ZULU AND XHOSA PRAISE POETRY AND SONG

 by DAVID RYCROFT 

Part of the following article has previously appeared in a paper presented to the Anthropology Section of the 124th Annual Meeting of the British Association for the Advancement of Science, held at Manchester, August/September, 1962.

The boundary between song and some forms of verse or declamation which are nowadays classified as "oral literature" is a blurred one and calls for co-operation between linguists and ethnomusicologists. Very little study has yet been made of musical characteristics found in the border-line art of praise-poetry or praise-singing which is practiced very widely throughout Africa. Among the Eastern and South-Eastern Bantu, what small evidence there is reveals considerable differences between the traditional style of delivery associated with Zulu izibongo praises,1 and that, for example, of the Heroic Recitations of the Banyankore, a Bantu people of East Africa. A. N. Tucker2 has recorded some of the latter and they were taken as the subject for a thesis by H. F. R. Morris.3 They are uttered with quite phenomenal rapidity and the overall intonation contour of each line is a gradually descending one, without the observance of fixed musical pitches.

Zulu practice, in the South-East, is totally different. In a number of recordings of iZibongo, by two different reciters,4 four recurrent levels of pitch resembling the "scale" of notes in a piece of music appear to predominate and might be said to serve as a basic tonal structure (in the musical sense) throughout the recitation. Rudimentary evidence of reciting conventions among other South-Eastern Bantu such as the Xhosa, Sotho and Venda seems to suggest that this four-note, quasi-musical style is practised only by the Zulu, in this area. Other peoples, from evidence available so far, seem to deviate less from normal speech when they recite praises. 

The four predominant notes used in the Zulu izibongo recordings mentioned, could be represented roughly by the Tonic Solfa symbols doh', te, sob, and Dob. Low Dob occurs only finally in a stanza and may tail off to lower, indeterminate pitch just as the final syllable fades into silence. Syllables taking one or other of the higher notes do not always main- tain absolutely level pitch. Glides to or from one or other of the notes, or from one to another, are frequent, but this is also the case in true Zulu song. 

My teacher, the late Dr. B. W. Vilakazi, left us with a long-standing riddle when he made the statement that "It lyric poetry was originally intended to be sung, then this quality of poetry still exists in Zulu. The poet has to tune his voice to some melody when he recites his imaginative descriptions".5 

He added the observation that tone in Zulu is "semantic" and that this "semanticism of tone, though wide in the spoken language is more apparent in the recitation of verse." From the last statement, one gathers firstly, that the "melody" in izibongo recitation does not violate the speech-tones. This certainly turns out to be true. There is no single, constant melodic sequence - pitch movement is conditioned by the words. 

But there are four possible notes, and Zulu, in common with most other Bantu languages, is a two-tone language (in the linguistic sense), using only two contrasting registers. Attempts to establish the upper two notes as variant realisations of High speech tones, while the lower two represent Low tones, prove fallacious. As I have described in some detail in an earlier papers, it finally became evident that only the highest note, doh', represented High speech tones. Low speech-tones take low Dob when final, but whether they take te or sob when non-final depends upon the initial consonant of the syllable. Unvoiced consonants require te for the syllable, while most voiced consonants demand sob. The process is operated quite mechanically by the Zulu reciter. There is, in fact, in the spoken language - as also in German and Chinese - an auto- matic voiced consonant/pitch-lowering correlation, as was first noted by D. M. Beach in 19247. This phonetic feature seems to be exploited and exaggerated in izibongo recita- tion. But this, no doubt, has a natural foundation in the fact that consonantal pitch- lowering has a more pronounced effect, in Zulu, when one talks at the top ot one's voice. From the recordings it seems that recitation takes place within a pitch range at least an octave higher than that of normal speech - judging by pitches used by the reciter when announcing the title of each izibongo. 

The validity of Vilakazi's claim that speech-tone contrasts become "more apparent" in recitation is borne out if we compare recited lines with normally spoken ones, as in Fig. 1. The upper transcription in each case shows recited pitch levels while the lower shows those of a normally spoken version of the same line. The multiplicity of pitches in the spoken versions results from the interaction of speech-tones, lowering consonants (underlined), and, thirdly, overall "sentence intonation" which, in most types of utter- ance in Zulu, confers progressive dropping of pitch, or "downdrift." 

