Jewish Mysticism

The Mystical Texts: Songs of the Sabbath Sacrifice and Related Manuscripts by Philip S. Alexander(Library of Second Temple Studies: T&T Clark) This essay provides an overview of a position I have worked out at greater length in The Mystical Texts: Songs of the Sabbath Sacrifice and Related Manuscripts (Companion to the Qumran Scrolls 7; London: T&T Clark International, 2005), to which the reader is referred for detailed documentation. The present article, however, is not just a summary of the book. The necessity of compressing and simplifying the case has led me, to some extent, to rethink and clarify my argument. A number of points (e.g., the anthropology behind Qumran mysticism, and the doctrine of predestination, which seems to be all over the relevant texts) now strike me as more important than I realised when I wrote the book. My purpose is to open a debate on what happens if we take certain Scrolls seriously as mysticism, and read them into the western mystical tradition. More

Shabbatai Donnolo's Sefer Hakhmoni: Introduction, Critical Text, and Annotated English Translation by Piergabriele Mancuso(Studies in Jewish History and Culture: Brill Academic)

Shabbatai Donnolo's. Sefer Hakhmoni 

Sefer Hakhmoni by the tenth-century, Jewish polymath Shabbatai Donnolo is one of the first texts written in Hebrew in medieval Europe and one of the most important documents of the "Hebrew Renaissance" of Byzantine Jewry in southern Italy between the ninth and the eleventh centuries. Written as a commentary on  Sefer Yesira (Book of Formation, an anonymous text probably written in Palestine between the third and the sixth centuries), Sefer Hakhmoni is in fact a much more complex work, consisting of biblical exegesis, astrology, medicine, a detailed analysis of the Neoplatonic idea of melothesia, and the correspondence between the elements of the microcosm and macrocosm. This volume offers the critical text, an annotated English translation, and a comprehensive introduction to Donnolo and his works.

Piergabriele Mancuso, Ph.D. (2009) in Jewish Studies, University College London, taught at Boston University's Center for Italian and European Studies, and he is now post-doctoral research fellow at Insubria University, Como. He has published several articles on early medieval Italian Jewry and on Shabbatai Donnolo's writings.

Excerpt: Donnolo composed Sefer Hakhmoni  in 946, sixteen years after the composition of Sa'adiah's Tafsir Kitab al-Mabadi and less then a decade before Dunash's commentary on  Sefer Yesira. The criticism which Dunash levels at Sa'adiah's interpretation of  Sefer Yesira throughout his own commentary testifies to the existence of an intense debate on  Sefer Yesira, which directly involved Sa'adiah, Dunash and probably also Isaac Israeli. While Sermoneta's definition of  Sefer Yesira's standing in the loth century as "the official philosophical text of Judaism" is probably an overstatement, his suggestion that Donnolo may have received the text of  Sefer Yesira from North Africa cannot be ruled out. The Neoplatonic ideas which, to varying degrees, inform the commentaries of Sa'adiah, Dunash, and particularly Donnolo, may well represent the lowest common denominator of an emergent exegetical tradition.'

This notwithstanding, the philosophical and exegetical scheme underlying Donnolo's commentary on  Sefer Yesira is primarily an elaboration of the Neoplatonic outlook which rabbinical tradition, without moulding it into a specific philosophical framework, had incorporated in its midrashic and aggadic traditions.

In all likelihood, Donnolo did not know Arabic and could not have had direct knowledge of Sa'adiah's commentary or the writings of Israeli, which in the Arabic speaking Jewish world, had contributed to the debate on  Sefer Yesira and its interpretations.' Both Sa'adiah and Dunash attribute to  Sefer Yesira's letters, sefirot, and the creative processes in which they are involved a symbolic meaning aimed, particularly in the case of Sa'adiah, at obliterating the apparently emanationist character of the process of creation described in the text in favour of what they took to be the biblical-

rabbinic notion of creatio ex-nihilo. In contrast, Donnolo inserts the events of creation as described in  Sefer Yesira into a coherent Neoplatonic scheme in which the sefirot and the combinations of letters are involved in a sequence of actions in the metaphysical world which are the necessary premise for the creation of empirical reality and material existence.

