פרקי רבי אליעזר

PIRQE RABBI ELIEZER
ELECTRONIC TEXT EDITING PROJECT

Is Every Medieval Hebrew Manuscript a New Composition? The Case of Pirqé Rabbi Eliezer*

Lewis M. Barth

TABLES
1. Pirqé Rabbi Eliezer Manuscripts and Fragments: Sorted by Type and Date
2. Libraries
3. Catalogues Of Collections Of Hebrew Manuscripts

Introduction

With an article published in 1986, Peter Schäfer launched a debate that has far reaching implications for the study of rabbinic literature in general and midrashic literature in particular.⁽¹⁾ Schäfer raised the most fundamental questions about the nature of the sources of rabbinic literature: the texts themselves.

The terms with which we usually work—text, ‘Urtext', recension, tradition, citation, redaction, final redaction, work—prove to be fragile and hasty definitions that must be subsequently questioned. What is a ‘text’ in rabbinic literature? Are there texts that can be defined and clearly delimited, or are there only basically ‘open’ texts, which elude temporal and redactional fixation? Have there ever been ‘Urtexte’ of certain works, with a development that could be traced and described? How do different recensions of a ‘text’ relate to one another with respect to the redactional identity of the text? How should individual tradition, the smallest literary unit, be assessed in relation to the macroform of the ‘work’ in which it appears? What is the meaning of the presence of parts of one ‘work’ in another more or less delimitable ‘work? Is this then a quotation in work X from work Y? And finally what is redaction or final redaction? Are there several ‘redactions’ of a ‘work'—in chronological order—but only one final redaction? What distinguishes redaction from final redaction? What lends authority to the redaction? Or is the final redaction merely the more or less incidental discontinuation of the manuscript tradition?⁽²⁾

Both implicitly and explicitly, Schäfer’s argument undermines the goal of identifying a “best text” or creating a critical edition of any rabbinic work in the classical sense. Chaim Milikowsky noted this as he attacked Schäfer’s view that there is no “Urtext.” “This [Schäfer’s] is indeed an extreme claim and one would expect it to be preceded by extensive textual analysis of at least several rabbinic works.⁽³⁾ Using the examples of two quite divergent manuscripts of Bereshit Rabba (Vatican 30 and 60), Milikowsky argued that “the question of recensional variation should not be identified with the question of redactional identity.”⁽⁴⁾ He notes that among the possible causes for recensional variations is that

... some traditions of Jewish scribal activity did not hesitate to ‘correct’ the text by emending it and by adding to it, thereby causing the text to receive the shape it—to the scribes’ mind—should have had originally. This would create recensional variation, but by no means should we call this a new redaction of the text, which had already received its fixed form centuries earlier.⁽⁵⁾

Based on his view of the relative stability of manuscript tradition (i.e. only one or two manuscripts reflecting “widely divergent readings”) Milikowsky argues for rigorous stemmatic analysis of the corpus of early rabbinic texts. He recognized that it might be impossible to reconstruct an Urtext of Bereshit Rabba; however that does not mean that an Urtext never existed.

In response to Schäfer’s question—‘are there texts that can be defined and clearly delimited?’—Milikowsky responds with an unequivocal ‘yes.’ While the form of a synoptic edition may be suitable for the publication of Hekhalot texts, “there can be no doubt that the traditional format of a critical edition with a single text printed is suitable for the vast majority of rabbinic texts.”⁽⁶⁾

At the conclusion of his article, Milikowsky seems to make a concession to the possibility of “open texts.” He notes that the editors-homileticists of later rabbinic texts, such as Midrash Tanhuma,

... did not see the texts they produced as closed. Every scribe allowed himself the freedom to re-edit and every scholar the freedom to reformulate, and it would indeed be wrong-headed to search for the original Midrash Tanhuma.⁽⁷⁾

The majority of the extant manuscripts of rabbinic literature—both early and late texts—were produced in the Middle Ages. In an article dealing with the notion of medieval Hebrew compositions as “open books,” Israel M. Ta-Shma has described the relationship of authors to their books:

A long and intensive review of the medieval Hebrew book indicates that quite often books were not meant by their authors to serve as final statements, but rather as presentations of an interim state of knowledge or opinion, somewhat like our computerized databases, which are constantly updated and which give the user a summary of the data known at the time of the latest updating. In a similar way, the medieval book was sometimes conceived of as no more than a solid basis for possible future alterations by the author himself. There were many reasons … for this profound phenomenon, which can give rise to serious problems as to finality, authorship and authority of a given text of a work.⁽⁸⁾

Malachi Beit-Arié, adds to the picture of textual transformations in his article “Transmission of texts by Scribes and Copyists: Unconscious and Critical Interferences.” He sets out to “challenge ... our overestimation of the textual evidence of the medieval manuscript and our concept of the nature of scribal reproduction of texts.”⁽⁹⁾ Beit-Arié, a paleographer and codicologist, calls attention to the “special circumstances under which texts were published in the Middle Ages.” These include:

• the length of time during which an author composed his text;
• the works the author might be reading during this period from which he might introduce new material into his composition;
• the fact that texts were copied at different stages of their creation;
• and that an author had no control over his text once it was disseminated.

In spelling out what he describes as the “dynamic state of the text and its transformational character,” Beit-Arié adds an additional dimension to the description of medieval authors’ attitudes toward their own works. In contrast to the modern concept of the exclusivity of intellectual property, medieval authors seem to have a “concept of collective ownership.” This is reflected in the invitation various authors extend to readers to correct mistakes in their copies and even to add material to them. Further, unlike scribes who worked in the context and traditions of Christian monasticism, Jewish scribes—particularly those who copied texts for their own use—exercised considerable editorial freedom in reproducing texts. (According to Beit-Arié, this was true for all geographic areas in which Jews lived except for Yemen—a point that will be significant for the study of Pirqé R. El.)

