Armenia had no written literature at all until the fifth century of the common era. When Mesrop Mashtots, a former royal secretary who had become a priest and then a hermit, sat down in the city of Edessa to fashion a script for his own language, he was not reviving something that had decayed. He was inventing it. The thirty-six characters he cut, arranged so that seven of them served as vowels, gave a people a written medium that would carry their law, their liturgy, their chronicles and their science for fifteen centuries. Within a few years he used it to translate the Hebrew Scriptures and the New Testament into Classical Armenian. That translation, worked on and revised over two generations, is the reason a script of roughly forty signs can still be read today by specialists in the most important non-biblical witnesses to the history of the biblical text, and it is why the Armenian Bible continues to appear in the footnotes of critical editions of the Masoretic Text. This essay follows that book from the scriptorium at Edessa to the printing house at Amsterdam, and asks what a translation made for a small Caucasian kingdom has to teach us about the transmission of scripture itself.
The Alphabet and the First Translation
The biography of the invention survives in the Life of Mashtots by his pupil and biographer Koriwn, who was himself the author of the first work of Armenian prose literature, a martyrology written in the 480s. Koriwn’s account is hagiographic in tone and plainly shaped by the controversies of his own century, but its details are consistent with everything else that is known about the early translation. Mashtots was, according to his biographer, in the service of the Arsacid court and was present at the baptism of the Parthian king Vologases, after which, as Koriwn puts it, the holy man withdrew to the desert, first to the Syrian city of Syma, and then to the city of Edessa. There he studied Greek and Syriac, and there he began to write.
The alphabet itself was a technical achievement with a practical edge. Mesrop did not invent it alone. Koriwn records that in the city of Samusata, in the district of Edessa, Mashtots encountered a Greek grammarian and calligrapher named Hrop’anos, who was able to cut and fashion the characters. The two men worked the forms out together, and Mashtots returned to Edessa to complete the scheme. The system as it survives has thirty-six letters, seven of which are vowel signs written above and below the consonants rather than beside them, and a small number of ligatures that assist with clusters. The design is a marriage of Greek letterforms and a consonantal structure borrowed from Pahlavi and Syriac writing practice. Whatever its immediate models, the result was legible to Armenian speakers within a generation and has remained in use, in one form or another, for more than sixteen hundred years, which is a record no other script of the early Christian world has bettered except perhaps the Greek and the Hebrew themselves.
The alphabet was not a private solution to a private problem. Armenian Christianity existed inside the Persian Empire, in a church that had absorbed a great deal of Hellenised culture, and the alphabet as Mashtots devised it was soon applied to the neighbouring languages of the Caucasus. The royal house of Iberia, that is Georgia, adopted a script of the same family, and the so-called Caucasian Albanian language received one too. That last alphabet survives in a single document, a palimpsest of the eleventh century preserved at the monastery of Saint Catherine at Sinai, and the fact that we know the script at all is due to a palimpsest tells you something about the fragility of these first experiments. Writing was not inevitable. It had to be made to work, and then it had to be kept.
The translation followed the alphabet closely. Koriwn says that Mashtots, assisted by two of his disciples from the districts of Ekegheac and Paghin, began with the Proverbs of Solomon, and that the book was copied out by the same scribe who had taught the disciples to write the characters. The choice is not accidental. Proverbs opens with an exhortation to knowledge, which suits a translator whose first task is to demonstrate that the new letters can carry an existing sacred text. Movses Khorenac’i, writing a century later, adds that Mashtots translated the twenty-two canonical books and the New Testament. A modern summary of the work puts the translation in the region of 396 to 430, with the earliest activity dated to roughly 406 and the first translation completed within a few years of that.
Meshtots was not alone, and the wider setting of the translation was confessional rather than merely cultural. The church in the Persian Empire had been through a violent and formative period in the previous generation, including the persecution of the Christians of the Persian dominions under Shapur II and the upheaval that the succession of the bishops of the Armenian church produced in the 340s. The translation of scripture into the vernacular of a people whose bishops had just been summoned, deposed and restored was therefore an act with teeth. It declared that this people had a scripture of its own, read in its own language, and that its bishops had a right to judge by it. That claim would occupy Armenian church history for the next thousand years, and it is the context in which the chronicle writers, the hagiographers and the doctrinal controversies of the following centuries were produced.
