The Role of Music

by H. Wakefield Foster

“All poetry begins with oral tradition” (Foley 2002: 19-20). As a corollary, perhaps it is justifiable to add that all music, too, begins with oral tradition. Just as we distinguish between oral and textual traditions, we should recognize that musical expression has roots as old and well established as speech. Singing flourished as a heightened form of speech well before systems of musical notation developed, and music is no less protean than speech in its ability to change shape and elude our grasp.

But music’s elusive, ephemeral nature and the culturally embedded metaphor of music and poetry as distinct entities—Muses, or Milton’s “blest pair of Sirens, Voice and Verse” (1) —have contributed to the notion of epic poetry as consisting of two separate (but not equally considered) components. Cultural tradition encourages a paradigm in which music and poetry are discrete art forms that commingle but rarely to bring us the felicitous collaborations of musician and poet—Mozart and DaPonte—or poet and musician—Gilbert and Sullivan. Such models suggest an essential subservience of one field to the other. But, as A. M. Dale writes, “For the Greek lyric poet, Voice and Verse were not a pair of Sirens; Verse was merely the incomplete record of a single creation, Song (1969: 166).” The musical aspect of epic poetry has long been regarded as a “poor relation” whose centrality to the performance event is undervalued or discounted, although its close kinship is universally recognized. This chapter explores the role of music and its inextricable unity with verbal art in South Slavic oral epic.

Textualization has been a culprit in our neglect of the music of epic. The role of music in oral traditions is more difficult to examine closely, since writing developed far earlier than musical notation and more people are “literate” in reading language than in reading music. Text has always been the more palpable relic of epic narrative, although recitation can revivify dormant metrical rhythms—the perceptible verbal remnants of musical performance. But an essential component of South Slavic traditional epic is the music to which the guslar, or poet, composes and sings his tale. A thorough inquiry into oral traditional poetry should include a close inspection of its musical setting.

In an effort to understand more fully how musical accompaniment not only enhances but completes the guslar’s art, I examine and discuss aspects of Salih Ugljanin’s performance of The Captivity of Djulić Ibrahim (Ropstvo Djulić Ibrahima, SCHS 1, No. 4) and Béla Bartók’s methods of transcribing that performance (SCHS 1: 437-62). I comment also on Stephen Erdely’s transcription of Mujo Velić’s performance of The Wedding of Ograšić Alija (Ženidba Ograšović Ale, 1995: 55-226). Lastly, in order to cast more light upon the relationship between poetry and music in South Slavic epic, I offer a musical transcription and analysis of the first 101 lines—including gusle introduction—from Halil Bajgorić’s performance of The Wedding of Mustajbey’s Son Bećirbey. (2)

A difficulty for modern scholars with respect to the “Homeric Question” (3) has been that the Greeks apparently saw little reason to examine, analyze, or otherwise explain the poet’s process of musical expression in the performance of epic poetry. We have little evidence that might inform us of the Greek traditional poet’s musical performance practice within the bardic tradition of lyre or cithara players. (4) Ancient scholars, for example, left us scant critical apparatus with which to discuss the music of a rhapsode’s performance. As regards Homer, we are left to surmise an interplay between the bard’s poetry and his musical art.

The fieldwork of Milman Parry and Albert Lord in the 1930s and later, however, provides modern scholars with a quite extraordinary time capsule in the form of high-quality, aluminum-disc recordings of performances by traditional poet/singers in the former Yugoslavia. (5) What interests me here is the larger dimension of expression that the musical setting brings to the guslar’s performance. Metrical rhythm forms a temporal grid within which the singer composes his song, and melody provides the heightened voice that qualitatively distinguishes epic register from ordinary speech. By listening to the CD that accompanies the second edition (6) of Albert Lord’s The Singer of Tales (2000), one is impressed by the fluid technical skill of the guslari and the many practice hours and careful thought to musical accompaniment that such skill represents.

Bartók’s and Erdely’s transcriptions offer comparands for our discussion. In Bartók’s score, Salih Ugljanin’s vocal compass far exceeds the range of the single-stringed gusle. Although the gusle’s range in this performance is a sixth (G to E), Ugljanin’s voice spans an octave and a fifth (C to G') but remains predominantly within the interval of a fourth throughout the song. The melody’s narrow range, however, retains vitality by means of frequent embellishment and quick tempo (the quarter note varies from 138 to 165 beats per minute). Ugljanin’s gusle melody, first stated in the instrumental introduction, occupies the interval of three half steps from G sharp to C flat, a very narrow tessitura:

The melody expands upward through three half steps and contracts quickly to the initial pitch. In contrast, the voice melody entwines itself around B natural by moving up and down by half steps from B natural to C and back to B natural, then from B natural to A sharp and back to B natural, pausing on B flat for the fourth beat and ending on C:

Bartók’s score of Ugljanin’s performance is invaluable as a “textualized” representation of the South Slavic epic singer’s musical art, and it offers carefully wrought solutions to a set of unique problems. Ugljanin’s singing and gusle accompaniment contain stylistic aspects or idiosyncrasies that are not a normal part of the tradition of art music. For example, Ugljanin sometimes sings out of tune. His performance contains several pitch discrepancies that would be unacceptable in the arena of art music, whether by an individual or between members of an ensemble. Although such pitch variants might be called “quarter tones”, conventional musical notation does not normally record intervals smaller than a half step. (7) Bartók, however, transcribes pitches that are egregiously higher or lower than the expected pitch (8) by inserting arrows (↑↓) above notes to denote the direction (sharp or flat) of a pitch discrepancy. Other textualization problems inherent to transcribing nontraditional musical forms concern the choice of time signatures and tempo markings.

