The Collection and Analysis
of Oral Epic Tradition in South Slavic:

The Fragment from Gioia del Colle

(page two)

Following the editio princeps of “Lo Balzino” (Marti 1977), Pantić worked from the following lines in fifteenth-century Italian orthography, which cannot be taken as an intact text in any known or positable Slavic dialect:

Orauias natgradum smereuo nit core
nichiasce snime gouorithi nego Jamco
goiuoda gouorasce istmize molimtise
orle sidi maolonisce dastobogme
progouoru bigomte bratta zimaiu
pogi dosmederesche dasmole slauono
mo despostu damosposti istamice
smederesche Jacomi bopomoste
Jslaui dispot pusti Jsmederesche
tamice Jatechui napitati seruene
creucze turesche bellocatela vitesco

Adding something to this text in seventeen places (I show his additions below in boldface), subtracting something in eight places (I show his deletions in square brackets [ ]), and interpreting the orthography differently from place to place twenty times so as to standardize the text phonetically (I show such orthographic interpretations in italics), Professor Pantić reconstructed the text as follows, with ten lines instead of Rogeri’s twelve:

Orao se vijaše nad gradom Smederovom.
Nitkore ne ćaše s njime govorit[h]i,
nego Janko vojvoda govoraše iz tamnice:
„Molim ti se, orle, sidi ma[o]lo niže
da s tobome progovoru: Bogom te brat[t]a jimaju
pođi do smederevske gospode da s’ mole
slav[o]nomu despo[s]tu da m’ ot[s]pusti iz tamnice smederevske;
i ako mi Bog pomože i slavni despot pusti
iz smederevske tamnice, ja te ću[i] napitati
črvene krvce turečke, bel[l]oga tela viteškoga.‟

I translate:

An eagle circled over Smederevo city.
No one desired to speak to it
save only Yanko, leader of troops, who spoke to it
from (where he lay in) prison:
“I pray thee, eagle, descend a little lower,
so that I may talk to thee. I have thee (as my) brother;
get thee to the noble folk of Smederevo, let them beseech
the famous despot to set me free from Smederevo prison;
and if God aids me and the famous despot sets (me) free
from Smederevo prison, I shall feed thee
crimson Turkish blood, white flesh of mounted warriors.”

Pantić’s reconstruction is a great improvement upon the error-ridden original notation by Rogeri de Pacienza, who by his own admission knew no Slavic and understood not a word of what he had recorded. How he recorded the poem is unknown; whether by his own hand as it was sung, or from a dictation repeated after the actual singing, or with the help of some other literate person who perhaps understood more of this foreign language than did Rogeri. What is clearer is who sang the poem; de Pacienza uses the third person plural in his Italian description of the scene, and says moreover that the whole company of men and women, children and adults alike, sang the song together at the top of their voices as they danced (saltando como caprii girava et insiem tal parol cantava). Professor Pantić has certainly helped to clarify the sense of what they were singing as they danced, which Rogeri de Pacienza did not know at all.

In another aspect of the poem however, the fifteenth-century Italian poet, who was at least very accustomed to counting syllables, may have understood something about the little Slavic dance-song which Professor Pantić did not observe. In regard to their meaning, Rogeri’s division of the lines from one another is impossible as he recorded them; Pantić no doubt correctly divided all of the first four differently. Yet Rogeri heard a syllabic meter in the Slavic poem, and wrote it accordingly even though he had to divide the lines strangely in order to compensate for the absence of words which he had missed. For bad as they are linguistically, metrically Rogeri de Pacienza’s lines are quite regular: no line varies by more than one syllable of length more or less than the length of the line before or after it. Professor Pantić’s reconstruction is radically different in this respect:

verse 1:  13 syllables
verse 2:  13 syllables
verse 3:  14 syllables
verse 4:  13 syllables
verse 5:  12 syllables
verse 6:  13 syllables
verse 7:  12 syllables
verse 8:  12 syllables
verse 9:  12 syllables
verse 10: 13 syllables
verse 11: 13 syllables
verse 1:  14 syllables
verse 2:  12 syllables
verse 8:  15 syllables
verse 4:  12 syllables
verse 5:  16 syllables
verse 6:  10 syllables
verse 7:  18 syllables
verse 8:  15 syllables
verse 9:  15 syllables
verse 10: 17 syllables

A fragment of bugarštica?