In recitation, this normal intonation feature is entirely absent until the last two syllables of a stanza where, as the pitch drops all at once, a wide interval is traversed which provides an effective concluding formula. Avoidance of the normal downdritt intonation of speech, and the maintaining of fixed levels ot pitch like musical notes instead, is no doubt what Vilakazi had in mind when he referred to tuning the voice to "some melody" when reciting. He was himself a leading Zulu poet. 

Fig. 1 Extract from IZibongo zikaShaka 8

U-Shaka ngiyesab' ukuthi unguShaka

U-Shaka kwakuyinkosi ya.semashobe-ni.

Upper transcription:  Recited pitch levels (Mr. John Mgadi);
Lower transcription: Spoken pitch levels (Mr. S. Ngcobo).

Translation: "Shaka - I am afraid for thou art Shaka!
Shaka - There was a king amongst the cattle tails!"
(i.e., A master of the cattle raid he was!)

The question remains, however, whether or not Izibongo recitation should be regarded as a species of song. The sequence of pitches is certainly not a free, musically determined melody. Use of one or other of the four notes is conditioned directly by an interaction of speech-tones, consonants, and stanza finality. Linguistic determinism here appears to be absolute, and this state of affairs stands in distinct contrast to what happens in items which are clearly acceptable to true song. In traditional Zulu songs, speech-tones and consonants certainly have an influence on the melodic rise and fall, but musical requirements are also in evidence and there is give and take between the two. As Hornboste I stated of African Negro song, generally:

"The pitches of the speaking voice, indeed, appear to determine the melodic nucleus; but they have no influence upon its inborn creative forces; these forces, and not any qualities of speech, direct the further course of the melodic development"9.

In the old dance-song of the Buthelezi clan shown in Fig. 2, only four notes are used - or three and their octaves. But these bear no relationship to those used in izibongo and the way in which High and Low speech-tones are set to the notes is quite different. Hign speech-tones are not realised always on the highest note. In the first two words of the initial phrase, High and Low speech-tones do consistently take dob' and sob, respectively. But in the next word, High tones take sob, and Low tones take ray and, finally, Doh. In the final word of the men's part, speech-tones are melodically over- ruled: the sequence should properly be High-Low-High.


Fig. 2 Buthelezi (Zulu) dance-song.10
Translation: "They set him up for one month: then they deposed him.
                   He is getting old now! Father is getting old!"

It seems to be permissible in this and other true songs, for High syllables at various points in the line to be realised on almost any note within the particular "scale" in use, provided that one or more lower notes remain available for the setting of intervening Low syllables. Occasionally, especially at the end of a line, speech-tone requirements may be entirely over-ruled. The descending melodic line which is characteristic of all such Zulu songs gives a suggestion of affinity with the overall downdrift intonation of normal speech, while the iZibongo convention of consistently maintaining the pitch height of High syllables stands in distinct contrast both to song and to normal speech. The use of exaggerated concluding formulae is also peculiar to Izibongo.

Regarding metre, fundamental distinctions could be cited between practices in song - where length is often distorted mercilessly for metrical ends' - and in Izibongo, where such things as regular "feet" are not to be found, but rather the natural ryhthms of speech.

Izibongo and song differ further in rate of utterance. An appreciably greater number of words per minute are uttered in recitation than in any true Zulu song. Words, chosen for their imagery, sound and aptness, are the very core of iZibongo. Their pitch setting could be said to be somewhat mechanical, despite the fact that a series of notes is used which resembles a rudimentary musical scale. In song, on the other hand, words often convey little actual meaning. Lyrics generally consist of a few short phrases which are constantly repeated, with occasional interpolations. Musical expression is paramount.

From this it would seem that Izibong do not fit, conveniently into the category of Zulu song. From the linguistic point of view they constitute a form of speech utterance with its own special form of overall intonation - possibly comparable with forms of "monotonic chant" in other cultures, such as mentioned by George List in a recent article12. From a musical point of view, Izibongo are excessively word-bound, allowing no freedom to Hornbostel's "inborn creative forces" of the melodic nucleus.

In contrast to this borderline category of musically stylised speech we find the clear prose folk-tale within whicn short crystallised items of true song occur - though tne teller may at times slip almost imperceptibly from tne one medium into the other and back again. In Africa, as elsewhere in the world, the song within the folk-tale often has magical power. In a Xhosa tale from the Cape, the river monster, Sinyobolokondwana, steals the clothes of the twin sisters, Wele and Welekazi, from the river bank while they bathe. One of the sisters manages to get her clothes back from the monster by singing the required song:
. Fig. 3 Xhosa folk-tale song'3
Translation: " Sinyobolokondwana!
                  Give back my clothes!
                 Bhakubha is a long way off;
               Mother will give me a beating."

