3. Composition
©2025 Ruth Finnegan, CC BY-NC 4.0 https://doi.org/10.11647/OBP.0428.03
How is oral poetry composed? Does the performer memorise a given text? Does he improvise in the heat of the moment? Does the poet rely basically on formulae learnt beforehand but combined in different ways on different occasions? Or is there no one single mode of oral composition?
These and similar questions have exercised many scholars and have caused much controversy. Because the nature of composition and the controversies concerned with it are fundamental to the study of oral poetry, it is appropriate to begin the substantive account in this book of the various facets of oral poetry by considering the processes and problems of composition.
3.1 Is memorisation the key factor?
On the face of it, memorisation would appear to be the correct description of what is involved. It seems that a singer or reciter going fluently through the delivery of a piece must surely have memorised it from already existing words, and that the piece is fully formed before he starts, only needing to be called to mind at the appropriate moment. The familiar model of a child learning off a poem or song and then performing it from memory at a competition or concert seems one that could reasonably be extended to cover the general relationship of the performer of oral poetry to the process of composition. He is repeating from memory a piece which has been composed prior to the performance, either by himself or, more likely, by others, perhaps years or generations earlier.
This interpretation is frequently used. It looks like a common sense view. In addition, it gains apparent support from the ‘folk theories’ of oral tradition discussed in chapter 2. If ‘oral tradition’ can be seen as something which grew communally and spontaneously or was inherited by its current bearers over a period of many years (or even centuries) then all the present-day oral poet has to do is to learn and remember this tradition. Even statements which avoid the extreme view of oral literature as a ‘fossil survival’ or ‘archaic relic’ from the past still give the impression that the contemporary performer does little more than passively receive and pass on material already formed. For example, if it is true that the Old Iranian epic was ‘handed down by men orally for some fifteen-hundred years’ (Boyce in Lang, 1971, p. 101), that Inuit literature was transmitted down from ‘days of yore’ (Thalbitzer, 1923, p. 117) or that Ewe poems in West Africa are ‘almost as old as the Ewe people themselves’ (Adali-Mortty in Beier, 1967, p. 3) then the present-day oral ‘poet’ need have little if anything to do with composition. Statements like this are widespread, and basically envisage the performer as a recipient of the oral tradition: his role is merely to memorise and deliver to contemporary audiences. As one analyst sums up this approach (in the context of the ballad), ‘Memorization … is the basic vehicle of oral tradition’ (Friedman, 1961, p. 114).
This view gains support from the statements of some local poets. Wakarpa, the old blind bard who recited Kutune Shirka, the famous Ainu epic about the Golden Sea Otter, ‘insisted’, according to Arthur Waley, that ‘he had merely repeated the epic as he had learnt it’ (Waley, 1951, p. 236). Similarly, Parry and Lord recorded statements by Yugoslav singers of oral heroic poems in the 1930s that they could repeat exactly the same song that they had heard from another singer, ‘word for word, and line for line’ (Lord, 1968a, pp. 27–8); and a Somali reciter often makes it clear to his audiences that the poem he is delivering was composed not by himself but by another named poet (Andrzejewski and Lewis, 1964, p. 46). Here, it seems, is definite evidence that the current performer is proceeding primarily by memorising an already formed text.
This view of memorisation and recall is still widespread. But as a scholarly theory about the basic process of oral composition it is now under fire from many scholars, and in specialist circles analyses in terms of ‘memorisation’ have become unpopular.
The critique has proceeded on three main fronts. First, there are the general doubts now held about the ‘folk theories’ of age-old oral tradition etc.; second, the growing awareness of the significance of current performance, and the effect of the audience and the context on the process of composition on a particular occasion—an awareness consolidated through our increased grasp of variability in oral transmission due to the extensive use of the tape-recorder; and third, a series of detailed empirical studies of actual oral composition allied to a wide comparative approach, resulting in the work of scholars such as Parry and Lord and the ‘oral-formulaic’ school of analysis inspired by their work.
The romantic background was discussed in the last chapter and doubts about the general validity of the underlying assumptions were raised there. The transmission of oral literature is taken up further in chapter 5. So beyond saying summarily that there is now much evidence leading us to doubt older assumptions that oral literature was naturally and inevitably formed through age-old transmission in unchanging form over generations, I will not pursue this particular line. The general emphasis on the active role of poet and audience in the examples and analyses discussed here in the context of composition, itself tends to undermine the plausibility of many romantic assumptions. More however needs to be said about the other two aspects.
First, one has to consider the effect which the situation may have on the performance. It is a striking characteristic of many performances of oral literature that the performer is affected both by his audience and by the occasion. These can affect his poem, sometimes radically, to an extent where one is forced to speak of composition by the poet/performer rather than memorisation, with minor variations, of a piece composed by others.
One aspect which can be affected by the nature and reactions of the audience is the length of the piece. Radlov’s account of the case of the Kirghiz singer of heroic poetry is often quoted:
The minstrel, however, understands very well when he is to desist from his song. If the slightest signs of weariness show themselves, he tries once more to arouse attention by a struggle after the loftiest effects, and then, after calling forth a storm of applause, suddenly to break off his poem. It is marvellous how the minstrel knows his public. I have myself witnessed how one of the sultans, during a song, sprang up suddenly and tore his silk overcoat from his shoulders, and flung it, cheering as he did so, as a present to the minstrel.
(Radlov, Proben, V, p. xix, translated in Chadwick, III, 1940, p. 185)
As Radlov says, the minstrel ‘knows how to represent one and the same picture in a few short strokes. He can depict it more fully, or he can go into a very detailed description with epic fulness.’ (Ibid., p. 182).