In a response that contrasts strongly with the position of Milikowsky, Beit-Arié concludes his own analysis with these words:

…owing to the circumstances of medieval publication, texts were disseminated at various stages of their creation and revision, and their authors were usually prevented from controlling what happened to them. Reproduction and distribution of texts were never institutionalized in Jewish societies, but were carried out by individual private initiative, to a large extent by learned people or scholars who themselves copied the books they wished to study or use. Encouraged by authors to correct their own mistakes, and being aware of the unavoidable corruption of texts by the unconscious mechanics of copying, copyists certainly did not view copying as mechanical reproduction, but as a critical editorial operation involving emendation, diagnostic conjecture, collation of different exemplars and even the incorporation of external relevant material and the copyist's own opinion. It seems that the copyist's main goal was to establish what Kantorowicz defines as a ‘richtige’, right, version, as opposed an ‘echte’, authentic, one. Consequently, most of our Hebrew manuscripts present texts not only corrupted by the accumulation of involuntary copying errors, but also distorted by editorial or even redactional reconstruction, by contamination from different exemplars and versions, and by the deliberate integration of related texts. What medieval copyists performed while copying was indeed what in modern theories of criticism is known as deconstructing the text and then reconstructing it. Therefore, many principles and practices of classical textual criticism, such as the establishing of genetic relationships between manuscripts, stemmatic classification, the reconstructing of archetypes and the restoration of the original, are not applicable to Hebrew manuscripts, not only because many of these represent horizontal rather than vertical transmission and so provide us with open recensions, also because their texts may have been affected by the intervention of learned copyists. What is the implication of this iconoclastic presentation of medieval transmission? Should we abandon the medieval manuscripts because they offer inauthentic, unstable texts, and have been corrupted by the free critical editing of learned copyist and the whimsical copying errors of preoccupied poor scribes? Of course not. These are the only sources we have. But we must use them with great caution, suspicion and skepticism, and above all refrain from establishing authentic texts, or even critical editions, and rather resort to the safe synoptic presentation of the transmitted texts, while proposing our critical analysis and reconstruction in the form of notes.⁽¹⁰⁾

The Case of Pirqé R. El.

For reasons that will become clear, Pirqé Rabbi Eliezer (Pirqé R. El.) is an excellent object for testing some of the contrary assumptions and viewpoints in the debate over the status of manuscripts and how they should be studied and presented. Pirqé R. El. was and remains an exceedingly popular literary work in Jewish religious circles.⁽¹¹⁾ Manuscripts containing Pirqé R. El. are numerous and notable for their wide geographical and temporal distribution. What is the size of the corpus?⁽¹²⁾ Pirqé R. El. is preserved in eighteen manuscripts that are “complete” or nearly complete; these are dated from the fourteenth or fifteenth centuries through the nineteenth century.⁽¹³⁾ There are thirty-one partial manuscripts dating from the end of the thirteenth through the nineteenth centuries. The rest of the material is comprised of fragments.⁽¹⁴⁾

What is to be done with such a massive textual corpus? Even if we were to restrict “publication” of a corpus this massive to manuscripts and fragments and exclude commentaries, printed editions and translations,⁽¹⁵⁾ it would be fiscally and physically impossible to presented it synoptically in printed form, because of the limitations on the size of printed books. Further, printing places finality on a corpus that may increase in size with the future discovery of manuscripts or fragments. An average electronic file of Pirqé R. El. manuscripts contains approximately 50,000 words of text.⁽¹⁶⁾ I estimate that the entire textual corpus of Pirqé R. El. will be larger than one million words of text, and when encoded will be occupy over ten megabytes of diskspace.

In addition to the encoding of textual material, digitization of manuscripts has become an important part of the scholar’s work. David Gilner, Director of Libraries at Hebrew Union College-Jewish Institute of Religion, arranged with IBM to digitize three HUC Pirqé R. El. manuscripts.⁽¹⁷⁾ One of the three manuscripts contained in seventy-six folios is preserved in slightly more than twenty-two megabytes. I estimate that the entire project—text, digitization, and scholarly information—will take up somewhat more than five hundred megabytes.

In light of significant technological developments in the electronic processing and presentation of manuscript materials—textual and graphic—Beit-Arié’s concept of a “safe synoptic presentation” of manuscript material is outdated. It is possible to consider a different form of publication in which “synoptic presentation” is only one possibility. ⁽¹⁸⁾

Presently available texts.

What is the present state of affairs regarding the texts and manuscripts of Pirqé R. El. that are available to the scholarly community? How shall we evaluate what we have in relation to the issues raised above?

The following five items represent the most important current “editions” of Pirqé R. El., three generally available and two unavailable.

A. RaDaL (1798-1855)

Perhaps the best known edition of Pirqé R. El. was edited by Rabbi David Luria (RaDaL) and published in 1852 with his commentary.⁽¹⁹⁾ Because of the availability and popularity of this edition it might be referred to as the textus receptus. Although the text is based on the tradition of printed texts descended from the Venice edition, 1544, the text is filled with omissions. It is unknown why Luria did not include some passages. The source of these lacunae may have been the censor or the commentator himself, as I. H. Weiss noted.⁽²⁰⁾ Nevertheless, the Luria edition—text and commentary—continues to be quoted in scholarly articles to the present.⁽²¹⁾

B. Horowitz (ca. 1855-1905)

In 1971, Makor published what is described on the title page as a Mahadura Madait [a Scientific Edition] of Pirqé R. El. prepared by Chaim Meir Horowitz.⁽²²⁾ This is obviously not a scientific edition but a facsimile edition of Horowitz’ working text. Horowitz used the Venice edition of 1544 as his “copy-text.”⁽²³⁾ The Venice edition is based on the first edition of Constantinople, (1514), but Horowitz provides numerous interlinear corrections to what he considers “mistakes” in the first edition.⁽²⁴⁾ Friedlander and Haag have noted that almost all later printed editions are dependent on this text, although often corrupted by editors or censors.⁽²⁵⁾ Horowitz took the Venice edition, gave each line a number and “corrected” it according to the readings of fifteen manuscripts, parallel texts and other editions.⁽²⁶⁾

In addition to the facsimile of the Venice edition with Horowitz’ corrections, the Makor volume also contains three significant lists prepared by Horowitz.

