1. The problem of the origins of Akritic songs is further complicated by our uncertainty as regards their relationship to the epic romance of Digenis Akritis, the oldest version of which (preserved in a 15th century ms. in Escorial) was composed, most likely in Cappadocia, in the first half of the 12th century; see Alexiou, S., ΒασίλειοςΔιγςήςΑκρίρης (Athens, 1985)Google Scholar, Introducton, p. 100 (and ed. minor, Athens 1990, p. 55)Google Scholar. It is recognized nowadays that the composition of the Digenis poem was based on a medieval oral-epic tradition that has to be extrapolated from such Akritic songs as The Song of Armouris and The Song of Andronikos'Son (recorded in 15th century mss., but also transmitted orally until quite recently), Byzantine narrative poetry in the vernacular, and the Digenis poem itself which, although probably created with the help of writing, displays all kinds of stylistic features - formulae, ‘repetition, orderly progression, recurrent details, mirrored scenes', a ‘tidy disposition of parts’ – showing that like other ancient and medieval epics it was composed ‘under the stylistic domination of an oral tradition', Fenik, B., Digenis: Epic and Popular Style in the Escorial Version (Herakleion-Rethymnon, 1991), p. 18Google Scholar. Whether the majority of Akritic songs are remnants of that tradition, as I think, or offshoots of the Digenis epic, as Alexiou suggests, preserved in either case within and by the mainstream of folksong tradition, seems difficult to decide at the present state of our knowledge. On the relationship of the Digenis poem with folksongs see Promponas, I. K., Άκριτικα (Athens, 1985Google Scholar; and Alexiou's, review in Έλληνικá 39 [1988], 189–195)Google Scholar, Sifakis, , Ariadne 5 (1989 - Festschrift S. Alexiou), 125–139Google Scholar; on the medieval Greek oral poetic tradition see Jeffreys, M. J., ‘The oral background of Byzantine popular poetry’, Oral Tradition 1/3 (1986), 504–547Google Scholar, and his many other contributions (many with E. M. Jeffreys) in , M. J. and Jeffreys, E. M., Popular Literature in Late Byzantium, Variorum Reprints (London, 1983)Google Scholar; Beaton, R., The Medieval Greek Romance (Cambridge, 1989), pp. 32, 40ffGoogle Scholar.

2. See now his collected papers on folksong, , τò δημοτικò τραγούδι, ed. by Kyriakidou-Nestoros, A. (Athens, 1978), pp. 169–207Google Scholar.

3. Although we have to admit that the ways a folk tradition uses to construct, deconstruct and reassemble themes, tales and stories (some of them real) into an ever-shifting, kaleidoscopic selfadjustment are very imperfectly understood, if at all.

4. On ancient and modern chelidonismata(swallow songs) and other ancient children's songs see now the masterful musicological treatment by the late Baud-Bovy, Samuel, Δοκί μιο για το ηλληνικά δημοτικό τραγούδι(Nauplion, 1984), pp. 2–12Google Scholar.

