1. See Kyriakidis, S., (Athens, 1926)Google Scholar; idem, , ‘Forschungs-bericht zum Akritas-Epos’, Berichte zum XI Internat. Byzantinisten-Kongress, II, 2 (Munich, 1958)Google Scholar; Grégoire, H., Autour de Ièpopèe byzantine (Variorum Reprints: London, 1975)Google Scholar; Kalonaros, P., (Athens, 1941), introduction; and Mavrogordato, J., Digenes Akrites (Oxford, 1956), introduction.Google Scholar

2. Mavrogordato, op. cit., p. xviii; Morgan, G., Cretan poetry: sources and inspiration, offprint from XIV (Iraklion, 1960), p. 54.Google Scholar

3. ‘Notes on Digenis Akritas and Serbocroatian epic’, Harvard Slavic Studies, II (1954), 375–83. See also idem, , The singer of tales (Cambridge, Mass., 1960), pp. 207–20.Google Scholar

4. See Notopoulos, J., ‘Originality in Homeric and Akritan formulae’, XVIII (1959), 423–31 Google Scholar; Papacharalambous, G., ‘Akritic and Homeric poetry’, XXVII (1963), 25–65 Google Scholar; and Lord, A. B., ‘Parallel culture traits in ancient and modern Greece’, BMGS, III (1977), 71–80.Google Scholar

5. Jeffreys, M., ‘The literary emergence of vernacular Greek’, Mosaic (Manitoba), VIII (1974), 169–93 Google Scholar. See the series of articles by Jeffreys, M.: ‘Formulas in the Chronicle of the Morea’, DOP, XXVII (1973), 165–95 Google Scholar; ‘The nature and origins of the political verse’, DOP, XXVIII (1974), 141–95; and by E. and Jeffreys, M., ‘Imberios and Margarona: the manuscripts, sources and edition of a Byzantine verse romance’, B, XLI (1971), 122–60 Google Scholar; ‘The traditional style of early demotic verse’, BMGS, V (1979), 113–39. A useful summary of the work of the Jeffreys in this area is given by Jeffreys, E., ‘The popular Byzantine verse romances of chivalry. Work done since 1971’, XIV (1979), 20–34.Google Scholar

6. Parry, M., ‘Studies in the epic technique of oral verse-making, I: Homer and the Homeric style’, Harvard Studies in Classical Philology, XLI (1930), p. 80.Google Scholar

7. op. cit., p. 122.

8. Cf. Finnegan, R., ‘What is oral literature anyway?’ in Stolz, B. and Shannon, R. (eds), Oral literature and the formula (Michigan, 1976), p. 160 Google Scholar: ‘… poetic language, whether “oral” or written is merely one kind of language, and the occurrence of repeated patterns in oral poetry need not in principle any longer be regarded as some puzzling case to be explained in quasi-mystical terms like “dependence on formulae” which seem to set it apart from other poetry.’ An attempt to redefine the formula in linguistic terms has been made by Kiparsky, P. (‘Oral poetry: some linguistic and typological considerations’, in Stolz and Shannon, op. cit., pp. 73–106), and his conclusion and that ofFry, D. K. (‘Caedmon as a formulaic poet’, Forum for Modern Language Studies, X, 3 (1974), 227–47)Google Scholar in a study of formulas in Anglo-Saxon poetry, take the redefinition of the formula one step further, with the suggestion that it is the stylisation of language into formulas that creates metre, and not the other way about. This is an idea which will need further investigation, but I am particularly struck by Kiparsky’s suggestion that formulas should be studied in oral prose (op. cit., pp. 87–8).

9. op. cit., p. 86.