The Commentary on Genesis 1:26

As pointed out by Sermoneta, there is a close relation between the commentary on  Sefer Yesira and this section of the work, where Donnolo, by analysing Genesis 1:26, and above all by resolving the apparent paradox of man created in the image of an imageless God, defines the principles of a philosophical system where he can find "satisfactory answers to the problem of the outcome of reality from the First Cause" (my translation, P.M.).4 At the same time he outlines an interpretative scheme—that of the correspondence between micro and macrocosm—which he consistently applies to the entire analysis of  Sefer Yesira.

In the opening part of this section Donnolo gathers a series of verses selected from the book of Proverbs, all of which hint at the existence of Torah prior to the Creation of the world. Drawing from the midrash but, as observed by Sermoneta, also perfectly in line with the Neoplatonism of Philo who identifies the world of eternal ideas with Wisdom, Donnolo affirms that two thousand years before the Creation, God, by gazing at and combining the letters of the Torah, arranged all the elements of the future Creation in front of Him, so that everything that was to take place in the empirical world would thus be a manifestation of what God had outlined.

Probably conscious that such a predeterministic explanation would deny man's free will and the value of religious choice, Donnolo states, again elaborating on the midrash, that together with all other elements of the preordained scheme of Creation, "God ... set out and established repentance [teshuvah] since before the Creation of the world." Donnolo, however, goes beyond the notion of repentance preserved in the midrash—a moral act where human beings exert their free will—considering it also an action which can have direct repercussions on the natural universe, for example, by making the constellations retrocede. Since the divine plan finds perfect expression in the eternal and immutable movement of the stars and constellations, Donnolo assigns repentance to the Moon, the planet which has a twofold nature:

Since in the first hour of the eve of Friday, the earth brought forth every kind of living being ... and since in the same hour it occurred to God to create man, who knows [how to distinguish between] good and evil, the moon was appointed over good and evil as well as over beauty and ugliness, and its importance to all creatures is that it sets them up for better or for worse. For this reason, the moon was appointed to govern the first hour of the eve of Friday. From this you learn that the creatures were granted permission to act either malignly or benignly.

What lies behind the development of the idea of teshuvah seems to be, as observed by Sermoneta, the notion of reditus (literally, "return"), which Christian Neoplatonism, especially John Scotus Eriugena (ca. 810-after 877) in his Divisione Naturae, had conceived of as a universal event, "the means whereby the whole of Creation ... will be transformed into God." Donnolo elaborates and "expands" the traditional Jewish meaning of teshuvah but, it is important to note, does extend its definition to embrace the whole of the idea of Eriugena's reditus. This, an eschatological event which takes all reality back to its source, is not an expression of human choice and man's free will but an expression of the ineluctable laws governing the universe. Therefore, while there indeed seem to be some points of contact between the teshuvah and reditus—especially in relation to the idea of the return of creation to a previous state—Donnolo's teshuvah is, either in relation to the traditional idea of teshuvah or to the Christian reditus, an original formulation of the idea of return, which harmonises the law of necessity and man's free will.

Once the ideal world had been established and man granted the ability to practice free will, God "immediately embarked on the creation of the [material] world with His great might." As observed by Sermoneta,29 throughout the text Donnolo consistently uses the expression "great might" (koho ha-gadol)—probably a calque from the original Greek in connection with the creation of the empirical world and the creation of man. It is, as Sermoneta suggests, "a hypostasis, an intermediate entity which allows the realisation of the creation; it acts as intermediary between the ideal and this world" [my translation. P.M.]. Wolfson, following Sermoneta (whom, however, he does not mention) observes that the same expression occurs with the same meaning in many other Hebrew and Greek sources such as the Jewish apocryphal Vita Adae et Evae § 21, Wisdom of Solomon, and Mekhilta de-Rabbi Shim'on bar Yohai.