What became of the other sister, who refused to sing properly, is another story. Two transcriptions of this little song have been shown in Fig. 3 - first of an initial recording without any rhytnmical accompaniment, and secondly of anotner recording in which tne singer accompanied herself with regular hand-clapping.

Singing in Negro Africa very frequently takes place, as we know, against some rhythmical accompaniment - whether this be provided by instruments, dance-steps, hand-clapping, or merely the repetitive movements of some daily task. In such rhythmi- cally accompanied song it has been observed that it seems to be a widespread African habit for word-stresses to fall not on, but between the physical beats.

A perfectly natural physiological foundation for this suggests itself in the case of work-songs in which heavy muscular effort is called for, or in strenuous acrobatic dancing. It is, of course, an instinctive human reflex to tense the diaphragm and hold the breath, by closing the glottis, at the actual moment of maximum exertion - in tact, even babies do so during defecation. At the actual moment ot this instinctive breath-holding during pushing, lifting, leaping and the like, the emission ot vocal sound of any kind is, of course, impossible. But immediately before or after the moment of exertion - or both before and after it: "buk - - - aaab" - sound of some sort is not only possible, but very probable. A Zulu work-song with the unfortunately all too topical text, "They arrest us!", which is repeated ad infinitum, demonstrates this point clearly. A transcription of the first few lines of this song appears in Fig. 4. The beat, heard as a heavy physical thud whenever it is given expression, always just immediately precedes the beginning of the phrase, during vocal silence.

Ba-ya-a-si-bo -- pha-, Percussion (thud of shovels) Ba - ya-si-bo - pha - -- etc
 Fig. 4 Zulu work-song14

Among the Zulu and Xhosa peoples of the extreme south-east, instrumental ensembles are not used at all as a basis for dancing. Dancers sing their own dance music and, particularly with Xhosa dance-songs, there is what seems to be a subtly calculated off- beat relationship between word syllables and the regular dance-step and hand-clap rhythm. This may be seen in Fig. 5, as also in the second version or the song referred to earlier in Fig. 3.

Fig. 5 shows what my informant called a "sour grapes" dance song - which she had heard during wedding celebrations, and which she thought must have been composed by an old maid.

Fig. 5 Xhosa dance-song15
Translation: "How fortunate I am to be unmarried -
                  I can still follow my own inclinations!"


Here it will be seen that word syllables seldom exactly coincide with a hand-clap, and often fall somewhere between the beats. One gains the impression of a rather loose relationship between words and clapping. This "near miss" relationship is not hap- hazard, however, but seems to be repeated with exactitude with each repetition of the song. My own theory, put forward in an earlier paper,16 is that in Xhosa singing, instead of the best being made to coincide with the release of a consonant - into a vowel, so that the onset of the vowel is on the beat, as is our own practice - it coincides with the initial closing or thrusting movement of the consonant (when this type of consonant occurs) so that the commencement of the vowel invariably occurs later, a little after the beat. This effect appears to be further exploited and exaggerated for stylistic purposes, and closure of the glottis - necessary in strenuous exertion - could, of course, also take place on the beat, during the consonantal closure, if required.

In passing, it may be ot interest to observe that, in America today, one of the most highly paid singers of "pop" and cabaret songs - with their currently favoured gimmicks of off-beating and deliberately loose word-phrasing-is Miss Miriam Makeba17, a South African of Xhosa extraction, who played the leading role in the original production of the musical, King Kong. This feature of non-coincidence between words and rhythm is, of course, not confined to the Xhosa. Richard Waterman coined the expression "off-beat phrasing of melodic accents in relation to percussion metre"s8 to describe what he found to be a common characteristic in West African music, thousands of miles north of the Xhosa. Apart from this point of similarity, however, there seems to be very little in common between the musical practices of tne West Coast and those of tne extreme soutn-east, where there are no drums or spectacular percussion ensembles.

Since 1947, an invaluable rallying point for African musical studies has been tne African Music Society and, later, the International Library of African Music, which together have their headquarters near Johannesburg, under the directorship of Mr. Hugh Tracey. The Society issues a journal entitled African Music, and Mr. Tracey has conducted recording expeditions throughout a large part of Africa south of the Sahara. Long-playing discs of the field recordings are available from the International Library.