Sahak Partev, who became Catholicos, and Eznik of Golb, who wrote the treatise that is the classical statement of Armenian Christian apologetics, both had a hand in the translation. Eznik’s Proof of the Truth of the Faith, written in the 430s, argues for the Trinity by way of the Incarnation, using the Gospel of John as its main proof-text, and it is one of the earliest documents in which the new Armenian Bible is quoted. It is worth pausing on how strange that is. Within a single generation of the invention of the script, the language was carrying sophisticated theological argument, and the scripture it argued from was already in wide use. The translation was not an exercise in missionary linguistics. It was the foundation of a written culture, and it was treated as such from the start.
One further early episode belongs here, because it shaped the canon. At the Council of Ephesus in 431, which ruled on the nature of Christ in a way that separated the churches of the Roman and Persian empires, Armenian delegates are said to have argued the case for the council’s decisions alongside the Greek fathers. The identification of those delegates with Eznik, Yovhan of Zich and Yeghes has been repeated often enough to become standard, although it rests on inference rather than on an unambiguous source. What is certain is that the Armenian church accepted the council’s verdict, and that the scholars who brought the translation back into line with the Greek text afterwards were the same men who had argued at Ephesus.
The Revision and the Hebrew Behind It
The single most important fact about the Armenian Bible for students of the Hebrew text is that it was translated from Hebrew. Not exclusively from Hebrew, and not without help, but the base text for the Old Testament was a Hebrew text closely related to the Masoretic Text as we know it, rather than the Greek Septuagint alone. This is a claim that has to be argued, because the Armenian canon includes books the Greek does not and the Armenian translation was certainly revised against Greek manuscripts. The argument runs on internal evidence, and it is strong.
Where the Hebrew can be checked against the Greek, the Armenian usually sides with the Hebrew. A classic instance is the numbering of the first book of Samuel, where the Greek and the Hebrew divide the opening chapters differently over the passage that records Samuel’s establishment as ruler, and where the Armenian, following the Hebrew, places it in the tenth chapter rather than the eleventh. In Joshua, in Kings and in the Prophets, the Armenian translators repeatedly render a Hebrew word with a term that only makes sense if they were looking at the Hebrew, and they correct the Greek where the Greek has plainly gone astray. Scholars have also noticed that the Armenian preserves, in a number of passages, readings that agree with the Hebrew against the Septuagint. The picture that emerges is of translators who knew the Hebrew, who had a Hebrew text in front of them, and who were not content to translate a Greek at second hand where the Hebrew was available.
That is a striking thing for the fifth century. Translators in the Latin West at this period were overwhelmingly working from a Greek text, and the Latin tradition did not reach the Hebrew until Jerome a century later. The Armenian achievement was to base a major Bible translation on the Hebrew at the very moment when the tradition of Christian exegesis was still centred on the Greek. Armenian translators may also have had access to the Greek of Aquila, the literal rendering of the Hebrew made in the second century and preserved in the columnar arrangement of Origen’s Hexapla, which is an argument that has been made often and that this essay does not need to settle.
The closeness to Hebrew shows itself in the language itself. Armenian absorbed a large share of Semitic vocabulary, most of it through Syriac but in a number of cases evidently directly from Hebrew, and the words the translators chose for Hebrew technical terms, for the names of the sanctuary and its furniture, for the sacrificial system and for the offices of the priesthood are a stratum of loanwords that can be read off the page like a glossary. When a translator reaches for a Hebrew loan rather than an inherited word, a Hebrew original has usually been sitting in front of him.
The revision itself is a second chapter of the story. When Eznik and his colleagues returned from the Council of Ephesus, they are said to have brought with them a Greek Bible obtained in Constantinople. Recognising the superiority of that resource, the Armenian scholars revised their translation to bring it into conformity with the Greek text, and the revision was complete by about 440. That single event explains a great deal of what looks odd about the Armenian Bible: it is a translation whose base text was Hebrew but whose final form was checked against Greek, and whose readings are therefore sometimes closer to the Greek than to the Hebrew and sometimes much closer to the Hebrew, according to which translator had his way on the day.