Time Signatures

For the guslar, a “word” may be what we term a colon, partial colon, single line, multiple lines, a scene, or an entire song (Foley 2002: 17-18). To define “word”, the guslar relies on various criteria that reflect the nature and requirements of his art. The guslar’s “word,” regardless of how large it may seem in contrast to our text-based notions, remains his basic unit of verbal expression. (9) In musical notation, time signature denotes the basic unit of time duration in a measure of music manuscript and describes how many units constitute the measure. Time signature symbolizes the relationship between note and measure, thus ordering the spacing of bar lines that in turn delineate rhythmic units. Any number of ratios may be expressed—2/4, 3/4, 1/16—but they are particularly valuable in suggesting a general realm of rhythm patterns and note lengths that the performer may expect to encounter. There are no time signatures in Bartók’s or Erdely’s transcriptions (excluding Ugljanin’s gusle solo) because, presumably, the guslar’s rhythmic and melodic phrasing has no precedent for those who would read the transcriptions. In other words, time signature broadly suggests musical form: 2/4 time might indicate a pavane, rondo, or march; but it precludes a galliard, minuet, or waltz. In his transcriptions of South Slavic epic song, Bartók faced the challenge of interpreting rhythmic groupings that had no archival body of traditional (art music) reference.

Neither Ugljanin’s nor Velić’s melody exhibits the symmetrical eight-measure phrasing that dominates European art music of the eighteenth and early nineteenth centuries. Melody here tends to remain well within the temporal boundary of ten to twelve quick beats. Bartók’s dotted (hypothetical) bar lines do not necessarily imply larger, overarching rhythmic units within a composition, nor do they presuppose symmetrical phrasing as in “Freude, schöner Götterfunken, Tochter aus Elysium. . .” However, on a practical level, one prefers to read musical notes that are separated by bar lines just as one prefers to read words separated by spaces.

Musical notation can easily represent the ten “beats” or syllables of the South Slavic epic line (deseterac) without recourse either to time signature or bar line. Despite Bartók’s use of dotted bar lines, Ugljanin’s verse is equivalent to three measures of 4/4 time, or twelve beats. Ten beats account for the ten-syllable verse; two remaining beats satisfy the end-line pause that the singer fills with gusle accompaniment. Stephen Erdely (1995: 11) dispenses with bar lines altogether, saying that they would “distort the metric formula since they imply accents of which the South Slavic epic ten-syllable is free”.

I am unsure what Erdely means by this comment. In Salih Ugljanin’s performance, one hears a recurring ictus on syllables one, five, and nine imposed by the colonic structure of the deseterac as well as frequent coincidence of word accent and metrical ictus that suggests Bartók’s three bars of 4/4 time. My perception here of recurring ictus is not influenced solely by Bartók’s transcription, however, since the audio recording conveys an audible ictus throughout. (10) Bartók’s score reproduces a preponderance of verses that consist of two and one-half measures of four aurally distinct rhythmic units (one syllable per unit) that correspond to the ten-syllable colonic division of 4-plus-6 (four syllables-caesura-six syllables), followed by two further beats of solo gusle at line end. It is true that some units are subdivided and contain triple rhythms or combinations of duple and triple rhythms (SCHS 1: 437, measures 3-7). However, the ninth syllable is very often quantitatively stressed, and syllables one and five are articulated respectively by line-end pause and word boundary between syllables four and five. The consistent presence of such articulating devices supports Bartók’s division of the ten-syllable line into three groups of four time-units each. This arrangement does not preclude other useful interpretations of time signature and does not suggest that word accents must coincide with metrical ictus, although a metrical configuration such as the South Slavic deseterac naturally exhibits a high rate of coincidence between musical and metrical ictus when it is set to three measures of 4/4 time.


Tempo, or the rate at which Velić and Ugljanin deliver the ten-syllable lines, varies throughout their performances. Velić has more sense for constancy of tempo than Ugljanin. But, in spite of the fact that they are performing different epics, their distinctive performances may be due more to musical idiolect than to their different songs. Bartók (1951: 89) refers to nonuniform tempo as a “rural characteristic”, a neutral description that avoids impugning the guslar’s technical command of the instrument. Perhaps the most revealing hallmark of the performance practice of late twentieth-century art music is its rigidity of tempo. Conductors and musicians alike demand steady metronomic tempi unless the score instructs otherwise. This flies in the face of nineteenth- and early twentieth-century performance practice, which cultivated highly individual performance styles, especially among pianists, violinists, and singers of art music. An often erroneous modern criticism of frequently varying tempi in art music recorded from the 1890s to the 1940s is that they are shaky or irregular because performers were rhythmically imprecise or unskilled at maintaining evenness of technique.

The authority of the musical score began to prevail over traditional performance practice early in the twentieth century. Igor Stravinsky crafted his compositions with exacting rhythmic detail and included frequent metronome markings in order to avoid what he considered misinterpretations by the performer. He strenuously voiced his desire that musicians “execute” or “present” his music rather than “perform” or “interpret” it (Stravinsky 1947: 132-33). His scores record mathematically precise note lengths and incorporate minute fractions of subdivision in an effort to supersede the musician’s role as interpreter of musical idiom. (11) Today the conductor alone, as he or she directs from the score (a textualization of the composer’s desired performance), is expected to provide what, if any, interpretation there may be. Subsequently, the score has taken precedence over performance; the conductor has taken precedence over performer.

Interestingly enough, in his transcription of Mujo Velić’s performance, Erdely eschews such controlling impulses. A comparison of his and Bartók’s transcriptions shows that Erdely (1995: 55ff.) approximates the guslar’s tempi to be between 132 and 138 (quarter note) beats per minute for the first 42 measures. From measure 42 to 232, he records the tempo as remaining steady at 104 beats per minute. By contrast, Bartók’s transcription contains ten tempo markings in the first 104 measures—eight are precise metronome markings; two are approximations (SCHS 1: 437-48). The most striking contrast, however, is between Bartók’s and Erdely’s rhythmic notation. The former strives for a mathematically precise textualization of actual note lengths; the latter provides a close approximation, a practice that effectively makes his score much easier to follow.

Bartók’s transcription is a masterful and stunningly accurate notation of the performance event. Nonetheless, it inevitably falls short of recording every aspect of Ugljanin’s art. For example, Ugljanin holds the second half of beat 3 in the third measure of the introductory gusle solo slightly longer than one half beat. Ugljanin’s exclamation “Ej!” at the start of the voce part (measure 1) is longer than one beat. Bartók’s approximations are logical and comprehensible; nevertheless, he deemed these particular rhythmic nuances unimportant to the accuracy of his transcription. Such fleeting, minuscule irregularities are personal stylistic aspects of Ugljanin’s art. Had he recorded the same song elsewhere, it would certainly be different within the larger context of his style.

