud. Rashi’s commentary, and sometimes Tosafot, occasionally appear in the margins of some, as in modern editions, and a few manuscripts include other compositions as well. 5

Most of the manuscripts contain corrections, rendered by either the original scribe or a later hand, and these sometimes attest to alternate readings that were known to the correctors.

Dating the Manuscripts

All known manuscripts – some sixty-eight in total6 – are written on parchment or paper.7 Some manuscripts state when they were created in a colophon, but many others can be dated only approximately, based on handwriting and other attributes.

The oldest extant manuscripts are from the eleventh or 12th century, with the oldest dated one was written in 1177; most were written in the 13th-15th centuries, though several from Yemen date to the 16th or 17th centuries. The last of these were produced after the printing press was introduced, but they were not copied from printed editions and therefore still witness the Talmudic text of earlier times.

Geographic Origins of the Manuscripts

The manuscripts and manuscripts fragments were written in many different places. Conventionally, their geographic distribution is assigned to the following categories:

  • Oriental, i.e., from the Middle East (specifically, from Iran to Egypt; see map);
  • Yemenite;
  • Sephardic, including the Maghreb (i.e., North Africa, including Tunisia, Morocco, and Algeria);
  • Ashkenazic (nominally Franco-German, but including other areas as well; see map);
  • Italy;
  • Byzantium, including Greece, the Balkans, and other areas (see map);

Below is a map illustrating the geo-cultural distribution of the different Hebrew codicological (characteristics as books) and paleographical (characteristics of writing) types.Click map to enlarge. From Malachi Beit-Arié, Hebrew Codicology, preprint Internet version 0.7 [December 2017], edited by Zofia Lasman [Hebrew], p. 68.

It is a remarkable and unfortunate accident of history that the majority of surviving, complete manuscripts come from regions other than the “Orient” (yellow), where rabbinic literature was initially composed. (We can determine the origin of each manuscript based on its style of handwriting—what is called “its hand.) The precise breakdown is as follows:

  • Ashkenazic or Italian hand  —   35
  • Yemenite hand  —  6
  • Sephardic hand  —   22
  • Byzantine  —    5

Noticeably, nearly all these surviving manuscripts are European or North African, while a few are Yemenite in provenance. None contain complete tractates copied in the Orient, i.e., Babylonia and its environs, where the Talmud originated, although pieces of manuscripts from this region have indeed survived.

The Singularity of Genizah Fragments

The availability of Genizah fragments, or remnants of manuscripts, has greatly contributed to textual scholarship of the Talmud. The complete manuscripts that have survived were produced at the earliest cases the 12th century, and in most the thirteenth or later, while a number of the Genizah fragments are older.

Although a precise date is unavailable for most fragments, some are known to have been created as early as the time of the Geonim, during the tenth and possibly even the 9th century, long before the first of the Rishonim.

Likewise, while most of the complete manuscripts were written in Ashkenazi lands, the documents from the Cairo Genizah come from Babylonian, Persia, the Land of Israel, Egypt, the coastal cities of Lebanon and Syria, and North Africa. Surprisingly, however, the Genizah also contains fragments from the other regions, including Ashkenaz.8

From Handwritten Manuscript to Printed Edition

With the founding of the printing press in the 15th century, the production of the Talmud became standardized. The earliest tractates were printed in Spain and Portugal, with the oldest in 1480, before the expulsion and forced migration of their Jews. Around the same time, the Soncinos, a family of printers in Italy (and later Istanbul) began printing selected tractates, beginning in 1483. Italy became a center for Talmud publication when an entire set was printed in Venice during the years 1520–1523.9

All editions produced since —including the classic Vilna edition of 1880–1886, which remains in use to this day10—are based on the incomparable Venice edition. These editions contain changes of various types, including printing errors and many purposeful corrections and emendations.

The Origins of Divergences

Although generally speaking manuscripts contain the same Talmud, there are differences, large and small, between them, and virtually no manuscripts are identical. The existence of many textual variants is familiar to us from the study of other classical Jewish works, such as the liturgical traditions of the various prayer books that exist: Ashkenaz, Sepharad, Edot ha-Mizrach, Yemen, and so on and so forth.

What gave rise to the discrepancies among the manuscripts and how are we to conceive of the relationship between these manuscripts and variants and the formation of the Talmud?