• “The list of the books from which I copied variant readings and which I used or mentioned in my book and my introduction, plus the manuscripts.” Horowitz lists two hundred twenty books in Roman Numerals in his edition. (These are numbered as 1-153, and 183-250; numbers 154-182 are lacking.)
• “The list of midrashim, targums and aggadic books whose expositions are quoted in Pirqé R. El., or which copied from it (i.e. Pirqé R. El.) that I used in this (edition).”
• “Expositions that are quoted in books that are no longer found but that utilized Pirqé R. El.”

Unfortunately, the lists give only the title of the works, leaving out publication information for printed texts or adequate identification for manuscripts.⁽²⁷⁾

C. Higger (1892-1952)

Michael Higger published a manuscript of Horowitz’ in the periodical Horeb in the 1940’s.⁽²⁸⁾ He states that his version of Pirqé R. El. came from a manuscript that Horowitz prepared based on three manuscripts in the Casanatensa collection.⁽²⁹⁾ This Horowitz manuscript is listed in the catalogue of J. Rovner, A Guide to the Hebrew Manuscript Collection ...of the JTS (1991), #R1859, and dated 1879. It is written in modern cursive script on full sheets of paper plus scraps of notes. According to Higger, Horowitz copied one manuscript⁽³⁰⁾ from Chapter 3 to the end and placed variants from the two other manuscripts between the lines.

Higger published the base text along with what he considered the most important of the variants that are placed either on the bottom of the page or appended at the end of a section. In addition, Higger indicates that he omitted such distinctions as haser and male, or differences between final mem and nun. He also notes that he occasionally introduced into the base text readings from the other two manuscripts.⁽³¹⁾ Higger also used a fourth, unknown manuscript for Chapter 6 and the beginning of Chapter 7. Following Horowitz, Higger laid out Chapters 1-2 (3) synoptically, that is in parallel columns, reflecting the readings of all three manuscripts.

In recent years, the Higger edition has been cited in scholarly literature with increasing frequency as if it were the “best text.” It is, however, an edition at a third remove. That is, it represents Higger’s editing of Horowitz’s copying of the initial manuscripts. Further, there is no indication of the relationship between the three manuscripts, except that they are found in the Casanatensa Collection. Nevertheless, its importance will increase because it is the text available in electronic format on CD-ROM.⁽³²⁾

D. Gottlieb (1910-1983)

Dr. Zev Gottlieb devoted his later years to the editing of important medieval Hebrew texts and is best known for his edition of Sforno’s Commentary to the Torah published by Mosad HaRav Kook.⁽³³⁾ Dr. Gottlieb worked many years on the preparation of an edition of Pirqé R. El. In a letter to Professor Perez Fernandez of Grenada dated 6-1-83, Dr. Gottlieb wrote: “I have a text, as near perfect as possible, listing all variants, big or small, of all MSS and first editions, with brief notes and sources or parallels in Talmud and Midrash, etc.” He indicated that he still had to “prepare the final revision of the apparatus of Notes and write a Preface.” On the basis of a contract with the Mosad HaRav Kook, Dr. Gottlieb's edition of Pirqé R. El. was to have the same format as their edition of Midrash HaGadol. Unfortunately, he died before finishing this edition.⁽³⁴⁾

Dr. Gottlieb’s typescript edition of Pirqé R. El. uses as its base manuscript JTS Ms. Enelow 866. This judgement can be verified by comparing his edition with a printout of the electronic version of Pirqé R. El. of Academy of the Hebrew Language (AHL) in which this JTS manuscript serves as the primary text. Dr. Gottlieb’s text of Pirqé R. El. is well prepared, with each chapter divided into basic units or subsections. The text has punctuation marks, and biblical references are listed as line notes in a footer on each page. The use of editorial judgement is evident on each page and the typescript contains later revisions written between the lines or occasionally in the margins.

Although the variant readings are in longhand, they appear to be in near final form. The variants were obviously prepared for a Mosad HaRav Kook type edition. The designation of the manuscripts and fragments is based on lists prepared by Dr. Gottlieb using the catalogue of microfilms of the Machon as is evident from the microfilm numbers and the dating. It is not possible to determine which is the last list prepared by Dr. Gottlieb.⁽³⁵⁾

The explanatory notes are similar to those found in the Mosad HaRav Kook's editions of medieval biblical commentaries. They are traditional in nature and generally serve to provide a brief explanation of some point in the text based on material drawn from Midrashic and Talmudic sources. Because there are spaces on some of the pages of the notes, it is clear that Dr. Gottlieb had not fully completed this part of the project.