It is not just the length of the Kirghiz minstrel’s song that is affected by his audience and situation.
Since the minstrel wants to obtain the sympathy of the crowd, by which he is to gain not only fame, but also other advantages, he tries to colour his song according to the listeners who are surrounding him. If he is not directly asked to sing a definite episode, he begins his song with a prelude which will direct his audience into the sphere of his thoughts. By a most subtle art, and allusions to the most distinguished persons in the circle of listeners, he knows how to enlist the sympathy of his audience before he passes on to the song proper. If he sees by the cheers of his listeners that he has obtained full attention, he either proceeds straight to the business, or produces a brief picture of certain events leading up to the episode which is to be sung, and then passes on to the business. The song does not proceed at a level pace. The sympathy of the hearers always spurs the minstrel to new efforts of strength, and it is by this sympathy that he knows how to adapt the song exactly to the temper of his circle of listeners. If rich and distinguished Kirghiz are present, he knows how to introduce panegyrics very skilfully on their families, and to sing of such episodes as he thinks will arouse the sympathy of distinguished people. If his listeners are only poor people, he is not ashamed to introduce venomous remarks regarding the pretensions of the distinguished and the rich, and actually in the greater abundance according as he is gaining the assent of his listeners.
(ibid., pp. 184–5)
However much the Kirghiz minstrel has learnt from other poets—whether ‘handed down by tradition’ or not—it is clear how far he himself takes part in the process of composition in the performance. It is impossible to describe his poems, as actually delivered, solely in terms of ‘memorisation’.
The Kirghiz possibly lay more stress on improvisation in their heroic poems, with correspondingly less on memorisation, than some other peoples. But this receptivity to the expectations of the audience and the demands of the occasion is widely documented for oral poetry. The Xhosa imbongi (praise singer) in South Africa is often fired to compose and declaim by some event which he observes: he responds to the situation, not to a memorised or long-deliberated text. This is how the imbongi Nelso Mabunu describes the process
Some people think perhaps an imbongi sits down and studies. That is not the thing: it’s an inspiration. When you see something, you know, it’s like a preacher in church when he preaches the gospel, you feel touched, then you feel like saying some words yourself, you know—that’s an inspiration. It’s nothing else and it can be nothing else. You can judge a recitation, you know, done by school children, I mean by a school child, something that he has learnt and he’ll recite. But singing, you know, praises for a chief or anything, it’s an inspiration.
Again the free-lance Hausa praise singer of Northern Nigeria inclines his praise poems to the needs of his own pocket as well as the circumstances of his temporary patron. The wandering singer arrives at a village and carefully finds out the names of leading personages in the area. Then he takes up his stand in a conspicuous place, and produces a praise song to the individual he has decided to apostrophise. It is punctuated by frequent demands for gifts. If he gets what he wants, he announces the amount and sings his thanks in further praise. But if he does not, his delivery becomes harsher and the song becomes interlaced with innuendo about the ‘patron’s’ birth and status. Sooner or later the victim gives in, and buys the singer’s silence with a cash payment or valuable gift (Smith, 1957). Here too, the skill with which the singer adapts his song to the circumstances cannot be explained solely in terms of memorisation.
Similar instances abound, from the adaptations of traditional themes to the circumstances of the moment by Akan dirge singers in West Africa (Nketia, 1955, chapter 4, especially pp. 66ff), to the attempts of a folk preacher in the American south to arouse a bored audience by breaking into a rhythmic passage of the ‘Four Horsemen’—a theme unrelated to the original subject of the sermon (Rosenberg, 1970, pp. 68f)—or the compositions of the Black American singer ‘Left Wing Gordon’: ‘Wing’s blues were mixed and of wonderful proportions. He could sing almost any number of blues, fairly representative of the published type with, of course, the typical additions, variations, and adaptations to time and occasion’ (Odum and Johnson, 1926, p. 211). East European scholars, with their studies of the Russian byliny from the nineteenth century and earlier, were among the first to emphasise the creative role of the poet as an aspect of performance, and the improvisatory rather than memorising element of oral poetry. As the Chadwicks sum it up ‘on the whole we must regard the free variety, which allows more or less scope for improvisation, as the normal form of oral tradition, and strict memorisation as exceptional’ (Chadwick, 1940, III, p. 868).
The use of the tape-recorder to provide accessible copies of a large number of renderings has made us more aware of variability in detail. Up to a point, such variability has long been recognised. Child’s classic collection set an example by giving differing versions of each of the 300 or so ballads he prints. Mary Hamilton appears in at least fifteen different variants, and Barbara Allen in three (to which can be added a dozen or more collected by Sharp in the Southern Appalachians early this century (1932, I, pp. 183ff)). The same is true with the famous ballads of Mrs Brown of Falkland in North East Scotland at the end of the eighteenth century, accepted as classic renderings. She produced two versions of The Lass of Roch Royal, in 1783 and 1800 respectively, which differ enough to make it clear that recreation rather than exact reproduction was involved (see Bronson, 1969, pp. 69ff). From these and other examples collected by earlier folklorists, linguists and others, it became clear that variability of detailed text is common in oral tradition—in what is usually termed ‘folklore’.
This in itself weakens the memorisation theory. Variability has sometimes been accommodated in the theory and explained in terms of faulty memorisation. So variability could be seen as resulting from misremembered versions of some forgotten original. And indeed it is reasonable to attribute some variants to the fact that singers may have forgotten musical or verbal phrases and filled the gaps as best they could; and this is supported by the likelihood that literary pieces get distorted over time or space. The words and tunes of ballads ‘could not possibly remain unaltered, considering the fallibility of human memory, which plays as many tricks with the unlettered singers of folk-songs as it does with the rest of us’ writes Gerould (1932, p. 163), just as ‘any composition travelling from mouth to mouth, from generation to generation, from country to country is bound to suffer from a certain amount of verbal corruption and degeneration’ (Coffin, 1950, p. 3).
