Performance-based Commentary

This Commentary is meant as a reading and listening companion to Halil Bajgorić’s performance of the Ženidba Bećirbega Mustajbegova (hereafter HB’s ŽBM). Toward that end, it concentrates on sound, morphology, and traditional structure as well as lexicon, syntax, context, and translation. The primary source for the song it glosses is the digital audio tape (DAT) of Parry no. 6699, records 4594-4612, sung in Stolac on June 12, 1935 and initially transcribed by Nikola Vujnović (NV) (see further Kay 1995: 231). The DAT was produced from the archival aluminum records by David Elmer at the Milman Parry Collection, Harvard University. The secondary source, which comes into play on those rare occasions when the DAT is inaudible, is a magnetic, analog tape of the performance, also generated from the aluminum records. Readers are encouraged to become an aural audience by listening to the sound-file of the performance.

Every syllable of the ŽBM’s 1030 verse lines has been audited and transcribed from these two tiered sources and compared against NV’s original transcription. A table of the discrepancies between the performance and the transcription, which range from simple differences of dialect through lapsus linguae (a slip of the tongue) and lapsus calami (a slip of the pen: a simple writing error) to larger and more complex kinds of variation, is available in the section on Nikola Vujnović’s Resinging (NVR) elsewhere in this eEdition. Of particular interest are the moments when NV, himself a guslar with sufficient literacy skills to serve as Parry’s and Lord’s amanuensis as well as interviewer, effectively remakes HB’s song in his own traditional idiolect. I have labeled such intrusions lapsus auris because they result from HB’s song being passed through the filter of NV’s “singer’s ear”; the phenomenon can be compared to Anglo-Saxon scribes’ recomposing of traditional poetry even as they copied the vellum manuscripts (see O’Brien O’Keeffe 1990). The Commentary makes frequent reference to the informative contrast between what HB actually sang and what NV heard and wrote (that is, “resang”).

For the most part I have aimed at a literal and consistent translation into English, so that the reader can readily track the recurrency of traditional phraseology and narrative patterning. In those rare lines that demand a freer or augmented rendering for the sake of clarity, I have provided the literal sense in the note to the given line. As a convention, I have regularly translated the South Slavic historical present tense as a simple past tense in English; while a degree of vividness is lost in this conversion, there is a corresponding gain in straightforwardness and narrative transparency. Let me also stipulate here that, although I have consulted all available lexical resources, some translations must remain provisional. Because the South Slavic oral epic register is a complex weave of multiple dialects, anachronisms, and borrowings that as a whole often diverges radically from the more streamlined, “standard” language of literary and other printed sources, even the most copious and wide-ranging lexicons and dictionaries fall considerably short of complete coverage. When one adds the idiomatic meanings encoded in the epic register, which go far beyond denotation to the implications of traditional referentiality (see Foley 1991, 1995a, 2002: 109-24), the task of making acceptable sense of HB’s way of speaking presents an ongoing challenge.

Specific features noted in the Commentary include expressive strategies, rhythm and meter, HB’s epic idiolect, consonants inserted as hiatus bridges (“performatives”), other adjustments to phraseology, similes and other figures of speech, and explanations of cultural patterns, religious imagery, and tangible items from material culture. Lexicographical problems are addressed as they arise, with special attention to performance-related deflections of both South Slavic epic vocabulary and the many Turkicisms that characterize the singing register employed by HB and other guslari from his tradition. Occasional parallels in usage and expressive strategy are cited from another of HB’s performances, Parry no. 6703: Halil izbavlja Bojičić Aliju (Halil Rescues Bojičić Alija), recorded by dictation to NV in Stolac, on June 13, 1935 (see further Kay 1995: 232), hereafter abbreviated as 6703.

Separate sections on Music and Performatives are also available elsewhere in this eEdition, as is a selective digest of traditional elements and their implications (the Apparatus Fabulosus, hereafter AF) that recur either within this performance or in the South Slavic Moslem epic tradition at large or both.

Resources and reference codes:

B = SerboCroatian-English Dictionary, Morton Benson, with the collaboration of Biljana Šljivić-Šimšić, 3rd ed. (Cambridge: Cambridge University Press, 1994)

Š = Turcizmi u srpskohrvatskom jeziku, Abdulah Škaljić, četvrto izdanje (Sarajevo: Svjetlost, 1979)

V = Srpski rječnik (istumačen njemačkijem i latinskijem riječima), Vuk Stef. Karadžić (Beč, 1852; rpt. Beograd: Nolit, 1975)

MS = Rečnik srpskohrvatskoga književnog jezika, 6 vols. (Novi Sad: Matica Srpska; Zagreb: Matica Hrvatska, 1967-76)

SAN = Rečnik srpskohrvatskog književnog i narodnog jezika, 12 vols.- (Beograd: Institut za Srpskohrvatski Jezik, Srpska Akademija Nauka i Umetnosti, 1959-)

P = Pravopis srpskohrvatskoga književnog jezika (sa pravopisnim rečnikom) (Novi Sad: Matica Srpska and Zagreb: Matica Hrvatska, 1960).

[Unless otherwise attributed, all translations from South Slavic, Latin, German, and French lexical sources are mine.]

1-2, etc. Singers often use expletives like Oj! And Ej! as attention-getters and (what amounts to the same thing) rhetorical devices to indicate beginnings and emphases. Sometimes they are extrametrical, as in line 1, while at other times they constitute part of the basic decasyllabic structure, as in line 2. They can be approximated by translating them as “Hey!” or “Yes!”, but I choose to maintain the original words in order to stress their performative function as something other than ordinary lexemes. Compare the initiatory Hwæt! (“Lo!”) that opens Beowulf and other Old English oral traditional poems; see further Foley 1991: 214-23.

1. Note the relatively rare performative [w] that precedes the expletive Oj!, presumably to foster ease of articulation, as performatives do throughout the epic singing tradition. HB uses an unusual variety of these sounds ([v], [j], [h], [m], [n], [l], [w], and very rarely [nj], [d], and [s]), customarily to avoid intervocalic hiatus and the attendant glottal stop between words or between syllables in the same word (for comparison with hiatus in Homer, see Foley 1999a: 73-74, 85, 88). See further NVR and the section on Performatives elsewhere in this eEdition.

Here and throughout this performance it is crucial to recognize that although HB sings both 11- and 9-syllable lines, neither type is truly “long” (hypermetric) or “short” (hypometric). Rather the “extra” syllables occur outside the melodic and rhythmic frame of the line, while the “missing” syllables are actually vocal rests within that same frame. This phenomenon has major implications for the identity and dynamics of the poetic line, which is far more than an ordered sequence of lexical items (see further the section on Music in this eEdition). In addition to maintaining the basic integrity of the decasyllable as an expressive medium, the vocal rests are a species of the “right justification” that characterizes oral epic phraseology in South Slavic and ancient Greek (see further Foley 1990: 82-84, 96-106, 129-55, 178-96). Each such line is marked in the original-language text (* * for extrametrical syllables and ## for vocal rests) and commented upon in the note attached to the individual line. Lines with initial extrametrical syllables are as follows: 1, 12, 13, 77, 93, 148, 160, 223, 347, 526, 630, 692, 713, 773, 830, 847, 914, and 966. Lines with initial vocal rests of one or two syllables are as follows: 111, 212, 431, 641, 745, 854, 886, and 911; cf. 1001 (this last instance internal).

2. Unlike wOj! in line 1, vEj! is rhythmically and melodically part of the ten-syllable increment. See further lines 511 and 514, with notes.

4-7. A four-line capsule that memorably describes early morning and, like Homer’s “rosy-fingered dawn,” acts as an initiatory marker in the narrative, signaling not only “day” but more fundamentally the onset of a new narrative segment or episode. See further the AF.

5. HB sings savata while NV transcribes as sahata, restoring the expected form via lapsus auris. Disparities such as this are tabulated in NVR.

8-15. This is HB’s Coffee capsule. Cp. line 222-24 and see further the AF.

10. Here (with džev॒z॒u) and throughout his transcription NV uses underlining to indicate either uncertainty or his conviction that a form is somehow nonstandard.

11. HB sings kavu and NV aspirates > kahvu. Cf. line 222 (with note) as well as the note to line 249.

12-13. In both lines HB uses a performative plus run-up glide (*hI*) to lead into the initial sound of the first metrical element (jednu). Both instances are extrametrical, occurring before the metrical and musical pattern of the decasyllable. See further the note to line 1. Cp. line 223, with note.

13-14. HB sings ćejif- while NV transcribes as ćeif-; see line 224, where the same disparity occurs. Š gives ćeif as the first form of this Turkicism.

16. The proverbial observation that “A bachelor has no maidservant” acts as a boundary following the Coffee capsule. See further line 100, with note, and the AF.

19. See the note to line 484 and the AF.

20. HB devoices niz to nis before kulu, and NV does not restore the standard form. It is well to note that this deflection is a natural and regular change usually obscured by (print-centered) orthographical convention; NV thus is doing no more than faithfully reflecting what HB actually sang. On the idiomatic force of this Position change line, see the AF.

21-49. This is an occurrence of the widely attested typical scene of Readying the Hero’s Horse; see further Foley 1991: 67, 125-27; 1999a: 84, 128, 133, 300n33. See further the AF.

25. HB fronts the final sound in gori (< gore), apparently under the immediate influence of svali, which then becomes a partner in the common traditional pattern of in-line or leonine rhyme. NV does not restore the standard form, but does mark his awareness of the unusual form with underlining (gori॒). See further lines 194 and 207 (with note).

26. NV adds palatalization, hearing zlatalja for HB’s zlatala via lapsus auris.

27. NV first writes svog (“his”), then crosses out the second word and substitutes dok (“while”), reading “A dok dobra konja timarijo” (“And after he groomed [his] fine steed”) and reflecting what HB actually sang. Here lapsus auris could have yielded a slight refashioning of the line, in the process changing the line from a dependent to an independent unit (at least nominally, since the additive, paratactic structure of the epic register programmatically blurs that distinction).

29. HB sings djibretom (cf. djebre, Š and SAN) and NV does not restore the expected form.

30. This and seven additional occurrences of bači (72, 210, 640, 644, 645, 667, and 700; cf. also zabači at 37 and prebačijo at 453) instead of baci argue that the lexically nonstandard form is in fact a regular feature of HB’s traditional idiolect. NV transcribes consistently (except for line 210, where his baci probably amounts to lapsus calami) as bači.

34. Lit., “Then he tightened it so that he did not overbalance it.” Here (as sometimes elsewhere) NV transcribes by joining a proclitic to the next word; see further NVR.

35. A snaffle-bit is a restraining device consisting of two bars jointed at the center. HB sings djemo’, with initial palatalization and deletion of [m] before studenijem; NV deletes the palatalization and restores [m] via lapsus auris.

37. HB deletes the expected [n] from *Zlatnu and sings Zlat’u, an instance of lapsus linguae perhaps attributable to the influence of the acoustically similar vilicu in the preceding line or to the mirroring of either the acc. s. of the name of Zlata, Bećirbey’s betrothed (Zlatu, e.g., 262) or the dat. s. of zlato, the word for “gold” (zlatu, e.g., 459). See further the note to line 30.

40-42. On the traditional idea of a horse prancing without guidance from a rider, see the AF.

40. Here and throughout his recorded epic repertoire (but not in the register of speech used in his conversation with NV) HB pronounces sj as [š] rather than [sy], whether in this word (šede < sjede) or elsewhere. Since it thus amounts to a (regular) peculiarity of his singing dialect or idiolect, I transcribe the remaining instances below without further comment.

41. The semivowel [w], as here between the two elements in Po avliji, appears to be part of the general articulation of [o] or [u] before another vowel (compare vodu o’skočijo at line 137), and so I do not transcribe it as a performative. If, on the other hand, [u] is used initially as a run-up glide, I transcribe it as a full syllable, positioning it between asterisks to mark its extrametrical character. See, e.g., lines 630, 692, 713, 773, 847, 914, 966; compare also lines 130-31 and the appended note.

42. Here (twice) and at lines 895 and 896 HB sings prez (for the standard bez, which occurs nowhere in this performance), and NV transcribes in all four instances as prez without any indication of the nonstandard form. Cp. line 412, where HB sings brez, maintaining voicing but again with intrusive [r]; NV transcribes as brez on that occasion, with the underlining signaling the nonstandard form.

44-49. Most similes in the South Slavic epic tradition are a single verse or two in length, but here HB provides an extended comparison between a horse so proud and well-trained that it prances independently about the courtyard and a young shepherdess roaming the upland pasture clad in her hood and motley jacket and carrying a lunch her mother packed for the day’s nourishment. As in the Homeric epics, this simile memorably juxtaposes the world of heroic achievement and the domestic, bucolic world that knows little or nothing of battles and heroes. See further the AF.

44. HB sings piški, a difficult word that I take as a deflection of pišljiv (“valueless, insignificant”; therefore “careless”) through addition of the common adjectival suffix –ski to the root. The lack of agreement (one expects *piška) may be explained by HB’s reflex to preserve the original vowel in the second syllable of pišljiv(a), adjusted metri causa via apocope. Note that NV transcribes as the unpalatalized and uninflected pišliv.

46. HB handles numbers in a systematic fashion, reducing multiples of ten from -deset- to -des’ (dvades’ at 310, 509, 510, 563, 1019); trides’ at 81, 82; cf. the full forms at 710 [trideset] and 1028 [pedeset], where they fit metri causa. Numbers in the teens, on the other hand, are reduced from -naest to -n’es’ or -’es’ (dvan’es’ at 93, 94, 357, 614, 687; četer’es’ at 269, 544, 590, 865) and -n’ejes’ (petn’ejes’ at 303) or -najes’ (dvanajes’ at 320 and 395 [where it partners with bešlija to form a second-colon formula]). See espec. the note to line 544.

49. The palatalization of nje (< ne) seems to be due to the influence of the immediately preceding word joj, not at all an uncommon “leakage” of palatalization from one word to another (cf. back-palatalization in South Slavic, which proceeds in the opposite direction). Some instances of this phenomenon may be interpreted as simple lapsus linguae, while some appear to be built into the epic register as a natural phonological dynamic. See further the note to line 52, with note, where proximate phonological leakage may again be operative. HB adds initial [š] and sings šćerka; NV underlines the first sound (š॒ćerka) but does not restore the standard form ćerka.

51. HB apocopates haljinama > haljinam’ (metri causa) and NV mirrors his articulation.

52. Cf. 6703.170: “Dobro svlači a bolje oblači” (“She doffed her grand garments and donned grander ones”). The flexibility of the phrase allows for a movement from “poor” (slabo) to “grand” (dobro) in the present passage (ŽBM, line 52), perhaps more appropriate for the lone hero Djerdelez Alija; or for a progression from “grand” (dobro) to “grander” (bolje), perhaps more suitable for the maiden Zlata in 6703, who proceeds to put on masculine clothing to disguise herself as a hero. Note also the rare performative [d], which bridges the hiatus between dobro and oblači. It may be understood as lapsus linguae, under the influence of the initial [d] in dobro, but we should not miss its performative function; without this or another such bridging sound, HB would have to resort to a glottal stop, interrupting his sung vocalization. NV deletes the [d] and restores to oblači. Cf. line 1026, with note.

53-99. This is an occurrence of the typical scene of Arming the Hero; see further the AF and Foley 1991a: 67; 1999a: 15, 84-86, 94-98.

54. The expectable tkate becomes skate, under the local influence of skovate, its grammatical and syntactic mirror-image and rhyming partner, later in the line. With this phonological modification, they also match alliteratively. Such acoustic parallelism between cola is typical of the epic decasyllable, and is functional within the register as a compositional device and a key to audience reception (see further Foley 1999a: 83); occasionally, as here, the same tendency toward acoustic patterning leads to lapsus linguae.

55. HB simplifies *Već > Vet’, and NV does not restore the expected form.

56-57. In both lines HB sings djumišlije while NV transcribes as djurmišlije, with the r added above the line in both instances. Š lists djumišlije as the only possible form; NV seems to have added the excrescent [r] later, perhaps to agree with his own idiolectal pronunciation.

60. The noun varčevi is transcribed by NV as arčevi here, but as harčevi at 6703.377. Š gives the first form as harač, but also includes arač in the head entry. Whether we take the [h] as lexical or not, the fact remains that HB substitutes [v] as a hiatus bridge in the ŽBM. Such is the instability of an orthography that reflects traditional processes and involves archaic vocabulary that may be unclear to the transcriber (even to NV, who is also a guslar).

61. HB swallows the final [m], thus singing palco’ instead of palcom; NV restores the full form. Note the morphological match with zlatom at the end of the next line, an example of the not infrequent phenomenon of grammatically based rhyme, most often at end-line.

61-62. These two lines are sung as a couplet, with no musical break between them. On such features, see further the section on Music elsewhere in this volume.