For Professor Pantić, the explanation for this great metrical irregularity (and for much else too that is peculiar about the poem) is to say that it is a piece of bugarštica, i. e., a kind of South Slavic oral traditional epic of which several dozen manuscript and printed examples have survived associated with various dates between the mid-sixteenth and mid-eighteenth centuries. It is, however, a form of epic—if it is a form of epic—which no one has reported from any live singer since more than two hundred years ago, after which time it had apparently died out. Never before the recent productive attention to the Slavic debris in Rogeri de Pacienza’s unique autograph copy of “Lo Balzino” preserved in the municipal library in Perugia had any serious investigator suggested a connection between the bugarštica-tradition and Serbian epic singers, that entire tradition being known theretofore to have existed only on the eastern littoral of the Adriatic Sea.

But is the little dance-song from 1497 really an epic bugarštica? Two aspects of it speak decisively against any such notion: the one is metrical, as already observed (more concerning it hereafter), and the other is the fact that never in five hundred years, with hundreds of collectors collecting millions of lines from the tradition—never has there been a single other report of oral epos being sung by a group of men, women, and children whilst leaping about “like goats” (come caprii) in a strenuous dance such as that at Gioia del Colle. The very idea is ludicrous on its face, for it supposes what is self-evidently a physical impossibility. Short dance-songs of a few lines’ length have of course been reported in teeming abundance everywhere in the Slavic Balkans, not only where epic has been found but also where it has not, and there is a long-established name for this different genre, which, unlike epos, has indeed been universally known and sung by people of both sexes and of every age group that is able to dance. Such a song is called a poskočica in Serbian (meaning literally a “jumping” or “leaping” song, from the root skok-/skak-) and by other equivalent names in the other Balkan Slavic languages.

Nothing in the spectrum of human social kinetics is farther removed from the hopping, jumping, leaping, and consequent breathlessness of the South Slavic ring-dance in all its forms than the long-winded, quietly seated, leisurely singing and listening of the oral epic tradition. On that ground alone, Rogeri de Pacienza’s scrap of Slavic poetry is not, and never could have been, an epic. But even if it were metrically indistinguishable from epic, to call it epic merely for that reason would require us also to call Anacreon an epic poet just as Homer was because both composed in dactylic hexameters. This we obviously cannot do.

Having recognized that much, we are still left however with the formal issue that Professor Pantić has usefully posed: for if indeed it is not epic (which it certainly is not), is the little piece from Gioia del Colle nevertheless truly indistinguishable in its form from the epic prosody of the long-extinct bugarštica-type? This question brings us around once more to the metrical peculiarity already noted in Professor Pantić's reconstruction, and to certain other equally striking anomalies thereto related.

What is the bugarštica’s meter?

For the past hundred years, the academic custom has been to say of the bugarštica-meter that it displays certain tendencies toward regularity of syllable-count without, however, being entirely confined to those tendencies. In general, the earlier the date of such texts, the more frequent is the irregularity, and the later the date, the stricter the regularity. Some of the few surviving bugarštica texts are literary compositions, not oral traditional, but others afford good evidence of an erstwhile oral traditional prosody, and one can sufficiently recognize in them the characteristic metrical patterns of that extinct singing tradition.

The metrical regularity found in them consists in a hierarchy of features with a descending order of significance. The first and most consistently observed feature is division of the ‘line’ into hemistichs by a word-boundary falling approximately mid-way in it, a juncture which may accordingly be called a medial caesura, provided it be understood that the term does not necessarily imply any audible pause in delivery of the line as for inhalation, about which nothing is known. The first hemistich, the one before the medial caesura, when it is regular, is further divisible at a word boundary within it into two syllabically measured cola, either 4 + 3 or 4 + 4.

The second hemistich similarly, when it is regular, consists of either 4 + 4 or 5 + 3 ≅ 3 + 5. Thus, two different placements of word-boundaries were widely practiced in each hemistich, of which one was dominant and the other a recessive or secondary alternative in each half-line:

Dominant schema:   4 + 3 ‖ 4 + 4

Recessive schema:    4 + 3 ‖ 5 + 3 ≅ 3 + 5

So, for example,

(4 + 3)  Ma se bješe | žalostan ‖ s grešnom dušom | razd’jelio  (4 + 4)

So the miserable wretch gave up his sinful ghost

shows the dominant schemas in both hemistichs, while

(4 + 3)  Bez glave je | ostavi ‖ usred zelene | planine  (5 + 3)

He left her headless corpse in that green mountain wilderness

shows the dominant schema in the first hemistich, and the recessive one in the second hemistich. Essentially the same construction also produced the line

(4 + 3)  Kad je došla | mlađahna prid ‖ starca | despota Đurđa  (3 + 5)

