A critical Edition of the Armenian Bible: a progress Report

“A Critical Edition of the Armenian Bible: A Progress Report,” an invited paper given at the international conference “Where the Only-Begotten Descended: The Church of Armenia Through the Ages,” convened by Kevork Bardakjian at the University of Michigan, Ann Arbor, Apr. 1–4, 2004.

A Critical Edition of the Armenian Bible: A Progress Report

Almost immediately scribes began, consciously and unconsciously, to revise the Greek translation toward the Hebrew. This was particularly tempting so long as scribes were bilingual: it was easy to adjust the Greek toward the Hebrew, to make it more “faithful” to the Hebrew. Other translations came into being: for example, that associated with the name of Theodotion was often “closer” to the Hebrew and became popular, to the extent that his translation of Daniel almost entirely displaced the Septuagint translation. Aquila and Symmachus are names attached to two other translations.

The discovery of the Dead Sea Scrolls and the media attention they have justifiably attracted continue to educate even laypeople about the transmission of the biblical text. Many of us have heard about the faithfulness with which the now dominant form of the text was copied and recopied, i.e., the Masoretic text, but also about the existence of different forms of the biblical text in the period when the Septuagint was translated. For example, we now know that the shorter form of the book of Jeremiah, known previously from the Old Greek translation, is based upon a Hebrew original that was shorter—the book circulated in two forms!

In the Great Divorce, when the church and synogogue separated, the church took the Septuagint, from which it had been preaching since its earliest days. Consequently the synagogue placed an even greater emphasis upon the Hebrew. So the stage was set for the arguments of the second century about whose text more accurately preserved the Word of God. From our perspective, we would immediately judge “the original” to have priority, but it was not all that clear at the time because the Septuagint was also regarded as “inspired”. Of course, neither party had “the original”; both sides were the recipients of texts that had been copied and recopied, with the developments that are inevitable in such transmission over time.

The third century saw Origen (d. 251) prepare the Hexapla which, among other things, permitted a minute comparison of the Hebrew and Septuagint texts of his day. It was a huge text critical endeavour, even if Origen did not intend it to be so. The name of Lucian of Antioch, who lived early in the 4^th century, is connected with a major revision of the Greek Bible, at least the Old Testament, along stylistic lines. And then at the end of the 4^th century Jerome gave the western church the Latin Vulgate, in which he translated from Hebrew into Latin, bypassing the Greek. These revisions and new translations were about obtaining a more faithful, “better” translation. One should mention as well that Jerome’s Hebrew-based translation was not entirely successful in displacing familiar, Septuagint-based translations into Latin.

I put the word “better” in quotation marks because what one means by “better” depends upon the criteria that one brings to the question. On the one hand, a translation might be judged to be better if it is based upon “good” manuscripts, perhaps those which have been known to be in the possession of an old community for many years; on the other hand, a translation might be judged to be “better” if it is more easily understandable, or is done with greater attention to style and intelligibility. Such an interest in a “better” translation has been a part of the Armenian biblical tradition from the very beginning.

The story of the Bible in Syriac translation is less complicated than that of the Greek, at least for the Old Testament; the Syriac translation of most books may be as early as the first and second century. The translation is from Hebrew, but the translators may have made use of the Septuagint sometimes. The name given to this translation is the Peshitta. In the New Testament, as we might expect, the situation is more complex: Tatian’s Diatessaron, or harmony of the Gospels, a 2^nd century effort, may have been compiled in Syriac; the Peshitta New Testament represents a revision of an older Syriac translation. Indeed, this revision was being undertaken, on the basis of a Byzantine Greek text, in the period of the Armenian translation of the Bible, i.e., the early decades of the 5^th century.ⁱ We see the same look westward for a “better” text that we see in the Armenian tradition.
Koriwn

Koriwn’s Life of Mashtots is our major historical source for the translation of the Bible into Armenian.ⁱⁱ In it we see an interest in translators obtaining the training necessary to accomplish good work, but we also see an interest in basing translation work upon a “better” form of text. Where did that interest come from? Well, we can see that already in the 3^rd century there was an awareness in various quarters that there were different forms of text and that some were to be preferred over others. Koriwn informs us that, following the Council of Ephesis (431), seminarians working under Mashtots‘’ tutelage returned to Armenia from Constantinople with “reliable” copies of the Scriptures. The form of text which those copies represented had the imprimatur of the church in Constantinople; it was a sort of official form of text that they brought home. The adjustment of existing Armenian translation work toward the text of those manuscripts resulted in a “better” translation in Armenian. Or so it was thought. Where we can compare the earlier and later translation work we see that the later work also bears a distinctly different style, so that “better” included more than simply a different textual basis.