63. HB sings vA, whereas NV transcribes as Pa (“Then”). This is the first instance of many such substitutions, due to lapsus auris and the special vulnerability of line-initial proclitics to idiolectal variation (the other instances are at 63, 69, 106, 107, 114, 115, 219, 382, 439, 443, 467, 485, 495, 515, 518, 520, 537, 553, 555, 557, 559, 609, 633, 634, 642, 709, 744, 783, 920, and 1012). This is to say nothing of the much more common shift of HB’s vA (HB) to NV’s A (“And, But”). All of these alternatives – as proclitics that serve to open the decasyllable line and thus occupy the most variable position – have about the same illocutionary force in epic performance. See further NVR. HB apocopates the dat. pl. plećima > pleći’ (metri causa plus swallowing of the [m]: plećima > plećim’ > pleći’) and NV does not restore either sound.

65. HB sings zlatane, adding a medial syllable metri causa to fill out the decasyllable. NV does not restore the standard zlatne, transcribing as zlatane without underlining.

66. A čelenka (Š) is “a type of plume, fashioned of gold or silver in the shape of linked feathers, sometimes decorated with precious stones. It is found on the front side of a [hero’s fur] cap or turban. It also served as a decoration for bravery.” Apparently Djerdelez Alija was such a notable hero that even his shirt was made of this unusual and precious material.

67-71. HB describes a single weapon-belt (mukademom pasom) in 67, but seven belts in 70-71. It seems best to imagine a single belt with seven loops or cinctures; thus the translation of these lines.

67. NV writes HB’s mukademom as two words, muka demom, perhaps on the basis of perceived morphemic constituents (the etymologically unrelated muka [“torment,” e.g.]).

68. Here and throughout his performance HB idiolectally clips the gen. pl. adjectival inflection -ih > -i, and NV mirrors his articulation without a diacritical signal.

69. See the note to line 63.

70. HB sings kajisera (< kajser, Š), apparently mixing this word with the phonologically similar (but etymologically unrelated) kaiš (“belt made of hide,” Š), a lapsus linguae that ekes out another syllable. NV mirrors HB’s misspeaking and does not restore the expected form.

72. See the note to line 30.

74-75. Here HB sings vandžar (74), with performative [v] displacing initial (and lexical) [h]. But compare line 75, where the standard form handžar- is retained, presumably because in the latter case a terminal consonant precedes the word, obviating the need for a hiatus bridge. This sequence is evidence for the situational nature of performatives as well as for the intermediate status of [h], which because its articulation involves a stoppage of voicing is often replaced by a performative. For further evidence, especially on the added difference between preceding front and back vowels, see lines 775 (handžarom), 897 (jandžara), and, more generally, the section on Performatives.

77. Note the relatively infrequent usage of [w] and [l] as performatives. The asterisks (*u*) designate an extrametrical element, a genuine syllable that temporally precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is a [u]-glide run-up to the initial sound in wOko. NV leaves the [u] untranscribed. See further the note to line 1. The use of performative [l] seems to be stimulated by the anticipated alliteration with ljuta.

78. On pleći’, see the note to line 63. Although none of the available lexical resources lists the hapax legomenon, I posit the compound fermen-čelebiju (“nobleman’s vest”) on the basis of its constituents fermen and čelebija (see further Š, MS on the two roots), with acc. s. inflection of the Turkicism resulting from the implied “he bore.” NV transcribes the initial element as ferman (“letter, proclamation”), a common term in the epic tradition that occurs much more often than fermen.

79. Toke are metal plates or buttons sewn onto the front part of an open vest, intended as ornamentation and sometimes as armored protection for the chest and thighs (Š).

80. An oka is an old measure of weight: 1.283 kg.; four okas thus equals about five kg. Here the expected partitive gen. pl. (oka; more common grammatically than nom. pl. oke) is deflected to voke by leonine rhyme with toke plus performative [v]. NV fails to correct and transcribes as oke. See further the notes to lines 46 and 544.

81. See the note to line 82.

82. HB sings Trides’, an appositive or terrace with the preceding line. NV transcribes Tri od, a clear case of lapsus auris (probably influenced by the structure of line 80). See further the notes to lines 46 and 544.

83. Lit., “So each button was [made of] kvarta,” where a kvart is defined (SAN) as “an imprecise measure of volume, weight, or length.” See further line 85, where kvart can be directly translated as “measure” without awkwardness.

84-85. The section “grocem, -- zlata” is inaudible on the DAT. Here and in all such cases below, I have restored the inaudible words and syllables from two sources: (a) the magnetic tape of 6699 and (b) NV’s transcription, made directly from the aluminum records but filtered through his own idiolect. See further NVR. HB sings pod grocem, with the expected instr. s., while NV transcribes as pod groce, ostensibly with acc. s. (though pod takes acc. only when motion is implied, as is not the case in this instance). Strictly speaking, lapsus auris leads to a grammatical error here, although such departures from conventional inflection are not rare in oral performances.

89. The curious rureta appears to be a deflection from sureta (= slika, or “picture, image,” Š).

93. The asterisks (*hI*) designate an extrametrical element, a genuine syllable that temporally precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is a [i]-glide run-up to the initial sound in Jedan. NV leaves the hI untranscribed. The excrescent [a] in kalapak (as against the expected kalpak) likewise does not make the line hypermetric, since the [a] functions like a grace note to the melody rather than as a fully fledged syllable in the verse. It may result from the sound-patterning in [a] throughout line 93. As noted above, this and other “long” and “short” lines are thus effectively decasyllables, since the vocal and instrumental patterning and structure remain constant, with the major syllables aligned with the notes and timing that constitute the usual rhythm and melody. Since such “extra” sounds (or their opposite, the vocal lacunae) are entirely performance-based phenomena, representing them as yielding a hypermetric or hypometric verse would be inappropriate; such a practice would amount to falsifying the performance-text. In this particular case, kalapak may be in part stimulated by the similar metrical and phonological shape of čelinaka at line-end. On “long” and “short” lines, see further the notes to lines 1 and 2 above; on “grace notes,” see also the notes to lines 106, 130-31, 525, 713, and 968. See further the notes to lines 46 and 544.

94. See further the notes to lines 46 and 544.

96. Note the infrequent usage of [l] as a performative, perhaps motivated here by anticipation of the medial sound in nearby palu.

100. The proverbial observation that “A bachelor has no maidservant” serves as a boundary line following the typical scene of Arming the Hero. See further line 16, with note, and the AF.

102. HB palatalizes and sings Naponaše; NV restores the expected unpalatalized [s] but divides the word: Na ponase.

103. A good example of how performance influences metrical realities: HB sings sv’ odaje, while NV transcribes hypermetrically as sve odaje (typically without noting any intervocalic performative).

104. In a rare case of apparently unmotivated lapsus calami, NV transcribes HB’s momak as mamak. NV’s transcription of HB’s vavliju (acc. s.) as avlije (acc. pl.) seems to be stimulated by the tendency toward end-colon, leonine rhyme (here with sidje).

106. HB sings vA vone, with performatives and a grace note that does not make the line hypermetric, while NV transcribes as Pa on, restoring the expected form of the pronoun but modifying the A (“And, But”) to Pa (“Then”); see further the note to line 63. As a consistent idiolectal feature, HB palatalizes to pušća (< puštati), and NV does not restore the unpalatalized form. HB varies the formula “two spirited greyhounds” (dva hrta zelena; 106, 114, 203, 207) with “two spirited wolves” (dva zelena vuka, 169), its structural and expressive equivalent; see further the AF. The difference in word order stems from the difference in accent (cf. P): hrta carries a short falling tone and thus follows the SBL (shorter-before-longer) rule of decasyllable phraseology, while vuka bears a long falling tone and can optionally migrate to line-end under the IAD (initially accented disyllable) rule. On these and other metrical placement rules, see further Foley 1990: 177-98.

107. See the note to line 63.

108. HB sings jama for the expected jami; NV transcribes as jama, with the underlining indicating uncertainty or a nonstandard form (but querying the root vowel rather than the odd inflection).

111. The code ## marks an initial vocal rest of one syllable. As explained above (note to line 1; also 93), this is not a metrical flaw and therefore not a hypometric line. Rather HB pauses vocally and picks up his singing one position (that is, one syllable-beat) later in the melody. Interestingly, NV transcribed this pause as Pa, which HB did not sing, presumably in harmony with the preceding line: both begin Pa ondaka in his transcription, an emendation encouraged by their overall syntactic and lexical parallelism and the common poetic strategy of pleonasm or terracing (effected here via lapsus auris). More evidence that HB himself was implying Pa, even though he did not articulate it, may be derived from the performative that begins “vondaka,” which indicates that the singer senses the need of a hiatus bridge to eliminate a glottal stop and smooth the vocal transition from the “missing” Pa to the explicit ondaka.

112. Here and elsewhere (e.g., 117) HB’s articulation of the masculine past participle in –ijo at line-end approaches –ija ([o] > [a]). Throughout the performance HB shows a general tendency to interchange [o] and [a].

114. On the “two spirited greyhounds/wolves,” see the note to line 106 and the AF. See also the note to line 63.

116. Within šever HB shifts internal [h] to performative [v], apparently on the grounds that [h] does not sufficiently bridge what he takes as word-internal hiatus, while NV restores the standard form šeher via lapsus auris. See further the note to line 174. HB apparently construes Sarajevo as acc. s., while NV recomposes via lapsus auris as Sarajevu (dat. s., with no corresponding inflection of šeher within the colonic formula or single “word”), an equally viable but slightly variant construction.

117. See the note to line 112.

118. I construe the initial participle from ustitrati (“to dance,” MS: zaigrati), a common enough description applied to a rider’s control of his or her horse; cp. vigrajući at lines 440-41. HB sings vUstipraći and NV transcribes as Ustiprać॒i, failing to restore anything close to a standard form. On the usage vokrenovo (< okrenuti, with two instances of performative [v] and deflection of [u] > [o] under the influence of proximate syllables in the same word), compare line 835, where HB sings okrenuvo ( < okrenuo, again with performative [v] to bridge hiatus within the two-syllable ending), and line 863, where he sings vokrenuvo (for the same reason, with an initial performative to avoid hiatus with the preceding word, konja). Taken together, these three examples illustrate the rule-governed pliability of the epic register, which has many built-in strategies available to meet the exigencies of vocal performance.

123. Note the plasticity of the performance idiom (and especially its relative independence of standard-language dialect), with both palatalized (njema) and unpalatalized (nema) forms of a single word in the very same line.

126. Lit., “He thought everything, he thought through on one [thing].” See further the AF.

127-38. As do many heroes in a multitude of songs from both the Moslem and Christian traditions of South Slavic epic, Djerdelez Alija here depends on the special abilities of his horse when his own options seems to be exhausted. See further the AF.

127. See the note to line 602.

130-31. In these two lines HB uses a [u]-glide run-up within two words: Huoće (130) and vovakuo (131). These are again grace notes to the melody rather than full syllables and therefore do not make either line hypermetric. They reflect an unusual, intra-word application of a common enough strategy usually employed at word- or line-beginning to start the articulation of [o]. Cp. kabuanica at line 713, as well as the note to line 93. NV treats the interrogative enclitic particle li in line 130 as grammatically linked to the verb Hoće; this reflects a recurrent transcription policy that permits (but does not mandate) the attachment of enclitics and proclitics to the preceding or following words, respectively. See further NVR.

134. NV transcribes HB’s jA (“And, But” plus performative [j]) with Ja (ostensibly the 1st s. nom. pronoun, “I”). Here lapsus auris leads to an outright error. HB apocopates nositi > nosit’ (metri causa) and NV mirrors his usage.

135. Lit., vali means “but” rather than “and.” The contrastive sense here seems to be that Alija’s mount was a fine animal, but one that nevertheless knew in a human fashion how to assist its master in traversing the surging river Drina. A close, even anthropomorphic relationship between a hero and his horse is a staple of South Slavic epic; see further the note to line 127-38, as well as the gloss to those same lines in the AF.

137. On the (untranscribed) semivowel glide between lexical elements in vodu o’skočijo, cf. the note to line 41. HB deletes the [d], singing o’skočijo in place of the standard odskočijo, and NV mirrors except for the shift of [o] > [a] in the verb stem (n. b. the standard-language doublet -skočiti/skakati).

138. HB sings “Na vonu ga stranu prenosijo” (“[And] bore him to that [opposite] shore”, while NV transcribes “Na drugu ga stranu prenosijo” (“[And] bore him to the other shore”). These are virtually synonymous phrases that illustrate the role of NV’s “singer’s ear” (lapsus auris) as he slightly remakes the song in transcription.

141. Although he characteristically deletes the instances of performative [j] and [v] from HB’s jodpočinuvo, NV mirrors his voicing of expected [t] > [d].

144. To avoid hiatus unsatisfactorily bridged by [h], HB sings vajvana in place of hajvana. Expectably, NV transcribes as hajvana. See further the notes to lines 74-75.

145f. Raca is glossed by MS as an “insulting Hungarian name for a Serb.” According to available lexical sources (SANU), Rac is a masculine singular noun that identifies a region where Serbians live, equivalent to Šumadija in line 146. HB, however, seems to assume a neuter plural (Raca). The negative image of Rac(a) is emphasized by the application of kleta (“accursed”) to Šumadija (though cp. the lexicalized formula “niz Markovac kleti,” “down by accursed Markovac,” as discussed in Foley 1991: 244-47) and by the notation that “a hero’s good fortune guided him” in line 148.

145. HB sings niz, while NV transcribes by devoicing the final sound (nis). This difference may reflect NV’s own habit of devoicing a terminal consonant before an unvoiced consonant (here [t] in ta), an articulatory commonplace in many registers of the language.

148. The asterisks (*hI*) designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line and thus from the point of view of performance does not render the line hypermetric. It is an [i]-glide run-up to the initial sound in junačka. See further the note to line 1.

149. The Lika is understood as a semi-mythical borderland in South Slavic Moslem epic, a liminal territory presided over by the treacherous Mustajbey. See further the gloss to lines 51-99 in the AF.

150. On the noun-epithet formulas employed to name Mustajbey, see the note to line 804.

151-56. On this idiomatic description of the bey and his entourage, see the recurrence at lines 193-96 and the AF.

152. HB sings Šejir, whereas NV transcribes as Sejir, restoring the standard form; see further line 532. Here and twice more during the performance HB pronounces livode/-u instead of the expected livade/-u. This latter regularity may stem from more than the consistent pronunciation of this word as an isolated element, however, since in each case a sound-pattern supports the [a] > [o] shift: “vuz rosnu livodu” (283), “niz rosne livode” (569), and “Šejir polje čini ji livode” (152). NV transcribes all these occurrences as livod-, but at 283 he overwrites an original a with o, showing an awareness of the standard form.

154. HB sings Davulhana, while NV initially transcribes Dauhana and then, crossing out that first transcription, writes Davulhana, with underlining. Given that the most common form of this Turkicism is dabulhana (Š; davulhana and daulhana are also cited afterward, but not dauhana), it appears that NV recognized the word after a false start but remained confused by the performative [v], which he here transcribes (as he virtually never does) but also underlines to mark its singularity. Note that this is a collective noun (Š), which I translate here and at lines 196-97 as a plural, adjusting the grammatically singular verbs accordingly. See further lines 196-97, with note.

155. HB alters the expected form udaraju to vud’ruju, via performative [v], syncope of the second syllable (metri causa), and modulation of the vowel [a] > [u], apparently under the assonantal influence of first- and last-syllable [u]. NV transcribes as udruju, with underlining marking his awareness of the nonstandard form.

156. The [m] performative is relatively rare. Here it generates a homophone with mu, the dative singular of the personal pronoun on (“he”); however, under the compositional (and receptional) rules for the epic register, which license the insertion of a whole range of performatives, no confusion would result. Thus NV does not transcribe the [m]. See further lines 196, 259, 279, 314, 341, 446, 664, 854, 965, 968, 969, and 1008, as against 303 (with note); at line 196 it may be that the performative is fossilized within the formula: cf. “kako mu veziro” (156) vs. “kako mu vezira” (196). Here HB sings veziro, with apparent lapsus linguae (but cf. the instability of [a] - [o] elsewhere in the performance [e.g., lines 112 and 160, with notes], while NV restores the expected vezira via lapsus auris.

157. HB sings “’Vako pazi” (“So he spied”), whereas NV transcribes “A opazi” (“But he caught sight of”). These close equivalents effectively illustrate lapsus auris.

160. The [i]-glide run-up to jev’, itself equipped with performative [h], is extrametrical and does not make the line hypermetric. See further the note to line 1. HB apocopates to jev’ (< evo) metri causa and NV mirrors, without performative [j] as usual. HB sings ozdala, with his idiolectal [o] > [a] shift; NV does not restore the standard form ozdola.

163. This kind of grammatically and syntactically balanced line, with the “young man” in one colon linked to his “bay horse” in the other and with the same quality of “brave” (dobar) ascribed to each, is a common pattern in South Slavic epic. Whether it can be established as formulaic (on the basis of the rest of HB’s recorded repertoire) or truly proverbial (either inside or outside the epic tradition) is of less significance than its general pattern and recognizably proverbial-sounding structure. See further Foley 1994.

