H. Birnbaum & S. Vryonis. Aspects of the Balkans. Continuity and change (1969)

The Turkish conquest was instrumental in the transformation of a tradition of ballads, short narratives, and local feuding songs into epic poetry. Before the arrival of the Turks the ingredients were there, themselves the amalgam of at least two traditions, Slavic and Greek, stored up at the ancient crossroads between East and West. At the end of the 14th century among the Balkan peoples there was a very respectable tradition of ballads and short narrative songs, some telling of supernatural beings and others recalling ancient heroes, some of whom we no longer know. With the coming of the Turks, in the mountains of Bosnia, Hercegovina, northern Albania, and elsewhere in the peninsula, epic poetry lived again for a brief and final span in the Balkan peninsula.

This paper is in essence a study of the Moslem oral epic songs and their relationship to the Christian traditional songs in the Balkan peninsula. Oral traditional epic normally belongs to the dominant group. It is characteristically optimistic and its heroes are almost invariably victorious. It was, therefore, to be expected that the Slavs who were singing songs before the coming of Turkish rule and the culture it brought with it would continue to sing after its arrival, especially the group which became Moslem. Only now they would sing of Moslem heroes, of the Turkish Empire. Since it is almost axiomatic that protest songs lead only a precarious if noble existence in a totalitarian regime, if they exist at all, some might wonder not that the MOSLEM songs developed during the Turkish rule, but that CHRISTIAN songs continued to be sung, or even survived at all.

So much attention has been given to the anti-Turkish songs that it may seem a little hard to visualize oral epic and ballad in the Balkans before the arrival of the Turks. The songs of the hajduks and klephts have no substance or meaning without the Turks, and the Moslem coun-

terparts would not, of course, have existed without the hajduk songs. This whole genre depends on the Turks. The Kosovo 'cycle' is patently impossible without the Turks. In short, most of the historical or pseudo-historical oral narrative traditional songs are post-Turkish invasions. This is especially true of the Slavic Balkans, but it applies equally well to Greece, except that we can look further back in Greek tradition than we can in Slavic, namely to the events on the eastern borders of the Greek Empire in the early Middle Ages, which gave rise to the songs about Digenis Akritas. [1] These songs, as such, had practically disappeared by Turkish times, although their basic patterns remained. The songs of Digenis sung in the last century and to some extent even today have little if anything to do with the Digenis of the medieval ballads and epic. [2]

There was, nevertheless, a highly significant body of ballads and short narratives in oral tradition dealing with 'mythological' subjects. The term is generally used to indicate any songs concerned with non-historical people or with the supernatural. The category becomes a vague catch-all for those songs hard to classify in the more historical pigeon-holes. They have been somewhat downgraded as the interest in the historical rose to front rank. Yet I should like to suggest that they and not the historical songs are the central core of any oral narrative tradition, either ballad or epic, and that the historical songs are a later development. It is in just this respect, it seems to me, that the Moslem oral epic songs have been most valuable in the Balkans; for, I hope to show, they have preserved and developed some of the mythic patterns more elaborately than the Christian tradition, which turned its sights more and more to the hajduk and klephtic songs and hence to history. Although the Christian tradition did retain some of the mythic songs rather well, others were preserved with less and less enthusiasm, leaving aside more and more of the mythic elements. The richest of the mythic material in Christian tradition collected around the figures of Marko Kraljević among the Balkan Slavs and even Albanians, and to some extent also that of Miloš Obilić among the Serbs. Notable here are the songs of the slaying of dragons or dragon figures (in Greece attached to Digenis and Constantine) and of relations with female supernatural such as the vile or samodivi. Here too belong the justly famous stories of human sacrifice required for building a bridge or other edifices (The Bridge of Arta type),

1. See especially the Introduction to John Mavrogordato, Digenes Akrites, edited with an introduction, translation and commentary (Oxford, 1956).

2. See for example Lucy M. J. Garnett, Greek Folk Poesy, 2 vols. (Guilford, 1896). annotated translations from the whole cycle of Romaic folk-verse and folk-prose.

and some tales of unfaithful wives (as in "Banović Strahinja"). This kind of song one can rightly place at the core of Balkan oral ballad and epic tradition before the coming of the Turks. The 'mythological' narratives formed the main repertory of the singers, Slavic, Greek, Rumanian, or Albanian, at the moment when a group of them became Moslems and began the transformation of part of that repertory into something close to true epic. We should keep all this in mind as we examine a number of the aspects of Moslem oral epic in the Balkans.

First, looking at the geographical distribution of Moslem oral epic, we find it significant to note not only where it developed, but also where it did not develop, or, at least, where we do not know of its development. We find that it is, or was at one time, practiced from the Bihać area in northern Bosnia in an almost continuous region south to Janina in the Epirus. It seems to have shunned the Adriatic coast and thus to belong to the interior. I do not know of its practice east or north of Bosnia, but the Sandžak of Novi Pazar is still one of its strongholds. We do not seem to find it east of Albania in Yugoslav Macedonia or Bulgaria, in spite of the fact that there were and still are Moslems living in those regions, or in northern Greece east of the Epirus. We know that it has disappeared from some areas where it formerly not only existed but even flourished. Thus, while at the present time Moslem songs have vanished from central Montenegro (e.g. around Kolašin), we know that before 1912 Kolašin was famous for its singers, and the richness of the tradition in the Sandžak of Novi Pazar was increased by the influx of singers from there. The Milman Parry Collection at Harvard contains a mine of information about the pre-Balkan Wars tradition in and around Kolašin. The gap in the continuity at the present time in Montenegro is, therefore, a comparatively recent thing. In Serbia north of the Sandžak there used to be Moslems, but I know of no evidence that they ever sang Moslem oral epic songs.