There is a point beyond which we fail to reproduce completely any performance event. Bartók’s transcription is far more detailed than Erdely’s, who prefers instead to discuss broader aspects of performance such as melodic variation, general rhythmic formations, and motivic structure (1995: 29). This is not to say that Bartók’s transcription is overfilled with information or that Erdely’s approach is reductionist. Differences in their transcriptions are analogous to the variety of approaches to textualization of “the words” in the field of ethnopoetics.(12) Unfortunately, Erdely supplies no recording of his transcription as do the editors of the revised edition of Albert Lord’s The Singer of Tales.

Within the field of art music, tempo markings often carry stylistic implications. For example, tempo vivo and tempo agitato are each quick, but vivo implies a lightness of accent and articulation that would be inappropriate to the style of an agitato tempo. Drawing on the terminology of nineteenth-century pianists’ performance practice, Bartók uses the term parlando rubato to describe the guslar’s freedom of tempi (cf. Erdely 1995: 31). A fluid tempo that allows a measure of flexibility in the placement of notes is perhaps the performer’s most effective expressive tool. Parlando rubato (to be performed in the varied tempi of speech) describes the guslar’s ever-flexible rate of delivery that defies precise metronome markings. The Polish pianist Ignaz Paderewski, whose art epitomizes late nineteenth-century pianism, defines the term “rubato” thus:

There are in musical expression certain things which are vague and consequently cannot be defined; because they vary according to individuals; because a musical composition, printed or written, is, after all, a form, a mould: the performer infuses life into it, and he must be given a reasonable amount of liberty, he must be endowed with some discretional power. In our modern meaning discretional power is Tempo Rubato. Lexicons give the literal translation of it as “robbed, stolen”. We recognize that the very essence of Tempo Rubato is a certain disregard of the established properties of rhythm and rate of movement. The French translation mouvement dérobé is the best of all. It implies the idea of fleeing away from the strict value of the notes, evading metric discipline. (Cumming 1971: 109-10.)

Instinctively, the guslar exercises his discretional power to accelerate or decelerate delivery rate, although he undoubtedly has no understanding of rhythmic notation or the trained musician’s notion of metric discipline. Indeed, nothing in the guslar’s art relies on systems of codified textualization. Erdely defines tempo parlando as approximately 66-69 beats per minute (1995: 31). He adds that as Velić accelerates, the tempo fluctuates from 69 up to 102 and then to 132 beats per minute. With few exceptions (for example, tempo recitativo), a “standard” tempo in art music permits no such generalization of metronome markings. There is no constant tempo in the transcriptions. It is better, perhaps, to say that they breathe, as it were, with the developing plot.

In Greek Metre, An Introduction, D. S. Raven (1962: 21) writes, “There is no regular ‘tempo’ in Greek verse as there is in English.” This is a perplexing statement. Raven certainly means that Greek poetry consists of polyrhythmic verse—containing both duple and triple rhythms—whereas English poetry normally does not. The implication is that regular “tempo” requires an equidistant beat, or ictus, which is not the case. For example, because of substitution, many verses in Sophocles contain both triple and duple rhythms within the same metron. Thus, in addition to iambic (triple) rhythms, an iambic trimeter may contain spondees or dactyls (duple rhythms) that preclude an equidistant beat within the otherwise purely iambic verse.

Raven’s remark is relevant to our discussion of tempo in South Slavic epic because it reveals not only a misunderstanding of the meaning of musical tempo but also an unawareness of the interrelationship between tempo, ictus, and rhythm. Ictus reflects tempo and, at the same time, delineates rhythm patterns. Rather than being a function of rhythm, ictus is the aural device that extracts rhythmic shape from a succession of sounds; its placement depends entirely upon the metrical framework imposed by the performer or poet/composer. This is not to say that ictus must be audible at all times but that its location (dependent upon the verse’s metrical configuration — hexameter or deseterac, for example) has been internalized by both performer and listener, so that the verse’s rhythmic shape is always in the mind’s eye. By analogy, just as bar lines mark off rhythmic groupings for the eye, ictus provides the ear with rhythmic articulation. Guslari freely employ both triple and duple rhythms in the South Slavic epic (SCHS 1: 437). They instinctively manipulate tempo, (13) and their beat, or ictus, is frequently not the equidistant pulse that Raven expects.

In the arena of art music, nineteenth-century performance practice freely admitted fluctuating tempi. Rubato is nowadays a somewhat pejorative term used more often to criticize the occasional nonuniform tempi still accepted in musical theater. Notation, however, is unable to account fully for every tempo variation in art music. Bartók comments:

Some believe that the essential difference between art music and folk music is the continuous variability of folk music as against the rigid stability of art music. I agree with this statement but with the qualification that the difference is not one of contrast, but one of degree. . . . We must realize that even performances of the same work of art music by the same performer will never occur twice in absolutely the same way. . . . Although perfection cannot be attained in transcribing . . . folk music, we must always endeavor to approach an ideal of perfection. (Bartók 1951: 19-20)

This shortcoming of musical notation afflicts as well the canonized composers of art music. Extreme rhythmic flexibility and abrupt tempo fluctuations were idiomatic of Chopin’s performance of his own mazurkas. Sir Charles Hallé, a British musicologist and friend of Chopin, wrote in 1846:

I observed to Chopin that most of his mazurkas when played by himself appeared to be written not in 3/4 time but in 4/4 time, the result of his dwelling so much longer on the first note of the bar. He denied it strenuously, until I made him play one of them and counted audibly four in the bar, which fitted perfectly. He laughed and explained that it was the national character of the dance which created the oddity. The more remarkable fact was that listeners received the impression of a 3/4 rhythm whilst listening to 4/4 time. (Schonberg 1963: 144.)

Hallé marvels that Chopin’s audience (remarkably) received the impression of 3/4 rhythm, although they actually heard, he insists, 4/4 time. His insistence reveals a non-native listener’s unfamiliarity with the highly idiomatic (and untextualized) performance style of the mazurka, an eighteenth-century folk dance that originated in the Masuria region of northern Poland. Indeed, Chopin’s listeners heard a 3/4 rhythm that bore a style appropriate to the mazurka. Ironically, however, neither Hallé’s eyewitness account nor Chopin’s response to Hallé allows us to “hear” Chopin play; we are left, as before, with only the static textualization of his music: the score.