Zohrapian

How did Zohrapian choose the manuscript that he printed? Zohrapian describes in detail the manuscripts that he used for his edition. The first manuscript described is the one whose text he chose to print, namely, Venice 1508. This manuscript, he says, was written at the beginning of the 14^th century, and is therefore old; it is complete and carefully copied; corrections appear in the margins; it has colophons at six places, the first of which connects the manuscript with Georg, who may be the scholar Georg Skewrats‘i; the name of the scribe is Hovhann. Unlike his description of several other manuscripts, he makes no mention of any apparent comparison with the Latin Vulgate, which was likely a factor in his choice of Venice 1508.^iv Finally, we know that Zohrapian chose this manuscript as a base text from all the manuscripts that were available to him in Venice at the time; further, he made use of all the others.

Several of Zohrapian’s considerations continue to be important in the choice of manuscripts for editions of texts: the age of a witness is important, though not always, as is completeness; the care with which a manuscript has been copied, information about provenance, and freedom from contamination from other textual traditions are also to be born in mind. However, as important as these characteristics may be, modern text critics try to determine on the basis of minute comparison of manuscripts—and against, in the case of sub-versions like the Armenian, the parent text—which witnesses preserve the purest form of text among the textual families that emerge. This scientific approach was only developing at the end of the 18^th century and is associated with the name of the New Testament scholar, J. J. Griesbach. “Critical” editions are based on this kind of an approach.

Diplomatic versus “critical” editions of texts

A “critical” edition of a text is a more daring and, at the same time, a more useful type of edition because it aspires to present the original text as fully as possible. The witnesses chosen for use in the edition are arrived at on the basis of a critical examination that determines textual groupings and which of these groupings preserves the purest form of text, i.e., the form of text freest from the corruptions that inevitably occur in the process of copying and recopying. The text, whose starting point may be a single manuscript, is eclectic, i.e., it represents the product of innumerable instances where secondary readings are relegated to the apparatus. The printed text does not exist in any manuscript; it is, to some extent, hypothetical, as close to the original as we believe we can come. That critically-established text can then be used for other scholarly endeavours, such as comparison with its supposed parent text or texts. The Nestle-Aland and United Bible Society editons of the Greek New Testament are examples of critical edtions.

Ter-Movsessian

Stone and Ajamian

Shahé Ajamian published a brief modus operandi that included the preparation of complete lists of manuscripts and the deposit of microfilms in Yerevan and Jerusalem. He committed himself to Isaiah; H. Amalyan would do Maccabees and A. Zeitunyan the book of Genesis. The method in each case was to be identical. The collation of sample passages from the manuscripts against Zohrapian’s text would reveal the textual families for the particular books. In the case of Isaiah, in 1975 Ajamian stated that he had collated 65 manuscripts, from which 12 would be selected for the critical edition. He proposed using Jerusalem 1930, dating from 1323, or a Yerevan manuscript dating from 1384 (M4113). His description of the project is optimistic; the methodology that would base an edition upon a dozen manuscripts chosen from those extant seemed to offer achieveable results.^ix

In connection with the Isaiah edition, Michael Stone published a milestone article in which he set forth the process by which manuscripts J1930 or M4113 had been selected from among some seventy witnesses. Because of the state of knowledge of the version at the time, he counselled in favour of a diplomatic edition. In his study the choice of witnesses is arrived at by a process of exclusion: first, 1400 is used as a “cut-off date,” so the number of witnesses is reduced by half. The elimination of manuscripts attesting various kinds of corruptions brings the number down to 13, subsequently reduced to two upon which an edition might be based. This article was important because, for the first time the selection of manuscripts was placed upon a scientific basis. It was, however, a “pilot study,” as Stone says.^x