From Bihać to Gacko, however, the two traditions, Moslem and Christian, have lived side by side throughout the centuries, and this seems also to have been the case in northern Albania ; but to the best of my knowledge, in the Epirus, as in Montenegro, the Moslem tradition disappeared with the passing of Turkish rule there. Austria was more tolerant, it would seem, than Montenegro, and hence the Moslems were able to keep their songs and their culture, as they did also in northern Albania. Yugoslavia, although at times discouraging Moslem singing, has also on the whole been tolerant of the old Moslem traditional epics, and so the Moslem singing continued after World War I not only in

former Austrian territory but also in the areas that were previously part of the Turkish Empire, specifically in the Sandžak of Novi Pazar. The geographical distribution of Moslem oral epic in the Balkans is thus a reflection of historical realities.

It is to be noted that this same area I have delineated as the area of Moslem songs is also that of the one-stringed bowed instrument known as the gusle. Or, at any rate, it is its area in the Balkans, because the gusle or a very similar instrument is found also in central Asia and in sub-Sahara Africa, and perhaps in other regions of which I have no reports. Epic poetry in the Balkans is sung to the accompaniment of three different musical instruments, none of which is, I believe, native there. It is also sung without accompaniment. The instruments are the gusle, known under the name of lahuta in northern Albania, the tambura in northern Bosnia, which is the same as the çifteli in northern Albania, namely a plucked instrument with two metal strings, and the g'dulka in Bulgaria, called the ćemane in parts of eastern Yugoslav Macedonia, and the lira in Greece.

The regions where the g'dulka (ćemane, lira) are used are also the areas in which narrative songs can be unaccompanied and where women as well as men sing them. Unaccompanied singing belongs to both men and women. Whereas the men sometimes use the g'dulka, the women seem never to use instrumental accompaniment. With the exception of the Epirus (and my knowledge here is not very certain) these are also areas in which there were no Moslem songs recorded. These are regions, too, where the songs are strophic in form, as compared to the stichic technique of the gusle tradition. One should point out as well that the lira, unaccompanied singing, and women singers are, or were, to be found on the Dalmatian coast and the islands of the Adriatic. In the 16th century we have the invaluable evidence of Petar Hektorović of Hvar in his Ribanje of unaccompanied singing in the 'Serbian manner' (sarbskim načinьm). [3] And the Matica hrvatska has published some very good songs from women singers from the islands. [4] In these areas were to be found refugees from the southeast. Finally, in Vuk Karadzic's day there were women singers in northern Serbia, [5] where, once again were to be found

3. Petar Hektorović, Ribanje i ribarsko prigovaranje, with an introduction by Ramiro Bujas (Zagreb: Jadranski institut Jugoslavenske akademije znanosti i umjetnosti, 1951), line 519.

4. For example Hrvatske narodne pjesme, Vol. II, Nos. 19 and 20, "Ženidba Marka Kraljevića", from Anica Begin of Luka on the island of Šipan. Her songs are in the collection of Glavić Baldo Melkov.

5. For example the blind Stefanija in Srem, but born in Serbia, four of whose songs he published (see the introduction to Srpske narodne pjesme, Vol. IV [Belgrade, 1958], p. xxi), or blind Živana who lived in Zemun, but was born in Serbia and had even been in Bulgaria, six of whose songs he published (Ibid. p. xviii).

people who had come at an earlier period from the south. In other words, the Moslem tradition of epic singing is surrounded by a markedly different, though clearly related, tradition of singing epic or, at any rate, narrative songs. There may be no connection between the type of singing tradition and the religious faith of the singers or the content of the songs. It may be only chance that where Moslem epic songs live, and perhaps have lived, men singers, the gusle, and a stichic arrangement of lines of ten or eleven syllables are characteristic; whereas, where the g'dulka type of instrument exists, or where there is unaccompanied singing, or where women as well as men sing, there is not, and perhaps never was, any Moslem oral epic tradition.

It is a curious fact that, while the gusle is the chief instrument for epic accompaniment in the Moslem epic area, and in the formerly Moslem area of Montenegro, from Bihać to Tiranë, another instrument, the tambura, makes its appearance in the northern and southern parts of that area, but nowhere else, that is to say, in northern Bosnia and northern Albania. It is to be noted also that the tambura is (or was) used exclusively by the Moslems in northern Bosnia, but it is used in northern Albania (or was in the late thirties) by the Christians and only for the short songs in eight syllables about more recent historical events. Whatever the case for the Albanian çifteli may be, it is demonstrable by formula study that the tambura in northern Bosnia is later than the gusle and that it has not been there long enough to remake the formula structure of the first half of the line to its own image. In spite of the intrusion of the tambura, the gusle is the typical instrument of Moslem oral epic song. [6]

gusle

stichic

men singers

sometimes long

Moslem and Christian

(epic)

g'dulka, lira, ćemane, unaccompanied

strophic

men and women singers

usually short

Christian

(ballad)