In general however, the widespread existence of variants prima facie sheds doubt on the concept of exact memorisation as the key factor in oral composition: even the early collector Sharp, who recorded only in writing, rejects the idea that variants are ‘corruptions in varying degree of one original’ (1972, p. 14, first published 1907). But the implausibility of attributing variability largely to faulty memory has been much increased by use of the tape-recorder. This has helped to show that variability is not just a feature of lengthy oral transmission through time and space but is inherent both in different renderings of one literary piece within the same group and period and even in texts by the same person delivered at no great interval in time. In such cases, memorisation of basic themes or plots is involved, but a generalised explanation of the oral poetry in terms of particular texts exactly memorised does not easily fit the abundant variability demonstrated in tape-recorded (as well as dictated) texts. When one adds this general evidence of variability to the specific cases mentioned earlier, showing the creative role of the poet responding to the audience and occasion, it becomes extremely difficult to continue to hold the theory that memorisation is the sole factor at work. Clearly there are many occasions when the performer takes a hand in the process of composition.
How much freedom the composer-performer has is not always clear in the published accounts. The degree of ‘composition’ as against ‘memorisation’ probably differs both between different cultures, and between different genres in the same culture, and between poets. The Chadwicks, for instance, contrasted the emphasis on memorisation among the Turkomans with the ‘high development of extempore composition’ of the Kirghiz (Chadwick, III, 1940, p. 184), and the amount of strict memorisation involved in the ritualistic poetry of an established religion is likely to be more than in a light-hearted topical song.
Even within one genre, there may be more, or less, improvisation. There is an instructive contrast between different kinds of work songs among Texas prisoners. The songs which accompany critically-timed tasks, such as a team of men cutting down a tree, give little scope for change or development in the singing, though ‘one can interject names in formulaic lines—names of guards, fellow-workers, people one has known or heard of, and so forth—and there are chorus lines and repeats that give the singer time to think up another verse’ (Jackson, 1972, p. 34). But songs for less rigorously timed group-work like cotton picking or sugarcane cutting give more opportunities for solo songs and for lyrical and ornamental development by the leader (ibid., p. 33).
Individual poets differ in the emphasis they lay on memorisation as against creation. In the study of American oral forms, it has become common to distinguish between ‘passive’ and ‘active’ traditors—those who largely reproduce what they have heard, having memorised it as best they can, and those who actively participate in composition or recomposition. In the nyatiti lament-songs of the East African Luo, too, there are different degrees of personal creativity by the singer, depending less on external situation than on the ability of the singer. Some perform from a relatively fixed repertoire—or from a set of basic structures, which, once learnt, can be modified to suit the circumstances of the funeral to which the singer has been summoned: he adds an ‘uncle here and a grandfather there, together with any knowledge he may possess of the attributes of the deceased’. But a gifted Luo singer creates a more individual and developed song, particularly when he is emotionally involved. His artistry is appreciated by his listeners. ‘The skill and beauty with which the musician is able to improvise at such moments is a measure of his musical and poetic stature’ (Anyumba, 1964, pp. 189–90).
These variations between cultures, genres and poets are in themselves interesting (see also chapter 5, where the topic is related to transmission and distribution). But the important issue is the extent of this variation. The blanket term ‘memorisation’ is too general to cover the manifold ways in which poets may proceed in different contexts. And ‘memorisation’ is not always the most appropriate description for the process of performance/composition observed in the instances mentioned here, and many similar ones that have been observed.
To these doubts have been added the theories and findings of the ‘oral-formulaic’ school, initially concerned with analysis of ‘formulaic’ language in the Homeric epics, supplemented by research on comparable forms in twentieth-century Yugoslav oral poetry, and now influential in the whole study of oral composition.
3.2 Composition-in-performance and the
oral-formulaic theory
There had long been controversy over the composition of the Homeric poems, the Iliad and Odyssey. The main battle was between the ‘unitarians’ who posited a single author, and the ‘separatists’ who held that the poems were a composite construction of different lays and/or strata. There had always been difficulties about either position. How could one decide questions about the composition of poems written down in Greece two and half millennia ago and perhaps composed much earlier? How explain the composition of such long poems (the Iliad 15,000 lines, the Odyssey 12,000—surely too long to be memorised?—in an age when most people were illiterate? And what was the best explanation of the recurrent lines and phrases, some repeated many times throughout the poem?
Against this background, analysis of the ‘formulaic’ nature of Homer’s style and its possible relationship to the oral composition of epic poetry was developed. Its initial exponents in the West were Milman Parry and his pupil Albert Lord—hence its designation as the ‘Parry-Lord theory’ although the foundations had in fact been laid by earlier writers like Murko on Yugoslav oral poetry and a number of Russian studies.