5. See his Greek Papyri. An Introduction (Oxford, 1968), p. 81Google Scholar.

6. This is inevitably too general a statement, and hence impossible to document properly. Conditions must have differed greatly from place to place, in different periods. Widespread literacy has been suggested only for the citizens of democratic classical Athens; but even there inhabitants of country districts, women and slaves were much less literate (Harvey, F. D., ‘Literacy in the Athenian Democracy’, REG 79 [1966], 585 ff.CrossRefGoogle Scholar; see also Turner, E. G., Athenian Books in the Fifth and Fourth Centuries “² 1977”Google Scholar, Woodbury, L., TAPA 106 [1976], 349ff)Google Scholar. For the very limited literacy of Sparta see Cartledge, P., JHS 98 (1978), 25 ffCrossRefGoogle Scholar. However, only the Egyptian papyri allow us to form a reliable picture. Turner, referring to the contents of some of the largest archives of papyri found in Egypt, writes: ‘Are we to accept the depressing conclusion that the ordinary man cared little for literature, even if he could read it? These archives are undoubtedly a useful corrective to common optimistic pictures of a widespread literary culture’ (op. cit., p. 78). But things are even more ‘depressing’ than this quotation suggests because the archives in question span a period of six hundred years (the Zenon Archive, 260–240 b.c, the Abinnaeus Archive, 340–350 a.d.) and belonged to men of the upper social strata (Zenon was a high ranking government official, Abinnaeus a cavalry officer with the Roman army, Aurelius Isidorus of Karanis was a landowner in the early fourth century who kept an archive of business documents which he could not even read himself). The extent of illiteracy among the middle and lower classes of the population (which however were multi-ethnic) is discussed by Youtie, H. C., ‘άγάμματος. An aspect of Greek society in Egypt’, HSCP 75 (1971), 161–176Google Scholar (- Scriptiunculae II, pp. 611–627); see also his other contributions in GRBS 12 (1971), 239–261Google Scholar (- Scriptiunculae II, pp. 629–651) and ZPE 19 (1975), 101–108Google Scholar (= Script. Poster. I. pp. 255–262). A balanced survey of literacy in ancient Greece (with adequate attention to the concept and degrees of literacy) is offered by Thomas, R., Oral Tradition and Written Record in Classical Athens (Cambridge, 1989), pp. 15–34CrossRefGoogle Scholar.

7. It should also be mentioned in this connexion that the transition from aoidos to rhapsoidos was not instantaneous but may have lasted for quite a long time. We lack direct and unequivocal evidence for this but, although what has happened in other traditions proves nothing for Greece, Radlov's extensive field research in Central Asia in the previous century is very instructive: it documented the distinction of shair, creative singers, as a special category among the bakshy, their mostly repetitive colleagues, in Uzbekistan; a similar distinction existed among the Kirghiz manaschi(singers of the 'Manas’ epic cycle) between the creative jomokchuand the repetitive yrchy, while the akyn of Kazakhstan corresponded with the former (Chadwick, N. K. and Zhirmunsky, V., Oral Epics of Central Asia, Cambridge, 1969, p. 327)Google Scholar. Creative and repetitive singers, therefore, coexisted in Central Asia, and their distinction does not seem to have been so clear-cut as the difference we assume between aoidoiand rhapsoidoi, who are supposed to have functioned during successive phases of the Greek epic tradition. I submit that long after the establishment of rhapsoidoi as the official keepers of Homer's poetic legacy oral singers, more repetitive perhaps than creative, must have kept the tradition of epic singing alive in a minor capacity, away from the pomp of public festivals and their rhapsodic performances.

8. Athen. 14.620b, Ach. Tat. 3.20, Petron. Sat. 59.3–7.

9. Neither the Renaissance nor the Scientific Revolution had been felt inside the borders of the Ottoman Empire; the effects of the Enlightenment were limited to a handful of Greeks and other subjects of the Ottomans who had studied or lived abroad; the conservatism of the Eastern Orthodox Church and Ottoman despotism combined to maintain the medieval character of society up to the eve of the Greek Revolution.

10. But this does not mean that all clergymen were literate; on illiterate abbots in Constantinople itself see Merkelbach, R., ‘Analphabetische Klostervorsteher in Konstantinopel und Chalkedon’, ZPE 39 (1980), 291–294Google Scholar. These clerics must have known by heart the innumerable services of an Orthodox monastery over which they had to preside.

11. See Eideneier, H., ‘Leser- oder Hörerkreis? Zur byzantinischen Dichtung in der Volkssprache’, Έλληνικά 34 (1982–1983), 119–150Google Scholar; idem, Kmsopateras, Neograeca medii ævi III (Köln, 1988); , E. M. and Jeffreys, M. J., ‘The Style of Byzantine Popular Poetry: Recent Work’, Okeanos: Essays presented to Ihor Ševčenko, Harvard Ukrainian Studies 7 (1983), 309–343Google Scholar; Panagiotakis, N. M.,' ‘Μηλητἠματα πηπι ∑αℵλίκη’, Kretika Chronika 27 (1987), 7–58Google Scholar.