6. For another opinion see Alois Schmaus, Studije o krajinskoj epici (Zagreb, 1953).

There are in reality not many colleclions of Moslem songs from this whole area, [7] as compared to the number of collections of Christian songs, but they are of excellent quality. The oldest collection of Moslem songs that I know of is that of Vuk Vrčević, which still remains unpublished in the archives of the Serbian Academy of Arts and Sciences in Belgrade. The perface bears the date of March 1869, in Trebinje, Hercegovina. It contains thirty-seven songs. Next in time are the two volumes published in 1888 and 1889 by Kosta Hörmann in Sarajevo, containing seventy-five songs in all. In 1966 fifteen more songs from Hermann's unpublished manuscripts were edited with introduction and notes by Djenana Butorović and published by the Zemaljski muzej Bosne i Hercegovine u Sarajevu, thus bringing the published material from Hermann's collection to ninety songs. Hörmann's first two volumes are the only ones of Moslem songs to have reached more than one edition ; for they were given a second edition in 1933. In 1898 and 1899 the Matica hrvatska published two volumes of songs from the amazingly fine collection of Luka Marjanović made between 1886-1888; fifty songs were published in all, but 240 songs remain still unpublished in the manuscripts which are now in the archives of the Yugoslav Academy of Arts and Sciences in Zagreb. A splendid collection of twenty-three Moslem songs was made by Jovan Perović, a lawyer in Vienna, and acquired by the Serbian Academy in Belgrade in 1930, but it remains unpublished.

To the best of my knowledge, the most recent collection containing a large number of Moslem oral epic songs is the Milman Parry Collection at Harvard, begun by Professor Parry in 1933 and continued after his death in 1935 by myself and, since 1962, Professor David E. Bynum of Harvard. It now has approximately 1500 epic songs, both Moslem and Christian, collected between 1933 and 1967, and many thousands of lyric songs. In 1954, thirty-two songs, all from Moslem singers from Novi Pazar were published. Recently the Parry Collection has been enlarged by the addition of microfilms of the most important parts of the archive collections of the Serbian Academy in Belgrade and the Yugoslav Academy in Zagreb. These microfilms include all the Moslem collections in those archives and many, if not most, of the Christian collections as well. In 1937 I collected some one hundred epic texts in northern Albania, many of them Moslem, and these are now in the Parry Collection. We have, therefore, at Harvard at the present time the largest collection of published and unpublished Balkan Moslem oral epic in

7. See also Schmaus' survey of the collections and studies, op. cit., pp. 103ff.

existence. Only the unpublished materials in Sarajevo and in Albania are lacking. (The Parry Collection has also recently been greatly enriched by the acquisition of the James A. Notopoulos Collection of Greek oral poetry from 1953-1954, but none ofthat material is, of course, Moslem. I also added, with the help of the Bulgarian Academy, some Bulgarian songs in 1958 and 1959, but none Moslem.)

The major published collections of Moslem oral epic songs, therefore, are three, Hörmann, Marjanović, and Parry, covering songs recorded from the 1880's to the present. It should be noted that none of these collections as published is a 'popular' one, but that they are all learned publications intended for the scholarly world. Their contents trickled only slowly back to the people, if at all. Some small paperbacks of two or three songs, usually from Hermann's collection, were spread in the thirties into the villages. One of these is notable : it is a jekavski version of the ikavski text of the song of Smailagić Meho, published by Friedrich S. Krauss in Dubrovnik in 1886 under the title Smailagić Meho, pjesan naših Muhamedovaca. This may be the first published Moslem oral epic in the Balkans. Behind each of the three major collections is a varying number of still unpublished manuscripts, and in the archives yet others, most of which I have had an opportunity to see.

Beneath this pyramid of published and unpublished collections lie the countless songs of the tradition going back in time possibly to the 15th, probably to the 16th, and certainly to the 17th century. If this dating be true, then behind our collections stretches a period of perhaps three hundred years during which the Moslem songs, as such, came into being and reached a height of great development.

The Moslem songs have not been extensively studied. The collections, especially that of Marjanović, have useful introductions. Two monographs should be mentioned, that of Alois Schmaus for the Bosnian epic (Studije o krajinskoj epici, Jugoslavenska akademija znanosti i umjetnosti, Zagreb, 1953), and Stavro Skendi's for the Albanian epic (Albanian and South Slavic Oral Epic Poetry, Memoir of the American Folklore Society, No. 44, 1954). Schmaus's splendid work distinguishes the songs of the northern Bosnian border from those sung by the Moslem population of the southern districts, which he designates as the "Hercegovinian type", and he studies the structure of the songs of the north. He traces their spread into the southern areas and very interestingly observes the influence of the highly developed structure of the songs of the Krajina even on the Christian songs in those regions where the two traditions coexisted. Although I have collected, both in 1935 and in 1964, in the

Krajina and know the material there, the remarks in this paper are based on songs from the southern areas as well and possibly even more than on those from the north, which was covered so well by Schmaus. What I have to say should be looked upon as complementary to Schmaus's analyses and comments.