Unfortunately, Dr. Gottlieb did not produce the Preface as he had hoped. Consequently we do not have detailed descriptions of the manuscripts or fragments that he used or an explanation of the choices he made. The surviving lists, which contain both identifications of manuscripts and Sigla do not discriminate within categories of manuscripts or fragments. For example, the lists identify ten Yemenite manuscripts. These include two manuscripts that have been dated to the nineteenth century. Regrettably there is no indication whether these or other Yemenite manuscripts reflect an early tradition of Pirqé R. El. or contain material copied from printed editions.⁽³⁶⁾

E. Academy of the Hebrew Language=JTS Enelow 866

The last text of Pirqé R. El. to be noted here is the electronic edition of the Academy of the Hebrew Language.⁽³⁷⁾

The following are the manuscripts used by the Academy in their electronic edition:

• 00 New York, JTS Enelow 866 (1654) Yemen Complete
• 01 New York, JTS Lehman 300,1 (1596) Yemen Complete
• 02 Cincinnati, HUC 75, 14,1 (14/15 th or 16 th century) Mizrachit, Iraq Complete
• 03 Moscow, Ginzburg 111,1 (1468) Ashkenaz Partial
• 04 Warsaw, 240 (no date) Sefardi/Italian Partial

According to Dr. Mordecai Mishor of the Academy, the manuscripts were encoded by the Team of the Historical Dictionary of the Hebrew Language Project of the Israel Language Academy using the tagging system of the Academy. The tags include indications of many of the phenomena that should be indicated in transcribing a manuscript. It is important to note that the purpose of this electronic version is to provide information for the Historical Dictionary of the Hebrew Language and not to produce a reading text.

F. Evaluation of Texts.

In spite of the problems previously mentioned, Dr. Gottlieb’s text contains the only existing “critical edition” of Pirqé R. El. Every page reflects thoughtful decisions based on judgement and experience. However, Dr. Gottlieb’s “critical edition” does not satisfy present expectations of academic textual editing. There is no indication that the text resulted from stemmatic analysis, nor was a genealogy or pedigree of the manuscripts produced. Though incomplete and intended for a broad and perhaps traditional rather than academic audience, this work should be published, with or without the notes or should perhaps be made available on microfilm.⁽³⁸⁾

The AHL text represents a “diplomatic” edition of Pirqé R. El. because it is an edition based on one manuscript. It differs perhaps from a pure “diplomatic edition” as it contains errors in the base manuscript along with corrections from the four other manuscripts, each correction immediately following the problematic reading.⁽³⁹⁾ These corrections are specially designated and identified in the text itself, rather than in notes, as in a printed version. The corrections range from individual words or brief phrases to several lines. However, not every lacuna in the base manuscript is corrected.

In spite of these reservations, both the Gottlieb and AHL texts are better than Luria, Higger or Horowitz because they present Pirqé R. El. as it appears primarily in one manuscript.

Nevertheless, what we do not know about the textual history and transmission of Pirqé R. El. is still greater than what we know. The following represents a brief list of questions regarding Pirqé R. El.:

• What are the relationships between all the extant manuscripts and fragments of Pirqé R. El.? (At present there is no stemma or genealogical representation of the relationship of Pirqé R. El. manuscripts.)
• Do the manuscripts represent more or less the same tradition of the text or different revisions or perhaps “editions?”
• Specifically, as there are a large number of Yemenite mss, is there a different Yemenite tradition of Pirqé R. El. than that found in Iraqi, Ashkenazic and Italian-Sefardic manuscripts?
• Is JTS Enelow 866 in fact the codex optimus and if so, why?

Related to the topics of textual history and transmission is the question: how do the early secondary witnesses relate to the manuscript tradition of Pirqé R. El.? Generations of scholars beginning with Zunz have provided lists of quotations of Pirqé R. El. that date back to the Gaonic period and that also demonstrate deep connections between Pirqé R. El. and piyyut. ⁽⁴⁰⁾ It is generally agreed that the earliest citation of Pirqé R. El. is found in Pirkoi Ben Baboi.⁽⁴¹⁾ What do these sources reveal of the text as represented in citations and allusions?

Finally, there is the much tougher question: what is the textual status of Pirqé R. El.? Is it a relatively stable document represented in the manuscripts and fragments? Is it a reflection, however worked over, of the composition of an author? Or is it the end redaction or redactions of a textual tradition by unknown individuals each of whom has recreated Pirqé R. El. in his own image? These questions are not new, but they still have a connection to the work of textual editing and are a specific example of the issues raised earlier in this paper. These are the types of questions that are reflected in literary scholarship on Pirqé R. El.

Nature of Pirqé R. El.: What is it and how does that relate to issues of textual criticism?

Single Authorship, Compilation; Last Redaction; Authorship? Significance of terms and meaning for Pirqé R. El.

I previously offered the judgement that Dr. Gottlieb had produced a “critical edition” of Pirqé R. El. By this I refer to the general sense in which that expression is understood. It is, to borrow the description of G. Thomas Tanselle, “...a new text that incorporated the results of editorial judgement regarding variant readings and errors.”⁽⁴²⁾ Tanselle also writes “The job of a scholarly editor, therefore, can be stated as the exercise of critical thinking in an effort to determine the final intention of an author with respect to a particular text.”⁽⁴³⁾ In another place Tanselle refers to “the text intended by the author.”⁽⁴⁴⁾

Although the issue is debated in theories of textual editing, Tanselle’s repeated emphasis on the role of the editor in attempting to reconstruct the author’s intentions is important. Are the expressions “author” and “author’s intentions” useful categories in thinking about Pirqé R. El.? Pirqé R. El. has been variously described as a “midrash,” as a work which retells biblical narrative and as a pseudepigraphal rabbinic work. It clearly transcends any of these descriptions. How are we to label the person or persons who produced Pirqé R. El.? To what stage of the tradition do we refer when we talk about it? The following terms have been typically used to describe the creator of Pirqé R. El.: ba’al Pirqé R. El. and mechaber Pirqé R. El.. These appear, for example, in an important article by Yaakov Elbaum, “Rhetoric, Motif and Subject-Matter -- Toward an Analysis of Narrative Technique in Pirke de-Rabbi Eliezer.”⁽⁴⁵⁾ In terms of our discussion, Elbaum’s analysis both assumes and argues for the notion that that there was a single author of Pirqé R. El. who produced a unified work. In addition, the article assumes that the Luria and Higger texts, with occasional reference to “other manuscripts,” preserve and transmit this work more or less as the author intended.