Milman Parry was an American classical scholar who became interested in the formulaic phrases in the Homeric epithets, on which he published his study of L’Epithète traditionelle dans Homère (Paris, 1928). He noticed, like others before him, the apparently formulaic nature of recurrent descriptions of many of the people in the poems: ‘swift-footed Achilles’, ‘many-counselled Odysseus’, ‘glorious Hector’, ‘grey-eyed Athene’, and so on. These formulaic epithets, which to a modern reader may appear an irritating repetition or a not very meaningful trick of ‘epic style’, can also be seen as playing a significant part in the composition of the poem. For these various epithets fit exactly the constraints of the hexameter metre in which the poems are composed: ‘Achilles, son of Peleus’ (Πηληιάδεω Άχιλῆος) exactly fills the second part of a line. These ‘Homeric epithets’ are often combined with other formulaic phrases—repeated word-groups—which have the right metrical qualities to fit the first part of the line. So a whole line can be rapidly and easily constructed by the oral poet, built from a ready-made diction. On other occasions, other combinations are possible from similar metrical units, to give the necessary sense. Thus ‘in composing [the poet] will do no more than put together for his needs phrases which he has often heard or used himself, and which, grouping themselves in accordance with a fixed pattern of thought, come naturally to make the sentence and the verse’ (Parry, 1930, p. 77).
This principle can be taken further. A single hexameter line is a relatively independent unit which usually coincides with a sentence or fairly self contained phrase. Whole lines can thus be repeated in this ‘formulaic’ way, and this happens in the Homeric epics. Of all the lines in the Iliad and Odyssey about one third recur at least once (some many times). The extent of this device is usually illustrated from Parry’s analysis of the first ten lines of the Odyssey where epithets, phrases or lines which recur elsewhere are underlined (a solid line where the formula re-appears unchanged elsewhere, a broken line where similar phrases occur).
Ἅνδρά μοι ἔννεττε Μοῦσα πολύτροπον ὃς μάλα πολλά
πλάγχθη ἐπεὶ Τροίης ἱερὸν πτολίεθρον ἓπερσε·
πολλῶν δ’ ἀνθρώπων ἵδεν ἂστεα καὶ νόον ἓγνω,
πολλὰ δ’ ὅ γ’ ἐν πόντωι πάθεν ἃλγεα ὃν κατὰ θυμόν
ἀρνύμενος ἥν τε ψυχὴν καὶ νόστον ἑταίρων.
ἀλλ’ οὐδ ὤς ἑτάρους ἐρρύσατο ἱέμενός περ·
αὐτῶν γὰρ σϕετέρηισιν ἀτασθαλίηισιν ὃλοντο
νήπιοι οἵ κατὰ βοῦς Ύπερίονος Ήελίοιο
ἣσθιον· αὐτὰρ ὁ τοῖσιν ἀϕείλετο νόστιμον ῆμαρ.
τῶν ἁμόθεν γε θεὰ θύγατερ Διὸς εἰπὲ καὶ ἡμῖν.
(Parry, 1930, p. 120)
Sing in me, Muse, and through me tell the story
of that man skilled in all ways of contending,
the wanderer, harried for years on end,
after he plundered the stronghold
on the proud height of Troy.
He saw the townlands
and learned the minds of many distant men,
and weathered many bitter nights and days
in his deep heart at sea, while he fought only
to save his life, to bring his shipmates home.
But not by will nor valour could he save them,
for their own recklessness destroyed them all—
children and fools, they killed and feasted on
the cattle of Lord Helios, the Sun,
and he who moves all day through heaven
took from their eyes the dawn of their return.
Of these adventures, Muse, daughter of Zeus,
tell us in our time, lift the great song again.
(translated R. Fitzgerald, 1965, p. 13)
It can be deduced that the amount of repetition is very great in total, and ‘formulaic’ thus came to seem the most important characterisation of Homer’s style.
The repetitions and the use of the ‘Homeric epithet’ had often been noticed before. But Parry took the further step of using this formulaic style to prove that the Homeric poems were orally composed. It was the need of the oral poet, he argued, for fluent and uninterrupted delivery throughout a lengthy performance that made the formulaic style both necessary and suitable. The poet had a store of ready-made diction already tailored to suit the metrical constraints of the hexameter line. By manipulating formulaic elements from this story—the ‘building blocks’—he could construct a poem based on traditional material which was still his own unique and personal composition. The poet had at his disposal this series of traditional patterns built up over the years (so there was something in the theory of multiple authorship), but he was not passively dominated by them: he used them to create his own poems as he performed them.
Having come to this conclusion about the composition of the Iliad and Odyssey Parry took the imaginative step of going outside classical studies proper to try to find proof that his interpretation of them as oral compositions was correct. He turned to the study of Yugoslav epics, and in the 1930s, accompanied by his pupil and collaborator A. B. Lord, he collected and studied many oral heroic poems then being composed and performed by oral singers in Yugoslavia. These studies were to provide apparently irrefutable proof of Parry’s findings. Yugoslav oral bards composed on the same principles.
Here is a description of the process taken from the classic product of this research, A. B. Lord’s The Singer of Tales (1968a, first published in 1960). Lord starts from the striking fact that Yugoslav bards can perform long epics of thousands of lines (Parry recorded several with over 10,000 lines) and, with only short pauses for rest, can do this with uninterrupted fluency and remarkable speed—often at the rate of ten to twenty ten-syllable lines a minute. How does he do this?
‘Since, as we shall see’, writes Lord ‘he has not memorized his song, we must conclude either that he is a phenomenal virtuoso or that he has a special technique of composition outside our own field of experience. We must rule out the first of these alternatives because there are too many singers; so many geniuses simply cannot appear in a single generation ... The answer of course lies in ... the special technique of composition which makes rapid composing in performance possible’
(Lord, 1968a, p. 17)
This special technique involves building on repeated formulae, ‘ready made phrases’ which the singer knows and can use without hesitation to fit the metrical requirements of his line.