Schmaus divides the Christian epic into that of the "Dinaric zone" and that of the "peripheral zone". In the latter he includes Macedonia, western Bulgaria, Srem, Slavonia, and parts of Croatia and Dalmatia (pages 94-95). He says that in the "peripheral zone" the epic quickly degenerates (opada), becoming the property of blind singers and women, and abandoning the creation of new songs. If his implication is that the songs and singing tradition on the periphery are degenerate forms of the same tradition that one finds in the center, in what he calls the "Dinaric zone", I believe I must reluctantly take exception to his statement, or at least suggest a modification of it. In view of what I have said earlier about geographic distribution, it seems to me rather that one has on the 'periphery' in the south, i.e. in Macedonia and western Bulgaria, either simply an extension of the area of the ballads (Schmaus lists — p. 94 — the Bulgarian hajduk ballads, the Greek klephtic songs, and the short Albanian songs on more recent events) or a transitional region in which the two traditions, that of ballad and that of epic, or that of 'g'dulka' and that of 'gusle', meet and mingle. On the northern and western peripheries, however, it seems to me one meets with transplants of the 'balladic' or 'g'dulka' tradition, because of the movements of people at the time of the Turkish invasions, from south and east to those regions, and perhaps then a mingling of the two traditions.

One way of approaching the problem of the effect of the Turkish invasions on the oral epic of the Balkans would be to eliminate all the Turkish elements from it and to observe what remains, or, perhaps better, to observe what one has put aside. On the level of vocabulary and formula one could thus take away all the words in the formulas, and all the formulas, that are of Turkish origin. At 'stallion', and its compounds, dorat 'chestnut', djogat 'white horse', alat 'sorrel', would go, as would the 'Bedouin mare' bedevija. The horses that would remain would be zelenko 'gray', vranac 'black', malin 'strawberry roan', and šarac 'piebald'. Only one of the Turkish words for horse listed above has an equivalent in Slavic; djogat 'white horse' is equivalent to labud 'swan white'. In the Moslem songs the 'white horse' djogat is primarily that of Mujo of Kladuša, and he is clearly a very special, magic horse. The djogat is to

Mujo what Šarac, the piebald, is to Marko Kraljević. Mujo's horse is never labud 'swan white'.

Many elements in horse culture are expressed in Turkish words, suggesting that while horses may have been important in the epics before the arrival of the Turks, they grew in importance and in richness of description afterwards. For example, sedlo 'saddle' remains in use, but its adornments are Turkish. A particular form of stirrup, which is also a spur, bakračlija, makes its appearance. Actually, stremen 'stirrup' has a Turkish counterpart in uzendjija, which becomes much more frequent in the poetry than stremen, perhaps because it fits more easily than the two-syllable noun without an epithet into the metrical context. Other ornaments can be added to this list, but the principle is clear; in the description of horses and of equine equipment the Slavic epic was greatly enriched by the Turkish invasion. I presume that not only the words but also the articles themselves, when there are not doublets, came into the Slavic Balkans at the same time.

There are a number of words concerned with houses, their architecture and their furnishings, which are also of Turkish origin. One thinks of ćošak 'corner, drawing room', čardak 'porch, villa', kula 'tower', kapija 'gate', avlija 'courtyard', merdiven 'steps', minderluk 'dais, couch', pendžer 'window', džam 'glass', odaja 'room', šikli odaja 'a beautifully decorated room', and so forth. In the area of governmental administration and social hierarchy it is not surprising that innovation in vocabulary was widespread, as these were represented in the epic poetry. Sultan joined car for the top post, and matching kralj 'king', and knjaz 'prince', on the Christian side, for an older period of foreign governments, one finds the whole apparatus of vezir 'vizier', paša 'pasha', aga 'agha', spahija 'spahi', beg 'bey', beglerbeg 'bey of beys', dizdar 'fortress captain', and so forth. The list here is very long. Many people of differing social status would be included: kmet 'village chief', tefterdar 'secretary', kahvedžija 'coffee server', kiridžija 'caravan driver', surudžija 'outrider', and ćehaja 'representative of the vizier or pasha'. The whole world of Turkish administration and social structure was reflected in the Turkish words that the metrical formulas of the epic poetry absorbed.

Closely related here were the lexical units from the religious world, from Islam. The Moslem religious hierarchy is represented in the poetry, from šeh ul islam 'the high pontiff of Islam' to the hadžija 'pilgrim', hodža 'priest', mula 'religious teacher', and the mujezin 'the one who gives the call to prayer from the minaret'. The angels were joined by houries, hurije, and the manastir 'monastery', and crkva 'church', by džamija

'mosque'. To the liturgy, liturgija, and prayer, molitva, were added sabah 'dawn', 'the first call to prayer' (and it was combined with the Slavic zora 'dawn' to form a metrically useful formula sabah zora), ićindija 'the third call to prayer', akšam 'nightfall' (for which Serbo-Croatian seems to have no simple equivalent but used prvi mrak 'first dark' or zalac sunca 'sunset') and jacija 'the fifth call to prayer'. Molitva 'prayer' and moliti 'to pray' have their counterparts in the frequently used dova and dovu učiniti 'to make a prayer, to pray'.

This last expression is an example of an interesting series of half caique formulas composed of a Turkish word plus a Slavic translation of another Turkish word. The Turkish idiom consists of a noun such as dova 'prayer', above, plus the verb etmek 'to do or make'. The epic formula keeps the noun and translates the verb, in the proper person, number, and tense. Thus one has juris učinio 'he attacked', or surgun učinio 'he exiled'. These formulas are very helpful in the verse, because they provide a six syllable formula for a Slavic four syllable one; for example, juris učinio, instead of udario. The formula repertory was enriched and made more flexible by many of these doublets of Turkish words and expressions. Although the effect may have been most marked in the Moslem songs, it was also very considerable as well in the Christian poetry. The formulas became the property of the singers of both faiths, if they wished to use them. The epic formulas were thus adapted to reflect the cultural, social, religious, and political reality of their times. They grew in numbers and the epic style was enriched.

