Some time ago, in a comparative study of European balladries, the problem of the melodies halted me till Professor Cdsar Barja told me that my problem had already been solved by Professor Sigurd B. Hustvedt, who sent me a short time later his valuable pamphlet, A Melodic Index of Child’s Ballad Tunes (Univ. of California Press, 1936). I have made an extensive use of his system, having introduced a few modifications, and venture to think that my experiments may serve as (so to say) laboratory confirmation. His business has been with English and Scottish ballads; mine has been international.Order now
Briefly speaking, Professor Hustvedt’s proposal is as follows. Melodies are composed of notes named by the letters of the alphabet CDEFGA B, and they rise and fall by semitones. He proposes that we write down those letters, and number the semitones. The first process transcribes
the tune; the second describes it; and from the description we may assign its place in a catalogue.
It will save time and the heat of useless controversy, if it be at once understood that these proposals are strictly auxiliary. The staff notation is doubtless the proper form for definitive musical publications, just as the association of a trained musician with a literary researcher is a sine qua non for an ambitious publication. This is, for instance, how John Meier’s magnificent Deutsche Balladen pursues its stately way through the press. But there are many purposes for which collaboration is not needed or feasible. The staff notation is expensive, elaborate, and not understood—alas!—by so many of us. Some simpler notation, provided it be accurate, may have many advantages in auxiliary work.
I. Transcription.—The notes named by letters are arranged in octaves, of which the principal one rises from ‘middle C.’ We need some means of distinguishing the various octaves. Professor Hustvedt’s proposal is that the octave from ‘middle C’ shall be in brevier lowercase type, with brevier capitals for the octave below, and nonpareil lower case for the octave above. I have not cared to use these distinctions because they are too technically suited to the printing press. One’s cards will be either handwritten or typewritten, and it seems desirable that the symbolism should be one common to script, typescript, and print. A typewriter has redandblack or redandblue ribbon, and I have for some time used the red for italics. The same effect is got in script by underlining. I can get five octaves from the series: large capitals (underlined in typing, doubly underlined in script), small capitals, roman, italics, italic capitals (red in typing, underlined in script). A roman lowercase c is ‘middle C,’ and my octaves run:
C—B c—b c—b c—b C—B
These notes sufler some accidental modifications. They may be af fected by sharps, flats or naturals. Professor Hustvedt uses the usual musical signs for this purpose, placing them before the letters. They are written before the note on the staves, but they are pronounced after the letter in current practice. One says ‘B flat’ not ‘flat B.’ Hence one might misread such a formula as ‘CbB.’ Also, I do not find these signs on my typewriter; but I do encounter an acute accent, which is an evident equivalent of a sharp, a grave accent for a flat, and a tittle for naturals.
Hence 6 ic are, for me, ‘middle C sharp, flat, and natural,’ respectively. It is a convenience to place these signs immediately above the note they modify, because we find that further signs are needed in East European balladry to define the notes. The system of fixed intervals is Western.
In the Balkans a note may be a little above or a little below the Western semitone. This can be shown by placing a vertical tick above or below and immediately following the letter: c1 and Ci are, then, notes a little above or below ‘middle C,’ and they may be sharped or flatted if necessary. It is not possible to do more than this much; these quartertones and thirdtones of the Byzantine and oriental tradition occur in passages too fast to allow of exact determination. An ‘r’ means ‘rest,’ whatever the length of the rest.
Thus we have the names of our notes. We have still to transcribe our other information: key, rhythm, length of notes. We do so: the key may
be, say, ‘1 sharp,’ and the rhythm threefour time. These numbers may be placed side by side: 34. In the complicated rhythms of the Balkans, it may be advisable to separate the integers by a line: as 9/16 or 11/16. The ictus is marked by inserting vertical bars among the notes.