164. On the Tsar’s medals as a metonymic attribution, see the AF.

169. On the Two spirited greyhounds/wolves, see the note to line 106 and the AF.

170. Note that the second colon of the line – “pa ji’ pripitomijo” – is seven syllables long, but that HB avoids hypermetricity with a performance-based strategy: he sings the final two syllables (-mijo) as a resolved single syllable. NV seems to have transcribed the line correctly (minus the performative [j], of course) after a second try; initial A and mid-line i both appear to be added after the first pass.

173. On the role and meaning of the whole-line formulaic phrase [X] began to shout, which HB employs 36 times in a wide variety of circumstances within this performance, see the AF.

174. HB sings Hajde, whereas NV transcribes Ajde. These forms amount to two idiolectal variations of the same verb form, and can both be translated “Hurry!” They present another case of lapsus auris, based as usual on personal usage; see line 175 (where HB again sings Hajde and NV again transcribes as Ajde) and the note to line 373. Here and on four other occasions (358, 408, 423, 431) HB sings Nuvane/-a in place of the expected Nuhane/-a, illustrating once again the volatility of [h] in performance and its tendency to be replaced by a performative. NV restores Nuhane-/a via lapsus auris in all five instances. See further the note to line 116.

175. See the notes to lines 174 and 373.

177. HB sings “vAko bude” (“If he be”), whereas NV transcribes “Oko bide” (nonsensical). Oko is a preposition meaning “around,” but it governs the genitive and no such object can be found; bide does not seem to be a viable word. Lapsus calami appears to be the only explanation.

178. Via lapsus linguae, HB sings “bjega vu Grbave,” with the preposition u (“in”) instead of the usual and formulaic phrase with od (“of, from”), and with genitive instead of the expected dative inflection (as if he were understanding od plus genitive). NV corrects the grammar by transcribing “bjega od Krbave” but introduces an error by devoicing the initial sound in Grbave (lapsus auris > Krbave); cp. line 435, with note. As for meaning, “bey of Grbava” is an alternate formulaic phrase for “bey of the Lika.” Note the metrical equivalence of “bjega vu Grbave” and “bega od Lijeke” (cf. 173: “beže vod Lijeke”). The etymology of Grbava seems to indicate “the humped place” (linked to the Lika as borderland?), though how transparent the toponym may have been to the singer and native audience is hard to say. The palatalization of bjega is a situational phenomenon and is not heard elsewhere in this performance. Such occasional sound-shifts are a feature of the functionally pliable singing register.

180. HB deletes initial [h] from the Turkicism hizmet (Š) and sings ’izmet; NV mirrors. See also line 202, with note.

186-87. On the traditional signification of giving and receiving a selam, see the AF.

186. A selam (Š) is a Moslem salutation in which the speaker turns his or her head first to the right and then to the left and utters a prayer (“Greetings and may God’s peace be with you!”). Whether HB (or any of the Moslem singers in 1930’s Stolac more generally) was using this term in that strict sense or in a somewhat less specifically religious (and largely formal and ritualistic) manner is open to interpretation, but the word and gesture are at least idiomatic.

187. The word njemu is inaudible on the DAT. HB sings prifatijo, and NV does not restore the standard form prihvatijo.

191. The word mu is inaudible on the DAT. Djulić is here called momak (“young man”), a label more customarily applied to a Turkish hero or young noble than to a servant. The attribution is a case of lapsus linguae, but not a serious error; the assignment of speeches is clear enough.

194-97. See the note to lines 151-56.

194. See the notes to lines 25 and 207.

195. The word bajraktara is inaudible on the DAT.

196-97. On the two occurrences of Davulhana, see the note to line 154. Here NV transcribes Daulhana in both instances, clearly and without crossing out.

196. Another instance of the relatively rare [m] performative. NV showed no confusion with the homophone mu, transcribing as u without the performative. See further the note to line 156. The translation is filled out formulaically (understanding sviraju, “were playing”) on the basis of lines 154-56.

198-99. The section “otkitijo, / Pa” is inaudible on the DAT.

198. HB sings riječ (“word”), in the singular, which means “unit of utterance” in the epic register; see further Foley 2002: 11-21. I translate here as “words,” but the true meaning is closer to “speech-act.”

201. On Šćadijaše, see line 205 and the note to line 838.

202. See the note to line 180.

203. On the Two spirited greyhounds/wolves, see the note to line 106 and the AF.

205. On Šćadijavu, see also line 201 and the note to line 838.

207. NB mirrors HB’s reduction of Ležite > Lež’te (metri causa) in his customary way – without marking the deletion. He also transcribes HB’s fronting of dole > doli without indication of the nonstandard form; cf. HB’s gori (< gore) at lines 25 and 194. On the “two spirited greyhounds/wolves,” see the note to line 106 and the AF.

209. HB apocopates > kretat’ (< kretati) metri causa, and NV mirrors.

210. Although NV regularly transcribes forms of baciti just as HB sings them (with [č] in place of [c]), here he writes baci. Whether this is lapsus calami or, less likely given the regularity of NV’s treatment of this word elsewhere, a correction to a standard form via lapsus auris is open to question.

211. An interesting accommodation by singer and transcriber alike, and a case of lapsus auris. HB sings dizg’e (< dizgine) to shorten the word one syllable (metri causa), whereas NV transcribes dizgin, suppressing the final syllable for apparently the same reason. HB sings prifatiše, and NV does not restore the standard form prihvatiše.

212. The code ## marks an initial vocal rest of one syllable. As explained above (note to line 111), this is not a metrical flaw and therefore not a hypometric line. See further the note to line 1. NV inserts an initial A (“And, But”) that HB did not articulate; this is another instance of NV’s own singing idiolect overriding the performed reality (lapsus auris). HB sings vadati, with his frequent modulation from [o] > [a], and NV does not restore the expected vodati.

214. On this Position change line, see the AF.

216. HB sings “Kak’ upade,” whereas NV transcribes “Kako upade,” normalizing the form (lapsus auris) but making the first colon and therefore the line appear hypermetric. Except for the elision that HB institutes metri causa, NV’s transcription of agam nearly fills out the expected dat. pl. agama, correcting HB’s vaga’ via lapsus auris. On Giving and receiving a selam, see the AF.

217. The word Tuka is inaudible on the DAT.

219. See the note to line 63.

220. NV silently corrects HB’s metnuse to metnuše, restoring the expected palatalized form.

221. Ko’ vodžaka has undergone a two-stage process: loss of the [d] in Kod followed by interposition of the performative [v] to bridge the resulting hiatus.

222-24. See the note to lines 8-15 and the AF.

222. Here HB aspirates > kahvu and NV follows suit. But see also HB’s unaspirated kavu (NV kahvu) at line 11, which shows that neither form is consistently idiolectal. Cf. the note to line 249.

223. HB uses a performative plus run-up glide (*hI*) to lead into the initial sound of the first metrical element (jednu). The connective is extrametrical, occurring before the metrical and musical pattern of the decasyllable, so that the line is not hypermetric. See further the notes to lines 1 and 12-13.

224. See the note to lines 13-14.

225-43. On this instance of Identity of a stranger, see the AF.

227. Here the difference between what HB sang (the connective ji) and what NV transcribed (the preposition u) amounts to a difference in syntax as well as word choice. Both options are grammatical, however, and collectively they illustrate one kind of lapsus auris.

228. NV’s underlining in teb shows his awareness that this apocopated form (metri causa) is not customary. Interestingly, however, he does not mark the apocopated infinitive upitat’ similarly; in general, NV’s recognition of nonstandard forms seems intermittent.

229. Lit., “Grant [it to] me -- say who you are and how you are.” The form dji ( < i) seems to be generated by adding both a performative ([j]) to bridge hiatus and an excrescent [d], perhaps under the acoustic (alliterative) influence of Daj earlier in the line. NV restores the expected i.

234. Š lists both dajidža (HB) and daidža (NV), which amount to alternate spellings of the same sound-shape, as viable forms of this word.

235. HB syncopates > Saraj’vu, shortening the expected Sarajevu by one syllable (metri causa), and NV does not restore the (unmetrical) standard. See further line 298.

236. HB shortens ostao > ost’o and vrhu > vrh’ to fit the second colon (metri causa), and NV mirrors.

241. This is another instance of HB shortening metri causa (došao > doš’o) and NV mirroring the adjusted form.

242. The form si reveals a rare instance of [s] used as a performative. The intrusive [n] in barnjaka may well be an idiolectal usage fossilized formulaically; compare line 246 (“krpu si barnjaka” vs. “krpa ni barnjaka”). The motivation for si may be the palatalized sibilant in krojiš, via lapsus linguae. Throughout these two lines NV restores the expectable forms (lapsus auris).

243. Note the vocative bajraktare standing in as a pseudo-nominative metri causa.

246. See the note to line 242.

248-51. On the Tsar’s medals, see the AF.

248. HB apocopates > Vid’ (metri causa), avoiding hiatus in the process, and NV does not restore the full form Vidi.

249. The palatalization in njisi seems to be prompted by the contiguous njiha (itself expanded by HB < the expected njih, ostensibly to fill out the first four-syllable colon, and thus the decasyllable, metri causa). NV writes lahko, aspirating HB’s lako, although he transcribes the same word as lako at line 429; cf. their treatment of kavu/kahvu at lines 11 and 222, with notes.

252-77. On the traditional nature of The fiancée problem, see the AF.

256. HB sings Da (“So that”), while NV recomposes via lapsus auris as Ja (“I”). Both constructions are viable, the difference being that NV’s line specifies the otherwise implied subject of sam -- opravijo (“I dispatched”), while HB’s Da links the action of line 256 to the preceding phrase in what amounts to a continuation rather than a freestanding verse.

258. HB sings vU, whereas NV transcribes I. This amounts to a difference in syntax and slight variation in meaning, with both forms viable (using I leaves “Turčina, kanidžkog ajana” under the aegis of the preposition Do (in the preceding line) rather than NV’s U, but the sense is close and the gen. and acc. animate objects will be morphologically identical). NV’s intervention is again a case of lapsus auris.

259. Lit., “And with him is the noble Zlata.” Both the action that follows and the traditional pattern of the song indicate that Zlata is the daughter of the Kanidža champion; thus the augmented translation here. Note the relatively infrequent use of the performative [m] in mu; once again NV shows no confusion with the homophone mu, transcribing as u. See further the note to line 156. NV’s substitution of plemena for HB’s plemenita is another case of lapsus auris, leading here to a grammatically and logically viable – but metrically deficient – mishearing: the substituted phrase, “plemena Zlata,” is a syllable short of the six-syllable colonic form (as well as a unique departure from the performance norm).

259-60. Zlata and Zlatija are metrical bi-forms; I translate both of them here and throughout as “Zlata.” On the formulaic phraseology associated with this character (espec. plemenita Zlata, “noble Zlata”), see further the two-syllable instances of the name (259, 262, 265, 657, 772, 776, 798, 931, 940, 969, and 974) versus the three-syllable instances (260, 313, 593, 604, 659, 676).

260. Lit., “Zlata could have been for the vezir.”

261. Here NV analyzes the melded word Kamoli, reduced by apocope to Kamol’, into its constituents (kamo [“if only”] + li [emphatic particle] = “not to mention”). Whereas most modern dictionaries and lexicons (e.g., B, MS) represent this item as a single word, some older sources (e.g., V) break it into two words. As always, NV’s handling of proclitics and enclitics is variable; see further NVR.

263. HB sings Tamo (“There”), whereas NV transcribes as Samo (“Only”). This relatively small but finite shift in meaning results from lapsus auris.

267. HB sings svedi, wrongly inflected (3 pl. = svedu; svedi is nonsensical) and possibly cued by vodili in the prior line; NV transcribes as sung, without correction.

268. HB sings dočeka j’ (“[he] was waiting for her”), while NV transcribes dočekaj (ostensibly the imperative 2s, but unworkable as such here). In this case lapsus auris has led to an outright error in transcription, rather than an alternate form.

269. See the note to lines 46 and 544. Also, this line begins the sequence of HB’s 14 uses of the word hiljada (“thousand”) in various cases and constructions throughout his performance. Twelve of these occur in an unmodified form with [h], or hiljad-: 269, 310, 320, 392, 393, 395, 401, 510, 531, 563, 590, and 1018, of which all but two (510: “pa bilo hiljada”; 531: “Na hiljade”) follow a terminal consonant. These two examples amount to more evidence for the initial sound [h] as a consonant that does not necessarily require replacement by a performative. But see also lines 929 (“po jiljadu”) and 1019 (“ostalo viljada”), which reveal the phonological precariousness of initial [h] in the singer’s register.

271. The patron, or mušterija, is the person responsible for a betrothed young woman after the marriage is arranged and before the ceremony takes place. Here the original patron is slaughtered by Baturić ban and his army, so the task falls to Mustajbey, father of the bridegroom Bećirbey.

272. HB vacillates between the palatalized (djevoj-) and unpalatalized (devoj-) versions of this word; see further 302, 312, 572, 578, 584, 599, 733, 734, 737, 770, and 923. NV mirrors his articulation except for two cases, reversing the palatalization in lines 734 and 770. Note that these two apparent discrepancies may actually be instances of lapsus calami, given that NV’s graph for [dj] is a crossed d (one may have been incidentally crossed and the other inadvertently left uncrossed).

279. Note the relatively infrequent use of the performative [m] in mu. NV shows no confusion with the homophone mu, transcribing as u. See further the note to line 156.

283. See the note to line 152.

286. NV does not correct HB’s prifatiše to the standard prihvatiše.

287-88. These lines are inaudible on the DAT.

287. On this Position change line, see the AF.

288. HB’s hegbe and NV’s heljbe are alternate forms of the same word, meaning “saddlebag” (Š).

289-90. On Giving and receiving a selam, see the AF.

289. HB sings Kak’ (apocopated < Kako) and NV mirrors his usage. The word jim is inaudible on the DAT.

290. NV does not correct HB’s prifatijo to the standard prihvatijo.

292-93. These lines, along with the first two syllables of line 294 (vA za-), are inaudible on the DAT.

292. Oaths of this sort are handled variously by HB and reflected variously in NV’s transcription. That is, neither the singer’s vocalization nor the transcriber’s interpretation is absolutely consistent. Here HB sings Boga ti (which I translate “By your God”), whereas NV transcribes as a single word, Bogati. HB’s practice here and later in the performance indicates at minimum that the oath has not become entirely lexicalized, as in the standard modern-day form bogme, a form derived from bogami (etymologically, “by my God”), which B renders as the weak intensifier “really.” NV’s practice here and later in the transcription – one word at 292 and 933, two words at 637 – also reveals an awareness of the root constituents. See further lines 637 (HB: Boga ti, “By your God”; NV: Boga ti), 797 (HB: bogme, which I translate as “by God” because even the apparently lexicalized alternative may bear etymological force in the singing idiolect; NV: bogme), and 933 (HB: Boga mi, cp. 292; NV: Bogami).

293. Here, as often throughout the poetic tradition, the 3rd person pres. of biti (whether as an auxiliary in a compound past tense or as a freestanding copulative), is elided metri causa; NV mirrors HB’s usage.

297. Lit., “There he is from over there, from proud Bosnia.” HB sings vEt’, apocopating metri causa, and NV does not restore the full form Eto.

298. See the note to line 235.

299-300. Within the same speech Mustajbey turns rather abruptly from identifying the young stranger as Djerdelez Alija to a direct address of his brother Meho, whom he sent to ask for Zlata’s hand in marriage. This type of shift is not uncommon in traditional oral epic, where the individual performance is always contextualized by the audience’s prior experience.

302. See further the note to line 272.

303. One might construe Ma as mA, that is, as A (“but”) with performative [m], as does NV by transcribing as A, but idiomatic sense here seems to demand the interjection ma. See further the notes to lines 46, 156, and 446.

304-70. On the traditional structure and implications of the letters written by Mustajbey to his guest-allies (Catalogue I), see the AF.

304. HB apocopates > pisat’, shortening the line metri causa, and NV mirrors.

307-9. The section from -selami through vu is inaudible on the DAT.

307. The initial palatalization in njega seems to induce the “extra” [j] in ovakoj. On this line as a ritualized speech introduction, see further the AF.

309. NV shows no confusion with the homophone vodi, from voditi (“to lead”), deleting the performative [v] and restoring the [h] to yield hodi (via lapsus auris). See further lines 340 and 341 (with note).

310. Two more instances of shortening metri causa (dvades’ < dvadeset and hiljad’ < hiljada). From this point on, apocope that leads to this kind of syllabic adjustment will be noted in NVR but not separately described in the Commentary. See also the notes to lines 46 and 544.

312. Typically, NV fills out the swallowed consonant, transcribing as the standard kanidžkom in place of HB’s sung form, kanidžko’. See further the note to line 272.