Without judging whether or not this is correct, it is possible to turn the notion implied in the designation mechaber Pirqé R. El. into the following hypothesis:

A single author composed Pirqé R. El., although we do not know when or where he lived. An examination of the manuscripts will demonstrate that he produced a unified work, the Urtext of which, though changed by scribes in the course of transmission, can be reconstructed, at least hypothetically. In addition, through the use of either stemmatic or genealogical analysis, we can describe the relationships of the manuscripts and fragments, and demonstrate the transformations of the text from generation to generation.

On the other hand, Haag, for example, uses the term “final redaction,” to describe what we have of Pirqé R. El. This implies the existence of an “end redactor,” and is based on the view that it is impossible to know anything about an original Pirqé R. El. Haag presents a list of source material found in Pirqé R. El. and argues—as many others have done before and since—that it is impossible to draw conclusions about the author, time, or place of origin of the work. He writes:

Mit einiger Wahrscheinlichkeit lässt sich nur aussagen, das die Endredaktion nach den arabischen Eroberungen erfolgt sein muss, worauf Anspielungen auf Mohammed oder das Reich Yishma’els hindeuten.⁽⁴⁶⁾

After reviewing evidence for dating, he writes:

Immerhin ist man sich heute darin einig, dass die von den oben erwähnten Gelehrten berechnete Abfassungszeit (8.-9. Jahrhundert) als Zeit der Endredaktion unseres Midrasch angesehen werden kann.⁽⁴⁷⁾

And regarding time:

...dass die Endredaktion in der ‘Luft des Islam’ stattfand.⁽⁴⁸⁾

Again, if we were to turn this into a hypothesis to be tested it might be as follows:

Pirqé R. El. is a name applied to similar but not identical works that developed through the ages the traditions of which have a long history. The manuscripts and fragments represent these separate works. An examination of manuscripts and fragments will demonstrate that that there never was a unified work. What we have represents Pirqé R. El. type material as reformed and reshaped by scribes in the course of transmission. The texts that we have reflect final redactions of these works. Through the use of either stemmatic or genealogical analysis, we can describe the relationships of the manuscripts and fragments, and demonstrate the transformations of the text from generation to generation.

Note that with either hypothesis, there is the same methodological conclusion for textual analysis. Through the use of either stemmatic or genealogical analysis, we can describe the relationships of the manuscripts and fragments, and demonstrate the transformations of the text from generation to generation. There are now examples of this type of analysis in the work of Milikowsky on Seder Olam, Brody on The Sheiltot and Sussman on the Jerushalmi, Tactate Shekalim, and Kirschner on Baraita De-Melekhet HaMishkan—all very different types of texts.⁽⁴⁹⁾ The specific type of possible statistical analysis will vary for every composition, and the problem of “contamination” has to be factored in. Nevertheless, the goal and the general method are the same.

Is there a way to achieve these goals, using new techniques for textual encoding?

Goals of the Pirqe R. El. Electronic Text Editing Project

As suggested above, electronic text editing has transformed the way scholars now think about producing textual editions. Peter Robinson’s Chaucer: The Wife of Bath’s Prologue: on CD-ROM (Cambridge University Press, 1996), is the first world class example of the possibilities of such electronic textual presentation. Its advantages over the traditional “critical edition” and the more recent synoptic editions of medieval—and for our concern Hebrew—texts are clear. The Chaucer: The Wife of Bath’s Prologue: on CD-ROM contains:

…word by word collation of all [fifty-eight] witnesses, digitial images of every one of the 1200 pages (manuscript and early printed edition) we have transcribed, transcriptions of the glosses…, descriptions of each witness…, and spelling databases grouping every occurrence of every spelling of every word in every witness by lemma and grammatical category.⁽⁵⁰⁾

The hypertext presentation of the transcriptions permits the reader to see varient reading in context, rather than view the single word varients of the “classical editions.” Hypertext encoding permits even greater manuverability within the corpus of textual material. Robinson writes:

To click on any word in [the] base text brings up a window containing a complete record of all the readings in all the witnesses at that word. Further mouse clicks lead to the transcription of each witness, to images of the pages, to the transcription of the glosses, to and from the witness descriptions and the transcription introductions and into and out of the spelling databases for each witness and for all the witnesses together.⁽⁵¹⁾

Finally, through the use of cladistic analysis, Robinson and his team were able to evaluate the relationships between all the witnesses.⁽⁵²⁾

There is a significant utilization of computers in contemporary textual research on Hebrew manuscripts. However, no electronic Hebrew text project presently provides the resources to be found in the Chaucer CD-ROM.⁽⁵³⁾ Unfortunately, significant technological issues remain for the SGML/TEI encoding of Hebrew (as well as Arabic, Farsi, Urdu and Yiddish) texts, i.e., for texts written in “right-to-left” languages.⁽⁵⁴⁾ Nevertheless, we now have a model for works such as Pirqe R. El. “The Pirqe R. El. Text Editing Project ” is an attempt to provide a beginning example of this type of work.

ENDNOTES

⁽¹⁾ Peter Schäfer, "Research Into Rabbinic Literature: An Attempt to Define the Status Quaestionis," JJS 37:2 (1986): 139-52. His ideas were criticized by Chaim Milikowsky, "The Status Quaestionis of Research in Rabbinic Literature," JJS 39:2 (1988): 201-11. Schäfer responded in an article entitled "Once Again the Status Quaestionis of Research in Rabbinic Literature: An Answer to Chaim Milikowsky," JJS 40:1 (1989): 89-94. This interchange was followed from differing perspectives in a collection of conference papers published in Bulletin of the John Rylands University Library of Manchester, 75:3 (Autumn 1993), 33-52. [Theme Issue, Artefact and Text: the Re-Creation of Jewish Literature in Medieval Hebrew Manuscripts, ed. Philip S. Alexander and Alexander Samely.]. For our topic, the most important papers in that collection are Israel M. Ta-Shma, “The ‘Open’ Book in Medieval Hebrew Literature: the Problem of Authorized Editions,” 17-42; Malachi Beit-Arié, “Transmission of Texts by Scribes and Copyists: Unconscious and Critical Interferences,” 33-52; and Philip S. Alexander, “Textual Criticism and Rabbinic Literature: the Case of the Targum of the Song of Songs,” 159-174. I am grateful to Marc Bregman for bringing this volume to my attention.