The Commentary on Sefer Yesirah

In the first part of the commentary Donnolo discusses at length the problem of the sefirot, which he defines simply as something that lies beyond human understanding and of which there can be no positive knowledge. The sefirot represent "the depth of the beginning which preceded the Creation of the world, or the depth of the end, which will follow the final end of the world". They are things about which "no man can speak [about them], or satiate his eyes with seeing and fill his ears with hearing them" since "their beginning is God and their end is God." The sefirot are "fixed in His great might," koho ha-gadol, by means of which, as Donnolo indicated in the conclusion of the previous section, God had concretised the forms of the ideal world. Knowledge of the sefirot, which are said to be part of God, would therefore amount to knowledge of God's mind: "The fact that it is written that they have no end teaches you that there is no sage in the world who can know, understand and delve into the mind of God ..."

Just as Donnolo discusses at length God's unknowabilty and the invisibility of His image in the commentary on Genesis 1:26," so he stresses in this section the unknowabilty of the sefirot and their ungraspable images. This represents a major departure from Sa'adiah's and Dunash's numerical understanding of the sefirot, and constitutes the first manifestation of their theosophical understanding, which will later become an exegetical cornerstone of the Kabbalah.

What is probably more important to underline, however, is not only the distance between Donnolo's exegetical position and those of Sa'adiah and Dunash, but also the discrepancy between the general rationalistic character of Donnolo's interpretative scheme and his theosophical position with regard to the sefirot. It seems that Donnolo is not able to apply to the sefirot the same rationalistic principles that inform his interpretation of  Sefer Yesira as a whole. This, as suggested by Wolfson, may well have been due to the influence of Jewish mystical and esoteric traditions—especially the study of the Hekhalot and the Merkavah literature—well attested in 10th-century southern Italy. The problem with this suggestion, however, is that there is no evidence of any mystical interpretation of the sefirot that might have been available to Donnolo in his southern Italian milieu, where the esoteric tradition appears to have been confined to the study and transmission of the esoteric traditions concerning the Hekhalot and the Merkavah literature alone, a body of writings which is represented only feebly in Sefer Hakhmoni , and where the sefirot, a neologism of the author of  Sefer Yesira, are not mentioned at all.

The process by which, two thousand years before the creation of the world, God had envisaged the forms of the ideal world, was the combination of the twenty-two letters of the alphabet which make up the Torah. According to  Sefer Yesira, the origin of the physical elements was the three "mothers" (aleph, mem and shin) which generated the three "fathers," the primary elements of air, water and fire, out of which everything came into existence. The totality of creation on earth was derived from the interaction of the three primary elements, but what led to the differentiation of individual items of Creation was a process of permutation by which every single letter of the alphabet was combined with another and endowed with a "crown," which put it in charge of specific functions in the empirical world:

He made the aleph king over the air [ruah], bound to it a crown, engraved it at the beginning of the word, which is speech, combined the letters with one another, and turned them into two words. He named the first aleph, mem, shin and the second aleph, shin, mem. He then formed breath [matt] out of His spirit [mho], and with it He formed the atmosphere in the universe, moisture in the year, and the torso in mankind, each one of them male and female, the male with aleph, mem, shin, the female with aleph, shin, mem ...

In the commentary on Genesis 1:26 he had already defined a series of such relationships, for example, between the parts of the human body (microcosm) and the universe (macrocosm), which perfectly matched the principles of traditional Ptolemaic astrology and Hippocratic medicine but not the series of micro-macrocosm correspondences of  Sefer Yesira. These involved, for instance, the order and relation of the planets to the hours of the day, the relationships between the months of the year and the signs of the Zodiac."

Donnolo does not intervene in the text of  Sefer Yesira, which he reproduces with only a few minor additions. Instead he offers right next to it a completely reformulated version of the subject. In g44 of  Sefer Yesira, for example, the correspondences between the planets and the days of the week do not follow the scheme of traditional astrology, where the sun corresponds to Sunday, the Moon to Monday, and so on. Donnolo reproduces it in full, but proceeds immediately to correct it:

With bet were formed Saturn, the Sabbath, the mouth, as well as life and death. With gimel were created Jupiter, Sunday, the right eye, as well as peace and war ... Even though Saturn, the Sabbath, the mouth, as well as life and death were formed with the letter bet, Saturn does not govern life or the mouth but only the Sabbath ... Even though Jupiter, Sunday, the right eye, well-being and calamity were created with the letter gimel, Jupiter governs only Thursday ...