But when the young woman came before the old man Despot George

The recessive schema in the first hemistich and the dominant one in the second hemistich abolish the usual asymmetry of the line, and yield lines in which the first and second hemistichs are completely interchangeable:

(4 + 4)  Otidoše | govoriti ‖ vrli Turci | Mostarani  (4 + 4)

The fearsome Turks from Mostar then began to speak

Finally, there is the infrequent but still often enough attested construction with the recessive schemas in both hemistichs:

(4 + 4)  Jutro rano | podranile ‖ budimske | mlade djevojke  (3 + 5)

The young maids of Buda rose early on the morn

Besides the foregoing metrical components of the bugarštice-form, there were also cadential refrains which seemingly only some of the bugarštice-singers added to their lines from time to time. These refrains did not conform metrically to any of the habitual schemas of either cola or hemistichs, but were—deliberately, indeed—sui generis. They varied in length between five and six syllables, and tended to occur (when they occurred at all) as a kind of pause after some even number of lines: by far most commonly two, occasionally four, rarely six or eight, and in a few instances at the end of ‘runs’ as long as twelve lines (if the manuscript evidence, which is all that exists, tells us anything like the truth of the actual singing, which it very well may not regarding such details as this). Examples are, closing a quatrain:

(4 + 3)  A sad mu je | od rana ‖ i bolesti | potamnjelo,  (4 + 4)

a punice moja (6)

But now it is grown ashen by reason of his sickness and his wounds,

O mother-in-law of mine

And closing a couplet (much the most usual pattern):

(4 + 3)  Vino da mi | popiješ, ‖ pehar da ti | na dar bude,  (4 + 4)

moj Šajnoviću (5)

So drink the wine and keep the cup as apophoreton,

my good man Shainovich

And closing a decastich:

(4 + 3)  Ali kralju | svijetli ‖ odgovori | sestri svojom,  (4 + 4)

on kralju budimski (6)

But the illustrious king replied to his sister,

he the king of Budim

It should be noted too that the number of syllables in the refrain was independent of the number of lines in the stanza which it closed. Of the two metrical varieties (five or six syllables), the six-syllable refrain was the more frequent.

‘Bad’ lines and the tendency toward metricity

Now in an oral epic tradition it was inevitable—since human beings made it and not automata—that singers sometimes composed ‘bad’ lines. In fact, the more fluently and rapidly a singer composed, the more certain it was that he would eventually produce atypical lines; and the more his verses were recorded, the larger the evidence of such ‘irregularity’ must loom in a pristine record. The process of dictation, which is slower than singing, reduced the total number of such deviations from habitual meter, but by no means eliminated them. In singing, a bard often simply aborted a bad line, leaving it unfinished. If he was aware of having spoken confusedly (for, having his mind concentrated upon what comes next in his story, he sometimes would not even notice), he might elect to make the line over again ‘correctly’ (i.e., in the habitual way), or else—it was truly unpredictable—he might simply leave it in its partly formed and imperfect state and pass on to the next thought.

But on the other hand he sometimes also formed unusual lines completely, quite as though there were nothing exceptional about them, and never so much as noticed their unconventional features. Common instances of this kind include both hypometricisms and hypermetricisms, when a singer unwittingly ‘omitted’ something (which, if confronted with the fact, he might genuinely believe and insist he had actually said) and so produced a ‘short’ line; or else he added something, often by conflating similar or related formulas into an unconventionally expanded line, which would typically contain all the expected metrical units and some surplus of others in addition. These, when they are fossilized by writing, become the mysterious ‘long’ lines that may puzzle metricians but never even entered the singer’s consciousness as somehow different from all his other lines. For traditional epic singing was a biological process, not capable of recursive inspection by those whose process it was, and like every other biological process it was not perfectly efficient and never completely conformed to rule. In fact it had no real rules, but only habitual tendencies, and these mere tendencies are all that we can properly invoke in speaking about the meter of an oral epic tradition.

Consequently there is no more certain indication of intervention in a oral traditional epic text by a literary editor, regardless of whether the text was dictated or actually recorded as sung, than the complete or near absence in it of prosodic irregularities, for there never was an oral traditional text of any length and substance that was not endowed with a certain share of such irregularities at birth. This is no less true of the bugarštice than of any other form of oral epos.