The enrichment was in more than vocabulary, however. Much of the ceremonial elegance and brilliance of the Balkan epic seems to have entered with the Turkish Moslem culture. For example, in the Christian poetry and still maintained by the less ornate practitioners of the Moslem songs, messages are sent by simply saying, as in Vuk Karadzic, Srpske narodne pjesme, Volume III, No. 46, "Kostreš harambaša": "Knjigy piše care od Stambola, / te je šalje ka turskoj Udbinji / na Turčina od Udbinje Zula" ('The sultan in Stambol writes a letter / and sends it to Turkish Udbina / to the Turk Zula of Udbina'). In many of the Moslem songs the writing becomes more of a ceremony, and the elaboration entails a number of Turkish words, indicating that it is closely connected with Turkish culture. Thus the writer calls his ćatip 'scribe', or, if he be a higher official, his ćatib-baša (ćatibaša) 'chief scribe', and he in turn takes divit i hartija 'writing desk and paper' (the first word is Arabic and the second Greek, but both probably stem from Byzantium-Stambol), and he puts jazija 'writing' (from Turkish jazmak 'to write') on the empty

paper. If the writer is the sultan the ceremony of the sending is increased. He seeks a Tartar messenger, whose name is frequently given, and the Tartar takes menzilske konje, or simply menzile 'post horses', and the departure and journey is punctuated by tatar vrisnu, surudžija pisnu 'the Tartar shouted, the outrider sounded'. The letter itself may be not only a knjiga but also a burjruntija 'decree', a ferman 'firman', a katal ferman 'a decree of execution', or (from others than the sultan) a mahzar 'petition'. The letter is written with ceremony, carried with ceremony, and received with ceremony. The continuity is apparent. Messages were sent in writing before the Turkish invasions and continued to be so sent afterwards as well. The change is in the elaboration and ritual ceremoniousness of the telling, which in turn reflects a reality of Turkish culture. The words may have been Persian or Arabic in origin, and the ceremony may owe much to the richness of Byzantium, but certainly they entered the Balkans first in Turkish dress. One could add numerous examples of this kind of expansion so common in the Moslem songs and sometimes imitated or borrowed in the Christian poems. It contributed greatly to the sense of epic breadth in the Moslem songs and in the Christian songs when it is used.

Two passages of elaborate description from Avdo Medjedović's 12,311 line version of "Smailagić Meho" (Parry text No. 6840) will illustrate my point. The first (lines 783-837) describes, in the words of the hero's uncle, the horse sent to young Meho by the sultan; the second (lines 1579-1676) tells part of the dressing of the hero by his mother.

Among the results of added ornamentation and detailed development of themes is perhaps the most obvious and striking characteristic of the Moslem songs, namely, their length. On the average they are considerably longer than the Christian. But this is a general statement; there are some exceptions. First, the Moslem raiding songs of the same local scope as the hajduk tradition, indeed, the other side of the coin, are short, of approximately the same length as the Christian songs of hajduks. It seems that in this kind of song the two traditions vary little either in length or content. Second, there are some Christian songs that are the same length or longer than some Moslem songs. For example, in the first volume of the Hörmann collection there are only three out of thirty-nine songs that are longer than the "Ženidba Maksima Crnojevića" 'The Wedding of Maksim Crnojević' in Vuk II, and there is one other that is only twelve lines shorter (No. XXVI, "Mustajbeg lički izbavlja svoga brata" 'Mustajbey of the Lika Rescues his Brother', 1214 lines compared to the 1226 lines of "Maksim Crnojević"). But "The Wedding of Maksim Crnojević" is one of a comparatively small number of exceptions. Moslem songs can frequently reach two to four thousand lines, and Christian songs do not reach that length. "Smailagić Meho" of which I spoke earlier as perhaps the first published Moslem song is 2,160 lines, and the longest version of it in the Parry Collection has over 12,000 lines.

As I said earlier, it seems to me that the core of an oral epic tradition consists of songs based on mythic patterns rather than on history. One of the most significant contributions of the Moslem oral epic songs to Balkan culture is its development of several of the oldest and most widespread mythic patterns to truly epic proportions. "The Wedding of Smailagić Meho" develops the mythic pattern of the coming of age and assuming of power of the young hero. In ancient Greece it is paralleled by the story of Telemachus in Homer's Odyssey on the heroic level and by the assumption of power by Zeus in Hesiod's Theogony on the mythic level itself. In Middle Greek it is to be found in the epic of Digenis Akri-tas. I know of no Christian oral epic in the Balkans that has developed this theme to the degree attained by "The Wedding of Smailagić Meho", especially in the long version of Avdo Medjedović. The revolt of youth against its parents is beautifully portrayed, and the investiture of young Meho with the command of the Border as well as his acquiring a

Mythic patterns are the basic structural frame of stories about supernatural beings, usually called gods. Since the stories were associated with ritual and ceremony and were probably themselves intended to fulfill magical functions, they and their patterns had a significance beyond that of fantastic entertainment or even history, although one might think of them as spiritual history or a kind of supra-history. As religions change and one set of gods or supernatural replaces another, the supernatural elements shift, men and women take the place of the gods, who are no longer worshipped, and an outward aspect of history appears. But beneath a pseudo-historical dress the skeleton of mythic pattern persists, for some sense of its significance still clings to it, keeping its parts together.