The first critical edition of a biblical book in Armenian was Michael Stone’s edition of IV Ezra; the apocalyptic book had appeared in an appendix in Zohrapian’s Bible. Stone’s edition, published in 1979 with an English translation, betrays considerable advancement beyond the pilot study of Isaiah. Here all available witnesses but one, a total of twenty-one, are collated, regardless of date, divided into families, and presented in terms of stemmatic relationships.^xi The text of the critical edition stands closest to manuscript Yerevan 1500, copied before 1282 by Mkhit‘ar of Ayrivank‘. Under the same cover, Stone printed as well a diplomatic edition of IV Ezra. The translation itself he believes goes back to the 5^th century.

Stone’s influence reaches into several other editions of biblical books. Three of them were doctoral dissertations; two of them present diplomatic editions. First, Lipscomb prepared a critical edition of four works in the apocryphal Armenian Adam Cycle. There are only six manuscripts extant, of which he employed four.^xii The second and third both involve canonical books. The second was my own, in which I presented a diplomatic edition of the book of Deuteronomy, using as base manuscript Venice 1007, which dates from 1332; Jerusalem 1925, dated 1269, and the British and Foreign Bible Society (London) manuscript, a 17^th century witness, were also collated from the same purer text group. An analysis over against the critical Greek text of Deuteronomy revealed that there is no textual foundation in the Peshitta and that its closest textual relations are with a Byzantine type of Greek text quite influenced by Origen’s Hexapla.^xiii

Cowe’s diplomatic edition of Daniel was a 1983 doctoral thesis, prepared for the Hebrew University in Jerusalem.^xiv It is an exhaustive piece of work and pursues the possible residual element of the pre-Ephesus “Arm 1” along the lines of Lyonnet’s search for Arm 1 in the Gospels;^xv that is, aside from the manuscripts themselves, Cowe makes use of patristic sources, the lectionary, and especially the Georgian version—the Armenian is the earliest parent text of the Georgian. As for the manuscripts, he collated sample passages for some 120 and, from the five text groups that emerged, selected 15 for use in the edition, which prints the text of Yerevan 287, dated 1258. Once again Jerusalem 1925 is recognized as preserving an early text type and is fully collated as one of the other fourteen manuscripts. Cowe’s textual analysis concludes that the translator of Armenian Daniel worked from a Lucianic Greek text and the Peshitta simultaneously to produce what we call Arm 1. The post-Ephesus revision of this text, “Arm 2,” was done on the basis of another type of Greek text, not hexaplaric and, among the great Greek uncials, most closely related to Vaticanus. Finally, among many other interesting insights, Cowe, like Stone in his volume on IV Ezra, makes some tentative suggestions about the local texts which the groups of manuscripts represent.^xvi It is a rich volume.

Cowe followed up his work on Daniel with work on the book of Ruth, whose textual groupings bear a similarity to those in Deuteronomy. Here too the parent text is a Greek text of a particular type, with a hexaplaric component.^xvii

Zeytunian’s Pentateuch and Amalyan’s Maccabees

The initiative of 1969 bore its first recognized fruit with the publication of Andranik Zeytunian’s critical edition of Genesis in 1985.^xviii His presentation of the text is preceded by a major preface, written by Levon Ter Petrosyan, which surveys the Armenian translation of the Bible and whose most important contribution may be the exposition of the significant role that the Bible has had in the history of the Armenian people. The Genesis volume was followed by editions of the other books of the Pentateuch; the last volume, Deuteronomy, appeared in 2002.^xix Each successive volume is more attractive than its predecessor, until we reach Deuteronomy, which is a very handsome book, in a larger format, on fine paper, bound in hardcover, with a dust-jacket. The “form” of the editions of Exodus–Deuteronomy was set with Genesis: “the die was cast,” so to speak, and it seems that there was no possibility of change. At least no changes were introduced.

These volumes took their form in the Soviet period, when Yerevan was more isolated, not only politically and ideologically, but also academically, from major centers of text-critical enterprise. Exodus–Numbers appeared in the 1990s amidst all the economic and social disruption that followed the collapse of the Soviet Union, and the emergence of the Republic of Armenia. I mention this fact, because the mere appearance of these editions is a tribute to the tenacity and stamina of its editor, Zeytunian.