Donnolo encounters similar problems in another long passage of  Sefer Yesira (4: 54), which defines the relationship between the zodiacal signs and the months as follows: "With he were formed Aries, Nisan, the liver, sight and blindness ..." thus linking the month of Nisan (corresponding to March-April of the Julian and Gregorian calendars) to the constellation

of Aries, and similarly for all the other months and zodiacal signs. What  Sefer Yesira describes is not completely erroneous in terms of Greco-Roman and Ptolemaic astrology but certainly not sufficiently accurate, lacking the necessary information with which to define precisely the beginning and the end of the period governed by each sign. Nisan, for instance, is correctly linked to the sign of Aries but without specifying that the relationship begins only from the time of the vernal equinox (roughly corresponding to the second third of March) and lasts for thirty days. Donnolo reformulates the entire section to bring it in line with the principles of Ptolemaic astrology: "Even though with he [the constellation of] Aries and [the month of] Nisan were formed, [the constellation of] Aries does not govern all the days of the month of Nisan but only from the vernal equinox for thirty days and ten-and-a half hours."

By the same token, Donnolo redefines all the relationships between the organs and their function in the human body which, as outlined in the  Sefer Yesira, did not correspond to the Hippocratic-Galenic medical tradition. The liver, for example, which Hippocrates identified as the source of blood, appointed to enable man to see and hear, appears in  Sefer Yesira in connection only to sight and blindness, while hearing is attributed to the bile, without any further specification." Donnolo reformulates the entire set of relationships, allotting to each part of the body the function that is normally ascribed to it in Hippocratic-Galenic physiology: "Even though the liver, sight and blindness were formed with he, the liver governs sight, hearing, and mercy, since blood is generated by the liver."

With this section Donnolo concludes his analysis of the processes which in  Sefer Yesira were said to have originated in the triad aleph, mem and shin, the three "mothers" from which the three primordial elements—air, water and fire, sources of everything in empirical world—were derived.  Sefer Yesira refers at this point to another triad, parallel to that of the three mothers, made up of the Dragon (tli), the sphere and the heart, and representing three dimensions—the world, the year and the human body—in which, by the combination of the three letters, empirical reality had found expression.

While in order to understand the last two elements of this triad—the sphere and the heart—and above all the dominant role they are given in  Sefer Yesira, ("The sphere in the year is like a king in his kingdom. The heart in the body is like a king at war") Donnolo does not need to depart from the principles of Hippocratic-Galenic medicine and Ptolemaic astrology, in order to explain the nature of the Dragon and the role it is given by  Sefer Yesira ("The Dragon in the world is like a King upon his throne"), he embarks on a long and complex analysis. This amounts to what is undoubtedly one of the most original and interesting elements of his astrological scheme—not completely compatible with the Ptolemaic system, and probably derived, as we will shortly see, from a Byzantine cosmological tradition, transmitting a certain trace of Gnostic influence.

In the introductory section of Sefer Hakhmoni  as well as in Sefer ha-mazzalot Donnolo ascribes to the term tli two distinct meanings. Firstly, it is the imaginary line connecting the two points known as "lunar nodes," along which the orbit of the Moon intersects the terrestrial ecliptic. This is the meaning of tli in the calendrical table which Donnolo inserted in the introductory section of Sefer Hakhmoni . Secondly, in Sefer ha-mazzalot Donnolo employs the term tli as a synonym of axis mundi, the central bar around which, according to the Ptolemaic tradition, the earth and all other celestial bodies rotate. As observed by Sharf, in both definitions the tli appears as an "inert" astrological and astronomical concept which certainly cannot properly express the power with which  Sefer Yesira endows it when it states that all the letters are "adhering to the Dragon."