The seams are visible. A collation by Conybeare of the Armenian against the Syriac and the Hebrew in Jeremiah found the Armenian following Hebrew and Syriac together in the large majority of cases, the Hebrew alone in a smaller group, and the Syriac alone in only a handful. In Genesis the standard comparison lists around thirty places where the Armenian carries an addition traceable to the Syriac, together with a smaller number of places where a word or a phrase present in the other witnesses has been dropped. The Armenian translator was working with more than one document and did not always resolve the mixture cleanly. The result, as one recent survey puts it, is a synthesis in which the Syriac and the Greek base texts were poured into the mould of the Armenian language and culture.
What the Armenian translation preserved, in the process, is a Hebrew text that in a number of books is older than the form of the Hebrew we now call the Masoretic Text. In Job, in the Psalms, in Proverbs and in Ecclesiastes especially, the Armenian preserves readings that differ from the medieval Hebrew and that occasionally stand closer to the Septuagint than the Masoretic reading does. Critics of the Masoretic Text, from the seventeenth century onwards, have appealed to such witnesses. Pierre de Lagarde collected them in the volumes he called Antiquites nouvelles, resultantes de l’etude des textes armeniens, published in 1899, in which he printed the oldest Armenian manuscripts of the Old Testament with a critical apparatus and the Hebrew and Greek in parallel. Lagarde’s work was controversial, and the textual conclusions drawn from it by more radical scholars were rejected, but the underlying observation that the Armenian carries an independent and sometimes older form of the Hebrew text has never been seriously shaken. It is one of the reasons the Armenian Bible occupies a permanent place in the discussion of the history of the Hebrew text, and one of the reasons the Aleppo Codex and the Leningrad Codex, dominant as they have been, are not the whole of the story.
The Canon and the Third Column
The Armenian Bible has a canon of twenty-two books, counted in the Hebrew manner, and it numbers the Psalms according to the Hebrew rather than the Greek, so that Armenian Psalm 10 corresponds to Greek Psalm 11 and the two numbering systems run a verse apart across the whole book. The order of the Hebrew books is the Hebrew order. Around that core sits the deuterocanonical collection, which in content follows the Septuagint and includes Tobit, Judith, the additions to Daniel, the Prayer of Manasseh, the Epistle of Jeremiah, Baruch, and the three books of Maccabees, together with a second book of Esdras distinct from the Greek one. Several of these are known to us from the Armenian alone, which makes them among the rarest of textual witnesses in the history of the Old Testament. The Prayer of Manasseh survives in Armenian, Ethiopic and Syriac, and in the last two it is separated from the book of Daniel, as it is in the Armenian too, with one manuscript variant placing it before Daniel.
The additions to Daniel and to Esther are the sharpest illustration of what the Armenian canon can preserve. The Armenian carries the story of Susanna, the story of Bel and the Dragon, and the story of the youths in the fiery furnace; it carries Mordecai’s dream and the commendation of Mordecai at the close of Esther. None of these is in the Hebrew Bible, and the Greek witnesses are scattered and late for two of them. In the case of the additions to Esther the Armenian also preserves a long passage describing the dreams of Mordecai and of Esther themselves, which is of considerable interest for the history of the Septuagint text of Esther and for the peculiar shape of the book as it eventually emerged.
The New Testament canon of the Armenian church is more complicated than the number of books suggests, because it includes a book that does not exist in the Greek. Third Corinthians circulates in the Armenian, in some Syriac witnesses and in a single Coptic manuscript, and it was for a long time regarded as the survival of a lost apocryphal correspondence. The scholarly consensus has moved. Most specialists now treat it as a fifth or sixth century expansion of a genuine exchange between Paul and the Corinthians, produced in a Syrian environment, which entered the Armenian canon along with the Syriac Pauline material it accompanies. It was dropped from the Armenian canon at some point before the modern printed editions. Revelation was a late arrival in the opposite direction. It does not appear in the earliest Armenian lists, and its inclusion is the work of the later tradition, so that the Armenian Bible is one of the cases where the direction of transmission runs from the Greek East toward Armenia rather than the other way round.