⁽²⁾ Schäfer, "Research,” 150.

⁽³⁾ Milikowsky, "The Status Quaestionis,” 202.

⁽⁴⁾ Milikowsky, "The Status Quaestionis,” 204.

⁽⁵⁾ Milikowsky, "The Status Quaestionis,” 204.

⁽⁶⁾ Milikowsky, "The Status Quaestionis,” 207-208.

⁽⁷⁾ Milikowsky, "The Status Quaestionis,” 209-210.

⁽⁸⁾ Ta-Shma, “The ‘Open’ Book,” 17. Ta-Shma’s image of the perpetually revisable computer database is apt, as will be discussed below.

⁽⁹⁾ Malachi Beit-Arié, “Transmission,” 33.

⁽¹⁰⁾ Malachi Beit-Arié, “Transmission,” 50-51.

⁽¹¹⁾ This composition covers biblical material from Genesis through Esther and contains citations from nearly every book of the Hebrew bible. Yet the historical biblical narrative is often interrupted with chapters dealing with specific topics—charity, resurrection of the dead, acts of loving kindness, etc. The text also includes material on the calendar, passages drawn from hechalot literature, parallels to Islam, and allusions to characters or events found in the Pseudepigrapha or in Targum Pseudo-Jonathan, which are not preserved in main stream midrashic or talmudic texts. Already in the nineteenth century, Leopold Zunz pointed to hints of a liturgical midrash on the Eighteen Benedictions as part of the underlying structure of this text. See Leopold Zunz, HaDerashot B'Yisrael, trans. Chanoch Albeck. [Hebrew] (2nd edition, Jerusalem, 1947), 134-140, and notes, 417-424. Significant disagreement remains regarding the date and provenance of Pirqé R. El.; a reasonable consensus of scholarly opinion would place it in the eighth-to-ninth centuries, and probably in the Land of Israel. See also my comments, Lewis M. Barth, “The Ban and the “Golden Plate”: Interpretation in Pirqe d’Rabbi Eliezer 38,” 625-627, notes 1-4, in The Quest for Context and Meaning: Studies in Biblical Intertextuality in Honor of James A. Sanders, eds. Craig A Evans and Shemaryahu Talmon (Leiden, 1997).

⁽¹²⁾ I have prepared a database of all Pirqé R. El. manuscripts and fragments that are in the public domain. It was based initially on the card catalogue of microfilms of Pirqé R. El. manuscripts in the Machon l’tatzlumey kitvey yad, Hebrew National Library, Jerusalem and has been augmented or modified after I examined several manuscripts or fragments found in the Klau Library, Hebrew Union College, Cincinnati, the Library of the Jewish Theological Seminary, New York and the Cambridge University Library. In the process of examining catalogues and original documents, I was struck by the number of manuscripts of Pirqé R. El. and by their wide geographical and temporal distribution; therefore these details of the lists have also been entered into a database. The information—whether checked or as yet unverified—can be easily sorted for further evaluation. Each manuscript type has been designated as C=Complete, P=Partial, E=Printed Edition, F=Fragment, G=Genizah Fragment or L=Likkutim; these are mixed categories indicating the relative extent of material contained in the manuscript, or provenance or type of collection the manuscript contains. The quantitative difference between Partial (more than a few chapters) and Fragmentary (fragments) is intentionally loose and there may be some shifting of specific items. In some instances the extent of Pirqé R. El. material is not clear.

The chart of the database found at the end of this article utilizes “manuscript type” and “dating” as sorting criteria. Several of the items in the list have been directly examined; others need to be evaluated to determine whether the brief descriptions are accurate, the dating valid and the extent of Pirqé R. El. material correct. I have no doubt that additional Pirqé R. El. manuscripts or fragmentary material will surface and that all descriptions, dating, etc., may be subject to modification. The criteria for dating must be determined when there is no colophon or other reliable indication of time or provenance.

⁽¹³⁾ The various manuscripts contain Pirqé R. El. in fifty-two to fifty-four chapters. Depending on the manuscript, the text contains two hundred seventy-five verses from the Hebrew bible. These verses are cited as a whole or in part over one thousand four hundred and seventy-five times.

⁽¹⁴⁾ I have based the brief database printout on a survey of entries in the catalogue of the Machon, plus printed catalogues, citations in books or articles and referrals by colleagues. In particular I want to thank my colleague Marc Bregman and also Stefan Reif, Director of the Tayor-Schechter Genizah Research Unit at Cambridge, for many of the references to Genizah fragments. [One matter needs to be noted regarding microfilm identification. The Machon Collection contains a number of microfilms of JTS manuscripts. J. Rovner has also published a five-volume work, A Guide to the Hebrew Manuscript Collection ...of the JTS (1991), one section of which deals with microfilms of rabbinic manuscripts. It is difficult to determine the relationship between the description and numbering in the Rovner-JTS Catalogue and the description and numbering of the Machon JTS material. There appear to be some JTS manuscripts, microfilms of which are not available at the Machon and JTS may have more Pirqé R. El. material than the Rovner volumes indicate.]