Once a singer in the bugarštice-tradition was fluent in making first and second hemistichs of both the dominant and recessive kinds, it was effortlessly easy for him by negligible inversions of word-order to anticipate in a first hemistich, for example, some part of what otherwise would be second-hemistichic phrasing, and to substitute in the place of an anticipated phrase that part of the first hemistich which the anticipated words would supplant. Thus, instead of

(4 + 3)  *I sa mnome | ni plinca ‖ nije veće | razdilijo,  (4 + 4)

Nor shared booty with me anymore

which would conform exactly to the dominant traditional metrical schemes in both half-lines, the line which we actually have in the sixteenth-century text is

(4 + 4)  I sa mnome | nije veće ‖ ni plinca | razdilijo,  (3 + 4)

Nor anymore with me shared booty

Such lines show us incidentally that syllable-count stood much higher in the hierarchy of conventional metrical tendencies than did the placement of accents (which was chiefly trochaic in this tradition, just as in the tradition of the shorter, ten-syllable epic lines).

If, in analyzing such a verse as this, one is guided by literary notions of metrics, it will seem highly irregular, for constructions of 4 + 4 ‖ 3 + 4 are obviously deviations from the metrical norms of the bugarštice; yet at the same time such lines are neither so utterly rare nor so artless as to be automatically attributable either to scribal error or to poetic incompetence in their makers. Many an analyst of the bugarštice has been driven by such appearances of metrical and other prosodic lawlessness to invent cabbalistically mysterious (and completely imaginary) invisible supra-segmental accentual forces as explanation for such lines, or else to abandon all faith in any governing prosodic forces whatever and to declare that the bugarštice were simply an early kind of free verse. Nothing could be farther from the truth. It is indeed very difficult—perhaps even impossible—by isolated cudgeling of one’s own brain to conceive of how such poetic mutations can happen. One must pass many attentive hours listening closely to how the oral epic bard actually makes his lines (before any editor gets at them) in order to recognize, for example, the powerful metrically refractive force of such habits as word-order inversion, which is very common in all forms of the South Slavic oral epos.

Another easy and (from the traditional point of view) perfectly ‘lawful’ process whereby the bugarštice-singers made ‘abnormal’ lines was by first composing a first hemistich and a second hemistich of the usual kinds to form a line of a common type, but then, rather than making a new first hemistich at the head of the next whole verse, enchaining instead an additional series of two or more further second hemistichs, with the single initial first hemistich of the ordinary kind thus made to stand as a sort of incipit to a whole couplet or more of multiple lines, of which all but the first line would appear to be ‘irregular’ in meter. Thus we find in the manuscripts such unusual “first hemistichs” (which would however be, and are in fact, perfectly ordinary second hemistichs) as the following:

(5 + 3)  više košulje | nosaše ‖ vezenu l’jepu | mahramu  (5 + 3)

Over the shirtdress wore a lovely ’broidered shawl

or again

(5 + 3)  i još mu ide | djevojka ‖ ove r’ječi | govoriti  (4 + 4)

These were the words the maiden said to him again

It is quite likely that the basic recessive type of first hemistich (4 + 4) arose in just this manner; and naturally too 4 + 4 is much more frequent than 5 + 3 (≅ 3 + 5) as an alternative to 4 + 3 in first hemistichs for the same reason that 4 + 4 is more frequent than 5 + 3 (≅ 3 + 5) also in the second hemistich itself.

Consequently, we are surely right to recognize a fundamental functional difference between the dominant schema (4 + 3) and the recessive schema (4 + 4) in first hemistichs: the shorter, seven-syllable schema, with its habitual heroic feminine caesura distinguishing it from all other types of hemistich, had the basic character of an incipit, while 4 + 4 and 5 + 3 or 3 + 5 were fundamentally mechanisms for the adding of more phrases to a poetic period that was already in progress.

The bugarštica meter defined

Academic confusion about the meter of the bugarštice-tradition has persisted for more than a century because the basic metrical unit of that tradition has heretofore always been supposed to be the entire bugarštice-line as written in the manuscripts. But as we have now seen, such confusion dissolves as soon as one recognizes the half-line, rather than the whole line, as the basic rhythmic determinant in the bugarštice, for all half-lines are formed by one sequence or another of only three simple, basic metrical components, two of which were obligatory and one optional. Those three components were, namely:

1.  an incipital meter:

— — — — | — — — ‖

2.  an octosyllabic continuative meter composed
of two either symmetric or asymmetric cola:

‖ — — — — | — — — —
‖ — — — — — | — — —
‖ — — — | — — — — —

3.  an optional desinent meter:

— | — — — — (—)
— — | — — — (—)
— — — | — — —
(—) — — — | — —
— | — — — | — —
— — | — — | — —

next page

previous page

Return to Topic

Return to Main Archive