The Dragon, explains Donnolo, was created out of water and fire," in the image of a reptile set in the fourth firmament, the abode of the Sun, to which are attached all the celestial bodies and the constellations. It can obscure the light of all the stars and conduct their movements around the earth. Its power, however, and particularly the movement which it transmits to the celestial bodies, derives from the sphere upon which God stretched the Dragon:

The [celestial] sphere turns the planets, the constellations and the luminaries. This sphere is set within the depth of the firmament, and the constellations are permanently attached to it, never straying from their fixed position. This sphere surrounds the firmament to the south, the north, the east and the west, and the Dragon stretches within it from end to end, like a cross-beam ... stretched out in the middle of the sphere, from the central mid eastern point to the central mid western point.

As suggested by Sharf, this definition of the Dragon does not correspond to any traditional astrological or astronomical concept but seems to be the reformulation of an old Gnostic theme, surviving in 10th-century Byzantium.64 According to this, the Demiurge had created a powerful Dragon, which was set to rule over the material world and the celestial bodies, and could obscure the celestial sphere as it moved through it.

Donnolo's discussion of the Dragon presents another difficulty. While speaking of the Dragon and its position in the sphere, he claims that the Wain (that is, the constellation of Ursa Major or the Great Bear) is close to the Dragon, that its extremities are attached to the Dragon's ring, and that it turns both the Dragon and the sphere in which all the constellations are set.

To the Wain, then, are ascribed the powers which Donnolo had previously attributed to the Dragon, and there seem to be many similarities between the two. This, explains Sharf, was probably due to an old astrological tradition, first attested in the Phaenomena (Visible Signs) of Aratus of Soli (3rd century BCE), according to which Arcturus—the brightest star of the Wain—was termed "the head and the tail." This is the same as the name of the lunar Dragon with which it was commonly confused. The astrological concept ensuing from this misunderstanding would eventually result in a composite notion of the Dragon, which was thought to be in the northern hemisphere and able to obscure the luminaries.

What Sharf does not explain, however, is why Donnolo, if he really considers the Wain and the Dragon to be one and the same, nevertheless makes a distinction between them by using two different names. One possible explanation comes from a passage in Sefer ha-mazzalot, in which Donnolo discusses the nature of Orion, a constellation of the northern hemisphere, and the Pleiades, a group of stars included in the constellation of Taurus. Orion, says Donnolo as he reports the opinions of the "sages of Babylonia and India," is the same as the Wain and the Great Bear which moves behind the Pleiades.69 Donnolo explains why the Wain, the Great Bear and Orion, even though they bear different names, are in reality the same constellation: the Wain consists of seven stars, two of them called Orion and five called the cords of Orion, as in Job 38:31: Can you bind the chains of the Pleiades (kimah), or loose the cords of Orion (ksil). The Pleiades, explains Donnolo, has seven stars called the chains of the Pleiades.

In order to explain why the Wain follows the Pleiades, Donnolo refers to an astrological tradition preserved in the Talmud," according to which when God decided to flood the world, He removed two stars from the Pleiades (kimah) by which He opened the gates of the sky. By the same token, when He decided to stop the flood, He removed two stars from the constellation of the Great Bear (`ash) with which He replaced the two stars previously taken from the Pleiades. This passage, says Donnolo, explains why the Great Bear, looking for its lost sons, chases the Pleiades, but also, and most importantly, why the constellations move. Donnolo refers again to the verse in Job 38:31, explaining that the ma'adannot ["chains"] and moshekhot ["cords"] mentioned in the verse refer to the twelve constellations (in Hebrew mazzalot) which Job mentions using a slightly different spelling in the following verse: Can you bring forth the Mazzaroth in their season? (ibid. 38:32)." This, argues Donnolo, shows that the Great Bear (that is, the Wain) is responsible for the movement of the constellations. All this, he adds, conforms with the hierarchical principle according to which "the constellations are like captains, the Dragon (tli) is like a king and the Wain is like the driver, and all are within the Spirit [of God]."

This might explain, firstly, why in his commentary on  Sefer Yesira, Donnolo refers to the Wain, even though the Wain is not mentioned in  Sefer Yesira, and secondly, why its role is apparently identical to that of the Dragon. As king, the Dragon has dominion over all the constellations and the celestial bodies, but its real power is exercised by the Wain, which acts in the universe on its behalf."