⁽¹⁵⁾ There is an excellent bibliograph of printed editions of Pirqé R. El. in Hans Jakob Haag, Magisterarbeit, PIRQE DERABBI ELI’EZER KAP 43: Aufbau und traditionsgeschichtliche Analyse University of Köln, 1978). I am grateful to Peter Schäfer for mentioning this study to me and to Dr. Haag for sharing it with me.

Haag’s work is a traditions-geschichtliche analysis of Pirqé R. El., Ch. 43. In addition, it contains extensive bibliographical information on manuscripts and printed editions. Of the latter, Haag lists forty-three entries beginning with the Constantinople edition of 1514 through a Jerusalem edition of 1973. If printing run numbers were available, we would have further evidence of what we know from citation of the wide popularity of this collection, beginning in the ninth century and continuing to the present.

Haag also lists eight translations beginning with that of Konrad Pellikan in Latin (Gagliardi, 1604). As is well known, the most important modern translation is Pirkê De Rabbi Eliezer, trans. Gerald Friedlander (London, 1916); [reprint, New York, 1965], based on a manuscript that belonged to Abraham Epstein of Vienna. The translation takes on additional significance since it is unclear if this manuscript has survived WW II. Since Haag completed his work there have been two new translations, one in French and one in Spanish. The French translation: Pirqé De Rabbi 'Eliezer: Leçons De Rabbi Eliezer, trans. Marc-Alain Ouaknin, Eric Smilevitch and Pierre-Henri Salfati (Paris, 1984). It is based on the Eshkol edition (Jerusalem, 1973), the Luria edition, the Friedlander translation and a manuscript belonging to the Gaster collection. The Spanish translation: Los Capítulos De Rabbí Eliezer, trans. M. Pérez Fernàndez (Valencia, 1984). On this see: Gunter Stemberger in Hermann L. Strack, revised by Gunter Stemberger, Introduction to the Talmud and Midrash, trans. Markus Bockmuehl (Scotland, 1991). Stemberger writes: "Spanish translation based on the Luria edition, using the Venice 1545 edition and the three MSS published by Higger (358).”

Finally, Haag lists twenty commentaries to Pirqé R. El., of which the Luria commentary is generally acknowledged to be the most important.

⁽¹⁶⁾ A lightly marked up SGML/TEI (Standard Generalized Markup Language/Text Encoding Initiative) file is approximately half a megabyte. For a complete bibliography on SGML and TEI, see Robin Cover, “The SGML/XML Web Page”.

⁽¹⁷⁾ HUC mss. 75 (database: 004), 2043 (database: 005) and an HUC Klau Library first edition, (Constantinople, 1544), (database: 001 [the database reference is to the first edition in the Bodleian Library) . These digitizations are now available for viewing on the Pirqé R. El. Electronic Text Project web site .

⁽¹⁸⁾ Beit-Arié’s view is reiterated in the “Introduction” to the John Rylands collection. “This new view of the textual history of these works and of their comparative lack of literary definition has led naturally to the decision that the proper way in which to present them in scholarly editions is synoptically (Philip S. Alexander and Alexander Samely, “Introduction, 5.)” For the most recent synoptic edition of a midrashic text, see Rivka Kern Ulmer, Pesiqta Rabbati: A Synoptic Edition of Pesiqta Rabbati Based upon All extant Manuscripts and the Editio Princeps, Vol. I (Atlanta, 1997).

However, electronic textual publication now allows a far more useful presentation of material than is found in the best printed synoptic editions. See below the comments on Chaucer: The Wife of Bath’s Prologue: on CD-ROM, ed. Peter Robinson (Cambridge University Press, 1996).

⁽¹⁹⁾ Pirqé R. El., ed. Rabbi David Luria, (Warsaw, 1852). It has been republished or photo-offset many times since. I’m grateful to Richard White for sharing with me his own electronic version of the Luria text.

⁽²⁰⁾ Weiss, Dor Dor V’Dorshav, (New York and Berlin, 1924) 3:293, n. 24. See also Haag, Magisterarbeit, 96.

⁽²¹⁾ The value of Luria’s commentary is universally acknowledged. As with the text itself, Luria’s commentary contains omissions; see J.S. Speigel, “Additions of the RaDaL to his Commentary on Pirqe R. El.[Hebrew]” Sinai, 77 (1975), 146-156.

⁽²²⁾ Pirke De Rabbi Eliezer, codex C. M. Horowitz (Jerusalem, 1972).

⁽²³⁾ Tanselle defines “copy text” as “that text judged to have presumptive authority, the one to be followed at points where no emendations are made;” see G. Thomas Tanselle, Textual Criticism and Scholarly Editing (Charlottesville and London, 1990), 19.”

⁽²⁴⁾ For a remarkable discussion of the relationship of manuscripts to early editions, see Yaakov Sussman, The Tradition of Study and the Tradition of Version of the Jerusalem Talmud… [Hebrew],” Mehkarim b’Sifrut haTalmudit (Jerusalem, 5743), 12-76.

⁽²⁵⁾ H. J. Haag, Magisterarbeit, 27; see M. Steinschneider, Catalogus librorum hebraeorum in Bibliotheca Bodleiana (Berlin, 1852-600), Nr. 4009.

⁽²⁶⁾ Regarding the presentation of the Venice edition in the Makor volume, Gunter Stemberger called attention to the fact that pp. 183 ff. present the pages of Horowitz’ text, and thus the Venice text, out of order. See Strack-Stemberger, Introduction, 357.

⁽²⁷⁾ These lists supplement the information in Ch. Horowitz “Mishnat R. Eliezer,” HaMagid, (1879). It appears that additional lists prepared by Horowitz may be missing.

⁽²⁸⁾ "Pirqé Rabbi Eliezer", ed. M. Higger, Horev, 7 (1943): 82-119; 9 (1944), 94-116; 10 (1947), 185-294.