The most ingenious suggestion made by Mr. Hustvedt has to do with the indication of length of notes. He says we may pick whatever unit is most convenient, and represent that duration by the space occupied by one letter. The most useful units in practice are crochets (U4) and quavers (U8). The unit selected will be the one which gives the briefest transcription of the melody: if there be no note shorter than a crochet, then one will transcribe in crochets, not quavers, etc. If we are transcribing in quavers (U8) and encounter a crochet, we may double the dura tion of our note by placing a point after it; a semibreve will then be shown as a letter plus three points. To halve the unit of duration, place notes in brackets. In comparing some Castilian and Andalusian settings of the same tune I have occasionally found it useful to use double brackets to indicate quartering the unit of length. The difference between the two styles is that the one is solidly faithful to the melody, but the other includes some swift trills.
The trills are swifter still in Balkan lyrical melodies, and they are also improvised. Sometimes the notes cannot be caught at all, but only the beginning and the end of the trill; at other times the series used on any one occasion is only generally similar to that used on another. One may indicate these passages by drawing a line above them, from the first letter to the last. The sense of that line will then be, that the passage is an improvisation passing rapidly from the the first substantial note to the last, and that either the intervening notes arc not known, or arc of a certain general nature. These Balkan difficulties, however, do not afflict the Occidental researcher. He does occasionally encounter old melodies which have no fixed rhythm. He may not be able to draw in the bars, or the bars may be variable; in the latter case, they may be indicated by a broken vertical line or by a colon (:).
II. Description.—Melodies rise and fall by semitones, which may be counted. The horizontal movement may be taken for granted so that all we want is a means of estimating rise and fall. From the letters of the alphabet we obtain the means of counting 24 falling semitones (excluding I and O, which are liable to be confused with numbers), but we are not likely to have business with more than the 12 that make an octave. We need not represent monotony, and so may use zero as meaning ‘ten.’ Mr. Hustvedt gets twenty rising semitones by using larger and smaller types. For the reason already given I prefer to use roman and italic numbers, or black and red ribbon, or plain and underlined written, numbers. So my rising semitones are 1—0 1—0, and my falling semitones are A—Z.
As this formula indicates only the relative positions of the notes, it is independent of key, pitch, and rhythm the chief variables of music. It is thus well fitted to give always the same description of tunes essentially the same, whatever their casual differences. The first complete musical phrase usually suffices to characterize a melody.
III. Cataloguing.—Professor Hustvedt uses his melodic formulas to make a catalogue, placing rising semitones before falling ones, and small
numbers before large. Thus his catalogue runs 11 … 12 … 13 . . etc., 1A . . . IB . .. , etc., A1 . . . , and so on.
IV. Musical phrases.—The letters of the Greek alphabet may be used to denote musical phrases, and may be combined to give the formula for
the succession of phrases in a tune. I propose to give some examples of the sort of demonstrations thatmay be made by means of the above symbols. They arc, as we have said strictly auxiliary. There is therefore no need to transcribe the wholecorpus of ballad melodies before launching into a particular investiga tion. The whole corpus is indeed available: whatever can be successfullyset down by the musician on staff notation can be transliterated by these symbols.
1. They help us to form an idea of one of the differences which are fundamental in European balladry. European ballads arise out of the lyrical aspects of narration or the narrative aspects of lyrics. In the music this corresponds to an antithesis between recitative and song. The recitative, strictly narrative style holds for heroic balladry in Russia, Yugoslavia, and Greece, and also for Spain. Song is found in the women’s songs of those countries, and in the balladries of France, Italy, Germany, Scandinavia, and England. In the recitative style, we find that everything is subordinate to the enunciation. This is pitched relatively high, with one syllable to each note, along a natural wavy line of notes;
frequently the singer’s voice rises initially to reach this level, and still more frequently it falls after the last accented syllable. Notes are not the same from line to line; only the general conditions remain the same.
