Source

For the Venda, song differs from speech chiefly because its words are arranged in a strict metrical framework. The spoken rhythm of the first word-pattern of a song may influence its basic meter, particularly in the choice of crotchet or dotted crotchet units, but in general the rhythms of ordinary speech are subordinated to the meter of the song. On the other hand, speech-tone has considerably more influence on the melodies of Venda songs than does word-stress on the meter.

Relative tone is an important element of speech tone: for instance, may mean 'head' or 'monkey', depending on its speech-tone pattern; and the speech-tones of a word may change according to its position in a sentence. Certain principles of speech-tone changes in Venda have been discussed in detail by Professor E. Westphal (1962). The Venda do not, of course, expect a song to sound like ordinary speech, and they generally understand the meaning of texts, even if the patterns of speech-tone are distorted by the melody. Patterns of speech-tone affect melodies not simply for the sake of intelligibility, but because the Venda seem to feel an aesthetic impulse to pay attention to speech-tone when setting words to music, in much the same way that a composer of Western classical music may draw inspiration for his melodies from the texts which accompany them.

Some Principles which Affect the Setting of Words to Music

(1) variations in speech-tone may create rhythmic variation in the metrical pattern by means of agogic accent (i.e. accents caused by a rise or fall in tone, especially in places where the melody is not accented)

The melodic and metrical pattern of the first word-pattern of a song is, as far as possible, reproduced in all subsequent word-patterns. But if the speech-tone rises or falls before or after the beat on which the melody usually moves, the melody may be adjusted to suit the speech-tone, and hence the original effect of the beat is changed by the agogic accent.

(2) speech-tone patterns influence certain parts of the melodies, chiefly the beginnings of phrases

This may be compared to the patterns in numerous Venda songs in which there are parts for solo and chorus. The melody of the chorus does not vary, but the soloist executes musical variations, that are influenced by speech-tone patterns, on a basic theme stated in the first word-pattern of the song. In a few songs, speech-tone patterns are sacrificed almost entirely for musical considerations, particularly in many modern songs composed for church and school functions.

(3) the principle of 'harmonic' equivalence regulates many of the alterations in melody that follow speech-tone patterns

Melodic changes from one verse to another may not be regarded as different melodies by the Venda, but as aspects of the same melody, which may be sung together or independently. The chorus of most Venda songs, especially story-songs, may be embellished by the addition of one or more parts, which move sometimes independently, but most commonly in parallel motion at intervals of a third, fourth, fifth, or octave. The movement of a second melody in parallel motion is not, however, always an exact replica of the main melody, as it must employ the same mode (note). For instance, the 'harmonic' equivalents of the pentatonic mode g' - e' - d' - a can be derived only from this series of tones, so that they could be as in the figure below:

Two series of tones which may be sung as 'harmonic' equivalents on the basic scale g' - e' - d' - a.

This evidence supports the view that the Venda draw upon scales for the formation of their melodies; if the music were purely melodic, without any relationship to a fixed 'store of notes' (Hornbostel 1928:34-35), there could be no objection to consistent parallel motion in fourths, with f' and b, rather than e' and c', sounding with c' and e' respectively.

Thus a may be sung instead of e' or d', and c' instead of g', in order to suit the requirements of the speech-tone patterns. It must be emphasised that although in such cases a falling minor third may replace a rising major third, and vice versa, or a falling fourth may replace a rising major second, such changes in the melodic line are not regarded by the Venda as musical variations: they are linguistic variations. When I disrespected these rules I was told that I was wrong and sounded like a Tsonga. I thought at first that my intonation was incorrect, and that my mistake was to descend on the wrong syllables of the text. If I observed these changes but sang some tones flat or sharp, my teachers said that I did not sing very nicely, but that I at least sang correctly. This was one of the many occasions on which I was reminded that relative pitch is more important to the Venda than absolute pitch. For this reason, I am more concerned with patterns of melody and am content in the present analysis with the approximate indication of pitch afforded by staff notation.

(4) descents in speech-tone and melody need not coincide, but a rise in speech-tone is generally accompanied by a rise in melody

The most common exception to this rule is found at those points where the chorus-like cadences of melodies begin: for instance, the speech-tones / - \ / - \ may be accompanied by the melody g' - e' - d' - c' - c' - a. It is not necessary for the music to rise to d' in accordance with the speech-tone, as there is a shift of centre in the melody - a phenomenon which has already been observed by Herzog (1934), Schneider (1944), and Jones (1959). The tones c' - c' - a are for the Venda, however, 'harmonically' equivalent to g' - g' - e, such that the sequence could be sung g' - e' - d' - g' - g' - e', in which case the melody agrees precisely with the speech-tone pattern.

(5) when a high speech-tone occurs at the beginning of a word-pattern, the accompanying melody usually begins on the highest tone of the tone-row on which the song is based

In many cases the highest tone is g' or e', but this does not mean that those same tones must accompany a consecutive series of high speech-tones: the speech-tones may remain high, whilst the melody descends by a step of a major second or a minor third. Similarly, a high speech-tone occurring later in a sentence may be accompanied by a much lower tone from the tone-row, because of musical factors, such as a shift in the centre of the melody.

(6) in songs whose tonal nucleus is e' - d' - c' - a a low speech-tone at the beginning of a word-pattern is almost always set to c', whether the high speech-tones are sung on g' or e'

Thus, / / - may be sung g' - e' - d', but in the same song \ / - is sung c' - e' - d', and not e' - g' - e'. There appears to be a good reason for this: most of the melodies lean towards the end of the phrase, with a tendency to descend as rapidly as possible. Thus, if a descending musical pattern is established with e' as its second tone (e.g. g' - e' - d' - c' - a), this pattern is as far as possible retained: if the first two speech-tones of a word-pattern are low-high, the second tone, coinciding with the high speech-tone, will remain e', and the previous tone will be adjusted to suit its low speech-tone (thus c' - e' - d' - c' - a). For similar reason the pattern \ / \ / is sung c' - e' - c' - d', and not c' - e' - c' - e' or e' - g' - e' - g'. Furthermore, it will be seen that if such word-patterns were sung in this second way, the melodies would never reach their cadence (very often d' - c' - a) in time, and the musical phrase would thus remain incomplete.

(7) most word-patterns of the songs end on a low speech-tone, and in many cases the pattern of the penultimate and last syllable is - \, ^ \, or / \

Similarly, the most common final cadences of the melodic patterns are falling minor thirds or major seconds.

(8) the melody of the first 'line' of many songs often corresponds closely to the speech-tone pattern of the words, which may even influence those parts that are subsequently unaffected by changes in speech-tone

Thus the pattern of the cadence/chorus may be initially influenced by speech-tone, although subsequently its musical characteristics predominate.

This suggests that the initial word-pattern shapes the melody and meter of the rest of the song, and may even influence the choice of tone-row. This system cannot apply, however, in songs that are dominated by the descending figure a' - g' - e' - d' - c' - g', which remains the same in spite of numerous changes in the speech-tone patterns.

Once the basic melodic pattern of the song is set, influenced partly by linguistic and partly by musical factors, it is probable that the melody and meter then exert some influence on the word-composition of subsequent 'lines'.

When the Venda compose songs, it seems that words and music interact in the following way: an idea framed in words is set to music that is partly influenced by the structure of the words; the music then influences the development of the original verbal idea.