  • Do the changes we find in the manuscripts date back to the initial stage of oral transmission, and are they related to the process in which transmitters incorporated changes of various sorts to the material they were reciting, sometimes deliberately and sometimes not?
  • Or perhaps the changes came about as the Talmud was being copied by hand, during the time of the Rishonim, with copyists closely studying the Talmud even as they made changes and corrected the text?
  • Or perhaps some changes were made by scribes, each in his own style, in the transition from oral to written transmission?

As a rule, there are no unequivocal answers to these questions. Still, by cataloging differences in the manuscripts we have much to learn about how the talmudic text before us came to be.

Of Scribal Errors and Corrections

It is important to distinguish between scribal errors and scribal corrections. The Talmud was copied by professional scribes and students in a centuries-long process. One manuscript was copied from another (sometimes more than one), and mistakes, known as scribal errors, crept in during copying. Sometimes copyists read the text incorrectly, while on other occasions, they accidentally omitted entire words, phrases, and whole lines. Thus, one class of variants derives from scribal errors.

On the other hand, other manuscript variants owe their origin to “corrections.” Specifically, sometimes there is evidence of copyists intentionally changing the manuscript as they judged appropriate on the basis of the intensity of their study or the extent of their learning. They would study a passage, analyze it, and intentionally emend it. As some Rishonim observed about problematic passages in  the Babylonian Talmud: “the quills of manuscript correctors have gone over it.”11  Many if not most of the distinctions that separate one manuscript from another resulted from such deliberate attempts to correct or emend the text.

The Search for the Best Text

Textual scholars of Talmudic literature try to arrive at the best version of the talmudic text – to the extent that such a thing is possible. This task has the potential to greatly enhance our understanding of the Talmud.

In many cases, reconstructing a more original text contributes to our ability to comprehend both the content of the Talmud and its flow. Even when it is impossible to definitively conclude that one textual version is preferable to others, some of the readings contained in the various manuscripts and editions can be identified as secondary: versions that do not reflect the original, but came about as scholars and copyists of successive generations studied, copied, and corrected the text before them.  It is then possible to study why and when these reading originated, and to study them as part of the process of the transmission of the Talmudic text.

Familiarity with different manuscripts also allows us to uncover the specific texts Geonim and Rishonim used, and thus to better understand these commentators.

A New Tool: The Hachi Garsinan Project

Recently, a new web-based digital platform, the Hachi Garsinan project,12  was launched. The website displays all variant readings of any given passage of the Babylonian Talmud. Upon completion, the project will encompass all textual witnesses of the Talmud Bavli: Genizah fragments, manuscripts, early printings, fragments discovered in the bindings of other books,13 and other material found in public libraries and private collections all over the world.  The current version of the site includes all early printings, 36 complete manuscripts, and all Genizah fragments.

As the technology moves forward, it will be easier than ever for learners and lovers of the Talmud to check and compare medieval versions of the Talmudic text, thereby gaining a better understanding of a passage’s meaning and development.


Dr. Menachem Katz is Academic Director  Emeritus of the Friedberg Manuscripts Project in Jerusalem. He also lectures at the Open University of Israel and at Chemdat Hadarom College. Dr. Katz spends much of his time poring over handwritten fragments from around the world and has published widely on the Jerusalem Talmud, Aggadic literature, as well as in the field of Digital Humanities. His latest book, A Critical Edition and a Short Explanation of Talmud Yerusalmi’s Tractate Qiddushin, was published by last year (Yad Izhak Ben-Zvi & Schechter Institute of Jewish Studies, Jerusalem 2016). 
  1. This piece will be followed by future posts that exemplify the importance and method of using manuscripts for understanding the Talmud. For the present, see my “Five Cups of Wine at the Seder?, 2018.
  2. This position has become the scholarly consensus, and has been exhaustively described by Professor Yaakov Sussmann:

    When all is said and done, the “Oral Torah” is precisely what this designation implies: a Torah that not only is oral in the way it was given, but also was exclusively oral in its transmission, study, and development throughout the rabbinic period … Only if we truly internalize the full significance of the fact that the world of the Rabbis, from the first of the Pairs (zugot) to the last of the Sevoraim, was an oral one in the most absolute sense, a world whose literary works were produced orally and transmitted strictly orally—only then can we properly relate to and appreciate their literature.

