DIGENES AKRITES EDITED WITH AN INTRODUCTION TRANSLATION AND COMMENTARY BY JOHN MAVROGORDATO FORMERLY BYWATER AND SOTHEBY PROFESSOR OF BYZANTINE AND MODERN GREEK LANGUAGE AND LITERATURE IN THE UNIVERSITY OF OXFORD AND FELLOW OF EXETER COLLEGE OXFORD AT THE CLARENDON PRESS Oxford University Press, Ely House, London W.I GLASGOW NEW YORK TORONTO MELBOURNE WELLINGTON CAPE TOWN SALISBURY IBADAN NAIROBI LUSAKA ADDIS ABABA BOMBAY CALCUTTA MADRAS KARACHI LAHORE DACCA KUALA LUMPUR SINGAPORE HONG KONG TOKYO FIRST PUBLISHED 1956 REPRINTED LITHOGRAPHICALLY AT THE UNIVERSITY PRESS, OXFORD FROM SHEETS OF THE FIRST EDITION 1963, 1970 PRINTED IN GREAT BRITAIN PREFACE MANY years ago my friend Petros Petrides, himself born at Nigde in Kappadokia, used to tell me about the stories of Digenes, which were to inspire his own dramatic symphony first performed in Athens in May 1940. When at last I started to read for myself this heroic poem of Twyborn the Borderer, who guarded the marches of the Byzantine Empire in the tenth or eleventh century, I was alone in the country, and found my way through the text as best I could; if some of the notes now published sound like an ingenuous soliloquy, it is from that early period they must have survived— excusata suo tempore, lector, habe; exsul eram. Later on I was back in London, and there enjoyed the advice and encouragement of Nor- man Baynes, of F. H. Marshall, and of Romilly Jenkins; somehow or other my translation and a few lectures were on paper in 1938; and not long afterwards, thanks perhaps largely to their continued interest, I found myself living in College at Oxford and on the same staircase as my friend R. M. Dawkins. Many were now the anxious days and black nights in the course of which we talked about nearly everything in the world. I do not remember that in fact we ever talked about the Warden of the Marches. But Dawkins had such a lively knowledge of the land and the people and the language of Greece, and such an affection for them all, ancient, mediaeval, and modern, that his talk and his example were a university; and this is the place in which I must say how much I owe consciously and unconsciously to his learning and judgement. It was some years later, and I was no longer living in Oxford, when the Delegates of the Clarendon Press gave me one of the pleasantest surprises of my life by undertaking the publication of this work. By this time most of my books had had to go into store; and as I had been brought up to fill every margin with notes—not to mention several cardboard boxes full of half-sheets of paper—many of my cherished references had now become inaccessible. Perhaps that does not matter as much as I thought it did. But it has taken a long time to reconstitute the commentary and do all that was necessary; and I am immensely grateful to the readers and experts of the Press whose patience and skill have been at my service on every page. 3 PREFACE The Greek text here printed is that of the Grottaferrata MS. of Digenes edited by Emile Legf and and published in Paris (by Maison- neuve) in 1892. A few passages to fill up gaps in the narrative have been borrowed from the Trebizond MS., edited by C. Sathas and Legrand and published in Paris in 1875; ajid from the Andros MS., edited by A. Meliarakes and published in Athens in 1881, these passages amounting to 117 and 26 lines respectively. A second edition of Legrand's text revue et corrigee was issued in 1902, but all the errors I have noticed appear in both editions. Legrand can never be sufficiently honoured for his great services to the study of post-classical Greek, and above all for his two large collections of Monuments pour servir d I'etude de la langue neo-hellenique, but much of his work contains small inaccuracies. Some light may be thrown on these by an affectionate tribute he pays to his wife in the preface to the third edition (Paris, 1900) of the Cretan pastoral The Fair Shepherdess. He there tells us that in order to lighten his labours she had learned to read and write the Greek characters and so for many years had copied out all his Greek texts for him: 'ma femme transcrivait avec une exactitude d'autant plus surprenante (ce qui peut sembler paradoxal) qu'elle savait moins la langue.' Of the slips in his Grottaferrata text and apparatus criticus a few erratic punctua- tions and accents have been silently corrected. Otherwise his text and apparatus have been exactly reprinted and any emendations I have been obliged to make are recorded only in my own notes. In his introduction Legrand recalls that on the publication in 1875, edited by Constantine Sathas and himself, of the Trebizond MS.