314. The [m] performative is relatively rare, here apparently induced by the phonological context: spremi and odma’ (whose terminal [h] NV does not restore). See further the note to line 156.

316. See the AF.

318. Lit., “They say you have [there are to you] my three precincts.”

320. See the notes to lines 46 and 269.

321. HB sings beratli, an adjective from berat, “tsar’s decree (order, charter) according to which ranks, honors, privileges, or other such benefits are awarded” (Š). Note that, although Š does not cite beratli as a Turkicism, he lists the noun that is its source; -li is the common and customarily indeclinable Turkish adjectival ending. If construed as declinable, secondary support for this form may be found in its phonological mirroring of the adjectival ending –ih, which HB and other singers conventionally deaspirate to –i’.

324. The last four syllables of this line are inaudible on the DAT.

325. The palatalized form mjeni is situational (cf. podji at line-end) and occurs nowhere else in the performance.

327. The section from spremi through piše is inaudible on the DAT.

328. If there were any doubt about the ubiquity of performatives and their ability to displace [h] in the interest of fluent singing, HB’s form Jercegovinu (for Hercegovinu) should set it to rest.

330. Like most noun-epithet formulas, Bišćević Aliji inflects only at the end of the phrase or whole “word” (reč), the colonic unit of utterance; cp. Topalović Husu at line 361. But contrast buljubaši Muji at line 338, where the epithet, an impermanent because situational attribution, is less firmly attached to the noun (thus not forming a true, unitary reč) and where such inflection does not compromise the overall metrical word-shape of the phrase.

331. See the AF.

333. Lit., “Gather [your] young Bišćevites [residents of Bišće] for me.”

340. See the notes to line 309 and 341.

341. Here, as in lines 309 and 340, HB sings vOdi (“Come”). But whereas NV transcribes the form as Hodi in the other two cases, here he writes Odi, deleting the performative without restoring the expected form that begins with [h]. This is a rare occurrence of NV not fully “repairing” the deflection caused by a performative; cp. line 541, with note. All three instances involve lapsus auris, but with two markedly different outcomes. Note the relatively infrequent use of the performative [m] in mu; NV shows no confusion with the homophone mu, transcribing as u. See further the note to line 156.

342. A selambaša (official greeter) is the person who gives a sign of peace (selam) to an army on the occasion of a celebratory march (Š). The term derives from selam (see the note to line 186) and the Turkish suffix -baş, “head, chief.” See also line 488.

344. HB sings sve (“all”), whereas NV transcribes as te (“these”). This is a simple case of lapsus auris, with two roughly equivalent and equally viable alternatives; NV’s “singer’s ear” selects a viable idiolectal equivalent different from what HB actually sang (itself also quite viable).

345. HB sings the grammatically correct form knjigu, whereas NV transcribes as knjiga. Whether this transcription is simply lapsus calami or lapsus auris under the acoustic influence of the proximate Onda is difficult to say.

347. The [i]-glide run-up (*I*) leads into the initial sound in još; it is extrametrical and thus does not render the line hypermetric. NV does not transcribe the I. See further the note to line 1.

348. See the note to lines 779-81.

349. An instance of the relatively rare performative [l], which NV does not transcribe; cp. lines 420-21 and 537, with notes. Immediately afterward, NV first transcribes HB’s tebi as tebe, then overwrites the second e with an i.

350. Lit., “Look well at what this letter writes to you.” HB sings sta (depalatalized), whereas NV transcribes as the expected form šta, correcting the error via lapsus auris. See further line 962, with note.

355. HB sings ninovu (lapsus linguae, possibly motivated by alliteration with line-initial Na), while NV restores the expected form vinovu, correcting the error via lapsus auris.

357. See further the notes to lines 46 and 544.

358. See the note to line 174.

365. HB sings vaga (gen. pl.), whereas NV transcribes as age (acc. pl.). This is an interesting case of lapsus auris, with two distinctly different – and yet roughly equivalent – syntactic possibilities. As HB sings it, vaga is a partitive genitive, “of the aghas,” dependent on the phrase that follows it: “što ji’ više jima” (“as many as there are of them”). Thus, even though the preceding line begins a partial terrace or pleonasm, with the same imperative verb (Kupi) plus vage (acc. pl.), HB is syntactically justified in his variation to vaga in 365. NV, on the other hand, seems to hear the terrace as the primary criterion for the syntax and therefore repeats the acc. pl. object, vage/age, apparently construing “as many as there are of them” as a paratactic add-on outside the basic syntax of the second Kupi.

367. HB sings vOću, whereas NV transcribes as Hoću, deleting the performative [v] and restoring the expected form. Cp. voćemo (HB) / hoćemo (NV) at line 478 and voćeš (HB) / hoćeš (NV) at line 494.

371-79. On this loosely configured traditional unit, Cannon signals, see the AF as well as the sequel to this passage at lines 381-89.

372. This Djulić, identified by Mustajbey as his own son (“moj rodjeni sina”), may be a different person than Djulić Nuhan, whom he earlier dispatched on an epistolary errand to the King of Pokraljo (358) and who will undertake the errand to Tale of Orašac as well (see 408 ff.). But the apparent familial address may also constitute a mark of synthetic kinship, essentially a term of endearment that can carry with it certain obligations (see further Foley 1992). Part of the apparent confusion lies in the fact that servants are very often (even typically) named “Djulić” in the epic tradition at large, so that the name effectively designates a generic type rather than a uniquely realized character.

373. HB sings vAjde, whereas NV transcribes as Hajde, deleting the performative [v] and restoring the expected form. There is some instability in both HB’s usage and NV’s transcription within the performance, a poetic-language variance that reflects the interchangeable Hajde / Ajde doublet in everyday speech, the intermediate status of [h] as a hiatus bridge, and the phenomenon of lapsus auris. HB also sings vAjde at lines 408, 409, 438, 492, 680, and 788; NV writes Hajde for all but 438 and 680, where he interprets vajde as ajde. But HB sings Hajde at lines 174 and 175, with NV transcribing as Ajde. In regard to the second colon of line 373, the preposition na normally takes accusative when motion is implied, but odžak (acc. s.) would leave the line a syllable short and clearly hypometric. The “inflection” to odžaka, ostensibly gen. s., is thus better understood as a performance-driven accommodation to the melodic and rhythmic integrity of the line than as a grammatical error. There may also be acoustic reinforcement from the next (paired?) line (374), mirrored again at 381, where the gen. s. inflection in “viz odžaka” is grammatically appropriate as well as prosodically apt.

380. On the traditional structure and implications of this Pivot line, see the AF. The performative n is relatively infrequent. NV first writes Nu, then U, and then crosses them both out in favor of Nu, the underlining apparently indicating that he did not understand this word (in fact, his solution of U is correct).

381-89. See the note to lines 371-79 and the AF.

382. See the note to line 63.

384. See the note to lines 936-37.

386. The term lumbarde designates “a type of old cannon” (SAN). The further appellation of matice (lit. “source,” rel. to “mother, queen”) seems to indicate that these are the large, main batteries of cannon in the towers.

387. HB sings Karavulam’, with performative [v] and elision of the final syllable metri causa, whereas NV transcribes as Karaulami, deleting the performative and filling out the inflection via lapsus auris. Similarly, the form predvodnice ( < prethodnice) results from a word-internal shift from the aspirate [h] to the performative [v] and back-voicing of [t] > [d]. This is a relatively complex change, but one that is straightforwardly licensed by the rules of the performance register. NV transcribes as predhodnice, with voicing of [t] > [d] but deletion of performative [v] in favor of restoring the expected [h].

388. HB sings Popucuju (for standard Popucaju), while NV transcribes as Prepucuju, varying the verbal prefix and not restoring the standard inflection. The lexical difference here, which arises via lapsus auris, is slight: popucati = “begin to shoot” or, more likely given the context, “to shoot in all directions one after another” (MS), while prepucati = “to shoot strongly for some time” (MS; only the reflexive form, not observed in NV’s transcription, approaches the sense of popucati).

390-401. See the gloss on Catalogue II in the AF.

392. See the note to line 269.

393. See the note to line 269.

394. The first three syllables are inaudible on the DAT.

395. See the notes to lines 46 and 544, as well as to line 269.

399. NV leaves vodi completely out of his transcription. This kind of lapsus calami raises logical questions about the accuracy of transcription of ancient and medieval texts or performances, as well as illustrates the necessity of after-the-fact annotation (as an informative, not a prescriptive or proscriptive process). See also the note to lines 936-37.

400. Also referred to as “Topalo’s son Huso” at 361.

401. See the note to line 269.

402. HB sings zila with extraneous initial voicing, whereas NV transcribes as sila, removing the voicing and restoring the expected form via lapsus auris.

405. Lit., “The company was all [there], only there was no Tale.” Instead of the expectable gen. s. Tala, HB sings the nom. s. Tale both here and in the following line; NV mirrors his usage. Such small errors of inflection are more common among proper names than among other categories of nouns and noun phrases. Cp. the noun-epithet formulas, which seldom inflect their first elements, assigning case endings to the end of the whole phrase (thus, for example, one does not expect the gen. s. inflection on budaline in line 406); see further the note to line 330.

406. Tale is here called a hal, “wretched one,” the reasons for which attribution will become clear below when he appears. He is also called budaline (< nom. s. budalina), which amounts to the augmentative form of budala, itself generally glossed as “fool, madman, insane one” (Š, MS, V, B; other occurrences at lines 411, 416, 425, 437, 449, 478, and 695). In keeping with the morphological signal of -ina, which denotes size or extensiveness (often with a pejorative coloring), and with Tale’s overall identity as a trickster whose appearance and actions are precisely the opposite of what one would expect in a major hero, I translate budalina in its various inflections as “great fool” (deferring capitalization in both the original-language transcription and the translation because this epithet is not a patronymic or formal title). One lexicographical source (SANU) provides a secondary gloss of budalina as “stubborn, obstinate person,” but offers citations only to Tale and Kraljević Marko. In narrowing the field of reference strictly to characters who demonstrate stubbornness, this secondary gloss seems reductive and tautological. Both Tale and Marko qualify as budalina on many other grounds as well. Hopefully, the more diagnostic and etymologically secure rendering of “great fool,” which names the fundamental trait of broad-based iconoclasm and often anti-traditional behavior that manifests itself in so many ways, conveys some of the powerful, effectively magical inversion encoded in these unique figures. Note specifically that Tale is both an indispensable part of any successful armed force (see line 412, with note) and a superficially unpromising – even counter-heroic – personality. On Tale’s heterodox character, see further the AF, gloss to lines 402-78; and Foley 1995a: 32-41. As for the “bad” ending of Tale in line 406, where one would expect gen. s. Tala, the pairing budaline Tale may well result from acoustic patterning in [e] overriding grammatical logic, a common phenomenon in South Slavic oral epic performance; cf. the same dynamic with budalinu Talu (with dat. s. Talu in place of the acc. s. one would expect, apparently on the basis of in-colon rhyme with the acc. s. budalinu; at lines 411, 416, and 425). See also the note to line 405.

408. See the notes to lines 174 and 373.

409. See the note to line 373.

410. Ibro, the hypocoristic form of Ibrahim, is another name for Tale.

411. See the note to line 406.

412. Lit., “Without him there is no journey there.” By “there” Mustajbey means to designate the site of the battle that is traditionally the precursor to wedding in this subgenre of South Slavic epic (see the gloss on the Wedding Song story-pattern at lines 1ff. in the AF). Note that this line is proverbially attached to Tale, and Tale alone, in many epics, and that it speaks idiomatically to the necessity of his presence and contribution in battle (see the gloss to line 412 in the AF). Notwithstanding his sorry appearance, he is a combatant of unmatched bravery and achievement (see further the note to line 406). NV transcribes as brez, with the underlining indicating his awareness of the nonstandard form. Note that he does not reduce it to the expectable bez via lapsus auris; see further the note to line 42.

413. See the note to line 484 below, as well as the gloss in the AF.

414. Djulić’s horse is so swift that it is referred to as a “hare,” not an uncommon metaphorical appellation in this tradition.

416. The phrase “on arriving” has no equivalent in the original, but I add it here to smooth a particularly abrupt transition. This shift of place would of course be contextualized within the performance register, with the fluent audience or reader supplying the transition from prior experience with the epic way of speaking. Compare, for example, Homer’s famous paratactic shifts, paralleled as well in Anglo-Saxon oral-connected poetry. HB sings budalinu (dat. s.), whereas NV transcribes as Budaline (nom. s.). HB has inflected the first element of a noun-epithet formula, something rarely done by the guslari (see the notes to lines 330, 405), but allowable here because inflection does not alter the metrical word-type by making it hypermetric or hypometric. NV, on the other hand, “hears” the more frequent form, with inflection coming only at the end of the entire phrase or “large word.” See further the note to line 406.

419. Since the normal acc. s. would have zero inflection (potok) and leave the line short by one syllable, HB sings potoke, with an extraneous [e], to eke out the decasyllable metri causa. NV mirrors his actual articulation without underlining. Cp. line 424.

420. The performative [l] in lokovanu, not transcribed by NV, is relatively infrequent; cp. lines 349, 421, and 537, with notes.

421. Perhaps under the influence of line 420, the performative [l], again untranscribed by NV, appears intersyllabically in jodvelo; cp. lines 349, 420, and 537, with notes. On dvanajes’, see the note to line 46.

423. See the note to line 174.

424. Cp. line 419, with note. NV mirrors HB’s usage.

425. HB sings Kada, whereas NV transcribes as Kad on; this is a case of lapsus auris that entails no shift in meaning. The slight variation amounts to NV using the monosyllabic form of Kada and a pronoun (on) that specifies the subject of the verb instead of the full disyllabic form with the subject implied in the verbal inflection. On the inflection of budalinu Talu, see the note to line 406.

427. NV does not correct the inflectional error in vokreće (where the expected form in 3rd pl. is vokreću). See also the note to line 46.

431. The code ## marks an initial vocal rest of one syllable. NV provides the “filler” word A, which serves as a common line-beginning, to eke out the ten-syllable line. See further the notes to lines 1 and 174.

435. See the note to line 178, the only other instance in which NV transcribes as Krbav- (as against the expected Grbav- at lines 189, 192, 231, 335, 340, 434, and 439).

436. See the note to line 46.

437. See the note to line 406.

438. The emphatic form vOću (< Hoću) doći (“I shall come”) seems to indicate a grudging acceptance on Tale’s part, an attitude wholly in character for this trickster figure who conventionally shows disrespect for authority. For this reason I have added “All right” to the translation, a phrase that has no specific equivalent in the original. See also the notes to lines 367 and 373.

439. See the note to line 63.

440-41. The actions of dancing the horse and singing reflect Djulić’s relief over having successfully (and safely) persuaded Tale to join the army. Traditionally, Tale can be a difficult and even dangerous individual when asked (or – perish the thought – ordered) to do someone else’s bidding.

444. The reflexive particle se is inaudible on the DAT.

446. The performative m is relatively infrequent; NV does not transcribe it. Cf. its use with A at lines 854-55 and 968-69 (with notes), and contrast line 303 (with note), where Ma seems to be the correct rendering. See further the note to line 156.

448. The phrase “samo Tale nema” is inaudible on the DAT. See the notes to lines 405, 406, and 416.

449. See the note to line 406.

453. HB sings prebačijo and NV does not restore the expected form, prebacijo. See further the note to line 30.

454. Lit., “He pulled it with the motley halter.” But Tale is astride the animal (450), so the pulling must have to do with forcibly guiding his (proverbially stubborn) horse.

455. HB sings ćorbe (with initial palatalization) and NV restores the expected form, torbe, via lapsus auris.

456. A whole-word instance of lapsus linguae, supported by flexibilities that occur through the epic register. In place of the expected form prosijaveno, a past participle agreeing with brašno (neuter acc. s.) and meaning “sifted,” HB sings proševina; NV also transcribes as proševina. We can derive HB’s articulation via reduction of the third and fourth syllables (ija > e reflecting the ijekavski > ekavski dialect shift) plus his idiolectal substitution of [š] for [sj] and his common enough fronting of [e] > [i] (e. g., gore > gori at line 25, with note). These changes, as well as the inflectional disparity, echo the near-homophone proševina, cited in MS as prosidba, “wooing, suit”; this phonological similarity between unrelated words may have promoted HB’s lapsus linguae or NV’s transcription or both.

458. See the note to line 46.

459-60. HB ends this first section of his performance by singing line 459 in an extended, ritardando fashion. On resuming after a rest he begins with an instrumental interlude of 11 seconds, after which he “backtracks” to near the beginning of the last logical unit – the description of Tale’s troops – as he restarts the performance. This is the guslar’s usual way of continuing the song after a break: reprising from the last narrative boundary with or without a proem for continuation. In other words, although a guslar may stop for a rest anywhere in his performance, without regard to narrative partitions, he usually backtracks to a structural seam as he recommences (see Foley 1990: 284-88).