    Yaakov Sussmann, “‘Torah she-be-‘al Peh,’ Peshutah ke-Mashma‘ah: Kocho shel Kotzo shel Yod” {Oral Torah as Such: The Power of the Tip of a Yod}, Mehqerei Talmud 3 (2005): 209–384, 355.For more on the orality of the Mishnah, see Michael Chernick “The Modern Study of Mishnah: Rabbi Dr. David Zvi Hoffman’s Approach”; on the orality of the Bavli, see Shai Secunda, “Why the Talmud is the Only Rabbinic Work from Babylonia.”

  3.   Leaves from slightly earlier, c. 1480 printed Spanish editions, have survived. However, these cannot be precisely dated.
  4. For the most up-to-date study, see Uziel Fuchs, The Geonic Talmud: The Attitude of Babylonian Geonim to the Text of the Babylonian Talmud (Hebrew; Jerusalem: Herzog, 2017).
  5.   For example, a manuscript to tractate Yevamot held in Moscow 594 includes both Rashi and Tosafot ha-Rosh while the famous Escorial manuscript to tractate Neziqin includes in the margins Yerushalmi Neziqin, as well as relevant rulings of the Rif and Rambam.
  6. See Yaakov Sussmann (ed.), Thesaurus of Talmudic Manuscripts (Hebrew; Jerusalem: Yad Ben Zvi and Friedberg Genizah Project, 2012), vol. 3, pp. 284–285, 6n72. Cf. Michael Krupp, “Manuscripts of the Babylonian Talmud,” in Shmuel Safrai (ed.), The Literature of the Sages I (Assen, Netherlands: Van Gorcum,1987), pp. 346–366, 351–366; see also pp. 346–350. Notably, four Genizah manuscripts are included here, including two in Sephardic, one in Byzantine, and one apparently in Italian script. MS Oxford of Sukkah (cat. no. 858), a Yemenite manuscript, is excluded, because whether it originated in the Genizah is very much in doubt.
  7. Some Talmudic fragments found in the Cairo Genizah are on other materials, like vellum and even leather.
  8.  See David Golinkin, Ginzei Rosh Hashanah: Manuscript Fragments of Bavli Rosh Hashanah from the Cairo Genizah (Hebrew; New York and Jerusalem, 2000), pp. 57-58; and Neil Danzig, A Catalog of Fragments of Halakhah and Midrash from the Cairo Genizah in the Elkan Nathan Adler Collection of the Library of the Jewish Theological Seminary of America (Hebrew; New York and Jerusalem, 1997), p. ד.
  9. Until that point, printers continued the scribal practice of producing only individual tractates, and at most, Talmudic “orders,” at a time.
  10. The Vilna edition, it should be noted, constituted another leap forward in the history of Talmud production. First, the quality of the printing was superior to many prior editions. Second, this edition included previously unpublished classic commentaries, culled from manuscripts found in the Vatican library and other European collections, and placed on the Talmudic page. In addition, the notes of some major Lithuanian sages were collected at the end of the volume, while the publishers also collaborated with more critically inclined scholars.
  11.   See Yaakov Sussmann’s introduction to the Academy of the Hebrew Language edition of the Jerusalem Talmud, Talmud Yerushalmi According to Ms. Or. 4720 (Scal 3) of the Leiden University Library with Restorations and Corrections, Jerusalem 2001, p. 18 n. 94; In addition, see Menachem Katz, The First Chapter of Tractate Qiddushin of the Talmud Yerushalmi: Text, Commentary, and Studies in the Editorial Process (Hebrew; Ph.D. dissertation, Bar-Ilan University, 2003), V. 1, pp. 29-30.
  12. The Hachi Garsinan project was founded by Dr. Dov Friedberg and is freely accessible through The Friedberg Jewish Manuscript Society portal   The site offers text search and a lexicon on all Bavli text-witnesses. It contains high quality digital images of all original text-witnesses, and many innovative tools, such as the ability to view versions side-by-side with precise interactive transcriptions of the text in the image.
  13. On the talmudic fragments found in European book bindings, sometimes referred to as the European “Genizah, see my “Introduction to Binding Fragments and other Fragments of Talmud Bavli Manuscripts,” presently on my website.
Print Friendly, PDF & Email