In spite of its shortcomings, musical notation remains a sophisticated and sensitive form of what Katherine O’Brien O’Keeffe calls the “performing text” (1998: 55-57). As we see from examination of the tempo instruction “rubato”, the musical manuscript offers directions for performance as well as help in understanding the performance. Foley writes that we can essentially learn to “reperform” oral poetry by employing methods of structural ethnopoetics and by striving to reperform not on our imposed terms but on the poetry’s inherent terms (2002: 107). Nevertheless, oral epic resides in the multisensory world of live performance and expects not to be read but to be heard and seen. Musical notation, when combined with the text, provides more clues to the nature of the essential “live” experience, though it cannot reproduce the experience itself.

The term “rubato” implies an inconsistency of tempo that allows rhythmic fluidity. Thus a score marked tempo rubato is not to be performed with mathematically precise rhythm. The nature and intensity of rhythmic flexibility, an essential component of the oral poet’s idiolect, is left to the artist’s discretion and musical sensitivity. A listener attuned to the traditional referentiality inherent in a particular musical genre, such as the mazurka, receives the impression of accurate style only when the performer steps beyond the limited instructions of musical notation and interprets the text. The guslar and his audience, steeped as they are in the traditional referentiality innate to the epic genre, bypass the middleman function of textualization. Bartók’s tempo marking of parlando rubato is our own encoded performance instruction that informs the reader of stylistic nuances that would assist in authenticating the performance experience.

Musical Transcription and Analysis

To examine more closely the interrelationship between poetry and music in the guslar’s art, I have transcribed a short excerpt (8 minutes and 25 seconds) from the performance by Halil Bajgorić of The Wedding of Mustajbey’s Son Bećirbey (Ženidba Bećirbega Mustajbegova) recorded acoustically onto aluminum disks in 1935 by Milman Parry and Albert Lord.(14) The challenge was to create a balanced textualization that would be precise and clear but also comprehensible and not overfilled with minutiae. Initially, my intention was to transcribe two or three short sections that particularly drew my attention. However, when I began transcribing from the top of the performance (after the gusle introduction), I soon became intrigued by Bajgorić’s musical canvas as it unrolled before my eyes and decided that it would be more revealing to go on from the beginning and widen the present vantage point. Consequently, I begin with Bajgorić’s gusle introduction and continue to the second break in the recording (after line 101). For the discussion that follows, I refer the reader to the transcription that appears elsewhere in this volume and as part of the E-companion at

The transcription reveals musical episodes skillfully woven together to form a cohesive and effective musical expression. (15) Although I listened frequently to the opening lines before beginning to transcribe, much of the singer’s musical adeptness escaped my notice until I could examine at length the textualized score of his performance. Bajgorić’s artistry includes the use of two-line symmetrical phrasing and emergent musical phraseology in which tonal/rhythmic variations evolve over several measures to become “new” sections.

When he sings, Bajgorić generally subordinates the gusle to unison (or nearly unison) accompaniment of the voice, except for brief end-line rhythmic motives that separate most lines of the poem. As did Stephen Erdely, I transcribe only the voice part and exclude gusle accompaniment. To avoid a cluttered appearance in the manuscript from numerous extra lines below or above the staff, I represent the music in treble (G) clef with the notation 8va basso; that is, Bajgorić’s actual pitches sound an octave lower. Margin numbers in the body of the song refer to lines of text. My choices for time signatures and subsequent bar line divisions are based on the South Slavic oral epic convention that places word break between syllables four and five and on the epic’s tendency for ictus to fall primarily on syllables one, five, and nine. I used also as a guide the fact that word accents in Serbian and Croatian fall predominantly on the first root syllables of words when those words occur in isolation from the poetic register. Therefore, my aim has been to represent the preponderance of coincidence between word accent and metrical ictus — thus my frequent alignment of measures’ initial downbeats with metrical positions one, five, and nine—without trampling on passages in which there is a clear conflict or tension between them as, for example, in line 17, Alijá.

Bajgorić accelerates through most of the excerpt, sometimes aggressively, and there are passages in which the tempo pulls back only to accelerate further afterwards, as in lines 7 or 11. As Lord notes (2000: 22), this is a common phenomenon: as the guslar warms up, he composes and performs more and more rapidly. I judged that it was most logical, therefore, to begin with eighth-note time signatures and move to quarter- and half-note time signatures as tempi accelerate.

I employ Béla Bartók’s tempo marking rubato parlando to account for Bajgorić’s rhythmic flexibility and frequent tempo changes. However, as demonstrated earlier, musical notation fails to account absolutely for all rhythmic aspects of the performer’s style; there is room for alternate interpretation. For example, in the 3/8 measure of line 3 in the song, Bajgorić leans to such an extent on the initial sixteenth-note triplet of the third beat (C sharp) that one could argue instead for a 4/8 measure (as in the second measure of line 4) in which he hurries through the fourth beat. Indeed, the second measures of lines 4-10 each consist of two duple rhythms. Elsewhere in the song, however, it is Bajgorić’s practice to transition more gently from one figure to another by first introducing a minute variation and then elaborating upon it. Debate over such questions subsides when we consider the rhythmic implications of the term “rubato”. Bartók’s rubato parlando effectively describes Bajgorić’s rhythmic freedom, the “rural characteristic” that places us firmly in the performance arena of South Slavic oral epic poetry.

Gusle Solo

The introductory gusle solo centers around B natural in the five-note “gapped” (16) scale A, B, C sharp, D, E; but I hesitate to establish the key as B minor, since there are no Fs, sharp or natural. F sharp and G natural (in brackets) do not occur in the transcribed excerpt; A and B natural occur only at the lower octave:

Because the scale’s tonal center is B and because C is sharped throughout, the tonality resembles the aolian natural minor key built on B natural. Therefore, I provide a key signature of one sharp (C sharp(17)) for the 27-measure gusle introduction and refer to subsequent gapped tonalities as “major” or “minor” with the aforesaid disclaimer. For the song, I supply accidentals throughout (sharps, flats, naturals), which does not imply the absence of a genuine tonal center.