460. Note that NV’s transcription of the interjection vO! as the conjunction A obscures the grammatical force of the particle as a cue for restarting the performance. This instance of lapsus auris makes a substantial difference in the representation and reception of HB’s song. On dvanajes’, see the note to line 46.

461. Note the correspondence of lines 461 and 459, identical except for the difference in the performative ([v] versus [j]) at the beginning of the second colon. This discrepancy neatly illustrates the nominal nature of performatives as well as marks HB’s backtracking.

463. Lit., “How much Tale infuriated it!” HB sings ražljutijo, with back-palatalization; NV mirrors his articulation and does not restore the expected razljutijo.

465. Oblje dvije (or the usual form obje dvije without intrusive [l]) is interpreted by V as a two-word variation on obadvoje. The evidence is scant, but obje appears to be indeclinable, with dvije an alternate genitive plural (under the aegis of the preposition s). The noun strane then harmonizes with the colonic sound-patterning with its inflection in -e instead of taking the genitive plural (strana) expected in this situation. NV does not restore the expected form.

467. Lit., “And first he arrived, he began to shout vulgarly.” See also the note to line 63.

468. This curse is just the beginning of a speech in which this quirkiest and most essential of heroes insults nearly everyone present – a typical gambit for a unique champion whose standardbearer customarily rides backwards on his mount with the standard held upside down and whose Islamic clerical companion drinks strong liquor and brandishes a copy of the Qu’ran in his (unholy) left hand. See further the gloss to lines 402-78 in the AF, as well as Foley 1995a: 32-41.

471. See the note to line 546.

473. Alovit means “powerful,” “spirited,” “strong.” A reasonable compromise for its usage here is “fearsome.” The performative becomes [k], perhaps in harmony with the initial sound in klanci, and in fact NV transcribes as kaloviti.

478. See the notes to line 367 and 406. As is very often the case, HB does not inflect within the noun-epithet “word,” singing budalina Tale without the voc. s. inflection budalino. NV transcribes as budaline, ostensibly an unmotivated gen. s.

479. HB sings vA vuktaše, whereas NV transcribes as Zahuktaše. This instance of lapsus auris presents a viable alternative form, but changes the meaning slightly from “And sounded” to “Began to sound.” NV’s devoicing of HB’s zile to sile may result from momentary confusion with the common noun sila (“army, force”); cp. line 402 (with note), where HB voices the initial sound in sila to yield zila and NV restores the expected form. All of NV’s other transcriptions of zile (155) and sila (390, 529, 888, 892, 912, 913, and 919) are accurate.

480. Note how performative [v] operates within a word to bridge two non-diphthongized vowels whose sequence would otherwise produce hiatus.

481. HB sings vAzurula, whereas NV transcribes as Hazurulah, very close to the standard form Hazurola (Š). Cp. line 583, where HB’s Hazurula (this time without performative [v] inserted for [h]) is rendered by NV as Hazurula (without the additional [h] at word-end). This plasticity in the singer’s vocalization and transcriber’s interpretation is especially interesting given the virtually identical second half-lines:”kita vi svatovi” (481) and “kita mi svatovi” (583) vary only in their performatives.

483. Lit., “Horsemen, girths in a bond!”

484. In an instance of lapsus auris that varies the literal sense only negligibly and preserves the idiomatic meaning, HB sings “vA vu noge,” whereas NV transcribes as “Pa na noge.” The literal difference is thus (approximately) “And onto their feet” (HB) versus “Then to their feet” (NV). Both whole-line phrases preserve the formulaic sequence “And the heroes jumped to their feet,” which projects the idiomatic force of “an honorable response to an unexpected or threatening turn of events that demands the principal’s immediate attention” (Foley 1991: 85). This implication, which arises from the poetic tradition at large, harmonizes with the immediate narrative situation, the beginning of Mustajbey’s army’s march toward life-or-death engagement in battle. For other instances, see lines 19, 413, and 603 (as opposed to lines 180 and 815, where the verb skočiti [“jump”] appears outside of its formulaic and idiomatic context; nor does the prefixed o’skočijo at line 137, applied to Djerdelez Alija’s horse, bear the special signification). In order to qualify as the value-added signal, this phrase apparently must include both skočiti and some multiform of the “to his/her/their [light] feet” (na noge [lagane]):

19. Skoči momak na noge lagane

413. Skoči Djulić na noge lagane

484. vA vu noge skočiše junaci

603. kAj! Djeveri na noge skočiše

485. See the note to line 63.

486. HB sings [svaki], whereas NV transcribes as svati, nom. pl. of svat (“wedding attendant”) but without the expected infix (svatovi), apparently under the influence of the frequently used svat, perhaps especially in the next line (lapsus auris). I choose to represent [svaki] as svaki (“each”), nom. s., taking it as the subject of the paired plural verb, with the lack of agreement partially masked by the sound-patterning in terminal [i]: thus the translation “each and every one,” meant to mediate between singular and plural. Such lack of agreement is of course not the kind of fatal flaw it would be in a written text; in fact, such small lapses are the price imposed by the recombinative flexibility and sound-dependence of traditional phraseology.

487. Note the vocative ending on svate, introduced metri causa to eke out the four-syllable colon and harmonizing in that way (and via leonine or in-line rhyme) with the much more common vocative and pseudo-nominative Osmanbeže that fills out the second colon in a noun-epithet formula. Both vocatives effectively operate as alternate nom. s. inflections. Cp. lines 554, 610, with notes; also the note to line 330.

488. On selambaša, see the note to line 342.

490-511. On this scene of Tabulating the army, see the AF.

492. See the note to line 373. Unusually, HB here allows the preposition na to govern the genitive case.

494. See the note to line 367.

495. See the note to line 63.

496. HB sings vE!, whereas NV transcribes as Pa. This amounts to a case of lapsus auris, with NV’s “recomposition” almost certainly influenced by the shift of vA to Pa in the parallel position in the preceding line (also lapsus auris in its own right [see the note to line 63] and fortified by the compositional habit, widespread among guslari, of paratactic echo of individual elements [Pa– – Pa] that often produces pleonasm or some other species of anaphora). Cf. lines 542-43 and 603, with notes.

502. On this Boundary line, see the AF and lines 606 and 730.

503. On this recurrent line, see the AF.

504. HB sings prošečemu, whereas NV transcribes as prosječenu, restoring the expected form by reversing HB’s regular idiolectal shift of [sj] > [š] and his nasal exchange of [n] > [m].

505. HB sings rasklapijo, while NV restores the expected form, rasklopijo, via lapsus auris. This is another instance of the [a] – [o] instability in HB’s idiolect. Cf. his standard articulation of Rasklopi in the next line.

506. See the note to line 505.

508. HB sings najzad’i, while NV transcribes as najzadnji, restoring the standard form. Cp. line 612, with note.

509. See the notes to lines 46 and 544.

510. See the notes to lines 46, 269, and 544.

511. Here the interjection vEj! is part of the decasyllable rather than an extrametrical syllable; cf. lines 2 and 514, with notes.

512. Lit., “When the first ones were to Mezevo.”

514. Here the interjection vAj! is part of the decasyllable rather than an extrametrical syllable; cf. lines 2 and 511, with notes.

515. See the note to line 63.

518. See the note to line 63.

520. See the note to line 63.

524. Lit., “All of the young men patron to the girl.”

525. The word jediinak includes a second [i] as a grace note, and does not make the line hypermetric. NV normalizes the sung tetrasyllable to the expected trisyllable jedinak. See further the note to line 93.

526. The asterisks (*jE!*) designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. NV does not transcribe HB’s jE!, beginning the line with Nije. See further the note to line 1.

527. NV’s Veg must be a case of lapsus calami, since HB clearly sings Neg’ and there is no way to construe NV’s transcription.

528. Lit., “There it was full in regard to the cemetery.”

529-30. In both of these lines [n] functions as a hiatus bridge, or performative, rather than, as it might appear at first hearing, an elided negative (ne > n’). Confirmation that [n] is a performative rather than the elided negative particle is offered by the fact that NV does not transcribe the sound in either 529 or 530. While much rarer than [j], [v], or [h], [n] does occasionally serve as a performative; cf. 567, 606, 607, 674, 850-51. Within a more focused sample, [n] seems to be linked non-formulaically with the verb izginuti, which does not occur in its full form without prefixed [n]; see 529, 530 (these two reinforced by a shared terracing structure), and 939.

531. See the note to line 269.

532. See the note to line 152.

536. Lit., “He led Halil up to his knee.” The phrase privodi Halila is inaudible on the DAT.

537. HB’s lapsus linguae, substitution of an alternate liquid consonant (lukom for rukom), is corrected silently by NV; cf. 349, 420, and 421, with notes, where [l] participates as a performative. The first two syllables of udaraše are inaudible on the DAT. See also the note to line 63.

541. Unusually, NV fails to normalize HB’s zazaftijo to the standard zazavijo by revoicing [f] > [v] and removing the excrescent [t]. NV was apparently confused about this verb, writing zastavijo (“began to place or put”), then crossing that out and writing the partial word zast before settling on zazaftijo. See further line 341, with note.

542. HB sings jO, whereas NV transcribes as A, a small shift caused by lapsus auris that nonetheless dilutes the force of the hero Mujo’s pledge to avenge Baturić ban’s kidnapping of the betrothed maiden Zlata. Note that NV transcribes O correctly in the next line, perhaps in part a response to the cumulative semantic and syntactic content. See further lines 496 and 603, with notes, as well as the note to line 544.

543. See the notes to lines 542 and 544.

544. As in line 542, NV’s deflection of vO to A by lapsus auris obscures the force of this third line of Mujo’s oath; see further Foley 1999b. The number 14 still stands by analogy, since multiples of ten (ending in -deset) are reduced to -des’ in HB’s idiolect. Numbers in the teens (ending in -naest), on the other hand, are reduced to -es’, e. g., četer’es’ here and at lines 269, 590, and 865, or to –n’ejes’/najes’. See further the note to line 46.

546. Uncharacteristically as a general rule but quite consistently for this particular word, NV twice fails to construe jAl(i) as Al(i) plus performative [j]. Cf. the same phenomenon at lines 618, 684, and 896 as against line 471.

549-50. On Getting up early, see the AF.

553-60. HB musically frames these paired lines (553-54, 555-56, 557-58, 559-60) with a couplet melody. On the musical dimension of the first 101 lines of this epic performance, see the section on Music.

552. The often extrametrical interjection vEj! is here performed as part of the ten-syllable rhythm and melody.

553. See the note to line 63.

554. Among all of the wedding attendants, in other words, Osmanbey occupies the prominent position of eldest witness, called both stari svat(e) (cp. 487, 610, with notes; the pseudo-nominative inflection in vocative, svate, provides a metrically alternate form at 487) and starosvat (a single lexical item), as here. Both lexical possibilities are sanctioned (MS). These three semantically equivalent but metrically and morphologically various forms of Osmanbey’s given title in the wedding party collectively provide a useful traditional flexibility. NV transcribes each of them verbatim, with word-division where appropriate (stari svate, 487 and stari svat, 610 versus starosvat, 554).

555. See the note to line 63.

557. See the note to line 63.

559. See the note to line 63.

561. Lit., “Even in this way Bećko appeared!”

563. In a case of lapsus linguae stimulated by the impulse toward in-line or leonine rhyme, HB sings hiljadu at line-end, adding an eleventh syllable and rhyming with broju at the end of the first colon. This impulse toward the common leonine pattern overrides the ten-syllable constraint as well as applicable grammatical rules, which call for a partitive gen. (hiljada), apocopated metri causa to *hiljad’. In fact, NV “hears” the latter, transcribing as hiljad via lapsus auris. Performatively, HB melds the last two syllables of the line, giving the overall impression of a (faux-)decasyllable. See also the notes to lines 46 and 544 as well as 269 and 592.

564. Lit., “When they were to white Kanidža.” HB sings bili (for fem. gen. s. bile), and NV does not restore the expected form. A number of factors apparently combine to foster HB’s lapsus linguae: (a) the homophone bili (pl. past participle of biti, “to be,” is commonly used throughout this and all other epic performances [10 times in this performance, 6 of those at line-end]); (b) that very homophone occurs at syllables 3-4 of this same line, the site that joins with syllables 9-10 to furnish the site for leonine or in-line rhyme; and (c) other words ending in –e are fronted to –i by HB (cf. gore > gori at 25 (with note), 194, and 947 and dole > doli at 207).

566. NV first writes Dočekali (3rd pl. past tense, and therefore yielding a lack of agreement with kanidžki vajane [“the Kanidža champion,” with voc. s. standing in as a pseudo-nom. s.; cf. lines 243, 330, 487, and 610, with notes]), then crosses out the li and adds i as a separate word, yielding Dočeka i, which corresponds to the DAT (without performative [j]).

567. The performative [n] is relatively infrequent. NV shows no confusion with the homophone ni, leaving [n] untranscribed. See the note to lines 529-30.

568-70. While it would not be impossible to construe these three lines as a narrative appendage to lines 566-67, the fact that the Kanidža champion is said to utter a selam makes it likely that the three-line segment amounts to a verbal command that he issued in order to make arrangements for his guests’ reception. On Giving or receiving a selam as a ritualized speech-introduction, see further the AF, gloss to lines 186-87.

569. See the note to line 152.

572. See further the note to line 272.

573. Lit., “Here it was fine for them in the shelter.”

574-75. Note the classic example of pleonasm or terracing, with the final four-syllable increment in line 574 (the verb prešedješe) repeated as the first colon of the following line.

574. HB sings Vazdi, perhaps under the influence of ljudi later in the line, and NV does not restore the standard form Vazda.

576. This line provides a good example of how the choice of performatives is based on phonological environment rather than the particular words to which they are attached. HB sings jigre because he is bridging from the high front vowel in the last syllable of the preceding word (momci), but vigraju when that same verb follows the lower middle vowel in se. See further the section on Performatives.

577. HB sings voda, whereas NV transcribes as hoda, replacing performative [v] with lexical [h] and restoring the standard form. See further the notes to lines 74-75, 174, and 373.

578. See further the note to line 272.

580. HB sings jogranu, whereas NV transcribes as ogrija. Both verbs are employed with sunce to mean “the sun rose,” with the slight discrepancy that ogranuti involves crossing a boundary (or granica) while ogrijati/ogrejati involves heating or warming. This is a good example of lapsus auris leading to an idiolectal variance without any truly salient difference in meaning. On this Dawn marker pattern, see the AF.

581. See the AF.

583. Cp. line 481, with note. HB sings kita mi, whereas NV transcribes first as kićeni (“decorated, ornate”) and then overwrites as kita i. Note that there is apparently some instability in the poetic tradition at large over the doublet kita i svatovi (“forces and wedding attendants”) and kićeni svatovi (“decorated wedding attendants”), in part because of the near-match acoustically and in part because within the Wedding Song story-pattern the attendants are also an army that is customarily pressed into fighting a battle for the betrothed woman. See further V, who observes the functional equivalence between these two phrases (a kind of lapsus auris?) and juxtaposes them (rather imprecisely) to the opening line of the Aeneid: “kita i svatovi, that is, kićeni svatovi, wie im Virgil: arma virumque.”

584. See further the note to line 272.

586. See the AF.

590. See the notes to lines 46, 269, and 544.

592. An 11-syllable line and a true hypermetric verse, caused by the combination of the ijekavski form dijeliti (instead of deliti or djeliti, each one syllable shorter; the four-syllable form is mirrored by NV) and the logically curious acc. pl. megdane. Since Mustajbey seems to be describing a group action rather than a separable series of duels, I have rendered megdane as the collective singular, “war.” Note that HB modifies the usual rhythm of the second colon in order to try to accommodate the extra syllable, right-justifying the metrics by placing stress on the penultimate syllable of the line: the -dan- of megdane. See further Foley 1990: 95-102 for a discussion of the typically stressed ninth position in the South Slavic decasyllable, as well as the note to line 563 above. HB here sings megdane, while NV transcribes as mejdane. The regularity of NV’s spelling of mejdan for HB’s megdan argues for an orthographic habit and, from the point of view of his transcription of HB’s performance, lapsus calami. See further lines 773, 782, 785, 786, 799, 814, 826, 840, 856, 857, and 978, together with NVR.

594. See the AF.

595. See the note to line 602.

597. HB says monci here instead of the expected momci, and NV does not normalize. This nonstandard form actually appears four times in the performance (597, 599, 600, and 896) and is always transcribed as monci by NV, one more than the standard form (576, 880, 1008, always transcribed as momci by NV).

599. See the notes to lines 272 and 597.

600. See the note to line 597.

601. See the AF.

602. NV transcribes both ’fala (127, 720, 1002) and ’Vala (here in 602) as fala. HB sings prijetelji, while NV first writes prijatelji (restoring the standard form of the stem but not the expected voc. s. ending in -u) and then prijetelji, with the a overwritten by e. HB apparently deflects –u > -i under the proximate influence of tebi and glavni, a relatively frequent type of lapsus linguae. Cp. line 595, where HB sings Prijetelju and NV transcribes as Prijetelju with no attempt at restoring the expected form. I have added “but no” at the end of this line to ease the narrative transition from Mustajbey’s response to the actions that follow.