Take the following passage, quoted by Lord from one of Parry’s recordings. The phrases found more than once in the perusal of about 12,000 lines from the same singer (Salih Ugljanin) are underlined
|
Jalah reče,/zasede đogata; |
With ‘By Allah’ she mounted her horse; |
|
|
790 |
Đogatu se/konju zamoljila: ’Davur, đogo,/krilo sokolovo! Četa ti je/o zanatu bila; Vazda je Mujo/četom četovao. Vodi mene/do grada Kajniđe! |
She implored the white horse: ‘Hail, whitey, falcon’s wing! Raiding has been your work; Ever has Mujo raided. Lead me to the city of Kajniđa! |
|
795 |
Ne znam đadu/ka Kajniđi gradu.’ Hajvan beše,/zborit’ ne mogaše, Tek mu svašta/šturak umijaše. Ode gljedat’/redom po planini Uze đadu/ka Kajniđi gradu, |
I know not the road to the city of Kajniđa.’ It was a beast and could not talk, But the steed knew many things. He looked over the mountains And took the road to the city of Kajniđa, |
|
800 |
Pa silježe/planinama redom, Pa ga eto/strmom niz planinu, I kad polju/slježe kajnićkome, Kome stati/polje pogljedati, |
And crossed one range after another, Until lo he rushed down the mountain, And when he descended to the plain of Kajniđa, Were anyone to look out over the plain, |
(Lord, 1968a, p. 46)
Lord comments:
From the chart we can see at a glance the number of repeated phrases that without any hesitation can be called ‘formulas’. These phrases we know by demonstration that the singer has come in time to use regularly. Even within the limited number of lines used in the experiment, that is, 12,000, one quarter of the whole lines in the sample and one half of the half lines are formulas. It is most significant that there is no line or part of a line that did not fit into some formulaic pattern. In certain instances the pattern was a very common one and there was no difficulty in proving the formulaic character of the phrase. In a few instances the evidence was not so abundant, but it was still sufficient to make one feel certain that the phrase in question was formulaic. A number of the formulaic expressions could very easily have been classified as formulas, had we relaxed our established principles and standards. For example, davurdogo in line 791 misses being a formula because the evidence lists only davur sturan and davur doro. But dogo, sturan, and doro are all terms for horses. We could thus have easily increased the number of formulas.
Had we gone beyond 12,000 lines, the number of formulas would have continued to mount, and had we included material from other singers it would have increased still further, until it became clear that almost all, if not all, the lines in the sample passage were formulas and that they consisted of half lines which were also formulas. In other words, the manner of learning described earlier leads the singer to make and remake phrases, the same phrases, over and over again whenever he needs them. The formulas in oral narrative style are not limited to a comparatively few epic “tags”, but are in reality all pervasive. There is nothing in the poem that is not formulaic.
(Lord, 1968a, p. 47)
Yet it also becomes obvious that a formulaic style need not mean lack of flexibility in the poet. He can select what he wishes from the common stock of formulae, and can choose slightly different terms that fit his metre (for example any of several two-syllabled words for ‘horse’) and vary the details. Here are some linked but distinguishable alternatives used by the different singers named.
|
Jalah reče, zasede đogata. ‘By Allah,’ she said, she mounted the white horse. |
|
|
Sulejman Fortić |
|
|
Jalah reče, posede dogina. |
‘By Allah,’ he said, he mounted the white horse. |
|
Jalah reče, posede hajvana. |
‘By Allah,’ he said, he mounted the animal. |
|
Đemail Zogić |
|
|
Jalah reče, sede na dorina. |
‘By Allah,’ he said, he mounted the brown horse. |
|
Jalah reče, posede hajvana. |
‘By Allah,’ he said, he mounted the animal. |
|
Sulejman Makīć |
|
|
I to reče, posede dorata. |
And he said this, he mounted the brown horse. |
|
Alija Fjuljanin |
|
|
A to reče, zasede hajvana. |
And he said this, he mounted the animal. |
(Lord, 1968a, p. 48)
If one formulaic phrase filling the second part of the line is a variant of ‘mounted his/her horse’ etc., the singer can construct the first part of the line as the demands of his story and his art require. He can construct all the following lines, for instance, on the same basic pattern.
|
Svi konjici konje zasedoše. |
All the horsemen mounted their horses. |
|
A svatovi konje zasedoše. |
And the wedding guests mounted their horses. |
|
Ta put hajduk šajku zasednuo. |
Then the hajduk mounted his mare. |
|
A Mujo svoga pojaše đogata. |
And Mujo mounted his white horse. |
|
Jalah Suka sede na menzila. |
With a cry to Allah, Suka mounted his post horse. |
(ibid., p. 51)
From these patterns, the singer can derive the lines he needs for a particular context as he simultaneously composes and performs. As Lord explains it, the singer is doing more than ‘merely juggling set phrases’. He employs them because they are useful and serve the needs of the moment, and he is free to adjust them as and when he wishes. ‘In making his lines the singer is not bound by the formula. The formulaic technique was developed to serve him as a craftsman, not to enslave him’ (ibid., p. 54).