The gusle introduction divides neatly into two sections. Each is in quick tempo (quarter note = 108-112 beats per minute). Section one, through the first beat of measure 11, begins with duple rhythms and ends with triple rhythms in a two-and-one-half measure ritardando (gradual slowing) and full stop after the tenuto (lengthened) eighth notes in measure 11. Section two (measure 12) begins a tempo (at the earlier tempo) with two sixteenth-note pickups. Because of audio distortion and crackling in measures 3-8, the smaller (cue) notes represent indistinct rhythms and pitches that are conjectural, although they seem to be expressed. The tonal center remains B minor. Bajgorić visits briefly with B major (repeated D sharps) in measure 12 before returning to B minor in measure 13 where D naturals are restored. Another lacuna follows from the second beat of measure 18 through measure 21. Bajgorić sets up his voice entry with an interesting vamp section in triple rhythms (measures 22-25). The vamp unwinds with two rallentando (decelerating) measures that cadence in B minor in measure 27. His voice entrance follows immediately on the fermata C sharp (prolonged indefinitely) in 28. Bajgorić “scoops” the C sharp; that is, he initiates the note at a lower pitch level and brings it somewhat higher during the course of the fermata. However, the fermata C sharp remains noticeably lower than those in the gusle solo. (18)

The Song

Bajgorić tends to compose musically in two-line phrases and often introduces an element of variation (in musical notes or rhythm) in the second phrase, which he elaborates in the next line to establish a new phrase in the succeeding line. For example, the opening four lines form a symmetrical pair of phrases:

The first four syllables of lines 1 and 2, Rano rani and vEj! Alija, each occupy the same pitches (line 1, D naturals; line 2, C sharps) and share the same rhythm. The remaining six syllables, Djerdelez Alija and careva gazija, are musically identical except for the anticipatory sixteenth-note passing tone (C sharp) in Djerdelez. In line 3, Bajgorić alters the third syllable by raising its pitch a half step to D natural. He then repeats and varies the passing-tone pattern from line 1 but descends a further half step (to C natural) and repeats the motive in the next beat while descending a further whole step (C natural to B flat). He cadences line 3 with a C natural to B flat, the same interval that he uses to cadence the next seven lines. In line 4, we begin to discern the emerging shape of Bajgorić’s dovetailed variations: he begins line 4 with identical rhythm and pitches that he used to cadence the preceding line (C natural quarter note, B flat eighth note) and retains the passing-tone motive and its varied repetitions, thus establishing a new rhythmic/melodic shape for subsequent lines.

From another perspective, we see different motivic threads that weave further designs into his musical fabric. By looking at lines 1-4 as a larger unit in itself, an overarching framework of descending pitch movement becomes apparent: line 1 begins and ends respectively with D natural and C sharp; line 2, C sharp-C sharp; line 3, C sharp-B flat; and line 4, C natural-B flat. Bajgorić is modulating from a tonal center of F sharp minor (lines 1 and 2) through transitional material (line 3) to a tonal center of G minor (lines 4-6 [see complete score at end of chapter]), returning briefly to F sharp minor in line 7, and back to G minor in 8-10. In line 11, he returns to the earlier F sharp minor tonality and restates a variant of the theme from lines 1 and 2. Lines 12 and 13 provide further variation.

Bajgorić provides an interesting unfolding effect in the transition from Alija’s awakening to the Readying the Hero’s Horse scene (lines 16-24):

Line 16, a proverb (“A bachelor has no maidservant”), is suddenly quicker (dotted quarter = 92), and for the first time in the song the singer introduces three consecutive beats of triple rhythms (6/8 and 3/8 measures). Proverbs traditionally serve as boundary or transitional markers in South Slavic epic (Foley 1994: 82). As an added signal, lines 17 and 18 subsequently consist entirely of triple rhythms. Poetic and musical devices are acting in unison to highlight the approach of new material in the narrative. Of particular interest is the melodic figure in the first measure of line 18. The second note is the first occurrence of an E natural in Bajgorić’s song and is a prominent expansion of his hitherto narrow range. The motive itself (sometimes altered rhythmically but always retaining the E natural) recurs throughout the excerpt, as it does here, to signal an approaching development or new material in the music or poetry. Line 19’s idiom, “Skoči momak na noge lagane” (“The young man jumped to his light feet”), serves the same purpose verbally and further reinforces the listener’s anticipation. Underscoring the transition, line 19 begins a new, faster tempo (quarter note = 126) that will carry Bajgorić into the Position change and Readying the Hero’s Horse scene that begin in line 20. The singer provides further dovetailing by expanding the eighth-note-and-two-sixteenths rhythm in line 19 (na noge lagane) to six sixteenths and one eighth note in 20 (nis kulu bijelu), then to eight sixteenths in 21 (u tople podrume) before contracting the figure to six sixteenths and one eighth note in 22 (konja kosatoga), and finally to four eighth notes in 23 (vod jasala), the new pattern for the subsequent two-line phrase in lines 23-24.

As Alija prepares to saddle his horse in 25-29, we hear wider intervals with more frequency: C sharp to E (25-27); C sharp to A (27); B flat to D (28); and A to D (29, the widest interval in the transcribed excerpt). Bajgorić treats his new melody from line 25 with two variations in 26 and 27. Line 25’s melody—subsequent statements are nearly identical to this version—is a variant of the signaling device first encountered in line 18. Curiously, it bears a close resemblance to a motive from Lerner and Loewe’s Broadway musical Camelot. To facilitate reference to Bajgorić’s signaling device, I will henceforth call it the “Camelot theme”. The new material it predicts is, in this instance, musical (19): line 29’s melody is reminiscent of line 19, except that here Bajgorić repeats the melody three times (at a quicker tempo) before initiating meaningful variation. With line 33, the process of motivic and rhythmic change at an increased tempo continues as line 33’s pitch variant (B flat for A natural) sounds again in 34, which is varied further by changes in rhythm and time signature. Restatement of the Camelot theme in 36 signals a new musical section and a significantly quicker tempo that begin in 37-38.

In the passage from lines 27 to 35, Bajgorić gradually contracts his range in the song. In 27 he spans a perfect fifth (A to E) and, after five lines in the wider range, retreats to the song’s earlier tessitura. In lines 29-32 he sings several perfect fourths (A to D), further contracting that interval to a major third in lines 33 and 34 (B flat to D); in line 35 he spans only a major second (C to D).