10. See Finnegan, R., Oral poetry, its nature, significance and social context (Cambridge, 1977), passim, and cf. Kiparsky, , ‘Oral poetry’, pp. 99–102 Google Scholar; Vansina, J., Oral tradition: a study in historical methodology (Harmondsworth, 1973)Google Scholar, passim; and Smith, J. D., ‘The singer or the song? A reassessment of Lord’s “oral theory”’, Man, XII, 1 (1977), 141–53 Google Scholar. For further discussion of this subject see Alexiou’s, M. (unpublished) paper given at the 1978 Colloquium, ‘Discounting formulas in medieval Greek “oral” poetry’, and Beaton, R., Folk poetry of modern Greece (Cambridge, 1980), pp. 40–55.Google Scholar

11. A striking instance in the Yugoslav tradition is cited by Lord in the attempted reconstruction of an archetypal version from which five singers said they had learnt their own. The divergences were such that the attempt proved impossible, although Cor Huso, the singer of the lost version, had been only a generation older than Lord’s informants. See Lord, ‘Tradition and the oral poet: Homer, Huso and Avdo Medjedovic’, Atti del Convegno Intemazionale sul Tema: La Poesia Epica e la sua Formazione, Accademia Nazionale dei Lincei, Quaderno 139 (Rome, 1970), pp. 13–30.

12. See M. and E. Jeffreys, ‘Imberios’.

13. This is the dating given by Politis, L.. See ‘L’épopée byzantine de Digénis Akritas; problèmes de la tradition du texte et des rapports avec les chansons akritiques’, Atti del Convegno Internazionale sul Tema: La Poesia Epica e la sua Formazione (Rome, 1970), pp. 551–81 Google Scholar (reprinted in idem, Paléographie et littérature byzantine et néo-grecque [Variorum Reprints: London, 1975]).

14. References are to the following editions: G - Mavrogordato (1956); E and A - Kaonaros (1941). Where appropriate, quotations have been checked with the readings of Professor E. Trapp’s synoptic edition, Digenes Akrites - synoptische Ausgabe der àltesten Versionen (Vienna, 1971). Trapp’s edition has not been used otherwise, as the exact stylistic features of each version as recorded in the manuscripts have not always been preserved by his emendations. Similarly, Trapp’s reconstruction of Manuscript ‘Z’ from four different texts is of no use for an analysis of possible oral features.

15. Singer, pp. 212–13. An unpublished survey, begun by M. Alexiou and D. Holton some years ago, adds a good many more formulas to those listed by Lord, as well as a number of ‘non-formulas’ and ‘clichés’ (Alexiou, ‘Discounting formulas’).

16. G VIII, 79–80; 116–7; 121–2.

17. (VIII, 126).

18. (VIII, 151).

19. The apparent formula system (VIII, 72) could well have been used in this line,

(VIII, 83)

to make a ‘formulaic’ line;

20. See G I, 54; 64; 310; 321.

21. G VIII, 79–80; 88; 102; 116–17; 121–2.

22. Singer, pp. 213–15.

23. For example, in the second half-lines: (1670).

24. E 1658. Other examples of formulas ‘borrowed’ from folk songs in this passage are (1709) and (1779).

25. For example:

(1666–7), and cf. 1673–4.

26. Singer, pp. 211–12.

27. For example: (4526).

28. We find: (4551).

29. See Curschmann, M., ‘Oral poetry in medieval English, French and German literature: some notes on current research’, Speculum, XLII (1967), 36–52.Google Scholar

30. Grégoire quotes as evidence G I, 13, the introduction to A, and E 1673, where the are addressed (‘Le tombeau et la date de Digénis Akritas’, B, VI (1931), 481–508, reprinted in Autour).

31. At G IV, 4 and 19 (pointed out by Pertusi, A., ‘La poesia epica bizantina e la sua formazione’, Atti del Convegno Internazionale sul Tema: La Poesia Epica e la sua Formazione [Rome, 1970], p. 530 n.).Google Scholar

32. (Iraklion, 1979).

33. Most recently by Oikonomides, N., ‘L’ “épopée” de Digénis et la frontière orientale de Byzance aux Xe et Xle siècles’, Travaux et Mémoires, VII (1979), 375–97.Google Scholar

34. Successive editors have noted borrowings from Achilles Tatios, Heliodoros, Pseudo-Kallisthenes among ancient sources, and Meliteniotes and Genesios among those of the Byzantine learned tradition.