The formulaic quality of Yugoslav epic style is not confined to units within a line, or forming whole lines. For the singer has to compose a series of lines, one after the other. ‘The need for the “next” line is upon him even before he utters the final syllable of a line’ (ibid., p. 54). So the singer builds patterns of sequences of lines added to each other in a series of parallel sentences. There is little ‘enjambement’ (or necessary run-over of sense from line to line): of 2,400 lines analysed, 44.5 per cent had no enjambment, 40.6 per cent ‘unperiodic enjambement’ (i.e. the sense was complete at the end of the line but the sentence continued), and only 14.9 per cent necessary enjambement. This ‘adding style’ makes the singer’s task easier, and is so obviously a useful device in simultaneous oral performance and composition that Lord claims it generally as ‘a characteristic of oral composition’ and ‘one of the easiest touchstones to apply in testing the orality of a poem’ (ibid., p. 54). This ‘adding style’ is well illustrated by the rapid, almost staccato style of this sequence:
|
Kud god skita za Aliju pita. Kazaše ga u gradu Kajniđu. Kad tatarin pod Kajniđu dođe, Pa eto ga uz čaršiju prođe, Pa prilazi novom bazdrđanu, Te upita za Alino dvore. Bazdrđan mu dvore ukazao. Kad tatarin na kapiju dođe, Pa zadrma halkom na vratima. Zveknu halka a jeknu kapija. |
Wherever he went, he asked for Alija. They said he was in the city of Kajniđa. When the messenger came to Kajniđa, He passed along the main street, Then he approached the new shopkeeper, And he asked for Alija’s court. The shopkeeper pointed out the court to him. When the messenger came to the gate, He beat with the knocker on the door. The knocker rang and the gate resounded. |
(quoted Lord, 1968a, pp. 54–5)
A more leisurely description may interrupt the rapid course of the narrative, where the apparently simple ‘adding style’ has an impressive cumulative effect.
(ibid., p. 55)
As well as formulaic phrases and sequences, the bard has in his repertoire a number of set themes which he can draw on to form the structure of his poem. There are stock episodes (some of them familiar from other epic poetry) like the gathering of an assembly or of wedding guests, a journey, writing a letter, and so on. The wording of the episode varies from singer to singer, and so does the way in which they are joined together, or combined with other themes. But these stock episodes, as well as wider themes and plots, like rescues, returns or captures of cities, are all there as a traditional resource on which the poet can draw to construct his own poem.
The oral Yugoslav poet, therefore, can base his composition on these known patterns of phrases, lines and themes, without necessarily restricting himself to them. Through this technique of composition he is able to carry on the simultaneous performance and composition essential in his art. There is no set text which he has to learn—so that memorisation in the sense of word-for-word recall is not involved; instead he learns the poetic vocabulary and structure appropriate for heroic poetry in the way a child learns a language: it is the basic resource and medium within which and with which he composes.
One of the most significant points to emerge from the study of Yugoslav oral poets is the absence of a fixed text—the primary text or archetype so often sought for in classical studies. ‘In a sense’, writes Lord, ‘each performance is “an” original, if not “the” original. The truth of the matter is that our concept of “the original”, of “the song”, simply makes no sense in oral tradition’ (ibid., p. 101). There is no correct text, no idea that one version is more ‘authentic’ than another: each performance is a unique and original creation with its own validity.
This is so even when the poet claims to be singing ‘the same’ poem as one he had heard, or to be repeating a poem in exactly the same form as he sang it before. In practice, the words and detailed sequences are likely to be different, even between separate performances by the same poet. This came home forcibly to the researchers in Yugoslavia who had hoped to overcome the difficulties of recording in writing the rapidly-delivered text by filling in parts they had missed in one performance from later renderings. This proved to be impossible, for the variations between performances were too great (see Lord, 1968a, p. 149).
So the ‘formulaic style’, far from being an inducement to passive receptivity by the singer, provides him with the opportunity to make each performance unique and his own. This comes out clearly in the comparisons of poems by different poets on the same basic theme. As part of their research, Lord and Parry recorded experiments involving the direct transmission of ‘the same’ poem from one poet to another (Lord, 1968a, pp. 102ff). In The Wedding of Relja of Pazar the basic story remained much the same in the hands of a second poet, but a number of details were omitted, some speeches were expanded and so was the marriage theme at the end. In another case the talented singer Avdo Mededović followed a performance by another poet, Mumin Vlahovljak, with his own rendering of the ‘same’ poem. He expanded the poem to nearly three times the previous length (2,294 to 6,313 lines), which in itself meant many changes and elaborations. There are also new episodes and changes in the order of events, and the whole is suffused with the insights and sensitivity of Avdo Mededović himself, a poet whom Lord considers outstanding among the Yugoslav singers he recorded.
The oral poet in Yugoslavia is always the ‘author’ of the epic he performs, by virtue of his simultaneous performance/composition. In this sense, each epic has a single author. But in another sense, there is also a multiplicity of authors: all those who contributed to building up the traditional patterns, the store of formulae and themes which the oral singer has at his disposal (ibid., pp. 101–2).
This in essence is the famous Parry-Lord theory: it shows how oral composition is possible during performance itself through the poet’s reliance on formulaic style and themes. We now understand how long oral poems can be produced without the poet having to rely on rote memory. The theory also cut through much controversy about the authorship of Homer by demonstrating that, like Yugoslav poetry, Homeric epics were oral compositions using a comparable ‘oral-formulaic’ technique.
It would be hard to overestimate the importance of this approach. It has had a deep influence on Homeric studies; though not all scholars accept the theory in toto, few can ignore it, and many works have appeared which, in various ways, apply the approach to analysis of the Homeric epics (e.g. Kirk, 1965, Notopoulos, 1964, Nagler, 1967). Some puzzles and controversies remain—like the problem of how the Iliad and Odyssey were finally committed to writing—but the idea that oral composition was in some way involved in the creation of the epics is now widely accepted.