There is no end-line gusle accompaniment between lines 34-35, 36-37, and 38-40. In addition, from line 29 to 45 there are five distinct tempo changes, four of which are increases. Line 35 begins a musical transitional section (“Then he warmed up his mount with a cold snaffle-bit”) and serves more broadly to underscore a transition in the text from the Readying the Hero’s Horse scene that began in line 21 to Alija’s own dressing scene that begins in line 51. Bajgorić sounds almost breathless in line 42; he hurriedly “speaks” the last two syllables (represented by hollow, diamond-shaped notes (20)). The in-line (leonine) rhyme of kandžije and binjadžije may be playing a role here, or breathlessness may be an aspect of his performance; but more likely, he begins to tire and is out of breath. The lack of intervening gusle accompaniment between lines 34 and 35 and the new, duple-rhythm phrase that first appears in 39 (all quarter notes) and the increasingly agitated tempo of the entire passage combine musically to tell us that something important is about to happen.

Bajgorić swallows syllables in 43 (-vu) and 45 (gun-ju). He is probably catching his breath, but what draws more attention is that the gusle becomes louder. Perhaps he compensates for syllabic omissions by increasing the volume of the gusle accompaniment. However, I suspect that Bajgorić’s volume increase serves as an additional expressive tool to underscore the section from lines 43-49, which comprises a capsule of verbal as well as musical expression.

A simile in the text (lines 44-49, “Like a careless young shepherdess . . . ”) caps the Readying the Hero’s Horse scene. From line 43, voice and instrument are in unison. The tempo change (meno mosso, a bit slower) here marks the beginning of the transition to Alija’s dressing scene that begins in line 51. The simile summarizes Alija’s caparisoning of the horse and serves as a transition between epic themes in the poem. Bajgorić further highlights the transition by omitting end-line pauses in 43-46 and by altering the melodic material of line 48. The simile begins in 44 with an urgent chromatic melody (D natural-C sharp-C natural in measure 160) repeated in 45 and altered rhythmically in 46-47, culminating in line 49 with a restatement of the Camelot theme:

Thus, in the same way that he has used it before (lines 25 and 36), Bajgorić builds anticipation by inserting the Camelot theme to signal an approaching development in his narrative.

Bajgorić displaces verse ictus in 50-51 to assist in establishing the arrival of the new section. Line 50 begins with melodic material drawn from the transition (motive from mms. 157 and 174) but altered rhythmically so that ictus is shifted back onto the fourth syllable (-ga, line 50; se, line 51) instead of the fifth (na, 50; sr-, 51) where the second colon begins. Lexical ictus on syllable four is conventionally forbidden (Foley 1990: 96 and 105); but, as here, musical ictus may be superimposed upon the verse’s metrical frame:

The traditional caesura between syllables 4-5 is unaffected. By maintaining duple rhythms and extending duration of the initial melodic pitches (C sharp in 50, and B natural in 51) to the next quarter-note downbeat, Bajgorić causes musical ictus to overwhelm metrical ictus. Brief rhythmic tension resulting from the conflict between musical and verse ictus is an additional variational device for Bajgorić. To restore metrical ictus in 52, he keeps the initial half-note B natural but omits the repeated quarter-note pitch, so that the first colon of his newly emerging melody consists of five beats carrying four syllables. The new melody and rhythm predominate through line 57.

Line 58 initiates a new musical section. The transition begins with the C naturals that occupy most of the second colon in 57 and act as an extended chromatic passing tone to the new melody’s initial pitch, C sharp (the melody of the previous section began on B natural). The absence of end-line pause between lines 57 and 58 provides a sense of urgency that Bajgorić intensifies by doubling his syllabic delivery rate in the first colon of 59. Identical words occupy the first cola of lines 58 and 59, but Bajgorić compresses their duration from four beats in 58 to two beats in 59. Thus, without increasing the tempo, he quite effectively increases the song’s momentum. (21) Line 61 restates the Camelot theme and proceeds, again without end-line pause, to line 62. The melody in 63, underscoring a change of topic—shirts—and derived from material in the second colon of 62 (the intervals C natural–B flat–D natural), returns at a pitch level one half-step higher than its earlier statement in lines 52-57.

Bajgorić employed the same transitional devices earlier when going into Alija’s dressing scene (lines 50-51). Here, however, he further shapes the melody by using “mirroring” (22) —that is, measure 223 consists of two eighth notes and a quarter note followed by their rhythmic inversion, a quarter note and two eighth notes—in the second colon of 63 and passing tones in both cola of 64 before settling on the erstwhile final form of the melody in line 65:

Note the syncopated opening figure (line 65) that retains the original intervallic shape of the melody stated at the beginning of the dressing scene, although at the higher pitch level. Line 65’s melody acts as a transition to the new epic theme at 67, Alija’s arming. Lines 65-70 are set to nearly identical music. The first and last two syllables of each verse are syncopated. As the narrative moves from loops to guns, to dagger, and to chains (lines 70-75), Bajgorić continuously varies the musical accompaniment. He foregoes syncopation in the first colon of 71, increases syllabic delivery rates in both cola of line 72 by means of rhythmic variation, and omits end-line pauses before lines 73 and 75, thus lending the passage unity or a kind of integument.

The music of 76-77 is a restatement at a higher pitch level of the chromatic material that provided a transition earlier between the Readying the Hero’s Horse scene and the dressing scene (44-45). Bajgorić leans heavily on the downbeat of 77 (C sharp), shifts verse ictus backwards to the fourth syllable in line 78 (-ći’), and imposes an ictus on the seventh syllable (-če-) in the second colon of 78. From what we have seen earlier of Bajgorić’s musical development techniques, we should now expect a tempo increase and restatement of the Camelot theme. Syllabic compression in line 72 and the chromatic material of 76-77 simulate and anticipate a quicker tempo, but the actual increase does not occur until 78. Bajgorić begins his restatement of the Camelot theme in 79 but sings only the opening figure, C sharp to E natural. The infrequency of E naturals thus far in the song allows him to distill the Camelot theme to its most potent aural component, the C sharp-to-E natural interval, without weakening its ability to serve as a signaling device. The interval itself, in effect, operates as an encapsulated sêma, a compressed element of the singer’s musical/motivic idiolect that acts here as an aural representation of the signaling device (cf. Foley 1999a: 3-4).