—the first to be discovered—the interest excited had led to the publication of the Oxford MS., which was included by S. P. Lam- pros in his Romans grecs en vers (Paris, 1880), and of the Andros MS., edited by Meliarakes (Athens, 1881). Lampros in his introduction (pp. xc—ci) describes the manuscript which he had seen in the autumn of 1879 in the library of the Greek monastery at Grotta- ferrata near Frascati. It is a manuscript of the fourteenth century written on 79 leaves of paper (5^ X 8£ inches—14x21 cm.) of which the first 73 contain the text here printed. Joseph Miiller of the University of Turin had already called attention to this manuscript, and had himself made a copy of it which he had shown to Legrand and had subsequently presented to the German scholar Wilhelm Wagner (editor of Medieval Greek iv PREFACE Texts, London, 1870; Carmina Graeca Medii Aevi, Leipzig, 1874; and Trots poemes grecs, Berlin, 1881). After Wagner's premature death in 1880 this copy had been returned to Miiller and has never been heard of again. Meanwhile Nicolas Polites in Athens, fired by the description given by Lampros, had asked the librarian at Grottaferrata to have the manuscript copied for him; and Legrand through the French Ministry of Foreign Affairs had asked for a loan of the precious volume which in due course reached him in Paris. Lampros and Polites withdrew in his favour, and Legrand with his friend Jean Psichari read over together the copy which he had made, and certified its accuracy, in May 1887. The most important thing about my translation is that it follows the Greek line for line, so that the original can always be referred to without delay. This has involved a good deal of compression but few, if any, sacrifices. On the rare occasions on which some Greek word has had to be omitted the fact has been recorded in the notes. This does not apply to particles which the Greek author uses almost indiscriminately in order to fill up his line. In order to empty the English line I have omitted Tors', 'howevers', and 'yets> by the bushel, and 'huts' by the load; thus not only lightening the metre but im- proving the style. In English a minimum of such words is a condition of good writing, as a minimum of gestures is a condition of good speaking. Secondly, this translation has been kept as literal as possible. First finished in December 1938, it had to be copied out. In the first ten weeks of 1939 it was not only copied out but completely rewritten from beginning to end, because almost every line as it came up for approval was changed so as to give a more literal representation of the Greek, even if it produced less tidy metre and less conventional English. One result of this insistence on literalness has been an occasional loosening of the metre; so that instead of five feet and ten syllables some lines (e.g. 3314) have three beats each followed by a varying number of unaccented syllables. Thirdly, I have never tried to improve the Greek original. (Well, hardly ever.) Clumsy Greek has often been turned into clumsy English. I have resisted the temptation to conceal silliness or bathos, PREFACE and to make the English line more poetical than the Greek line which it is supposed to represent. Generally speaking I have used a basis of ordinary conversational English; the fundamental rule for the structure of this groundwork being that there shall be no inversion of the natural order of words. To reinforce this natural order there has been an avoidance of con- ventionally poetical usages. I have never used the second person singular, except where God is addressed. The heroine, when she is referred to, perhaps with some mystical significance, as Kore, is called the Girl rather than the Maiden or the Virgin. On the other hand, although based on natural speech the diction is one definitely removed from that of common talk. It is what Aristotle in his Poetics calls exellagmene; it is changed over from the customary dialektos, the conversation of ordinary life. This translation in fact uses a written language and not a spoken language. And why not? Does anybody suppose that the Greek was not an artificial language? It is only as a result of contemporary fashions, English as well as Greek, that one has to defend the perfectly natural division of the written from the spoken language. This division connotes the artificiality of the written language. But all writing is an artifice and it can never be anything else. The words commonly used in streets and fields are often better than others, but they are not in themselves patterns laid up in heaven. Having laid down the ground, a written language free from inver- sions and poetical conventions, based on ordinary speech but clearly divided from it, we proceed to colour it with reminiscences of the Bible, of Shakespeare, of Milton, and of the Ballads; inevitable reminiscences, not because they are beautiful English, and recall beauty by association even when they have it not in the immediate use; but because these exactly represent the Greek of the ancient classics, which the redactor of this version was attempting to revive, with painful but with loving artifice, in the middle of the eleventh century. His is certainly an