Alongside the main text the Armenian manuscripts carry what is usually called the third column, a marginal apparatus of variant readings drawn from Hebrew and Greek sources, and an older layer of it records readings that have Hebrew equivalents. This marginal tradition, which survives in a small number of manuscripts and in some printed editions, is one of the most direct pieces of evidence that the translators had access to the Hebrew. It is also, in a sense, the Armenian Bible’s own Masorah: a body of annotation attached to the text, recording where the witnesses disagree, maintained for centuries in the same way that the Masorah Magna was maintained in the Hebrew codices. One Armenian manuscript tradition goes further still and preserves a set of marginal notes that reflect the anti-Jewish polemical passages that Christian scribes inserted into manuscripts of this period, a reminder that the translation was copied, read and annotated in communities where the boundaries between traditions were worked out in ink.
Armenian tradition calls the Bible the queen of translations, a phrase usually traced to Gregory of Narek, who wrote at the end of the fifth century and who is credited with a commentary on scripture that is itself part of the classical Armenian literary canon. The compliment is understandable. The Armenian Bible is not a derivative curiosity. It is a translation made from the Hebrew, revised against the Greek, preserved with a marginal apparatus of variants, and quoted by writers who were thinking about the Trinity. It belongs in the same family as the Septuagint and the Vetus Latina, and it is on that level that scholars now treat it.
The Manuscript Tradition
The earliest substantial Armenian Gospel manuscripts date from the ninth century, and they are among the finest products of the Cilician scriptoria. The Gospels of Queen Mlk’e, held in the Mekhitarist library of San Lazzaro degli Armeni in Venice as MS 1144/86, is usually dated to 851, though the colophon reckoning has been read in ways that place it in 862. Its colophon is an elegy on the queen who commissioned it, describing her as having ordered the gospels written in letters of gold and gold-leaf ornament, and the manuscript is a principal witness to the earliest form of the Armenian Gospel text. The Lazarian Gospels, Matenadaran MS 6200, were written in 887 and are named for the Lazarian family who held them at Etchmiadzin. The Gospels of the Priests, Walters MS 537 in Baltimore, is dated to 966. Matenadaran MS 2374, produced in 989 and formerly catalogued as Ejmiatsin 229, is a treasure of a different kind, because it preserves an unusual form of the passage about the woman taken in adultery at John 7:53 to 8:11, carries the note about the elder Ariston between Mark 16:8 and Mark 16:9, and has four illustrations and ivory covers that are thought to have come from its exemplar.
One further manuscript deserves to be singled out. Matenadaran MS 10434, dated 1069, carries a colophon stating that its exemplar had, by tradition, been made by Mesrop Mashtots himself. Such claims are common in manuscript culture and are rarely true. What matters is that the Armenian tradition wanted the oldest surviving books to descend directly from the translator, and that a colophon could make that claim and be believed. The genealogy of the text was part of its meaning.
The Eusebian canon tables, familiar from Greek and Latin gospel books, appear in the Armenian tradition too, and they carry a feature of their own. Because the Armenian alphabet serves also as its numerical notation, the section numbers in the canon tables are given in letters that are at once a locator and a piece of the text. A reader who cannot read Armenian can still find the passage, because the numbers are the numbers. The arrangement of these tables, together with the illustrations, suggests that they were part of the Armenian tradition from the time of the revision in the 430s, not something added when the manuscripts were being written in the ninth and tenth centuries.
Illumination and colophon are the two things a reader of these books is most likely to notice, and both of them point the same way. The headpieces and canon-table illuminations use the same repertoire of flat colour bands, birds, fish, rosettes and archangels that the stone carving and the metalwork of the period used, so that a manuscript page and a church wall in Cilicia can be read as one visual language. The colophons are documents rather than decorations. They give the date, the place, the scribe, the patron, the price in coin, and often a curse on anyone who takes the book away. They are written in the vernacular rather than in the literary language, which means they preserve the ordinary Armenian of their own centuries and can be used for linguistic study as well as for history. Taken together they allow the modern scholar to reconstruct the economy of a Cilician scriptorium in a way that is possible for very few medieval book cultures.