If isolated, scholars in Yerevan who work at the Matenadaran have an immediate advantage over those of us who work outside Armenia, namely, easy access to that facility’s vast repository of manuscripts. One is envious. And Zeytunian has made good use of the resources available to him. His edition of the Pentateuch enjoys the inclusion of fragments of manuscripts, lectionary manuscripts, and citations to be found in commentaries. This provides his work with a rich sweep of materials. We are in his debt.

All that said, these editions rest on a very major shortcoming, which is found on page 84 of the edition of Genesis or, perhaps one might say, not found there. On this page, in two sentences, one finds the statement of principles for the selection of manuscripts and resources. Whereas Ajamian had outlined a process for the selection of manuscripts for the critical editions of biblical books, that whole process stands out here by its absence. We learn that there are 106 manuscripts containing Genesis and that, for the edition, Zeytunian fully collated 43 manuscripts, dated up to the 16^th century. That is, more than half of the manuscripts were left out of the picture from the start. It is true that later manuscripts may generally be regarded as preserving a more developed form of text, but that is not always so and sometimes late manuscripts preserve an early, excellent form of text. Such is true of the 17^th century manuscript belonging to the British and Foreign Bible Society in London, at least in Deuteronomy and Job.

In dealing with a large number of manuscripts it may be legitimate to employ a cut-off date, for example, in the case of preparing an edition of a Gospel; but even there, by the use of a profiling method, one can quickly ascertain the place of a manuscript in the textual tradition. Stone used the cut-off date of 1400 for his pilot project on Armenian Isaiah; but that was only a pilot project. Further, after the cut-off date is determined, it is essential to subject the manuscripts under scrutiny to the question of stemmatic relationships, so that one can determine which manuscripts among the large number preserve a purer form of text. That information becomes useful, perhaps essential, in preparing a critical edition because one gains the insight that certain combinations of witnesses are more likely to retain the original text. A partial examination of the witnesses, of course, is likely to produce a partial stemma of the textual tradition.

This failure to reduce the witnesses to groups or families means that the critical apparatus for the editions is burdened with repeated lists of single manuscripts that belong together and could have been represented by a siglum, relieving the user of a lot of work sorting things out. And, on the issue of sigla, some of us used Zeytunian’s central designation list of Bible manuscripts, wherein an Arabic number designated a manuscript: e.g., 13 represented Jerusalem 1925. Zeytunian himself, however, employs another, idiosyncratic system of identification that involves a capital letter and a subscript number. This issue of confusing methods of manuscript identification was resolved, we thought, at a meeting of the Association Internationale des Études Arméniennes in 1989, where it was agreed to identify manuscripts by their location and catalogue number—so Jerusalem 1925 is J1925—and it is to be hoped that future editions out of Yerevan, as well as elsewhere, will reflect this transparent nomenclature.

Page 84, cited above, also fails to mention what base text stands behind the critical edition of the books of the Pentateuch. What was the starting point? Was it Zohrapian’s edition? Or is the text completely eclectic? One does not know.

Having made this one, major criticism, we can say that the text that emerges in Zeytunian’s editions is purer than that to be found in Zohrapian, even if the latter, in his apparatus, was able to anticipate a superior text in a majority of divergences in the textual tradition, at least in Genesis. There is still much work to do on the Armenian Pentateuch: 1. the stemmatic relationships of the witnesses need to be worked out, for all the extant manuscripts;^xx 2. the entire text needs to be printed in one volume with an introduction and notes in a European language, by a publisher whose books are readily available. When these things are done, biblical scholars will have a text that might be considered the proper representative of the Armenian version for the Pentateuch. Then questions about a possible residuum of Arm 1 can be posed—since that has not been done, and we will have a true replacement for Zohrapian.
Maccabees

The only other edition to fall into the Yerevan critical texts series is that for Maccabees, edited by Hayk Amalyan.^xxi The critical edition is based on some 40 manuscripts, selected from seventy-six on the basis of the methodology in Stone’s pilot study for Isaiah. In fact, 32 manuscripts were selected, dating up to the 17^th century; eight more of the “best” witnesses were chosen from the remaining 44, to bring the total to forty.^xxii The manuscripts fall into four text groups, designated I–IV, though these sigla are not carried into the apparatus of the edition where we meet the same long strings of letters and numbers that we find in Zeytunian, though the numbers are no longer subscript.