603. HB sings kAj! whereas NV transcribes as A. Though phonologically minimal, the illocutionary difference between the two is appreciable: Aj! is an interjection marking the onset of action (here the first of a pair of such performance keys in lines 603-4), while A is a simple conjunction (“And, But”). Cp. lines 496 and 542-43, with notes. Notice the very unusual performative [k] here, and compare the sequence of [n]-performatives in lines 606-7. On the traditional idiomatic implications of line 603, see the note to line 484 and the AF, gloss to line 603.

604. See the note to line 211. NV writes and crosses out izvodiše (“they led out”) before mirroring HB’s prifatiše.

606. The [n] performative is relatively infrequent. Cf. negbe in the next line, and the note to lines 529-30. On this Boundary line, see the AF and lines 502 and 730.

607. HB sings negbe (< hegbe), while NV transcribes as heljbe, restoring the most common form (as attested in Š). See further the note to lines 529-30 and 606.

608. The conventional pronunciation of prijatelje, as against prijetelji (line 602, with note), illustrates the occasional volatility of vowel sounds in actual performance, before reduction in many editions to a mythical (that is, textual) uniformity.

609. See the note to line 63.

610. In this line there is no need to use a vocative (pseudo-nominative) form of stari svat to eke out the colon metri causa because HB has resorted to the pseudo-nominative Osmane; cp. lines 487 and 554, with notes.

611. HB sings z redom, whereas NV transcribes as redom. This very slight difference, a consequence of lapsus auris, derives from HB’s use of a preposition (sa > z before a voiced consonant plus the instrumental case) to generate the meaning “in order” versus NV’s hearing of the disyllable as simply the instr. s. of red (“order”), which has a lexicalized adverbial sense of its own (“one after another, in order”).

612. HB sings zad’u (“rear”) and NV transcribes as zadu, failing to restore the standard form zadnju. Cp. line 508, where HB sings najzad’i (“last of all,” cognate with zad[nj]u) and NV does restore the expected form najzadnji.

614. “Pack-horse”: a sajisana (< seisana, Š) is a horse assigned the specific burden of carrying the bride’s trousseau in a wedding procession. Zlata’s procession is grand enough to require twelve such beasts of burden and to have the redoubtable Tale in charge of them. NV does not correct HB’s sajisana to the expected form seisana either here or at line 687; cp. the same situation with sajisane/saisane at line 682. See further the notes to lines 46 and 544.

615. In a departure from his usual practice, NV does not delete the performative and restore the expected form ruha but transcribes as ruva (after HB’s ruva), presumably on the basis of intra-colon rhyme with Suva.

616-22. On the Fiancé and future mother-in-law pattern, see the AF.

618. As at line 546 (see note), NV does not recognize the [j] as a performative and mistranscribes as Jali.

619. Lit., “Bećko stayed with his [future] mother-in-law.” I have made similar adjustments for clarity in the translation over the next three lines as well.

622. Lit., “A son-in-law must show himself [to be like this].” With lines 622 (šte) and 628 (štje), both deflected from se, HB seems to feel the pressure of proximate palatalization. Both instances may also involve lapsus linguae with the near-homophone šta. NV transcribes as se in 622; see further the notes to lines 624 and 628 and the AF.

623. Lit., “When Bećko proceeded farther.” I take the više (“more” or “farther”) as indicating the distance that Bećirbey has to ride to catch up with the rest of the wedding party, and therefore translate “When Bećko made up the distance.”

624. This is a difficult line. HB sings Na (“at, among”), whereas NV transcribes as Za (“behind”). In the second part of the line HB sings je vupustijo, whereas NV transcribes as je uputijo. I take NV’s rendering of uputijo as a correction made via lapsus auris, and interpret HB’s (and NV’s uncorrected) je as lapsus linguae for se (cf. the notes to lines 622 and 628). Restoring the verb phrase as the reflexive se uputijo, or “set out,” we can then make sense of both HB’s Na svatovim’ (“among the wedding attendants”) and NV’s Za svatovim’ (“behind the wedding attendants”). Having stayed with Zlata’s family to present them with gifts after the main wedding party had left (lines 616-22, rounded off with a proverb), the bridegroom Bećirbey (here and elsewhere named by the two-syllable alternative, Bećko) must catch up with his comrades (line 623: više polazijo; see the note to line 623) before he can set out either “among” or “behind” them.

625. I take napr’jed as a syncopated form of naprijed, shortened metri causa. The other possibility is to see it as a palatalized form of the ekavski equivalent, napred, but on the basis of the rest of HB’s singing idiolect this latter interpretation seems less likely.

627-722. On the scene of Blind beggar solicits the guest-allies and its incremental, intra-scene units, see the AF.

628. HB sings Kad štje šedi, whereas NV transcribes as Kade sjede. Given the preceding line and the consequent likelihood of a terrace (Kad se dade / *Kad se sjede) as well as prior, proximate deflections of se (šte at 622, je at 624), NV’s “hearing” is a clear case of lapsus auris leading away from the singer’s actual performance. HB here sings šljepac, while NV transcribes as the standard, initially unpalatalized form sljepac. Over the course of this performance HB vacillates between šljep- (with back-palatalization) and sl(i)jep-, whereas NV regularly transcribes as sl(i)jep-, thus restoring the expected form in every case. This is another indication of the volatility of phonology in actual performance as opposed to the standardization typical of texts and often of transcriptions of performances. Here and elsewhere I have translated š/sl(i)jepac and its oblique forms as “blind beggar” rather than in the literal sense of simply “blind man.” The justification for this augmentative adjustment is more than locally contextual: blind beggars are stock figures in South Slavic epic and would be recognized as such by an audience fluent in the traditional idiom. See further lines 630, 636, 642 (with note), 649, 653, 658, and 703.

630. The asterisks (*u*) designate an extrametrical element, a genuine syllable that temporally precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is a [u]-glide run-up to the initial sound in Oba; NV typically leaves [u] untranscribed. See further the notes to lines 1 and 628.

633. See the note to line 63.

634. See the note to line 63.

636. See further the note to line 628. HB sings za, whereas NV transcribes as sa, devoicing [z] > [s] and restoring the standard form; cp. further lines 649, 670.

637. See the note to line 292.

641. A nine-syllable line, stemming from a three-syllable first colon; in comparison with similar lines (e.g., 635), 641 seems to lack a proclitic (A, Pa, Kad, etc.). Nonetheless, the line is not hypermetric because HB uses a vocal rest in the first metrical (and musical) position in the line. NV has supplied a first vocal syllable (A) that HB does not sing. See further the note to line 1.

642. HB sings slijepac and produces an eleven-syllable line, while NV transcribes as sljepac, reshaping the line as a decasyllable. Performatively, HB compresses the seven-syllable second colon, slijepac govoraše, into the musical and rhythmic space of six syllables. See further the notes to lines 63 and 628.

646ff. Although line 646 identifies the next Turkish leader as Osmanbey, lines 650 and 652 call the same person Mustajbey. This discrepancy seems a forthright example of a singer’s “error,” no doubt contextualized and rendered negligible by a knowledgeable audience. Given the specific portrait of the figure – specifically, his heroic decorations and the medals awarded by the tsar (647-48) – the “correct” identification is probably Mustajbey.

647. HB sings bjevo, with performative [v] and deflection of terminal -u > -o (perhaps under the anticipatory influence of vordenovi), whereas NV transcribes as bjehu, deleting the performative and restoring the standard form.

648. On Tsar’s medals as a metonymic attribution, see the AF.

649. See the note to line 628.

652. HB sings Tad, whereas NV transcribes as A. This very slight difference (“Then” versus “And, But”) emerges via lapsus auris and illustrates the principle of right justification: metrical units, both cola and whole lines, are more variable at their beginnings than at their ends. The pattern [proclitic] zavika [noun-epithet formula] serves as a traditional sign marking direct discourse, so that the usual semantic force of “shout” is lexicalized to something less dramatic. See further the AF and Foley 1990: 96-106.

653. A clear instance of lapsus linguae, with HB singing za (“for”) and NV transcribing as na (“on”); NV hears the “correct” preposition, given both the sense of the prepositional phrase and the inflection of putu. See the note to line 628.

658. See the note to lines 628 and the AF.

661. NV first transcribes as Vazda, then crosses that out and writes A za, mirroring HB.

662. HB sings Šta, whereas NV transcribes as Da. In this instance both words mean “So that” and can introduce imbedded sentences, so there is very little to choose between them (only dialectal preference, a criterion that has questionable status within epic singing registers, which are conventionally multidialectal [see Foley 1999a: 76-80]). In this case lapsus auris leads to a virtual equivalent. Later in the line, HB sings bude vazda zdravo, whereas NV transcribes in rearranged order as vazda bude zdravo, a difference in idiolectal order. No prosodic or word-placement rules favor either configuration (see Foley 1990: 98-99, 176-78), so it is easy to understand why HB and NV would order the elements differently.

663. On the difference between HB’s Sta ( < Šta) and NV’s Da, see the note to line 662. Both singer and transcriber (re-singer) use a terrace-based or pleonastic structure in lines 662-63, but each prefers a different beginning to the embedded sentence. In the second colon HB sings lišće, whereas NV transcribes as lišce, not entirely restoring the expected form lice but leaving the back-palatalization intact even as he reverses the palatalization of [c] to [ć]. Again a case of lapsus auris, this particular “rehearing” tolerates one register-specific deflection but not the other.

664. HB sings jI, whereas NV transcribes as Ti. In other words, HB uses a simple conjunction to link the last two lines in a four-line capsule (661-64) while NV remakes the line by explicitly specifying the implied second person sg. subject. Lapsus auris thus leads to two different interpretations that are nonetheless comparable. Immediately afterward, HB sings mu (with the relatively rare [m] performative), whereas NV transcribes as u, showing no confusion with the homophone mu. See further the note to line 156.

665. Within the traditional register a “word” (reč/riječ) designates a unit of utterance: never less than a phrase, and more often a line, scene, speech, or entire tale. See further Foley 2002: 11-21.

667. A madžarija is “a golden Hungarian coin that can be found in circulation as a dukat even today [1979], and which is used as a female adornment” (Š).

670. See the note to line 628 and the AF.

672. HB sings t’ uzdrž’o, whereas NV restores the elided syllable and transcribes as te uzdržo. If NV’s reading were to stand, the line would be eleven syllables long and hypermetric, but HB clearly reduces te to t’, and does so metri causa. Note also that the influence of unvocalized [e] modifies the vowel of the first syllable of uzdrž’o from [u] > schwa; cp. the melding of sounds in Homeric synizesis, which is also at root a metrical, performance-based adjustment.

674. The [n] performative is relatively infrequent. See the note to lines 529-30.

679. See the AF.

680. Lit., “Gather, ban, hurry that we go!” See the note to line 373.

682. See the note to line 614.

683. HB sings bogazde, whereas NV transcribes as bogaze, deleting the excrescent [d] and restoring the expected form. The noun occurs in both feminine (bogaza) and masculine (bogaz) versions in the standard language, but because both would take an acc. pl. in –e we cannot determine which gender is implied here (and its recurrence in line 688 provides no further information; but see the notes to lines 684 and 688). NV’s transcription of krivu amounts to a misinflection based on taking bogazde as a feminine acc. s. or construing krivu from kriviti se, with the verb stem with e as its root vowel. What HB actually sang – krivi – appears to be a deflection from krive, paralleled in his singing idiolect by dole > doli, gore > gori, and especially bile > bili (see line 564, with note).

684. HB sings Preka (for Preko), whereas NV transcribes as Deka. This interesting case of lapsus auris involves NV’s “recomposing” the line using the verb dekati, an archaic folk term meaning “to urge a horse onward by [shouting] the word de” (V) and construing prvu as its object (presumably the horse). There is also possible influence via acoustic patterning with majku. But HB sings Preka (“through”), and its object prvu, though here in the accusative rather than the expected genitive, refers back to the first of the bogaze, which we may now understand as feminine (see the note to the previous line). See also lines 546 (with note), 618, and 896. At the end of this line NV does not restore the standard form psuje but mirrors HB’s ‘suje.

687. See the note to line 614. See further the notes to lines 46 and 544.

688. HB sings bogade (acoustically reflective of napade in the leonine-rhyme position), whereas NV transcribes as bogaze, replacing the excrescent [d] with the expected lexical [z]. See further the note to line 683.

692. Lit., “Both legs, then, on one side.” The asterisks (*u*) designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is a [u]-glide run-up to the initial sound in Oblje. NV leaves [u] untranscribed. See further the note to line 1.

695. See the note to line 406.

696. This suggestion displays Vide’s and Baturić ban’s ignorance of Tale’s characteristic irascibility and tightfistedness; they apparently have no idea how foolhardy and dangerous such an approach will prove. The traditional audience knows Tale’s character, however, and it can be assumed that they expect the evasion and thrashing to come.

703. See the note to line 628.

704. Since NV departs from his regular interposition of a j to indicate an intervocalic glide in this position (he writes osvoijo in place of osvojijo), the difference must be ascribed to lapsus calami.

709. NV first transcribes as A, then crosses that out and writes Pa. See further the note to line 63.

710. HB sings heksera for the expected eksera, adding what is apparently a non-performative [h] to the beginning of the word, and NV retains it. Given the fact that many Turkicisms appear in the guslars’ singing dialect in forms with initial [h], with some occurring variously in both aspirated and unaspirated forms (though not ekser, whose only alternative form is jekser [Š]), we can explain heksera as the result of analogy. See further the notes to lines 46 and 544.

713. The asterisks (*u*) at line-beginning designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is a [u]-glide run-up to the initial sound in Osta. See further the note to line 1. HB also uses [u] as a grace note in the second colon (kabuanica), which does not qualify as a genuine hypermetric since the extra syllable is accommodated within the usual melody; cp. Huoće and vovakuo at lines 130-31. This second [u] in 713 seems to be linked performatively to the run-up [u]-glide at the start of the line: cp. the other eight instances of kabanic- (in acc. s. and gen. s.) in this performance (629, 638, 643, 651, 660, 675, 701, and 721), none of which includes the grace note and none of which occurs in a line with an extrametrical start. NV transcribes neither [u] sound in 713, restoring the expected form kabanica and ignoring the extrametrical syllable. See further the note to line 93.

716. Presumably, these are the four corners of Baturić ban’s cloak, the garment on which he was accepting donations. Tale thereby manages to appropriate the ban’s financial holdings as well as his clothing.

720. See the note to 602.

723-24. Since the arrival is told so telegraphically, I have filled out the description by specifying some understood subjects of verbs. Negotiation of this terse account would of course not prove as difficult for an audience steeped in the epic tradition, since they know the story and are fluent in the compositional idiom.

723. HB sings Mezeva, with a faulty gen. s. or pl. (or acc. pl.) inflection, whereas NV transcribes as Mezevo (acc. s.) indicating motion that is not an explicit part of the syntax of the original expression (lit., “When the first ones were at Mezevo”). The syntactically proper inflection would be dat. s. (locative), but neither HB nor NV construes the phrase as such. See further line 728.

724. NV first transcribes as polja (lapsus linguae stimulated by the preceding word, Polju) and then corrects to HB’s pola. The first two syllables of pritisnula are inaudible on the DAT.

726. HB sings Namještili, whereas NV transcribes as Namjestili, deleting the superfluous palatalization and restoring the standard form via lapsus auris.

728. See the note to line 723.

729. HB sings Saštevili ( < Sastavili), while NV transcribes “Kad su bili,” a logical alternative that is mirrored elsewhere (550, 564; cp. 489). In effect, HB has deviated so far from the expectable pronunciation that NV hears an entirely different – though compositionally quite viable – hemistich. But note that HB’s Saštevili fills the first colon with a verb syntactically and nearly semantically identical to the second-colon iskupili, thus setting up a syntactic and semantic terrace between cola that is dissolved by the mistranscription. This is a prime example of the transcriber’s lapsus auris, which this time produces a lesser poetic construction.

730. On this Boundary line, see the AF and lines 502 and 606.

733-37. Note the volatility in the palatalization (733) and then nonpalatalization (734) of djevojke / devojku, apparently under the influence of palatalization or its lack in the two respective lines. In line 737 HB returns to djevojku, a heavily palatalized line overall. NV transcribes as djevojk- throughout this passage, but cf. his devojku (for HB’s djevojku) in line 770. See further the note to line 272.

734. See the notes to lines 272 and 733-37.

737. See the note to lines 733-37.

738. HB sings trubeta, whereas NV transcribes as turbeta, undoing HB’s metathesis (lapsus linguae) and restoring the expected form.

740. HB sings noj, apparently an idiolectal form, whereas NV transcribes as noć, restoring the expected form (noć’ < noću via apocope, metri causa).