A faster tempo at line 78 accompanies a rapid verbal cataloging of articles (vest, silver plates, and buttons) that enhance Alija’s heroic stature. Unison voice and gusle in 80-81 restate chromatic transitional material from 76-77. Tempo accelerates further in 82 with a syncopated variation of the transition motive. Bajgorić adds verbal/rhythmic variation by shifting ictus (as earlier, from syllable 5 to 4) and restoring it two lines later in 84. More variation in the transition motive follows in 86-92.

Lines 89-91 comprise a brief metanarrative aside to the audience that serves also to connect the present arming scene (57-99) to the scene that will begin in 101:

If things happened to him as in this image,
Whenever the young man got very thirsty,
Then let him drink wine from his reserve! (89-91)

With marvelously subtle musical skill, Bajgorić creates a rhythmic cadence for this comment. An incremental tempo increase in the fourth measure of line 88 masks an actual retardation in syllabic delivery rate. Compare line 87 (from the C sharp in measure 302— -ca na čekrku) to 88 (from the C-sharp half note in measure 307—rezervu nosi):

Bajgorić doubles note values of syllables 6-9 in 88 (pitches identical to those in line 87) and syncopates the initial articulation of the tenth syllable. (23) Without slowing tempo, he has rhythmically slowed the motive by lengthening note durations.

The excerpt’s final passage, from line 89, presents a compact summary of the principal musical elements and narrative devices that Bajgorić has used from the beginning to enhance the unfolding of his epic tale. The music is organized into sections that follow similar patterns of development by incorporating an established sequence of processes: a melody undergoes variation that dovetails into a transitional episode (frequently chromatic); two or more lines are then joined musically by omitting end-line pauses; the signaling (Camelot) theme sounds; ictus is briefly displaced and restored; tempo usually (but not always) increases; a proverb in the voice acts as a boundary marker signaling the end of one scene and the beginning of the next. The outline below of lines 89 to 101 summarizes in sequence the devices, both musical and poetic, with which Bajgorić develops his song:

Musical and Poetic elements/devices Line number
melody 89
rhythmic/tonal variation 90
chromatic transitional material 91-92
omission of end-line pauses 91-92, 93-94
signaling theme 95
ictus displacement and restoration 96, 97
tempo change (slower or faster) 97 (slower)
proverb or “idiom” 100 (“idiom”)
new musical or poetic material 101 (new epic theme)


We must learn to listen to all aspects of the guslar’s verbal art, including the musical signals that accompany and enrich his words. Halil Bajgorić has given much attention to the musical setting of his epic song. His music is neither a rigid frame over which orally composed verse has been attached, nor does it provide merely a generic, drone-like rhythmic and tonal support. The singer/poet relies on a melodic repertoire and a tool kit of variational devices with which he alters the musical landscape for expressive effect as he moves through his tale. Emergent thematic and rhythmic development, frequent tempo fluctuation, motivic and intervallic signaling devices, and conflict and resolution between musical and verse ictus are among the more vivid colors from Bajgorić’s musical palette.

The oral-compositional nature of South Slavic epic verse strongly suggests that its musical fabric, too, is a creation of the moment, crafted for the performance at hand and drawn from a broad, traditional repertoire of musical themes, motives, and devices. Rational interplay between words and music in the opening 101 lines of The Wedding of Mustajbey’s Son Bećirbey predicts that much poetic and musical artistry remains to be discovered within the Parry-Lord Collection and within South Slavic epic more generally. As an oral traditional epic poet/musician, Bajgorić relies on carefully formed organizational principles to weave voice and verse into a single creation, Song. (24)

Glossary of Symbols in Musical Transcription

= 108 Quarter note occupies one beat (tempo = 108 beats per minute).
= Denotes a change in time signature from one in which the dotted quarter equals one beat to a time signature in which the quarter note equals one beat.
"Fermata" is a prolongation of a note at the discretion of the performer.
sempre staccato Continue short articulation of notes (represented by dots over preceding notes) until advised.
meno mosso Slower tempo.
poco meno mosso A bit slower tempo.
Gliss. “Glissando” is a rapid slide down to the next note.
↓↑ Indicates sung pitch higher or lower than expected pitch.
Right-pointing horizontal arrow above the staff indicates an acceleration for the notes below.
piu mosso Much quicker tempo.
Subito Agitato An immediate change to an agitato tempo.
segue Continue without pause.
Diamond-shaped notes indicate pitch approximations for syllables that are spoken rather than sung.
5/4 time signature in which rhythms are grouped duple + triple.
3+2/4 5/4 time signature in which rhythms are grouped triple + duple.
Indistinct quarter note pitches.