Of these mythic patterns one of the most clearly discernible and most studied is the vegetation pattern of death and resurrection, the dying god who returns. There are many variations of it but the main outlines are not hard to see. The god or hero disappears for a relatively long period of time and is seemingly dead, but eventually he returns, or is sought after and brought back. During his absence there has been devastation, but upon his re-establishment, which is performed ceremonially, order is restored, prosperity returns, and frequently he re-marries. I have expressed this in terms of a vegetation myth, because it seems to me that that fits best most of the examples of the pattern, and provides at least typical significance. It is the pattern PLUS SIGNIFICANCE, whatever that significance may be, that is essential to my argument, or to this view of a central core of stories in Balkan epic song and elsewhere.

Songs which tell of the return of the hero after long absence to find his wife about to marry again, the central theme of Homer's Odyssey, are very common in Balkan oral epic and especially well developed in the Moslem tradition. Two songs in the Vuk collection, both very famous, come immediately to mind: "Marko Kraljević i Mina od Kostura" (Vuk II, No. 61, 336 lines) and "Ropstvo Jankovića Stojana" (Vuk II, No. 25, 151 lines); the first concerns an evil suitor and the second an innocent suitor, but the wife is faithful in both. Both of these songs in Vuk are from unknown singers, that is to say we do not have any information as to who the singer was. It is immediately apparent that both of these songs are quite short. The greater length of "Marko i Mina" is in part due to the fact that it is really two songs in one, namely Marko's fighting with the Arabs, his withdrawal and eventual return to the fighting — all included in the midst of the return song proper, if I may

use that term. In the second volume of the Parry Collection, giving songs from Moslem singers in Novi Pazar, there are five clear return songs. The variety in length is great, from 684 lines to 1480, but it is clear that the Moslem songs are longer, even in the Novi Pazar tradition in 1934, than the Christian tradition in Vuk's day, a hundred years or more earlier, when the Christian tradition was at its height. In Appendix III of the Singer of Tales [8] I have analyzed twelve other return songs from the Parry Collection, eleven of which are from Moslem singers and one from a Christian singer, who, as a matter of fact, however, is singing a Moslem return song. To them should be added the "Ropstvo Jankovića Stojana" sung by Ćamil Kulenović in Bihać (he was from Kulen Vakuf), a Moslem singer singing a Christian return song.

Our material is, therefore, more abundant on the Moslem side than on the Christian. There may be more Moslem return songs than Christian. In what follows I shall undertake to compare a Moslem return song with a Christian one. Our models for analysis of the Christian songs is the "Ropstvo Jankovića Stojana" in Vuk III, and for the Moslem songs the "Ropstvo Dulić Ibrahima" in Parry I and II, the longest text sung by Salih Ugljanin. In both these songs we are told that the hero was captured soon after his wedding, Stojan a week afterward, and Ibrahim on his wedding night. It seems also that they both have a companion or companions: with Stojan is Smiljanić Ilija, and with Ibrahim are thirty Turks. When the hero is Velagić Selim, he frequently has a blood brother with him in prison, Alagić Alija. On the level of narrative technique, we can remark at this point that in the Moslem poem we are advised of the background by flashback, whereas in the Christian poem the telling is direct by the singer himself.

Ibrahim and his companions are in prison, but Stojan and Ilija were converted to Islam and given palaces. The Christians, therefore, do not escape, but actually run away, having first stolen the keys to the treasury and the stables, timing their departure to coincide with Friday, when the Turks are in the mosque. Since there is no indication in the poem of ill treatment or any news of troubles at home, Stojan's and Uija's running away is left without any motivation. By line 35 they are back home! Contrast the Moslem song. Uskok Radovan is captured and put into prison in Zadar. There he finds Ibrahim and Turks, and in an almost ritualistic question and answer scene, Radovan tells Ibrahim what has happened at home during his long absence, including the fact that his

wife is about to marry Halil Hrnjica. Ibrahim screams in prison, attracts the attention of the ban's wife, through whose intervention Ibrahim is freed to return home, provided he bring back with him the bridegroom-to-be Halil. Ibrahim is bathed, shaved, and fed, then sent home, and after adventures with border guards, finally arrives at his gate. This has occupied 510 lines!

Let us now look at the opening of our two songs from the point of view of mythic pattern. The hero is captured in both, but in one story he becomes a member of his captor's society, whereas in the other he does not; he is, rather, cast into prison. The idea of buried-yet-not-dead persists symbolically in the Moslem story more clearly than in the Christian song, because complete incorporation into the world of the dead is impossible in the myth of return. The really dead cannot come back. Moreover, in the story which dealt with supernaturals, the actual death of the immortal god was impossible. He could only retire or withdraw; he could not die, by definition.