Founders of the African Music Society were a handful of white people in Africa who had grown to love indigenous African music and were concerned by the rate at which, in many parts of the country, this was being lost or diluted in the context of rapid social change and under the influence of imported Western styles. Rescue action in the form of a large-scale recording drive was envisaged so that these treasures might be pre- served. Such recordings, it was felt, should be given the chance to compete with foreign music in regional radio programmes and in the record shops. Should the present genera- tion of new African townsmen fail to be impressed, a body of authentic recorded material might still serve to inspire later generations who turn in search of their cultural heritage.

A "preservationist" attitude towards tradition is by no means widely held by those Africans who have deserted tribalism for a way of life they feel is more suited to the 20th century and who feel that music from their past is out of place. The raison d'etre of many of their traditional musical practices, interwoven as they are with social custom, is no longer provided in town life, or now institutions may pay the piper and hence call a new tune. Under the circumstances, however, they deserve hardly more personal blame than the Western man-in-the-street who relishes only "rock-'n-roll" and the "twist."

African musicians and scholars there certainly are, however, who do value their indigenous music. The eminent Ghanaian sociologist and ethnomusicologist, Professor J. B. Nketia writes:

"In contemporary Ghana, old and new forms of folk music exist side by side... For some time there has been a danger of... the older type of folk music being abandoned by literate and urbanised Ghanaians as Ghana gets more and more industrialised. Nationalism, however, is fostering a new pride in our tolk music, and efforts are now being made to preserve or encourage the practice of the best in the older type of folk music throughout the country"x9. 

NOTES: 

(1) For examples of texts with English translation, see E. W. Grant: "The Izibongo of he Zulu Chiefs", Bantu Studies, III, 1928, pp. 203-244. 

(2) Professor of East African Languages, School of Oriental and African Studies, University of London.
(3) H. F. R. Morris: The Heroic Recitations of the Banyankore, University of London Ph. D. thesis, 1957 (unpublished).
(4) James Stuart reciting: IZibongo !ikaSenZangakhona, Zonophone 4195; IZibongo ZikaSolomoni ka- Dini.ulu, Zonophone 4178, and IZibongo ZikaShaka, Zonophone 4175. (These, among other praises and folk-tales spoken by Stuart, were recorded in 1927) and John Mgadi reciting the I:ibongo of six of the Zulu kings, recorded on Gallotone GE 1001; GE 967, and GE 998.
(5) B. W. Vilakazi: "The Conception and Development of Poetry in Zulu", Bantu Studies, XII, 1938, p. 116. (6) D. Rycroft: "Melodic features in Zulu Eulogistic recitation", African Language Studies, 1, 1960, pp. 60-78. (London: Luzac & Co.).
(7) D. M. Beach: "The Science of Tonetics and its application to Bantu languages", Bantu Studies, II, 1923-6, p. 75.
(8) Diagram reproduced from D. Rycroft: op. cit., p. 67. Recording: Gallotone GE 967.
(9) E. M. von Hornbostel: "African Negro Music", Africa, I, Jan. 1928, p. 31.
(10) Recorded by Hugh Tracey, 1955. Issued on LP disc., "The Sound of Africa" Seies, AMA TR-12 (A, 1). Transcription reproduced from D. Rycroft: op. cit. 2, p. 26.
(11) See: D. Rycroft, "African Music in Johannesburg: African and non-African features", Journal of the International Folk Music Council, XI, 1959, p. 26.
(12) George List: "The Boundaries of Speech and Song", Ethnomusicology, VII, 1, January 1963, pp. 3-6. (13) Singer: Mrs. L. Nongobo Whyman, 1955. Recording: D. Rycroft.
(14) Singer: Mr. R. Kunene, 1959. Recording: D. Rycroft.
(15) Singer: Mrs. L. N. Whyman, 1956. Recording: D. Rycroft.
(16) D. Rycroft: "Stylistic Evidence in Nguni Song", paper at Symposium: Music and History in Africa and Asia, Royal Anthropological Institute, London, 1962.
(17) See, inter a/ia, LP recording "Miriam Makeba Sings", London Records, HA 2332.
(18) Richard Waterman: "African Influence on the Music of the Americas", in Acculturation in the Americas, ed. Sol. Tax, University of Chicago Press, 1952, p. 213.
(19) J. H. Nketia: "Changing Traditions in the Folk Music of Ghana", Journal of the Internat. Folk Music Council, XI, 1959, p. 34. * Khulwane means "big" and may be omitted. N.B.--The notation shows the nearest notes in the diatonic scale.