As observed by Castelli, the analysis of § 5 5 of  Sefer Yesira and the discussion of the Dragon is the last exegetically relevant part of Donnolo's commentary, while on the remaining chapters of  Sefer Yesira (§§ 56-63), which offer no more than a plain restatement of some of its previous subject matter, he confines himself to glossing, with no particularly original commentary.

From what we have seen so far, Donnolo adopts a very critical attitude towards  Sefer Yesira. He does not fail to identify and emend a series of statements which clash with some of the fundamental principles of his scientific background (e.g. the relations between the zodiacal signs and the months, between the planets and the days), although he does not dismiss the text and its authoritativeness. What, then, did  Sefer Yesira represent for Donnolo? A philosophical, a scientific or an inspired text and a segment of the holy canon of Revelation? Donnolo, it is important to note, does not express himself explicitly on this point but, unlike Sa'adiah and Dunash, he never rejects the idea that  Sefer Yesira was transmitted by God to Abraham. Where, how and in what circumstances did he imagine this transmission to have taken place? According to  Sefer Yesira, it happened when God made a covenant with Abraham, that is to say, following the biblical verse quoted in  Sefer Yesira (Gen. 15:6), at some indeterminate point after he left Egypt and came into the land of Canaan. Donnolo does not attempt to narrow this down any further, but what he says is significant, particularly as a means of understanding the way in which he viewed  Sefer Yesira: "It is written in the Book of Genesis which the Holy One—blessed be He—transmitted to our father Abraham and to Moses our Master on Mount Sinai: God said: "Let there be the light" [Gen. 1:3] ..."  Since according to both the biblical and the rabbinical tradition, it was Moses and not Abraham who received the Torah, of which Genesis is a part, on Sinai, the most plausible explanation of the entire passage is that by the "Book of Genesis" Donnolo refers not only to the biblical book of Genesis but to the account of the whole creative process, in both its metaphysical and its physical forms, including the account of  Sefer Yesira which, as the text itself claims, was transmitted to Abraham. This, after all, fits Donnolo's general understanding of  Sefer Yesira, which in his view describes, as we have seen, the metaphysical aspect of the Creative process, the formation of the ideal patterns out of which the empirical world came into existence, as described in the biblical book of Genesis.

Although he consistently, throughout the text, emends and reformulates  Sefer Yesira according to the principles of Hippocratic-Galenic medicine and Ptolemaic astrology, Donnolo's criticism does not go so far as to deny the veracity of  Sefer Yesira and, more specifically, the elements which, while not true in relation to the empirical world, are certainly valid in relation to the metaphysical world. To Donnolo,  Sefer Yesira does not represent a treatise on physical science—of which his critique would probably have been more explicit and much sharper—but the faithful description of the processes which took place in the metaphysical world prior to the processes described in the biblical book of Genesis, which are confined to the empirical manifestation of reality.

Sefer Yesira by A. Peter Hayman (Hardcover, Mohr Siebeck) This the first comprehensive critical edition of a text which was a fundamental influence on Jewish thought in the medieval period and has continued to fascinate scholars and students of Judaism to the present day. It was initially understood to be a philosophical text which had descended by oral tradition from Abraham himself. It purports to tell us how God created the world using the ten sefiroth (the Spirit of the living God, air, water and fire, and the twenty-two letters of the Hebrew Alphabet). With its English translation of the three earliest recensions and its commentary on the variant early texts of the work, this book is essential reading for anyone interested in the growth and emergence of the Jewish mystical movement. There are four appendices setting out what parts of the text are attested in each of the manuscripts and in what order, a hypothetical reconstructed text and the text of the tenth century Vatican scroll of Sefer Yesira with the probable added material underlined.