Two highly significant elements are kept in both songs at this point. The hero was newly wed, and his absence was of a certain length. The seasonal element in the length of absence is too obvious to need much explanation. The number of years — which should, of course, be translated into months — is twelve, nine, twenty, twenty-four; these represent an annual event, or the period of gestation of the human, or a play on words between "dvanaest" and "dvadeset", or a doubling of twelve. In essence the sense is a long time, long enough for the purposes of the particular form of the myth which is operative. The essential idea of the newly wed hero is sexual. Combined here are elements of fertility and death. A parallel in harvest ritual comes to mind. The last sheaf of corn is buried in the earth, to represent both burial of the dead and burial of life-bringing seed. It is interesting that, whereas the Christian song of Janković Stojan had forgotten the burial symbolism, the life symbolism was kept. Yet it should be pointed out that in neither of these two examples is this element carried to its proper and logical conclusion, namely the son who should be waiting at home for his father's return, or who might have been captured during his father's absence and therefore needed to be rescued. It is found in other Moslem variants (e.g. the Return of Četić Osmanbeg), [9] but I do not recall any Christian

9. Serbocroatian Heroic Songs, Milman Parry and Albert B. Lord (Cambridge and Belgrade: Harvard University Press and Serbian Academy of Sciences), Vol. I (1954) and Vol. II (1953), No. 32. The son is also present in No. 20, "Sultan Sulejman Captures Budapest".

The loss of the prison symbolism in the Christian song has deprived the story of the motivation for the hero's running away, in spite of the fact that he had become a Moslem and been given a palace. It has also deprived the Christian song of the elaborate recalling of the hero from the dead. This is done in the Moslem song by means of a messenger (Radovan), a ritual questioning, the intervention of a female who is closely associated with the king of the underworld (if I may so express myself), bargaining as to conditions, and partial, but not complete revivification, enough to get home on. All this is followed by the journey across the border, back to the warm land of the living and to kin and wife. This is told at some length in the Moslem song and is preserved in more or less great detail in all the variants. The messenger's information triggers the rest of the action, and that information is of devastation at home that needs attention.

One of the famous ancient representatives of derivatives of this myth is, of course, the Odyssey of Homer, and it might be useful to identify, where we can, some of the mythic figures in that song. Clearly Hermes is the messenger who brings to the weeping Odysseus, or to his captor or detainer, the news that begins his release. Hermes as messenger of the gods, and more significantly here as psychopomp, is the obvious person for the task, since he goes regularly between the worlds of the living and of the dead. So, indeed, is Uskok Radovan, who as a renegade is a professional border-jumper. The female intervener at this point in the Odyssey is Athena, who stands as favorite daughter in a special relationship to Zeus, whose will is paramount. I am afraid we must call Calypso Odysseus' jailer! The ban's wife in Stojan's poem plays the role of the female intervener extremely well in Dulić Ibrahim's case. Thus many of the details of the Moslem song fall into meaningful places in the mythic pattern with great ease, but they are missing in Stojan's story. The car and the carica are in Stojan's song, to be sure, and, perhaps, without the mythic pattern the sultan would have found himself alone; and doors are unlocked by keys, perhaps symbolizing vestigially the dungeon doors. But there is no messenger, there is no ritual bargaining for release, and the horses and means of return are stolen rather than provided by the captor, as in the other cases. In this opening of the return song the Moslem story is closer to the mythic pattern than the Christian tale.

the return journey itself of the later Balkan hero, only in the Moslem song is some sort of journeying related. Some point is made in it of passing the border, of recognition at the border, of a struggle averted or won.

When, however, the hero reaches home and the problem of recognition is faced, the two traditions are a little closer together. If the rites of recalling from the dead have become lost in the Christian tradition, the rites of incorporation into the land of the living after the border has been passed have survived in greater detail in both the Christian and the Moslem traditions. Yet we should note that the Moslem song is much fuller in this section as well. In The Singer of Tales I have discussed in some detail this section of the return song in the Moslem variants and in comparison with the Odyssey, [10] but did not comment on the differences between the Moslem and Christian traditions among the South Slavs. The "Ropstvo Jankovića Stojana" is rather simple in its scenes of recognition. After saying goodbye to his friend Ilija, Stojan arrives at his own vineyard where he finds his mother and asks her if she has nobody to help her in her old age. She tells him about Stojan's captivity and about her daughter-in-law's coming remarriage. Stojan, without a word goes to his house where he finds the wedding guests, who receive him well. When he has had a drink, he asks permission to sing, and he sings the little song about the swallow which had built its nest for nine years and this morning was beginning to 'unbuild' it. But a falcon has come from the city of the sultan to stop her from destroying her nest. The wedding guests did not understand the point of the song, but Stojan's wife did. She ran to her sister-in-law and told her that her brother had returned. The sister went to see him, and, when she saw, she kissed him and wept. The wedding guests asked what was to happen to them and all the money they had spent for the wedding; Stojan gave them his sister and many gifts, and they departed. In the evening Stojan's mother returned and Stojan's wife went out to meet her and told her that her son had returned. When the old mother saw her son, she died, and they buried her gloriously.

As in the earlier part of the song, the Christian version here has kept some essential elements. For example, someone must die at the time of recognition. In South Slavic, both Moslem and Christian, it is usually the hero's mother. There is thus a kind of exchange of the living for the returned dead. (In the Odyssey it is Odysseus' dog who dies! Odysseus'

mother is already dead; he had met her in the underworld.) In the Moslem songs an animal is among those who recognize the returning master, often, indeed, the first to do so; the animal is generally the hero's horse, but may be his dog or dogs. Animals are often thought to have a sixth sense about living and dead people; they are aware of ghosts, even when humans are not. I suspect that the animal recognition is basically connected with the question of whether the returning hero is alive or dead. One sees something of this in the reaction of the dogs of Eumaeus in the Odyssey. When Athena appears to Odysseus only he and the dogs can see her in her beauty and full height, and at the sight of her the dogs did not bark "but cowered away, whimpering, to the other side of the shelter" (Book XVI, line 163, translated by Richmond Lattimore).