⁽²⁹⁾ See database manuscript numbers 057, 058, and 059.

⁽³⁰⁾ Sacredote Catalogue 173,1=I.IV.10; database number 057.

⁽³¹⁾ These are noted in the variants.

⁽³²⁾ It is found in both the ATM and SHOOT, and possibly other CDROM diskettes of Rabbinic literature.

⁽³³⁾ Bei’ur ‘al HaTorah l’Rabbi Ovadiah Seforno, ed. Zev Gottlieb (Mosad HaRav Kook: Jerusalem, 5744=1984).

⁽³⁴⁾I am grateful to Dr. Aviva Zornberg and her family for permitting me to examine the material left by Dr. Gottlieb. The literary legacy includes the text, variant readings and notes for his edition of Pirqé R. El. I compared the material with notebooks, lists and notecards. It is clear that the material is in the state Dr. Gottlieb described in his letter to Fernandez.

⁽³⁵⁾ I have reconstructed a brief list of the Sigla which Dr. Gottlieb used. In only a very few case did I find unresolvable differences between the Sigla used in the various lists.

⁽³⁶⁾ My database list contains fifteen Yemenite manuscripts of which fourteen are dated from fourteenth through the nineteenth centuries and of these five are dated in the nineteenth century. One of the two Cincinnati manuscripts, HUC 77 (Database 004), contains several chapters copied from the Constantinople edition (Chs. 42-end, ff. 65a ff.

⁽³⁷⁾ I am grateful to the Academy for providing me with a copy of this text and concordance for research purposes. I want to express thanks especially to Dr. Mordecai Mishor for his assistance and the information he has provided.

⁽³⁸⁾ It is my hope that a Xeroxed copy of this work be made available to scholars through microfilm deposited at the National and University Library. To my knowledge, this has not yet occurred been done.

⁽³⁹⁾ The manuscripts used to correct the base text are from widely differing communities and all manuscripts are of relatively late date.

⁽⁴⁰⁾ In relation to this question, Drs. Israel TaShema and Jacob Sussman both informed me of the existence of 200 manuscripts of commentaries on piyyutim likely to contain early citations of Pirqé R. El., which are to be dated earlier than any of the known manuscript material. What is the image of Pirqé R. El.in these commentaries and would they be reliable witnesses to its text?

⁽⁴¹⁾ See Geniza Studies in Memory of Doctory Solomon Schechter, ed. Louis Ginzberg (New York, 1929), 2:544 ff. Lists of additional citations are found in Luria, Pirqé R. El., “Introduction,” paragraph 7, Horowitz, HaMagid (1879), 118 and 126, and Zunz-Albeck, HaDerashot b’Yisrael, 417, note 12. See also, Ch. Albeck, Midrash Bereshit Rabbati (Jerusalem, 1940), 18, note 2 and Zunz-Albeck, HaDerashot b’Yisrael, Ch. 6, note 61.

⁽⁴²⁾ Tanselle, Textual Criticism, 276.

⁽⁴³⁾ Tanselle, Textual Criticism, 29.

⁽⁴⁴⁾ Tanselle, Textual Criticism, 42.

⁽⁴⁵⁾ In Jerusalem Studies in Jewish Folklore, 13-14 (1991-1992), 99; 101, n. 8.

⁽⁴⁶⁾ Haag, Magisterarbeit, 7.

⁽⁴⁷⁾ Haag, Magisterarbeit, 130.

⁽⁴⁸⁾ Haag, Magisterarbeit, 15, quoting Bernhard Heller, "Muhammedanisches und Antimuhammedanisches in den Pirke de Rabbi Eliezer," MGWJ, 69 (1925), 54

⁽⁴⁹⁾ See: Chaim Milikowsky, "Seder Olam: A Rabbinic Chronography" (doctoral dissertation, Yale University, 1981; Robert Brody, The Textual History of the She'iltot (Jerusalem, 1991); Sussman, “The Tradition of Study,” and Robert Kirschner, Baraita de-Melekhet ha-Mishkan. Cincinnati (Cincinnati, 1992). These works treat very different types of texts.

⁽⁵⁰⁾ Robinson, Chaucer CD-ROM, 13.

⁽⁵¹⁾ Robinson, Chaucer CD-ROM, 14.

⁽⁵²⁾ See Peter Robinson, “Report on the Textual Criticism Challenge 1991.” Electronic publication.

⁽⁵³⁾ See Peter Robinson, The Transcription of Primary Textual Sources Using SGML (Oxford, 1994), and especially Robinson’s comments on the “Sol and Evelyn Henkind Talmud Text Databank, 19-20.

⁽⁵⁴⁾ See Lewis M. Barth, “CETH Report,” Los Angeles, 1995 and “Electronic Edition of the Midrash Pirqe Rabbi Eliezer: Creating an Encoding Manual,” ALLC-AHC ’96 ABSTRACTS, Bergen, Norway, 1966. PDF version. [The 1966 Joint International Conference: the Association for Literary and Linguistic Computing and the Association for Computers and the Humanities ]

*This chapter appears in Agenda for the Study of Midrash in the 21st Century, ed. Marc Lee Raphael (Williamsburg, Virginia: 1999), 43-62 [Library of Congress Catalog Card Number: 98-74497; © 1999] and is published in electronic form with permission of the editor. Parts of this chapter were delivered in a much earlier form at the World Congress of Jewish Studies, Jerusalem, 1994, and in the present form at a midrash consultation convened by Marc Lee Raphael at the Department of Religion, the College of William and Mary, 1998. I am especially grateful to Marc Bregman for suggestions and criticisms.

For information on how to purchase Agenda for the Study of Midrash in the 21st Century, ed. Marc Lee Raphael (Williamsburg, Virginia: 1999), please use the ORDER FORM or contact Marc Lee Raphael at mlraph@facstaff.wm.edu.