When Rjabnin recited “Dobrynja i Vasilil Kazimirov” to Gilferding, he began several lines with a rising ‘cd’ to reach V and continue about thatWhen Rjabnin recited “Dobrynja i Vasilil Kazimirov” to Gilferding, he began several lines with a rising ‘cd’ to reach V and continue about thatlevel until his last abrupt descent to ‘bb’; but he began other lines where he meant to continue, with ‘ff’ or *fe.’ A transcription shows clearly thefall of the ictus in the verse, which makes the very varying lengths of the lines approximately equal. Russian chanting is, however, too illknownto be a safe guide. Möller, in Das Lied der Völker (vol. n, Serb 29) gives this opening for the famous Serbian ballad, “Kosovka djevojka”:
2 fiats 24 U8</rba/rbag/b.a./b .. . /bbab/cbab/a.gr/gbab/cbab/a.gr/ ,<f.c./b.a./c.b./a.g./a. . . (gfg)/a . .. /
The second and third lines may be taken as normal. The line is a trochaic decasyllable, with a caesura after the fourth, and a major stresson the ninth syllable. These divisions appear very clearly in the music, where each syllable has its note; the ninth is of double length, and thetenth is followed by a rest. The singer has in mind a definite pattern, which is, however, essentially high clear speech. He maintains his linesat the level of ‘b’ (with ‘a’ and V as relief), and descends to ‘g.’ He is not bound strictly to the notes. No two lines are the same, and in the firstand fourth he applies lengthening at his own discretion. Very unlike this style is that of the women’s improvisations. Though not restricted tofixed sequences of notes in the tune, they show great variations in the length of note, and in its attachment to the syllables, and generally thereis a wider range.
“Ko pije vino za slave Bo2je” (given by Karadiif, I, 156) may be shown as:1 sharp 24 U8(/.rbc.ba/b .. . di./dcbdcbc&ft). . . /
Similar differences could be noted in Greek songs. In Spain the narrative recitative style is represented at its simplest ina tune which Salinas (1577) gives as very ancient. It is for the ballad “Retrafda estd la infanta”:
There are other ballads not much more complex than this. Among the most naive of thirteenthcentury cossantes we must number MartimCodax’s “Ondas do mar dc Vigo”: yet the music shows at once that this is not narrative but song:
The most primitive tunes are those of the first m the first group, and have been illustrated above in the transcriptions of “Retrafda estd la in fanta” and “Rosa fresca.” (In “Rosa fresca” the tune for the second octosyllable is for seven syllables only. There is no question in this ballad of a‘paragogice’ to even up the syllables; it follows the Franco Italian rule that hemistichs must have different cadences. The ballad is one of many which proclaim their French origin.) In these primitive tunes the identical octosyllabic musical phrases are grouped in pairs by the sensepause which coincides with the assonance. That leads to dif ferentiation in the musical phrases, giving the other two couplet forms.
It is to this state of affairs in the ballads ‘old’ in his day that Nebrija referred in 1492, when he declared that they were essentially trochaic tetrameters. The second group arises by repetition, so that a/9 gives aa/9/9. By differentiation we get the musical quatrain; and it was to this
condition of the romance that Encina referred in 1496 when he proclaimed that the Spanish ballads were quatrains. The fashion was, perhaps, comparatively new. One associates it with the French quatrain chansons populaires and the passage in Portugal from the couplets of the old cossantes to the quadras. Everyone knows how the quatrain music tended to impose quatrain verses on the ‘artistic’ balladpoets of the Golden Age.
The last group is a rough transcription of Matos Flecha’s setting of “Roma abrasada” (sixteenth century). This is not traditional music, but an artistic composition for chamber concerts. All those musical phrases are requred to pronounce two octosyllables. The polyphonic style dragged the romance out interminably; it was therefore cut. Hence the Cancionero de Amberes (c. 1545) and other early printed books of ballads show many truncated pieces. The purchasers were not minstrels, but gentlemen wishing to exercise their throats. The cuts have lost us an incalculable amount of the original texts, but have also occasionally given such examples of lyrical felicity as “Conde Arnaldos” and “El