The repositories are as important as the manuscripts. The Matenadaran in Yerevan was created in 1921 on the basis of the manuscript library of Etchmiadzin, was transformed into a Scientific Research Institute of Ancient Manuscripts by decision of the government of Soviet Armenia in 1959, and was named after Mesrop Mashtots in 1962. It holds roughly seventeen thousand manuscripts, written in Armenian and also in Greek, Arabic, Persian, Syriac, Ethiopian, Georgian and Hebrew, and it has the distinction of being the first scientific research institution in Armenia. In 1997 its collection was included in the UNESCO Memory of the World programme. The Mekhitarists of the Congregation of Mechitarists of San Lazzaro, the Armenian Catholic order that settled in Venice in the seventeenth century and later in Vienna, hold another very large body of Armenian manuscripts and, just as importantly, ran the printing presses that carried the Armenian Bible into the modern era. Between the Matenadaran at Yerevan, San Lazzaro in Venice and the Mekhitarist monastery in Vienna, the two principal centres of the diaspora, the great majority of the surviving Armenian manuscripts are held.
The Cilician scriptoria produced the books and the colophons. Haghpat, Sanahin, Gladzor and Ani were the great houses of the twelfth-century Cilician renaissance, when Armenian church building, miniature painting and manuscript production reached a level that later scholarship has been inclined to call the golden age of the medieval Armenian book. The story of these monasteries is also a story of refugees. Armenian artists and craftsmen moved south and east as the political situation deteriorated, taking their models with them, and the resulting dispersal is one of the reasons that the artistic styles of Etchmiadzin, of Vaspurakan and of New Julfa can be compared so precisely. A workshop survives in its books when the building that housed it does not.
Loss, Print and Recovery
The tradition nearly died several times. A great famine in the middle of the ninth century devastated the Cilician towns and monasteries and permanently altered the manuscript economy of the region. The Mongol conquest of the thirteenth century destroyed the political framework of Armenian kingship in Cilicia, and the sack of Sis in 1375 finished it. Armenians migrated to Georgia, to Russia, to Poland, to the Netherlands and to a great extent to India, and the scriptoria that had produced the manuscripts carried the tradition with them into a diaspora that was, in the aggregate, far poorer and far more scattered than the kingdom that had produced it. The story of Armenian manuscript preservation is in large part the story of what a handful of monasteries and a religious order managed to carry out of a collapsing world.
Printing arrived late and from outside. The first substantial printed Armenian books came out of the Venetian presses of the Mekhitarists, including a Psalter printed at Venice in 1565 by a scribe named Abgar. The first complete printed Armenian Bible is a more complicated affair. Mattheos Tsaretsi, sent to Europe by Catholicos Hakob IV in 1656 to establish a printing house, settled in Amsterdam after an unsuccessful attempt in Italy, and in a 1658 agreement commissioned from the Elzevir cutter Christoffel van Dyck 170 block punches and 240 matrices for large and small Armenian letters. The Bible was printed under the direction of Voskan Yerevantsi, and the printing began on 11 March 1666 and was completed on 13 October 1668, in two volumes of 1,462 and 1,436 pages with a two-column double pagination. Its text was based on a manuscript of 1295 made for Hetum II of Cilicia. Voskan himself admitted that he had altered the Armenian in a number of places, correcting it and harmonising it with the Latin Vulgate under pressure from the Catholic inquisitorial authorities, and the edition was accordingly criticised for its orthodoxy as well as for its textual quality. One of the specific consequences is visible in the text of 1 John 5:7, where the passage about the three heavenly witnesses appears in Voskan’s edition and is absent from the earlier manuscripts; it disappeared again from the later Mekhitarist editions.
That single verse is a compact illustration of the whole history. The clause was not in the Armenian manuscripts. It was in the Vulgate. A translator determined to remove every doctrinal offence found in his exemplars felt obliged to add it, and in doing so permanently altered the text of a book that his own tradition had received without it. Two centuries later the Mekhitarist editors, working from the manuscripts rather than from the Latin, removed it again. The Armenian Bible can be watched being made, corrected and corrected back.