35. See Crosby, R., ‘Oral delivery in the Middle Ages’, Speculum, XI (1936), 88–110 Google Scholar; Baugh, A. C., ‘Improvisation in the Middle English romance’, Proceedings of the American Philosophical Society, CIII (1959), 418–54 Google Scholar; idem, , ‘The Middle English romance: some questions of creation, presentation and preservation’, Speculum, XLII (1967), 1–31 Google Scholar; and M. and E. Jeffreys, ‘Imberios’.

36. ‘L’épopée’.

37. Contemporary references to and are found in Constantine Porphyrogennetos, De administrando imperio and De caerimoniis; in the anonymous De velitatione bellica and De re militari; and in Kekaumenos, Strategikon. For detailed discussion of the use and meaning of these terms see Pertusi, A., ‘Tra storia e leggenda; Akritai e ghâzi sulla frontiera orientale di Bisanzio’, XIVe Congrès International des Etudes Byzantines, Rapports II (Bucharest, 1971), pp. 27–72 Google Scholar; Huxley, G., ‘Antecedents and context of Digenes Akrites’, GRBS, XV (1974), 317–38 Google Scholar; and Oikonomides, ‘L’” épopée”’.

38. Al-Tabari writing about 915 mentions a raid on Ikritiyah in the summer of 791 ( Brooks, S., ‘Byzantines and Arabs in the time of the early Abbasids’, EHR, XV [1900], p. 740)Google Scholar. Ikrita is said to be the seat of the strategos of Chaldia, in northeastern Anatolia, by Ibn al Fakih al Hamadhani (c. 902), whose account of the Byzantine themata derives ultimately from the testimony of an Arab prisoner exchanged in 845. (Text and discussion in Brooks, ‘Arabic lists of the Byzantine themes’, JHS, XXI [1901], 67–77.)

39. The earliest is the Arabic romance Dhāt el Himma or Delhemma, in which a Greek king, al-Aqrftish is twice mentioned ( Canard, M., ‘Delhemma, épopée arabe des guerres arabo-byzantines’, B, X [1935], 283–300 Google Scholar; and idem, , ‘Dhul-Himma’, The Encyclopaedia of Islam [Leyden and London, 1956], pp. 233–9)Google Scholar. The only published text of this work, in Arabic, appeared in Cairo in 1909. In the Turkish Sayyid Batial, which derives in part from the same tradition, a Byzantine general, ‘accursed Akrates’ is a champion wielder of the mace, Digenes’ weapon ( Ethé, H., Die Fahrten des Sajjid Banhai, 2 vols (Leipzig, 1871), Vol. I, pp. 99, 174–91)Google Scholar, and in Book IV a character of this name is described as ‘Prince of the Maghrib’ (Vol. II, pp. 79, 82). Finally, in the Turkish Dānismendnāme (c. 1245), a Greek champion, Aklātīs fights with an Amazon and excels with the mace ( Mélikoff, I., La geste de Melik Dānismend, 2 vols (Paris, 1960), I, pp. 404, 409)Google Scholar. On the Arab and Turkish romances and their relevance to Digenes see also H. Grégoire, ‘L’épopée byzantine et ses rapports avec l’épopée turque et l’épopée romane’, Bulletin de la Classe des Lettres et des Sciences Morales et Politiques deľAcadémie Royale de Belgique, 5e Série, XVII (1931), 463–93; idem, ‘The historical element in western and eastern epics’, B, XVI (1942–3), 527–44 (both reprinted in Autour); and Pertusi, ‘Tra storia’, pp. 41–50.