But the influence of this approach reaches far beyond Homeric studies. Lord claims in the Foreword to The Singer of Tales that Homer in a large sense ‘represents all singers of tales from time immemorial and unrecorded to the present. Our book is about these other singers as well’. Indeed the extension of his findings beyond the Yugoslav and Homeric cases was clearly intended by Parry at the outset. He wrote explicitly that his purpose was to
obtain evidence on the basis of which could be drawn a series of generalities applicable to all oral poetries; which would allow me, in the case of a poetry for which there was not enough evidence outside the poems themselves of the way in which they were made, to say whether that poetry was oral or not … A method is here involved, that which consists in defining the characteristics of oral style.
It is not surprising that his approach has been more widely extended, and that many scholars have tried to apply a similar oral-formulaic analysis to texts of all kinds, from Old Testament poetry, Beowulf, or mediaeval European epic to recent compositions like modern Greek ballads, Gaelic poetry or the formulaic intoned sermons of the Southern States of America. One or two illustrations will demonstrate the ways in which the theory has been developed.
One application has been to Old English poetry: in particular to the Anglo-Saxon epic Beowulf (composed in England around the eighth century). It has been noticed for many years that elements of the style were repetitive, but Magoun now tried to demonstrate that in its formulaic character it resembled ‘oral poetry’, which ‘it may safely be said, is composed entirely of formulas, large and small’ (Magoun in Nicholson, 1971, p. 190). He illustrates this by, among other things, a chart of the opening lines of Beowulf in which ‘word-groups’ which appear elsewhere in Beowulf or other Anglo-Saxon poems unchanged are marked with solid underlining, while formulaic phrases (appearing in similar but not identical forms elsewhere) have broken underlinings.
|
Hwæt, wé Gár-Dena on géar-dagum þéod-cyninga þrymm gefrugnon, hú þá æðelingas ellen fremedon. Oft Scield Scéafing sceaðena þréatum, |
|
|
5 |
manigum mægðum medu-setla oftéah, egesode Eorle, syþþan ærest wearþ féascæft funden; hé þæs frófre gebád, wéox under wolcnum, weorþ-myndum þáh. oþ-þæt him ǽghwelć ymbsittendra |
|
10 |
ofer hran-ráde hieran scolde, gamban gieldan; Þæt wæs gód cyning! |
(Magoun in Nicholson, 1971, pp. 216–17)
Attend!
We have heard of the thriving of the throne of Denmark,
how the folk-kings flourished in former days,
how those royal athelings earned that glory.
Was it not Scyld Shefing that shook the halls,
took mead-benches, taught encroaching
foes to fear him—who, found in childhood,
lacked clothing? Yet he lived and prospered,
grew in strength and stature under the heavens
until the clans settled in the sea-coasts neighbouring
over the whale-road all must obey him
and give tribute. That was a king!
(translated M. Alexander, 1973, p. 51)
A number of the episodes can also be seen as set pieces and so’ formulaic’—like the banquet, the voyage, the funeral, the battle and so on. The overall result has been an increasing interest in Beowulf as an ‘oral composition’—as evidenced, it is argued, by its ‘oral-formulaic’ style—and a rejection of older theories that it is a fixed written document assimilated to the models of the literary works of contemporaneous Latin Christianity.
A similar analysis has been made of the oral-formulaic style of the Song of Roland, the Old French epic perhaps composed around the eleventh century A.D. In this epic, the decasyllabic line divides into two main sections or hemistiches (the first of four, the second of six syllables). These hemistiches form the main formulaic units. Phrases like Li reis Marsilie, dist li emperere Carles which fit a hemistich and are substantially repeated can be counted as recurrent formulae and form the basis for the poet’s original composition. It has been calculated that something like 35.2 per cent of Roland is ‘formulaic’ (Duggan, 1973, p. 34), or, taking the first hemistiches only, over 50 per cent (Nichols, 1961, p. 20). This is perhaps considerably less than the 80 or go per cent claimed for the Homeric poems, but enough in the view of many analysts to establish Roland as ‘oral-formulaic’ in style and composition.
This ‘oral’ style does not cramp the poet’s individual genius. Though formulae are fairly evenly distributed through the poems, they are more frequent in the universally acclaimed purple passages, and can be used by the poet to convey subtle shades of character and feeling (Duggan, 1973, chapter 5). In addition, there are larger formulaic units, in the sense of set themes and episodes, like the many councils in Roland and the recurrent stages within the council episodes (Nichols, 1961), which again the poet uses as a vehicle for his composition.
Similar analyses have now been made of a good deal of poetry. There have been a few studies of recently-recorded oral poetry, like the Yugoslav epics; but for the most part scholars have concentrated on poetry which has come down to us in written texts: other early Greek poetry (e.g. Notopoulos, 1960), the Hittite epic (McNeill, 1963), early Tamil heroic poetry (Kailasapathy, 1968), mediaeval English, French and German poetry (usefully surveyed in Curschmann, 1961), English and Scottish ballads (Jones, 1961, Buchan, 1972), and early Hebrew poetry as found in poetic passages like the Psalms, the Song of Solomon and elsewhere in the Old Testament (Whallon, 1969, Culley, 1967).
Since these are all texts about whose composition and performance in the distant past we have little direct information, the emphasis has often been on the deductions about these aspects that, it is claimed, can be drawn from ‘the oral-formulaic style’. Some scholars have been so impressed by the insights of the Parry-Lord approach that they have taken anything which can arguably be dubbed an ‘oral-formulaic’ style as proof that ‘oral composition’ was involved. Thus Magoun, writing of Anglo-Saxon poetry, claims an ‘oral-formulaic style’ as the ‘touchstone’ for differentiating ‘oral’ and ‘lettered’ poetry (Magoun in Nicholson, 1971, p. 194), while Nichols’s interest in analysing formulaic diction and enjambement in The Song of Roland was ‘in order to demonstrate the textual characteristics which argue its oral rather than literary character’ (Nichols, 1961, p. 9). This interest in laying bare the oral character and origin of a given text has also been a preoccupation in most oral-formulaic analyses of the Homeric epics.