The Mekhitarists carried the improvement forward. Zohrab published a New Testament in Venice in 1789 and a critical edition of the whole Bible in 1805, with further editions at Serampore in 1817 and at Constantinople in 1892, the last of which printed the readings of the Hebrew text at the foot of the page. A Vienna edition followed in 1929. The sequence matters because it shows the Armenian Bible moving from a single exemplar, harmonised with a Latin rival, to a critical text with a Hebrew apparatus, which is precisely the trajectory that the manuscript tradition had prepared and that the printed tradition finally realised.
In the twentieth and twenty-first centuries the text has been revised again, this time into the eastern vernacular. A modern Eastern Armenian revision, produced for the Translation Bible Society, moved the Gospel of John in 1999, the New Testament in 2005 and the entire Bible in 2012, working from the classical translation of the fifth century and bringing it into closer conformity with the Hebrew Masoretic and Greek received texts. The critical apparatus of the Matenadaran, its general catalogues and its digitised facsimiles have made the older witnesses accessible to scholars who have no Armenian of their own. The point of all this is that the Armenian Bible is no longer a national possession in any narrow sense. It is one of the standard witnesses in the study of the Hebrew Bible and of the Greek New Testament, and it is consulted routinely in the preparation of critical editions.
There is a certain irony in that. A book written to give a small kingdom its scriptures turned out to preserve, in its margins and in its main text, material that helps scholars reconstruct the history of texts far older than the kingdom and far larger than it. Mesrop Mashtots invented thirty-six letters so that his own language could be written. In doing so he put a version of the Hebrew Scriptures into the hands of scribes who would copy it for fifteen hundred years, and the version they copied is still doing work.
References
- The Text of the Gospels: Armenian New Testament manuscripts and traditions – Matenadaran and the Armenian gospel manuscript tradition, with shelfmarks and dates
- The History, Base Text(s), and Translation Techniques of the Armenian Bible – Levon Jinbachian, 2004, on the Syriac and Greek base texts and the Masoretic background
- Classical Armenian: Introduction, Part 2 – Projekt zur Erforschung der Geschichte des Armenischen in Göttingen, on the Bible translation and the earliest manuscripts
- Museum of Ancient Manuscripts Matenadaran – Mesrop Mashtots Institute of Ancient Manuscripts, Yerevan, on the institute’s holdings and history
- Old Printed Books – Matenadaran, on the first printed Armenian Bible and the Venice and Amsterdam presses
- The Early Armenian Merchant Community of Amsterdam – De & De Kerakart, on the establishment of the Armenian printing house and the first Armenian Bible
- The First Armenian Printed Bible of Voskan Yerevantsi at Naasr’s Mardigian Library – Banber, the Armenian National Library of America journal, on the Amsterdam printing of 1666 to 1668
- Versions of the Bible – an historical survey of the printed editions of the Armenian Bible from the Psalter of 1565 to the Vienna edition of 1929
- Along the Paths of the Spiritual Map: Armenian Manuscript Art – Biblioteca Nacional de Espana, exhibition leaflet, on the scriptoria of Cilicia, Utik, New Julfa, Vaspurakan and Crimea
- The Aleppo Codex: Masoretic text and apparatus – the Aleppo Codex as the reference witness for the Masoretic Text in current digital editions
- General Catalogue of Armenian Manuscripts of the Mashtots Matenadaran, Volume XI – Matenadaran, Yerevan, shelfmark reference for the continuation of the printed catalogue
- Armenian – Translation Bible Society, on the modern Eastern Armenian revision of 1999 to 2012 and the printed history of the Armenian Bible
- Armenische Studien – Paul de Lagarde, Göttingen, Dieterich, 1877, digitised full text on the Internet Archive; the volumes of Antiquités nouvelles that followed in 1899 gathered the Armenian Old Testament witnesses he used
- Armenian Bible – overview of the translation, the canon, the manuscript tradition and the printed editions, with a bibliography
- Mesrop Mashtots – the invention of the Armenian alphabet, the translation of the Scriptures and the Armenian school of Sahak Partev and Eznik
- Matenadaran – the Mesrop Mashtots Institute of Ancient Manuscripts, its foundation, its holdings and its role in the preservation and study of the manuscripts