744. See the note to line 740.

745. The code ## marks an initial vocal rest of two syllables. As explained above (note to line 1; cp. also 93), this is not a metrical flaw and therefore not a hypometric line. The first two syllables of the line are filled by instrumental music (thus preserving the decasyllabic format). The translation assumes a “formulaic emendation” (Kad se), based on, e.g., line 580 and on NV’s addition of A kad (via lapsus auris) at the start of the present line. On this Dawn marker line, see the AF.

749. The five syllables of vA bedevija are contained within the (four-syllable) melody of the first colon, making the line hypermetric but accommodating the extra syllable traditionally; NV does not transcribe the first syllable, reshaping the line as a regular decasyllable. HB’s construction is inspired by the very common tectonic figure of terracing or pleonasm, a partial repetition of the second colon of the preceding line in the first colon of the next line. The paratactic structure of terracing here overrides syllabicity. Additionally, the final element, svezana, is spoken especially quickly, as if to match the unusual rhythm of the first part of the line; cola are not seldom interactive in performance.

751. Lit., “Ej! In running she overtook a swallow.”

756-57. The accumulation of nominative endings in place of the expected accusatives suggests that HB sees this construction – direct objects of ima in a figure of apposition – as virtually copulative. It also argues for the primacy of parataxis: since the silna vojska of line 756 is (wrongly) nominative, it remains so in the terraced line that follows (with the same implied syntax and consequent inflectional error, and with topovi also incorrectly inflected as nom. pl.). NV transcribes as silna vojska in both lines, and as topovi in the latter.

758. The guslar does not specify which Turk first spots the messenger, but leaves the attribution impersonal and general (and 3rd person s., though I translate as “they”), a common enough narrative strategy.

759. This Position change line (see the AF) works idiomatically with the preceding line to fill the narrative gap between the description of Baturić ban’s camp and the delivery of the ban’s challenging message to Bećirbey. The scene shifts via the convention of a traditional signal.

760. The formulaic phrase knjigu šarovitu does not recur within the ŽBM, but is ubiquitous in the South Slavic oral epic tradition. On the traditional structure and meaning of Speaking via written letters, see the AF, gloss to lines 760-76, 777-82.

762. Lit., “Where is the bey [Bećirbey] of Mustajbey of the Lika?”

764. Lit., “Let him be a džebelija to the ban!” A džebelija is an armed cavalryman (Š).

766. HB sings štade, whereas NV transcribes as stade, deleting the palatalization and restoring the standard form.

767. HB sings da njemu, whereas NV transcribes as dade mu. Since both phrases mean “gave to him,” NV’s substitution amounts to a singer’s equivalent (lapsus auris) without appreciable semantic change.

768. HB sings Ka’ što ( < Kao što, “As” [conjunctive]), whereas NV transcribes as Kad, što (“When, that”?). Here lapsus auris leads to an untenable sequence that defies ready construal.

770. See further the notes to lines 272 and 733-37.

771. Lit., “You’ll never [go] beyond me living,” with živa (“living”) modifying mene (“me”).

773. The performance of this line has produced a chain reaction, starting with HB’s deletion of lexical [h] from Hodi to produce ‘Odi and continuing with the generation of a [u]- glide run-up that prefaces the (consequently unprotected) initial vowel in that verb. The asterisks (*u*) at line-beginning designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. NV does not transcribe the [u] sound, restoring lexical [h] to produce the expected form Hodi. See further the notes to lines 1 and 592.

774. HB sings zvalu, whereas NV transcribes as žvalu, restoring the lost palatalization and the expected form.

775. See the note to lines 74-75.

778. Lit., potpis means “signature” rather than “reply.” This usage may reflect Bajgorić’s (and his tradition’s) lack of firsthand familiarity with the technology of letter-writing, or possibly his sense of a signature as the metonymic equivalent of reply, or both. Cf. the generalized definition of the verb potpisati as “napisati što ispod čega” (“to write something below something else,” MS). See further Foley 1999a: 1-5 on the “signs” inscribed on Bellerophon’s tablet (Homer’s Iliad, Book 6, lines 166-80). On cultural variation in reading, see Foley 2002: 65-77.

779-81. Lit., djido/-a identifies “a mischievous one, a lively young man; a fanatic who calls forth either admiration or reproach; a hero” (Š). Other usages (348, 812) seem entirely positive (and are consequently translated as “stalwart hero”), but I construe these two instances in lines 779 and 781 as less than honorable, even mocking addresses of a foe who has already insulted the speaker, and thus choose to emphasize the negative aspect by translating “hothead.” This is a rare instance of a word with two clearly different, even opposite connotations in the epic register. Cf. the double meaning of xeinos as “stranger” and “guest” in Homeric epic diction.

781. HB sings hA si (“But you”), whereas NV transcribes as Hasi. In fact, Hasi is not viable; in this situation, lapsus auris leads to an outright error: NV fails to recognize performative [h] and misconstrues the opening two syllables. Instead of djida, as HB has it with nom. s. inflection and NV transcribes without correction, we would expect the vocative (djido, as in line 779). The ending in –a may be influenced by the acoustic patterning in [a] throughout the line.

782. See the note to line 592. NV uses a semantic equivalent, izaći, to transcribe HB’s izići. Cf. lines 786 and 799, where HB sings (and NV writes) izaći.

783-88. On the pattern of Securing a parent’s blessing, see the AF.

783. See the note to line 63.

785. The word babu is clearly lapsus linguae for banu, an inadvertent echo of the immediately preceding babo. Interestingly, NV recognizes the slip – or simply “hears” the expected form – and, correcting the singer’s “error,” transcribes as banu. See further the note to line 592.

786. See the note to line 592.

787. Š glosses vajir-dovu (hajir-dovu) as “a prayer to God for someone’s well-being or good fortune.”

788. See also the note to line 373.

789. HB sings pritte, and NV does not correct to the expected form pritiče.

790. NV crosses out an initial V before restarting transcription of this line with Iz. This sequence shows that he is in fact hearing the performative [v], but that after a moment’s consideration he understands it as not belonging to the lexical unit and therefore as ineligible for transcription.

792. Again the usually extrametrical vEj! is contained within a regular decasyllable; cf. line 2, e. g.

793. HB sings vu’vatijo, and NV does not restore the standard form uhvatijo.

797. See the note to line 292.

799. See the note to line 592.

800. HB sings prifati, and NV does not restore the standard form prihvati.

802. The interjection more, glossed here as “I order you,” is a complex signal connoting the speaker’s superior status or greater seniority (Š). In this case Bećirbey, eager to join battle with the ban over the maiden Zlata, is asserting his status as Mustajbey’s heir and Djerdelez Alija’s commander.

804. Lit., “And Mustajbey of the Lika began to shout.” I insert here a “meanwhile” that I take as implied by the paratactic form of the traditional narrative, although there is no explicit lexical basis in the South Slavic. Such intrusions are made very rarely, and only when a transition would otherwise seem too abrupt for a reader unfamiliar with the tradition; they are always marked in the notes to avoid confusion. NV transcribes as lički only after writing and crossing out beže three times. This is an interesting instance of lapsus auris, one that illustrates the power of the formulaic phrase as a “word” (see Foley 2002: 11-21): lički Mustajbeže and beže Mustajbeže are the two main noun-epithet phrases used to name Mustajbey in the South Slavic oral epic tradition. Within the ŽBM the various combinations and inflections are as follows (vocatives serving as pseudo-nominatives are also labeled “nom.”; all phrases are second-colon formulas unless otherwise indicated):

  • lički Mustajbeže (nom.), 18 occs.: 158, 217, 244, 252, 294, 301, 407, 521, 532, 558, 586, 595, 601, 650, 652, 804, 954, 962
  • ličkog Mustajbega (gen.), 3 occs.: 150, 606, 762
  • ličkom Mustajbegu (dat.), 1 occ.: 443
  • beže Mustajbeže (nom.), 2 occs.: 292, 477
  • bega Mustajbega (gen.), 1 occ.: 193
  • Beg Mustajbeg (nom., colon 1), 2 occs.: 151, 468
  • beg Mustajbeg lički (nom.), 3 occs.: 190, 225, 943
  • silan Mustajbeže (nom.), 1 occ.: 908
See further the note to line 908. On the traditional meaning of Mustajbey’s noun-epithet formulas, see the AF, gloss to lines 51-99.

810-69. On the pattern of Single combat, see the AF.

810. The scene shifts abruptly from praying and sacrifice to Bećirbey’s confrontation with Baturić ban. Given the preceding line, the subject of this line is probably Bećirbey, but HB is not explicit. See also the AF.

812. See the note to lines 779-81.

814. See the note to line 592.

815. See the note to line 484.

816. See the note to line 800.

817. HB sings med’jeda, and NV does not restore the standard form medv(j)eda.

818. Lit., filova = “of elephants,” therefore “ivory.”

821. Note the metathesis from the expected bedeviju to debeviju, perhaps the most obvious lapsus linguae in this entire performance. NV corrects the singer’s “error” to – or hears it via lapsus auris as – the expected form bedeviju.

826. See the note to line 592.

829f. Here and at lines 834 and 854 one or the other of the combatants “dances” his horse around before charging (or intending to charge). What seems to be happening is that a rider must “dance” his mount backward an appreciable distance before turning (see lines 835, 860) and proceeding to attack. See further the AF, gloss to lines 829, 834, and 854.

829. HB sings vA, while NV transcribes as Ja (“I”) via lapsus auris, specifying the implied subject of ću. Either possibility is grammatically viable.

830. The asterisks (*i*) at line-beginning designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is an [i]-glide run-up to the initial sound in Ja. NV does not transcribe the [i] sound (lapsus auris). See further the note to line 1. HB’s junak’ (deflection from voc. s. junače, minus the inflectional –e, effectively defaulting to nom. s. junak as a homophone, metri causa) is in the classic decasyllabic position (syllables 3-4) for direct address; NV does not restore the full form.

833. HB sings Bećirbeg ostade, whereas NV transcribes as Bećir bega stade, which in addition to the morphemic splitting is grammatically untenable as a phrase. Lapsus auris leads here to an outright error.

836. NV’s underlining (Bala) seems to indicate uncertainty about the verb balati (“to dance,” therefore “to feint” [MS]). HB sings koplje’, swallowing the inflectional [m], whereas NV restores it via lapsus auris by transcribing as kopljem.

838. HB sings Stadijaše, whereas NV transcribes as Šćadijaše, deflecting the verb away from the acoustic reality on the DAT. Cp. lines 201 and 205, where HB sings Šćadijaše/Šćadijavu and NV transcribes as Šćadijaše/Šćadijahu. NV seems to be hearing these verb forms in the same way – within his own singing idiolect – no matter what HB sings. The second syllable of golub and the first syllable of pobjegnuti are inaudible on the DAT.

840. See the note to line 592.

843. Lit., “You have trampled enough of these things with your legs.” HB sings ti, whereas NV transcribes as si. Either the pronoun (that HB sings), which seems more idiomatic in this situation, or the second person sing. of the verb “to be” – here as the past auxiliary – (that NV writes) is a viable interpretation; the variance emerges via lapsus auris.

845. NV aspirates (Manu > Mahnu) via lapsus auris; either form is viable.

846. See the note to line 817.

847. The asterisks (*u*) at line-beginning designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is a [u]-glide run-up to the initial sound in Ona. NV does not transcribe the [u] sound. See further the note to line 1.

849. HB sings vA, whereas NV transcribes as Tad. This slight shift from the conjunction A (lit., “And, But”) to the adverb Tad (lit., “Then”) is idiomatically negligible, since such line-beginning words all have about the same force in epic performance. Additionally, the earlier part of any metrical unit in the decasyllable (whether first colon, second colon, or the whole line) is more prone to variation than the later part of the unit, so we expect NV to “recompose” most frequently at these particular sites; see further Foley 1990: 96-106 on “right justification.” Cp. the many instances of HB’s vA > NV’s Pa (lit., “Then”); e.g., line 63, with note.

850-51. HB sings [naide] twice over these two lines. In both cases it is the sense of the ongoing narrative that differentiates the [n]-performative plus verb (najde) from the acoustically identical verb with elided negative particle (n’ ajde). Note that NV does not transcribe either [n], another sign that these sounds are serving as performatives. See further the note to lines 529-30.

854-55. On the two instances of performative [m], see the notes to lines 156 and 446.

854. The code ## marks an initial vocal rest of one syllable; the line is thus not hypometric. See further the note to line 1. NV restores a syllable metri causa and divides the syllables differently, transcribing Pa odigra as opposed to HB’s elided form of that verb (’digra). This rather complex instance of lapsus auris thus produces a very different “hearing” with approximately the same meaning as HB’s original.

855-59. On the role of vilas (“mountain nymphs”), see the AF, gloss to the Single combat pattern, lines 810-69.

855. See the note to lines 854-55.

856. HB sings čudnaga, probably under the influence of the penultimate syllable in megdana, that is, the [a] in the stressed ninth position, while NV restores the standard čudnoga via lapsus auris. We may construe the genitive phrase as the object of an implied Evo (“Here is”) or Eto (“There is”). See the note to line 592.

857. See the note to line 592.

861. Note the lapsus linguae in the second-colon phrase vimlu vučimijo ( < hinlu učinijo, with added performatives and nasal substitutions), possibly stimulated by the bilabials earlier in the line. NV transcribes as hinlu učinijo, restoring the expected form of the phrase via lapsus auris.

865. Given the plural subject, the verb should be zapucaju ([historical] present) or zapucaše (aorist). HB sings zapuca’, shortened metri causa, and NV mirrors the form. See the notes to lines 46 and 544.

866. The term salkum designates “lumbard” cannon; see further the note to line 386. The archaic terminology for cannon is not entirely recoverable. NV’s underlining indicates his uncertainty about this word (no alternate form of the Turkicism is given by Š).

867. Lit., “Here there was good fortune for Bećirbey.”

869. Lit., “The hero seized [a position] under [the range of?] the cannon.”

871. Instead of hearing HB’s hEj! as the common interjection plus performative [h], NV transcribes with an alternate interjection, Hej.

872-79. On the Heroic martyrdom pattern, see the AF.

872. Again I have inserted a “meanwhile” into the translation to bridge the scene change (see the note to line 804). HB sings buljubaše, while NV transcribes as buljukbaše via lapsus auris; Š lists these as equivalent forms of the Turkicism.

873. HB sings Pa, whereas NV transcribes as Ha (equivalent to hA?), a very slight difference caused by lapsus auris. See further the note to line 849.

874. Lit., “Wealth to he who....” The expression “Blago + dative,” here elided, is used as a familiar Moslem blessing. In this situation it clearly designates a spiritual wealth; thus the translation of “Eternal reward.”

875. HB sings ćabulu, whereas NV transcribes as kabulu, undoing the shift from [k] > [ć] via lapsus auris and restoring the standard form. Note that kabul is customarily indeclinable (Š) but bears a dative s. ending here, syntactically parallel to onome in the preceding line. NV’s rendering of HB’s ne vokrene as neokrene both deletes the performative [v] and affixes the proclitic ne to the main verb.

877. Note the lapsus linguae of intrusive [l] between prefix and root, apparently under the influence of the final [l] in the same word. NV’s transcription of izišle restores the expected form, removing the extraneous [l] via lapsus auris.

879. HB’s ‘vaćaju (deflected idiolectally < hvataju) is almost exactly mirrored by NV, who devoices the initial consonant. Note that NV first writes na after Što, then crosses out na and transcribes as faćaju. Performatives can vary from one instance of a word to another, since their function is to bridge hiatus and avoid a glottal stop rather than to participate in the lexical reality of any given word. Here (with šenite) and at lines 1007 and 1010 (both with šejitu), we see alternative deployments of performatives: the customary form is šehit-, varied once with a substitution of [h] > [n] and in the other two instances of [h] > [j]. NV normalizes all three to šehit-.

880-96. See the conventions associated with Perspectives from the battlefield in the AF.

880. Lit., “And the young men appeared (showed themselves) on their feet.” The form vuvuriše ( < zuhuriše) is unusual: the internal shift of [h] to [v] is common enough in the singer’s idiolect, but the initial consonant shift from [z] to [v] is, to my knowledge, unparalleled throughout his repertoire. It may constitute an acoustic linkage with the intra-word [v] substitution and therefore lapsus linguae. Here NV first transcribes as uguriše, then crosses the word out and writes zuhuriše over it, a delayed variety of lapsus auris. The first attempt, nonsensical in itself, shows that he initially heard the [v] as a performative and assumed it was added on before the first syllable. The second – and correct – attempt overrides both [v]-performatives in favor of the standard form of the verb, which requires a restoration of [z] in place of the initial [v]. See also the note to line 63.

881. HB sings sa četiri strane, whereas NV transcribes as na četiri strane. Via lapsus auris NV corrects HB’s apparent lapsus linguae and restores a formulaic phrase that recurs three more times in this performance and which makes good idiomatic and traditional sense. The assault thus strikes out toward (na) rather than from (sa) the four compass-points, just as do letters (305, 370) and cannon (388).