artificial language. There is only one characteristic of his style which he shares with the good conversationalist and the good letter-writer: that he often begins a sentence without knowing how he is going to end it. This is not the place to examine other characteristics of his style, notably his mixture both in vocabulary and in syntax of elements from every period, from Homer, from the Septuagint, and from vi PREFACE contemporary folk-song. This eclecticism shocks the grammarians, and it shocks the Athenian demotics; but it does not shock anyone who has a sense and a love of the Greek language as a whole, to whom no part is unclean. It does not shock me, and I have some- times even attempted to reproduce it in English. In passages where the Greek is peculiar, and possibly corrupt, or the words rare or remarkable, or the meaning doubtful, the original was sometimes quoted in the notes, without comment, so as to call attention to the fact that further elucidation was required. But many of these mere exclamation marks were afterwards left out in order to save space. The lines of this edition are numbered on two systems. The num- bers in the margins of the right-hand pages, that is of the English version, include not only Grottaferrata but also the passages borrowed from the Trebizond and Andros versions to fill up gaps in the narra- tive. This numeration, running from i to 3850, is used in the notes and in the Index of Greek Words. The Greek text of the Grottaferrata and other versions on the left-hand pages have each book numbered separately—(GRO Book I, 337 lines; Book II, 300 lines; Book III, 343 lines; Book IV, 1093 lines; Book V, 289 lines; Book VI, 845 lines; Book VII, 229 lines; Book VIII, 313 lines; TRE and AND, Books I to X); this numeration given on the left-hand pages is to be associated with the number of the book in Roman figures given at the top of the page, and both the line-numbers and the Roman book-numbers must be used for reference to the Greek text. (Note, however, that the TREbizond and ANDros versions, each divided into ten books, are not numbered by separate books but straight through from beginning to end—3182 and 4778 lines respectively.) J. N. M. CONTENTS INTRODUCTION xi FIRST BOOK 2 SECOND BOOK 24 THIRD BOOK 44 FOURTH BOOK 66 FIFTH BOOK 142 SIXTH BOOK l62 SEVENTH BOOK 2l6 EIGHTH BOOK 232 APPENDIXES A-D A. GENEALOGY OF DIGENES 253 B. CONSPECTUS OF VERSIONS AND EPISODES 257 C. LIST OF BOOKS 26l D. REFERENCES TO ACHILLES TATIUS, HELIODORUS, AND MELITENIOTES 265 INDEX OF GREEK WORDS 267 INDEX TO INTRODUCTION 271 INTRODUCTION i. DISCOVERY IN the middle of last century nothing was known of the Byzantine epic of Digenes Akrites; but the atmosphere had been prepared by the publication of several ballads of what is now called the Akritic Cycle (a name first used by Legrand in 1874). The discussion of these—(in particular a paper by Biidinger who had used the headline 'A Greek Mediaeval Popular Epic', although the Song in question, The Sons ofAndronikos, was only seventy lines in length)—had opened the way for further revelations of an heroic age of mediaeval Greece. (i) The discovery was made at Trebizond. Manuscripts could not be photographed at Trebizond, and after considerable correspondence the precious work was sent to Paris by post; and in 1875 Sathas and Legrand published Les Exploits de Digenis Akritas—epopee byzantine du dixieme siecle... d'apres le manuscrit unique de Trebizonde. The manuscript is said to be not earlier than the sixteenth century. There are several gaps in it. The poem is divided into ten books of which the first and the last are missing altogether; there remain 3,182 lines of it. This manuscript was afterwards returned to the hands of Sawas loannides, the Trebizond schoolmaster, author of a statistical history of Trebizond, who in May 1868 had received it from a monk at the monastery of Soumela. (For the present state of the monastery of Soumela see D. Talbot Rice in Byzantion, v. 72 ff.) He published another edition of it at Constantinople in 1887," and then deposited it in the library of the Filologikos Syllogos. All the archives of this in- stitution are said to have been removed to Angora by the Turkish Government and the fate of the Digenes manuscript is unknown; but in the opinion of Kyriakides (conveyed in a private letter 7 April 1936) it must, for the present at least, be regarded as lost. (ii) Before the learned world had had time to digest this, several 1 Epos Mesaionikon / ek tou kheirografou Trapezountos / O Basileios Digenes Akrites / O Kappadokes / ypomnematisthen ekdidotai / ypo / Sabba loannidou / ... en Konstantinoupolei/ . . . 1887. The text is fairly correct and is supplemented by about 700 lines from the OXF and AND versions. loannides regards Digenes as a fully historical figure of the Kappadokian aristocracy who lived from 936 to 969 against a background slightly adjusted from the chronicle of Kedrenos. xi