  1. John Milton, At a Solemn Music, 1-2.
  2. See the transcription and English translation by John Miles Foley elsewhere in this volume. The entire performance can be heard (as well as read in the original and in English translation) in the E-companion to this book at
  3. In other words, was “Homer” the single poet to whom we should attribute the Odyssey, Iliad, and Homeric Hymns, or should his name signify a series of oral traditional poets through whom the poems were composed and transmitted? For a complete discussion of the “Homeric Question”, see Turner 1997: 123-45.
  4. Warren D. Anderson (1994: 27-57) provides a thorough account of what we know of a bardic tradition among the Greeks. See also West 1992: 34-35 and 52-53.
  5. In June of 1941, after Bartók had secured a research position at Columbia University, he expressed his great admiration for the Parry-Lord collection of recordings in a letter to his son living in Hungary: “. . . The reasons for the invitation involved ... the transcription and study of the incomparable material of Yugoslavian ethnic (epic) music. In fact, this is the reason that I came here . . . . Such material is found only here, nowhere else in the world, and this is precisely what I lacked so much there in Europe . . . .” (Lenoir 1997: 381).
  6. Attached inside the back cover of this volume is a 28-track CD (produced by Matthew Kay and Thomas Jenkins, designed and programmed by Ivan Audouin and Alexander Parker) that contains conversations with guslari and excerpts from their performances. Track 5 contains the gusle introduction and lines 1-12 of Salih Ugljanin’s The Captivity of Djulić Ibrahim, the performance under examination here.
  7. It is very difficult for our ears, trained as they are to hear half- and whole-step intervals, to distinguish between quarter tones and faulty tuning. In his orchestral composition Seven Studies on Themes of Paul Klee (Number 5, Arab Village), Gunther Schuller inserts asterisks above certain notes and provides alternate fingerings for them with which the oboist is to provide microtonal (raised or lowered) quarter-tone inflections. When my teacher, Raymond Weaver, performed the Schuller with the Houston Symphony Orchestra, a music critic (who had not familiarized himself with Schuller’s score) accused him in his review of playing out of tune. After Weaver telephoned the critic and informed him of the composer’s instructions, he received a private letter of apology.
  8. “Expected” pitches here do not include quarter tones but are those that proceed by whole and half steps from whatever standard pitch Ugljanin models as his tonal center. The particular pitch chosen as the “standard” is irrelevant as long as the resulting intervals are built accurately upon that pitch. The notion of a standard pitch is misleading and does not reflect the reality of pitch discrepancies. For example, in 1917 the American Federation of Musicians established A=440 (vibrations per second) as the official pitch for the United States. At the same time, English and Continental orchestras were playing at pitch levels between A=435 and A=452.
  9. It would be illuminating to learn what the guslari considered the smallest unit of music, whether a single interval, melodic phrase, or rhythm pattern. Bruno Nettl (1977: 141) writes that cultures have their own units of musical thought that must be carefully considered when we talk about transmission.
  10. Track 5 of the CD that accompanies Lord 2000 contains the gusle introduction and line 1-12 of Ugljanin's performance.
  11. See “Stravinsky and His Interpreters” (Milstein and Volkov 1990: 126-48). From a performer’s point of view, Milstein implies the counterproductive effect of Stravinsky’s hostility to musicians (p. 148): “Experience shows that great interpreters can make the best music even better. That is the peculiar nature of music. For better or worse, it is a performer’s art”.
  12. The field of ethnopoetics offers an innovative method of textualizing oral narrative that allows a reciter to reproduce more accurately sounds from the original “live” performance that relay stylistic or emotional expression. See Tedlock 1999: xxi-xlv, and DuBois 1998.
  13. Bartók and Erdely transcriptions, passim.
  14. Recorded June 13, 1935, in the Stolac region of Bosnia (see Kay 1995: 4 and 231). To say that this recording was acoustic is to distinguish it from written or dictated recordings; Parry and Lord’s equipment included the latest technology of the day—microphones—with which they recorded using the electrical process. See the transcription and English translation by John Miles Foley in this volume; the transcription, translation, and most importantly the sounds of the performance itself are available as part of the E-companion to this book at For a biographical sketch of the guslar Halil Bajgorić and excerpts from interviews with him, see the chapter entitled “Portrait of the Singer,” elsewhere in this volume, as well as Foley 1990: 45-48.
  15. There is an absence of musical stanza formation, however, which George Herzog posits as a characteristic of heroic epic poetry at large (1951: 62).
  16. Robert Nelson, composer, pianist, and professor of music theory at the University of Houston, suggests describing Bajgorić’s scales as “gapped”, since the missing pitches are implied in the harmonic structure; i.e., traditional dominant/tonic and subdominant/tonic cadential progressions. Indeed, such “implication” would mirror other aspects of the epic communication.
  17. Normally (that is, when it is not represented by a “gapped” scale), the key of B minor is notated with a key signature of two sharps (F# and C#).
  18. While he may simply be singing flat, he may also have deliberately lowered the pitch because of C sharp’s relationship to the song’s new tonality (F sharp minor). It is standard practice among singers, wind, and string players to “real temper” scales in order that pitches be tuned according to their relationship to the tonic. For example, in real-temperament tuning, because the C# in the gusle solo is the 2nd step of the B minor scale (a leading tone), it would be pitched incrementally higher than in the song, where it is the 5th in the F# minor scale, a tone that does not respond well to pitch modification because of its resonance in the overtone series. We would be mistaken to dismiss the possibility that “untrained” singers manipulate pitch in the same way, although their explanation likely would be simply that “it sounds better”. Of course, as I said, Bajgorić may be unintentionally flat; he does, after all, raise the pitch somewhat.
  19. Later, Bajgorić employs the Camelot theme to predict new poetic material (line 95). The signaling device relates to what Herzog calls a “dramatic line”, a musical feature that he attributes to all Yugoslav heroic epic songs (Herzog 1951: 62).
  20. For pitch approximations such as these that are more spoken than sung, I use Sprechstimme (also called Sprechgesang) notation developed by Arnold Schoenberg (1912) to represent the pitches of “sung” speech.
  21. This technique, widely used in jazz improvisation, is called “double-time” or, more technically, “diminution”. For a discussion of musical techniques shared by guslari and jazz musicians, see Foster 2004. The seemingly paradoxical logic behind the term double-time—which effectively halves the time that it takes to perform a passage—refers to a doubling of the note value signified by the lower number in the time signature, which describes the type of note that will occupy one beat. For example, a 4/4 time signature in which the quarter note occupies one beat would, in double-time, become a 2/2 time signature in which the half note occupies one beat.
  22. “Mirroring”, another term from jazz improvisation, refers to retrograde rhythmic or melodic variation, a musical device similar to chiasmus which, of course, exists in the verbal component as well (Foster 2004).
  23. The doubling of note values, “augmentation”, is common also to jazz improvisation, where it is sometimes called “half-time” (Foster 2004). Syncopation of the last note in line 88 simply adds further rhythmic variation to an already augmented musical motive. The final C in line 88 is actually one and one-half times the length of the final C in 87, but rhythmic lengthening denoted by the term augmentation remains in effect.
  24. I would like to thank John Miles Foley, who first suggested that I transcribe and analyze some of Bajgorić’s performance and who provided invaluable advice and encouragement during the entire project. I also wish to thank the following people for their expert observations and comments: Paula Eisenstein Baker, James Ronald Kerbow, Martin Langford, William McKenney, Nancy Molavi, Robert Nelson, David Salge, Larry Slezak, Margaret Vandiver (in memoriam), Michael White, and Amelia and David Winn.