The edition begins with Zohrapian’s base manuscript, Venice 1508, which belongs to group III. The explanation of the relationship of the text groups to one another is not clear either here or in an article published in English a year after the edition appeared.^xxiii The type of Greek parent text that lies behind the Armenian is not known, Amalyan says,^xxiv though the edition of the Greek text he employed is the old provisional Greek text of Rahlfs and not the Göttingen critical edition. So perhaps this issue can be clarified by a comparison with the Göttingen text.

The edition of the books of Maccabees stretches toward seven hundred pages. Its publication early in the quest for a critical edition of the Armenian Bible reflects the importance of these books in Armenian historiography and interest.

The New Testament

The challenges of producing critical editions of the books of the New Testament are even more daunting than those facing the textual criticism of the Old Testament. There are hundreds of manuscripts, not just one hundred or so, and, for the Gospels especially, one faces the difficulty of separating Arm 1 and Arm 2. Most likely an editor will decide to prepare an edition of Arm 2, placing residual Arm 1 readings found in other sources in a separate apparatus. This problem awaits a brilliant mind.

Conclusion

Little apparent progress does not mean there has not been progress. In particular, because of the work, and sometimes the interaction, of people like Stone, Burchard, Zeytunian, Cowe, Ajamian, Alexanian, van Lint, Amalyan, Leloir, Ter Petrosyan, Mathews, Terian, Thomson, Weitenberg, Johnson, van Esbroeck, Sanjian, Outtier, Winkler, and others, we are now much better informed about the matrix in which the Armenian Bible came into being, about the historical sources that describe it, and about the manuscripts in which the biblical text is preserved. We are learning to avoid generalizations, I think, and have come to see that the Armenian Bible is really a collection of texts that have diverging transmission histories.

Over the last thirty years there have been developments in ancillary disciplines that will contribute to the realization of a critical text for the Armenian Bible. For example, the Leiden edition of the Peshitta Old Testament has been completed; critical editions of the books of the Old Testament in Greek continue to appear from Göttingen—e.g., that for Job only in 1982. Profiling methods have been developed that will reduce the labour in determining textual relationships among large numbers of manuscripts, therefore enabling an editor to select the witnesses for an edition more expeditiously.^xxviii The Oxford Collate Program is a powerful tool for the preparation of editions involving a larger number of witnesses and permits the manipulation of a vast quantity of information.^xxix Finally, several symposiums and conferences devoted to the Armenian Bible and the preparation of editions of texts have led us toward a reasonable approach to the preparing of editions of biblical books.

The energy expended in recent years, with such meagre results in terms of critical text, gives one all the more admiration for Zohrapian who, with almost none of the technology and little of the resources available today, saw to completion a still useful edition of the text of the whole Bible.^xxx

Bibliography

_____. “Commentaire sur l’édition critique du livre de la Genèse en Arménien,” in Armenian Texts Tasks and Tools, ed. Henning Lehman and J.J.S. Weitenberg. Acta Jutlandica LXIX:1; Humanities Series 68. Aarhus: Aarhus University Press, 1993. Pp. 57-64.

Michael E. Stone and Claude E. Cox. “Guidelines for Editions of Armenian Biblical Texts,” Bulletin of the International Organization for Septuagint and Cognate Studies 15 (1982), pp. 51-59 = Revue des Études Arméniennes, N.S. 17 (1983), pp. 627-633.

Zeyt‘unian, Andranik. “Some Principles for the Edition of a Critical Text of the Bible,” in Text and Context. Studies in the Armenian New Testament. Papers Presented to the Conference on the Armenian New Testament, May 22–28, 1992, ed. Shahé Ajamian and Michael E. Stone. University of Pennsyvlania Armenian Texts and Studies 13; Atlanta, GA: Scholars Press, 1994. Pp. 115-122.