This edition of the text of Sefer Yesira has been a long time coming. I first conceived the idea of doing it in the early 1980s when I was reading the text with my students in a course on Jewish Mysticism at the University of Edinburgh. The fundamental research for the book was carried out in 1985 in a visit to the Microfilm Institute of the Jewish National and University Library in Jerusalem, funded by a grant from the British Academy. My initial intention was to produce an edition, translation and both a text-critical commentary and a commentary on the content. In the event it turned out that this was too ambitious a project to be accomplished within one book and, in any case, competing priorities, especially from the pressures of university administration, preventing me from producing more than a series of one-off papers and articles on Sefer Yesira. I now plan a series of three books: first, this edition, second, a collected edition of my papers on Sefer Yesira, and third, a commentary on the content of the text. This book, therefore, is concerned solely with the text — with the manuscripts, the recensions, the individual readings within the paragraphs. Issues of introduction, date, place of origin, and what the text might mean, will be reserved for the later books, though I have already dealt with many of these in my published papers. Of course, no rigid dividing line can be drawn between these different approaches to a text and, inevitably, I will stray into discussion of the content from time to time, but I wish to stress that this is not my primary purpose in this book.

Right from the beginning of the emergence of Sefer Yesira' into the light of day in the early tenth century it was recognized that its text had not been transmitted without errors. Saadya Gaon, the earliest commentator whose text has been preserved,2 states at the end of his introduction to  Sefer Yesira: "we think (it best) to write down each paragraph from it (i.e.  Sefer Yesira) completely, then we will explain it because it is not a book which is widely available and not many people have preserved it from suffering changes or alterations." Writing not much later than Saadya in 955/6 C.E., Dunash ben Tamim says: "mais nous avons déja dit qu'il pouvait y avoir dans ce livre des passages altérés que le patriarche Abraham [n'a jamais enonces], [provenant] des commentaires en hébreu, auxquels des gens ignorants ont ajouté postérieurement un autre commentaire et la vérité se perdait entretemps."4 The most comprehensive of the early commentaries, written by Judah ben Barzillai frequently quotes different versions of the text and discusses variant readings of which he was aware. Like Dunash he attributes the corruption of the text (almost

certainly correctly) to the incorporation into it of marginal notes and commentary material.' By implication Saadya locates this added material in the second half of the work (his chapters 5-8) when he remarks that there is little new in them and he does not intend to devote much effort to expounding them. Dunash explicitly attributes to the work of commentators the material, mostly in the latter part of  Sefer Yesira, which details the precise connections between each letter of the alphabet, element, and part of the human body.

These observations by the early commentators are fully vindicated when we come to compare the large number of manuscripts of  Sefer Yesira that have been preserved since the Middle Ages. If we just take a word count of the three manuscripts which serve as the base texts for this edition we can see the extent of the problem. Ms A (Vatican Library (Cat. Assemani) 299(8), fols. 66a-71b) has 2737 words, Ms K (Parma 2784.14, De Rossi 1390, fols. 36b-38b) has 1883 words, while Ms C (Cambridge University Library, Taylor-Schechter K21/56 + Glass 32/5 + Glass 12/813) has 2066 words. Some manuscripts have far fewer words than Ms K — as few as 1300, while others range anywhere between this low figure and the full range of material seen in Ms A.

From the tenth century on, then, it has been recognized that  Sefer Yesira existed in a number of recensions — some form of standard text, a longer version which contained commentary material, and a version which completely rearranged the material and which was attributed to Saadya Gaon.' Since the nineteenth century it has become conventional to refer to these versions as the Short, the Long and the Saadyan Recensions. The complex textual state in which  Sefer Yesira has been handed down is implicitly recognized in the first printed edition (Mantua 1562) in which the Short Recension is printed as the main text (with commentaries) and the Long Recension as an appendix. The fundamental work on delineating the recensions of  Sefer Yesira and working out which of them lay before the early commentators was achieved by A. Epstein in his articles in MGWJ. However, his fundamental conclusions that the Saadyan Recension is no older than Saadya himself and that the Long Recension is really only a copy of the text which is embedded in Shabbetai Donnolo's commentary' have been invalidated by manuscript discoveries of which Epstein was unaware at the time. As we shall see, it is more likely that the recensions predate any of the known commentaries on  Sefer Yesira.