883-84. These lines, which are linked grammatically and rhetorically by a terrace structure, are sung as a couplet, without the usual interlinear instrumental break.

883. HB sings topovim’, whereas NV transcribes as topove, the latter a misinflection (although a misinflection that achieves metricality without apocope) that arises via lapsus auris. Note the terrace with the following line: na topovim’ / Na topovim’, with the latter member of the pair correctly transcribed by NV.

884. The word toprakmarim’ (assumed nom. s. = *toprakmar) is uncertain; NV’s underlining (toprakmarim) reveals his uncertainty as well. I base the translation “trenches” on the near-equivalent formula, “na topove i na toprakale,” cited in Š under toprakala (fem. s., apparently a collective noun), meaning “the trenches where cannon are dug in.”

885. HB sings Da, whereas NV transcribes as A. This instance of lapsus auris makes for a substantially different syntax, with HB constructing a dependent clause and NV recomposing the line as an independent clause.

886. The code ## marks an initial vocal rest of one syllable; the line is thus not hypometric. See further the note to line 1. NV supplies (apparently metri causa) an initial I that is not on the DAT, perhaps partly under the influence of the glide that begins jOvako.

891. Lit., “Watered/soaked the trenches with blood.” See also the note to line 884. NV’s underlinings (napojijo) indicate uncertainty, although it is not clear of what sort. The first o may correspond to the overwriting of j just afterward, and the second o may be connected with a faulty sense that toprakmare is the (plural) subject instead of the object of napojijo.

894. “Lit., “O dear God, unordered (chaotic) sadness!”

895. See the note to line 42.

896. See the notes to lines 42, 546, and 597.

897. See the note to lines 74-75.

899. HB sings odkitile, whereas NV transcribes as okitile, deleting the excrescent [d] (itself intruding on the basis of a hypothesized but nonexistent *otkitile with voicing of [t] > [d]) and restoring the expected form.

900. NV’s underlining (napojile) registers uncertainty, but its source is unclear.

901. HB sings po (“half”), whereas NV transcribes as pa (“then”). Both interpretations make for tenable lines, with a difference in meaning generated via lapsus auris. HB literally sings, “But when it was half four hours” (could be interpreted as either two or, more likely, three and one-half hours), while NV recomposes as follows: “But when it was, then, four hours.”

906. This line indexes how desperate the wounded hero is: he will strike out blindly against anyone who approaches, even his loving spouse. See further the AF, gloss to lines 880-906, on the pattern of Perspectives from the battlefield.

907. A rather abrupt switch of scene from the horrors of the battlefield to the relative comfort and safety of Mustajbey’s tent occurs without an explicit signal for transition. Part of the audience’s (and capable reader’s) fluency in the epic register consists of the ready contextualization of such transitions by reference to prior, parallel shifts. Cf. lines 804 and 872, with notes.

908-19. On the Darkness/mist lifts pattern, see the AF.

908. There seems to be some instability in HB’s identification of the Turkish leader responsible for the maiden Zlata and associated activities from this point onward in the song. One symptom of that instability is HB’s assignment of Osmanbey’s customary epithet (silan, “powerful”) to Mustajbey (NV mirrors without indicating any confusion); see the table of Mustajbey’s epithets in the note to line 804, as well as the note to lines 943-44 on an in-progress correction from Mustajbey to Osmanbey. (This is just the sort of potential confusion that textualization would locate and eliminate, but which matters little within the traditional performance arena.) At any rate, it is well to remember that Mustajbey is the leader of the composite force, the bridegroom Bećirbey’s father, and the bride Zlata’s mušterija (“patron”), while Osmanbey, a very important leader in his own right, is the appointed starosvat (“eldest witness”) for the marriage procession and event. I transcribe and translate exactly what the singer said, without any attempt to “correct” or normalize his performance.

909. The clipping of hladan to ’ladan illustrates how initial [h] is especially vulnerable to deletion in the singing register. NV does not restore the expected form.

910. HB sings Da’, which NV transcribes as Da but I take as a reduction of Daj (“give, grant” in the imperative 2 s.) on the basis of the preceding line, with which it forms a terrace (daj mi – – Da’ mi). HB’s [da] could also be construed as Da (“So that”) with a dependent clause embedded in the first imperative sentence, but the traditional, additive structure of the terrace seems to be the determining factor here; see further the note to the next line. Also in 910, HB sings pu’ne, whereas NV transcribes as puhne, restoring lexical [h] and the standard form via lapsus auris; cf. the same restoration at 915.

911. The code ## marks an initial vocal rest of one syllable; the line is thus not hypometric. See further the note to line 1. NV supplies an initial A not on the DAT, metri causa and via lapsus auris. Note that this line begins with Da (“So that,” here translated as “To” for simplicity), which was the other possibility for construing Da’ in the preceding line.

914. The asterisks (*u*) at line-beginning designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is a [u]-glide run-up to the initial sound in Od (mediated by an untranscribed semivowel, [w]). NV does not transcribe the [u] sound; he first writes Kod, then crosses it out and writes Od. See further the note to line 1.

915. See the note to line 910.

917. On this Simile, see the AF.

920-22. This word for “tombstone” (turbe, -eta) shows evidence of being construed as a collective noun in HB’s epic dialect. The dat. s. ending in -u in 921-22, where a plural is clearly meant and elision of the dat. pl. form is not employed, is unambiguous evidence of that grammatical flexibility. The form in 920 may also be grammatically singular but plural in force. Cf. the other three usages in this performance, all grammatically plural, at lines 522 (partitive gen. pl.), 727 (acc. pl. object of u), and 743 (acc. pl. object of u).

920. See the note to line 63.

922. Lit., “Then there was no one at the tombstone(s).” HB sings nikog, whereas NV transcribes as niko (with no indication of elision), either deflecting the gen. s. object of nema to an untenable nom. s. or, more likely, simply swallowing the final consonant (cf. the instability at line 836).

923. See further the note to line 272.

926. NV transcribes the first word as Samo (“Only”), whereas HB clearly sang vAma’! from Aman, an expostulation meaning “Mercy!” or “O God!” (Š). Lapsus auris here yields a major change.

927. The lexicalization of vA as a line-initial marker with little real semantic value is never more evident than here. To bridge the logical gap between lines 926-27, I insert “Because.”

929. See the note to line 269.

930-31. Lit., “Where [have they taken] your one hundred young men?” to justify the acc. s. stotinu. In the next line the implied structure is copulative, with the nom. s. plemenita Zlata.

932. HB sings Tad, whereas NV transcribes as A. This kind of variation of proclitic monosyllables at the beginning of metrical units is, as noted above, very common. See further the notes to lines 652, 849, and 927; more generally, the note to line 63.

933. See the note to line 292.

936. The word stotin’ is inaudible on the DAT.

936-37. NV’s representation of pe’ stotin’ and similar numbers varies (cp. lines 384 and 399, with notes), although the sound-image on the DAT is consistent. In some areas, especially numbers, NV’s orthography is itself unstable.

938. A phonologically unusual second colon, with intrusion of excrescent [k] and performative [v] in the common formula dva puna sahata (cp. line 5, with note); NV repairs the lapsus linguae by transcribing as sahata via lapsus auris. In regard to the initial colon, NV first transcribes the second word as sve (the correct rendering audible on the DAT), then crosses it out and writes se. Given the sense of braniti, which means to defend a place or thing rather than to defend against something or someone, versus braniti se (reflexive), which means to resist, I render as sve, understanding HB’s articulation as lapsus linguae with excrescent [v] for se, perhaps under the influence of anticipated sve ( = “all”) in the following line.

939. Note the relatively rare performative [n]. The fact that NV does not transcribe the [n] is further evidence that it is a performative rather than the elided negative particle (n’). See the note to lines 529-30.

941. Only the first two syllables of this line are audible on the DAT.

943-44. This pair of lines seems to represent an in-progress correction: HB sings a line introducing the lamentation, first assigning it (incorrectly) to Mustajbey and then “revising” by resinging it with the correct attribution to Osmanbey. NV transcribes without intervention. See further the note to line 908.

945. The word mili is inaudible on the DAT.

947. HB sings bilu, whereas NV transcribes as bili. Here lapsus auris leads to a grammatical error, perhaps under the influence of end-colon or leonine rhyme (gori / bili; see also the note to line 25).

948. Even a proper name is subject to accommodation with performatives: HB sings Valila, whereas NV restores the standard form Halila via lapsus auris. Cp. HB’s Valil in line 950, which NV transcribes as the expected form, Halil; also HB’s Valija at 1003, transcribed as Alija. Contrast HB’s Halile (mirrored by NV) at line 952, where [h] is not replaced by performative [v], presumably because the preceding Mujagin ends in a consonant; as noted throughout NVR, the combination of an ending vowel before word-initial [h] regularly behaves like vowel hiatus and calls for performative [v].

949. NV hears [ve] as performative [v] plus a reduced vowel, and transcribes on, the appropriate 3rd s. masc. nom. pronoun. The correct reading is ve’, a reduced form of već, plus ran’e’ ( < ranjen; cp. the full form in the following line, making for a kind of terrace or anaphora). HB sings lječi, and NV does not restore the standard form lježi. See also the note to line 948.

950. See the notes to lines 948 and 949.

952. See also the note to line 948.

954. HB sings jest’ ( < jeste), whereas NV transcribes as jes, failing to restore the full, expected form. This case of lapsus auris also turns, however, on the colloquial nature of jeste, which is often apocopated to jest’ or jes’ even in everyday speech.

957. Note the relatively rare performative [l].

959-61. Alija’s strange appearance – the combination of his blackness and nakedness – may be attributed to the energy and speed of his headlong pursuit of Baturić ban and the captive Zlata. The apparent contradiction between the horse’s lack of a mane (961) and its being called a “long-maned bay horse” (958) is only superficial: Alija’s horse (and other hero’s mounts) are regularly indexed by this noun-epithet formula (e.g., line 22), which as an idiomatic and traditional identification does not conflict with the singularity of the immediate context.

960. HB sings Nem’ (Nema with elision, metri causa), and NV transcribes as that same form without marking the elision. HB sings ruba, which I translate as “raiment” (V glosses ruba as roba) rather than lapsus linguae for ruha (“clothes, garments”), which is NV’s initial interpretation before he corrects to mirror HB. Also, NV restores the expected form haljina from HB’s valjina, with performative [v], via lapsus auris.

962. NV’s transcription of initial Kad makes contextual sense (“When. . .”), being just the sort of adjustment or emendation a singer or well-informed editor might make at this point. But HB sang Sta (depalatalized Šta), which I translate here as an interjection. See further line 350, with note.

965. The performative [m], left untranscribed by NV, is relatively infrequent. See further the note to line 156.

966f. It will be noticed that the action involving Djerdelez Alija versus Baturić ban begins as a chase viewed from afar by Mustajbey and company, then modulates without explicit notice into a directly experienced scene in which narrative and geographical distance is dissolved. So complete is the foreshortening that dialogue between the combatants becomes present and audible. Once again, the fluency of the traditional audience would smooth over what we might perceive as a narrative disjunction.

966. The asterisks (*u*) at line-beginning designate an extrametrical element, a genuine syllable that precedes the melodic-rhythmic pattern of the line, and thus from the point of view of performance does not render the line hypermetric. It is a [u]-glide run-up to the initial sound in Oklen. NV does not transcribe the [u] sound on the DAT. See further the note to line 1.

968-69. Note the relatively rare performatives in [m], neither of which is transcribed by NV. See further the notes to lines 156 and 446.

968. HB sings mA vone, whereas NV transcribes as A on. The excrescent schwa attached to the pronoun on seems to function to ease articulation, that is to avoid the sound [ng] created by bridging from on to goni. Like other intra-line excrescent sounds, it is not a full syllable, but rather a grace note; NV does not transcribe the schwa. See further the notes to lines 93, 156, and 969.

969. Similarly to line 968, HB sings mA von and NV transcribes as A on. See also the note to line 156.

975. Although Alija is not explicitly named in this line, I make the attribution explicit on the basis of the preceding and following action.

977. Lit., “And from there blood poured out.”

978. See the note to line 592.

981. Lit., “And he came then to hand-to-hand combat.” See also the note to line 984.

982. NV first transcribes as Jali, then crosses that out and writes Jami, mirroring HB.

984. An unusual line, with inflections uncertain (probably affected by intra-line assonance). I insert “Aiming” to smooth the translation. The doublet prsi/prši, here palatalized (but cp. line 981, e.g.) indicates phonological instability and reactivity to context; cf. the note to line 40. In this line HB sings prši, whereas NV normalizes to prsi. Cf. the proximate unpalatalized instances at lines 981 and 990, all transcribed as prsi by NV. As for Alija’s nakedness, see the note to lines 959-61.

988. Lit., “And your lead [will] bend back here.” Djerdelez Alija is boasting that bullets cannot penetrate his chest, that he is impervious to attack by Baturić ban’s pistols. HB sings jo’dje, and NV transcribes as odje, failing to restore the standard form ovd(j)e.

989. NV first writes zagrmiše, then crosses that out and transcribes as zagrmlješe, mirroring HB.

990. See the note to line 984.

993. This is a challenging line. Taking njome as an error for njemu, uždi from užditi (which via V = užeći, or Lat. accendo, here understood in the figurative sense of “launch” [V, q. v. and SAN, q. v.]), and puštimice from pustimica (“wooden missile,” V, here understood as referring back to nadžačina in the preceding line, the deflection from the expected acc. s. pustimicu motivated by the ever-present tendency toward leonine rhyme, on this occasion with njome), I translate as indicated.

995. Lit., “And placed him onto the green grass.”

999. On this Position change line, see the AF.

1001. The ## code marks a vocal rest within the melodic line. HB has elided the first syllable of Bećirbeže, but the line is not hypometric. See further the note to line 1. NV supplies the missing syllable via lapsus auris and transcribes as the expected Bećirbeže.

1002. See the note to line 602.

1003. See the note to line 948.

1007-19. On the pattern of Burying the dead and gathering the wounded, see the AF.

1007. See the note to line 879.

1008. Note the relatively rare usage of [m] as a performative, with the choice possibly influenced by the immediately preceding noun momci. NV does not transcribe the [m], showing no confusion with the homophone ma. See further the note to line 156.

1009. HB sings Šta vu rake (“So that in [their] graves. . .”), while NV transcribes via lapsus auris by making a formulaic substitution: Časkom rake (“Right away the graves. . .”).

1010. See the note to line 879.

1011. The performative [nj] in njizkupiše is unique in this performance; palatalization of [n], itself a performative, apparently occurs because of the high front vowel that begins the verb proper and in harmony with the medial sound in ranjeno. Note also the voicing of [s] > [z], and cf. line 1007. NV transcribes typically by deleting the performative [nj] and devoicing [z] > [s], thereby restoring the expected form.

1012. HB sings vA, while NV first writes A and then crosses out that transcription (correct except for performative [v]) and substitutes Pa, a virtual performance equivalent, via lapsus auris. See also the note to line 63.

1013. I insert “rigged with” to smooth the translation. The litters were apparently constructed by stretching cloth between two parallel lances, with the resulting stretcher-like apparatus then borne by four men, presumably one at each corner of the rectangular litter.

1014. HB sings meć’u (syncopated < mećaju) and NV mirrors his usage.

1016. NV deletes performative [v] from vO!, as expected, and substitutes an idiolectally alternate interjection, Aj. HB inserts a performative glide [j] between vO! and Odoše; NV takes the [j] as the final sound in Aj. HB sings votlanen, which derives from otale(n) (“from there,” cf. odatle), whereas NV transcribes as otalen (with underlining), the standard form of the adverb rendered via lapsus auris.

1017. I insert “in battle” to fill out the translation.

1018-19. On the variation between hiljada and viljada in these two lines, see the note to line 269.

1019. Here, as in many other places throughout the performance, the semivowel [w] at the end of the articulation of [u] (here in mu] acts as a natural performative to avoid hiatus. Since these do not qualify as excrescent consonants, I have left them untranscribed throughout. See further the notes to lines 46 and 544.

1020-29. On the Wedding and family coda pattern, see the AF.

1021. NV omits na, apparently via lapsus calami.

1026. Either an extremely rare usage of [d] as a performative or, more likely, lapsus linguae under the alliterative influence of djece. NV “hears” the expected sound and transcribes via lapsus auris as the expected form rodijo.

1028. See the notes to lines 46 and 544.

1030. This closing line or Postlude (see the AF) is marked by a change in tune and cadence similar to the ritardando effect at line 459, the last line before the singer’s break. In its generic function of closure and its wide applicability to a variety of songs, this kind of line serves as an “opposite bookend” to the similarly generic pripjev or “prologue” that one finds at the opening of many South Slavic oral epics. On the pripjev, see further Foley 1991: 67-75.