12. It should be added, however, that such oral variants have had no effect on the textual tradition of these works as they originated in the listening to the reading aloud of printed texts. The advent (and spread) of typography eliminated completely the possibility of the contamination of printed works by the parallel oral tradition of the same works or works belonging to the same genre. For oral variants of works of Cretan literature see Vlastos, P., Kretikos Laos 1 (1909), 70 ff.Google Scholar; Doulgerakis, E., Kretika Chronika 10 (1956), 244ff.Google Scholar; Detorakis, Th., Άνέκδορα δημογιά τραγούδια τές Κρéτης (Herakleion, 1976), pp. 98ff.Google Scholar; Papadopoullos, Th., Kypriakai Spoudai 41 (1977), 211ff.Google Scholar; Puchner, W., Ariadne. Scientific Yearbook of the Fac. of Philosophy, Univ. of Crete 1 (1983), 173 ffGoogle Scholar; Alexiou, S., Erotokritos (Athens, 1980), Introduction, p. 100Google Scholar, Erophile (Athens, 1988), p. 77Google Scholar; Svoronos, N., Ariadne 5 (1989 – Festschrift Alexiou), 331–337Google Scholar. The influence of oral tradition on the works in question is examined by Holton, D., Byz. and Mod. Greek Studies 14 (1990) 186 ffCrossRefGoogle Scholar. ('Orality’ in medieval and modern Greek poetry is the subject of many other papers published in the same BMGSvolume).

13. Alexiou, S., Erotokritos (Athens, 1980), Introduction, p. 100Google Scholar.

14. According to the Naval Intelligence Division of the Admiralty, Geographical Handbook on Greece (1306), quoted (without date) by Harvey (see n. 6), ‘at the end of the 19th cent, it was estimated that in most regions of Greece 99% of the women could neither read nor write’ (p. 623). It is significant to note in this connexion that women are thought to be more important carriers of traditional culture than men: see Kyriakides, S., Αί γυναίκης ηίς τέν λαογραφίαν (Athens, 1920)Google Scholar.

15. Motifeme, formed by analogy to phoneme, was introduced to folklore studies by A. Dundes (from Pike, K. L., Language in Relation to a Unified Theory of Structure of Human Behaviour, 1954)Google Scholar as an alternative term to Propp's, V.function (Morphology of the Folktale, 1958)Google Scholar: ‘From etic to emic units in the structural study of folktales', American Journal of Folklore 75 (1962), 95–105CrossRefGoogle Scholar.

16. See, for instance, 1. Kakridis, Th., Homeric Researches (Lund, 1949)Google Scholar; Page, Denys, The Homeric Odyssey (Oxford, 1955)Google Scholar and Folktales in Homer's Odyssey (Cambridge, MA, 1972)Google Scholar. On a different (though not unrelated) plane, Tsagarakis, O., in his monograph on Nature and Background of Major Concepts of Divine Powers in Homer (Amsterdam, 1977)Google Scholar, has pointed out several parallel expressions in religious usage showing the longevity of certain concepts of popular religion.

17. Interest in this field of research has recently been rekindled by I. K. Promponas’ two-volume work on Homer and modern Greek folksong (quoted in Abbreviations). He collects an impressive number of parallels but his method is wanting in that he makes no adequate distinction between vocabulary items and longer or poetic expressions.

18. But if we spell κατáхλια (as the editors of the song do) we will transpose the notion of hotness from the tears to the spring to which they are compared. However, a ‘thoroughly warm’ (rather than hot) spring is hardly satisfactory as a poetic image (unless it is taken as a case of hypallage).

19. Lexicon in Bacchylidem, s.v.; cf. τανύφνλλον at Nonn. Dion. 20.250.

20. I should like to thank the editors of Greece &Rome for inviting me to participate in their worthy effort to demonstrate the relevance of classical studies to contemporary Europe with the publication of this celebratory volume of the journal. I am also grateful to Mark Edwards, Odysseus Tsagarakis, and Nasos Vagenas for their valuable comments on this paper, which have improved both its form and its content.