The mother in Stojan's tale plays more than one role, as a matter of fact. She is also the faithful retainer (Eumaeus and Eurycleia in the Odyssey) who is in many of the Moslem songs, as exemplified in Dulić Ibrahim's story, the first to meet the returning hero. But what is notably lacking in Stojan's story is the very characteristic element of the deceptive story, characteristic because it too is concerned with the problem of the hero's identity, of his being of the world of the dead or of that of the living. Another missing element of importance in the Christian song is the competition, the games in which the hero engages with the wedding guests. Here too is a part of the story that shows the identity and the vitality of the hero, and is part of the remarriage theme with its testing of the bridegroom.

In short throughout the two songs it is noticeable that the Moslem tradition has preserved and developed exactly those elements that are most associated with the meanings of the mythic pattern. On the other hand, it is clear that the Christian tradition of the song has gone part of the way toward losing the mythic symbolism, which, although still present, is weaker. Even in a song which is so strongly mythic in its pattern as that of the return, much of its significance has been lost, especially in the first part.

In the Christian song of "Marko and Mina of Kostur" there are more mythic elements than in "The Captivity of Janković Stojan". Marko Kraljević has attracted and kept more of the older patterns and symbols than any other Christian hero in the Slavic Balkans, and in this poem, as in several others, he is set forth as a faithful cooperative subject of the Turkish sultan. The version of "Marko and Mina" in Vuk II is an especially interesting one. The withdrawal of Marko in anger at the slandering of him by the people close to the sultan and his ultimate

return to battle and victory is much like the pattern of Achilles' withdrawal in the Iliad. Yet the setting of the whole poem is a return song, and the pattern would call at this point, as we have seen, for some imagery or symbolism of death which is not death. It may be that Marko's withdrawal and return corresponds to the imprisoning of Ibrahim in the Moslem songs, and thus this Marko song here preserves in an almost uncanny way a mythic element. Indeed, in the song Marko is sorely wounded in the battle with the Arabs and heals himself by drinking, restoring himself in this way to vitality as if he had been dead. In this version he is freed after he receives a letter telling what had happened at home, but there is no mention of a female intervener as in Ibrahim's story. Marko, however, returns in disguise, relates a deceptive story to Mina about Marko's death. But the recognition scenes are weak and muffled, although at least one detail is very reminiscent of the Odyssey. Marko's wife brings him his sword, just as Penelope provides Odysseus with the bow with which he kills the suitors. But, as I said, Marko the King's Son is a more mythic hero than any other Christian Slav in the Balkans, and he was fighting with the Turks against the Arabs and against a Greek opponent in Mina of Kostur.

There is one mythic pattern, a very basic one, that seems not to be represented in the Moslem song tradition. This is the story of the killing of the dragon. It is true that Djerdjelez Alija, in legend, was a killer of dragons, and in the Pešter region his fight with the dragon is associated with the names of certain villages. But this is legend, not epic song. The dragon slayers in the Slavic Balkans are Marko Kraljević and Miloš Obilić, and in Greece Digenis Akritas. They are the Heracles's of later Balkan myth.

In the song of "Marko Kraljević and Mina of Kostur" Marko plays a role very like that of Djerdjelez Alija in the Song of Baghdad (Parry No. 1), that is to say, he comes to save the sultan's army in the war with the Arabs. In the well known song of "Marko and Musa" (Vuk, II, No. 66) Marko also is the sultan's champion against the rebel Musa. Like Djerdjelez, too, Marko is slandered and plotted against in "Marko and Mina" by traitors close to the sultan; this is true of Djerdjelez in "The Seven Kings Seek the Head of Djerdjelez Alija". The two Moslem songs I have mentioned, namely that of the taking of Baghdad by Djerdjelez Alija and that of the seven kings, are both songs on mythic subjects; the first has connections with the Iliad and the Trojan cycle, the second is concerned with the death of a substitute, and hence also has a relationship with the Iliad. Once again we see that Marko Kraljević has a very

special place in the tradition of Christian oral epic in the Balkans viz-a-viz the Moslem songs, particularly in respect to those with a mythic background. This subject needs to be studied in much greater detail than is possible in this paper.

Thanks to the conservatism of the Moslemized bards of the Balkans, the oldest mythic patterns of oral epic in the peninsula were preserved, elaborated, and strengthened. A significant and meaningful core of stories was protected from the movements towards excessive historicizing that influenced so many of the Christian songs at the end of the 18th and the beginning of the 19th centuries. The Moslem singers developed a fullness and breadth of narrative style that was to leave a mark not only on their own songs but also on those of the Christian singers who were their neighbors. Profound as the changes were that came over epic poetry as well as many other facets of life in the Balkans with the advent of Turkish rule, the Moslem tradition of song was such that it assured a basic continuity with the deep past, even as it aided in evolving from age-old materials a poetic structure truly worthy of the name of epic.