40. Devgeniy as the sole name for the hero in the Russian versions must derive from the Greek poem in some form, and the absence in these versions of the name Akrites or of any reference to the meaning of Digenes in Greek make it improbable that the Russian versions can tell us much about the origin of the Greek poem, as was once thought. (On this whole vexed question see Graham, H. F., ‘The tale of Devgenij’, BS, XXIX [1968], 51–91 Google Scholar.) Similarly, the reported occurrence of Digenes as a proper name in Crete ( Politis, N. G., I, [Athens, 1920], p. 253)Google Scholar and of Devgeniy as a Russian patronymic (A. Solov’ev, ‘La date de la version russe du Digénis Akritas’, B, XXII [1952], 129–32), both of the twelfth century, can more easily be attributed to the influence of the poem than to the independent existence of the name in either language.

41. MPG, LXXXIX, p. 117 C.

42. Quaestiones physicas et epistolas (Paris, 1835), p. 57, letter 43 (wrongly attributed to Theophylaktos of Ochrid in Dimitrakos, Lexikon, entry ).

43. Commentary, p. 150, line 27.

44. Theophanes Continuatus (CSHB, 1838), p. 603.

45. G II, 291. Trapp oddly has although it is quite clear from the context that digenes here is an adjective. The Emir’s son is not formally named until the beginning of Book IV.

46. Digenes abducts the General’s daughter for the illogical reason that he wants to ‘travel all alone’ (G IV, 374); after his wedding he expresses a ‘boundless desire’ to be alone, but seemingly with his bride (G IV, 956–9); and he twice refers to himself as the only child of his parents (G VI, 289; VII, 120).

47. See the important ideas of Dronke, P., Poetic individuality in the middle ages. New departures in poetry 1000–1150 (Oxford, 1970).Google Scholar

48. Kouyeas, S., IV (1912-13), 236–70.Google Scholar

49. ‘Antecedents’, p. 327.

50. See Konomis, N., XIX (1954), 249 Google Scholar; and Yangoullis, K. [Salonica, 1976], pp. 24–5).Google Scholar

51. Lusignan, F. E. de, Description de toute I’iste de Cypre (Paris, 1580), p. 221 Google Scholar; Kyriakidis, S., 2nd ed. (Athens, 1965), p. 82.Google Scholar

52. pp. 57–77.

53. See, for example, E 1280;

The juxtaposition of the colloquial and the learned does not fall into any observable formulaic pattern in the poem, and is nowhere repeated.

54. See Yangoullis,

55. op. cit., p. 66.

56. To this extent I am in agreement with Mavrogordato and Morgan. However it seems to me that this was a version prepared for oral transmission, and not the result of such transmission. Cf. the arguments of S. Alexiou, pp. 35–42.

57. The Cretan provenance of E was first suggested, on linguistic grounds, by Xanthoudidis, S. I [1912], 527–72)Google Scholar and is integral to Morgan’s argument that it was the product of oral transmission such as still exists in Crete. Kyriakidis, however, considered that its origin might have been Pontos or Cappadocia (‘Forschungs-bericht’), while S. Alexiou accepts that the dialect features in the text are Cretan. but concludes only that the scribe was a native of Crete (and not that E has any connection with the Cretan vernacular tradition).

58. Politis, L., ed. (Iraklion, 1964)Google Scholar, Act IV, Sc. 8; Vincent, A. L., ed. (Iraklion, 1981), Act III, Sc. 9.Google Scholar

59. Laourdas, V., [1786] (Iraklion, 1947), line 430.Google Scholar

60. Jeffreys, M., ‘The astrological prologue of Digenis Akritas’, B, XLI (1976), 375–97.Google Scholar

61. Pertusi, A., ‘Alcune note sull’epica bizantina’, Aevum, XXXVI (1962), 16–37.Google Scholar

62. Bouvier, B., (Athens, 1960), p. 87.Google Scholar

63. Lambros, S., Collection de romans grecs (Paris, 1880), pp. xcix–c.Google Scholar

64. As a common noun it is found today only in Crete and in parts of the Peloponnese where it is a technical term in carpentry (Dimitrakos, Lexikon, entry .

65. Beaton, , Folk poetry, pp. 165–6.Google Scholar

66. Petropoulos, D., I (Athens, 1958), p. 13.Google Scholar