Thus since the initial writings of Murko and Milman Parry in the 1920s and 1930s, the ‘oral-formulaic’ approach has become an established school which must be taken account of by the analyst of much classical and mediaeval literature in Europe and Asia, as well as of oral poetry else where. It has affected the work of scholars in classics, mediaeval and Biblical studies, and work in linguistics, literature, anthropology and history. It has introduced a comparative perspective into many specialist areas, so that ‘oral poetry’ is no longer a special phenomenon, to be looked for among far-off and exotic ‘primitives’, but a concept central to the pursuit of a number of traditional disciplines concerned with the development of civilisation over several thousand years.
3.3 How valid is the oral-formulaic theory?
No-one who has read The Singer of Tales—surely one of the classics in the study of oral literature—can fail to be profoundly influenced by its findings and insights. It is no longer possible to argue that the only way to explain lengthy oral poems is in terms of memorisation by the performer. This in itself cuts through a number of problems, with the Homeric (and other) epic poems and also with English and Scottish ballads where ‘the commonplaces ... freed the singer from memorization’ (Jones, 1961, p. 105). Equally important is the emphasis on the lack of a fixed and ‘correct’ version of the text in oral literature. The model of written literature with its emphasis on the text, the original and correct version, has for long bedevilled study of oral literature, and led researchers into unfruitful and misleading questions in an attempt to impose a similar model on oral literature. Others besides Lord and Parry had pointed this out (in particular earlier scholars from Russia and Eastern Europe) but, for English-speaking readers, no one has conveyed the original aspect of each unique performance by a poet as convincingly as Lord in The Singer of Tales. The point has had a profound effect on analysts of oral literature. So has the account of the relationship between tradition and originality in oral composition: of the way the oral poet makes use of traditional patterns to express his individual and original insights. As another scholar has put it, ‘all is traditional on the generative level, all unique on the level of performance’ (Nagler, 1967, p. 311), and the old polarity between ‘tradition’ and ‘originality’ no longer means direct contradiction.
But there are also drawbacks and difficulties. Some of these have caused controversy within the ‘oral-formulaic school’, which now encompasses differences of opinion and varying wings of thought.
First, there is the problem of how far an ‘oral-formulaic style’ is indeed a sign of ‘oral composition’. Some earlier analysts like Parry, Lord and Magoun claimed that it was an indisputable sign, and they have been followed by recent scholars such as Duggan and Kailasapathy. As recently as 1968 Lord was claiming that ‘A pattern of 50 to 60 per cent formula or formulaic, with 10 to perhaps 25 per cent straight formula, indicates clearly literary or written composition. I am still convinced that it is possible to determine orality by quantitative formulaic analysis, by the study of formula density’ (Lord, 1968b, p. 24). This idea has come under fire in recent studies. As Benson has demonstrated, a heavily formulaic style is characteristic not just of the Old English ‘oral’ epic of Beowulf but also of some written compositions in Old English, including Old English translations from Latin originals (Benson, 1966). If the style proves Beowulf to be ‘oral’, how can one explain its use in written composition? Benson concludes ‘To prove that an Old English poem is formulaic is only to prove that it is an Old English poem, and to show that such work has a high or low percentage of formulas reveals nothing about whether or not it is a literate composition, though it may tell us something about the skill with which a particular poet uses the tradition’ (Benson, 1966, p. 336).
Similar points have been made for other literatures. One recent example comes from research on Xhosa and Zulu oral poetry in South Africa. Opland has pointed to the formulaic style of’ traditional’ Xhosa oral poetry and the way the Xhosa imbongi can ‘compose metrical poems on the spur of the moment in praise of anything that inspires him, and in order to do this he relies on formulas’ (Opland, 1971, p. 172). Opland has recorded a number of examples, and can testify from personal observation to their oral composition and performance. So far, this looks like a typical instance of ‘oral composition using formulae’. But some of these Xhosa poets have also produced written versions of their poems—and these are equally characterised by a’ formulaic’ style. This applies to a number of Bantu poets in South Africa, Zulu as well as Xhosa. Opland writes
Literate Bantu poets are using the traditional praise songs as a basis for their poetry. The Zulu poet B. W. Vilakazi writes of his poetry that he passed through a period of imitating European models, but subsequently returned to traditional forms. Having grown up in the Zulu tradition he expressed himself most easily in a manner that was part of his cultural identity. Colleagues of mine in the Department of African Languages at the University of Cape Town have written praise poems on the highway traffic, and on the Apollo moon landing. Their poetry conforms metrically and stylistically to the traditional praise poems sung by the imbongi. It seems reasonable, therefore, that literate Anglo-Saxon monks could have written formulaic poetry.
A ‘formulaic’ style is not therefore inevitably a proof of ‘oral composition’. Theorists have now to accept that since there can be both an ‘oral’ and a ‘literary’ use of formulae one cannot necessarily discriminate between ‘oral’ and’ written’ on the basis of a’ formulaic’ style alone. In other words, the excitement attendant on the discoveries about the ‘oral-formulaic style’ led some scholars to an extreme application of its findings. Now that a certain reaction has set in against the more extreme claims, it has become clear that while the kind of diction found in Homeric or Yugoslav epics is a suitable, even likely, style for oral composition, it is not an infallible sign of it.
There is the further difficulty—less often stressed by exponents of this school—that the meaning of ‘oral composition’ is by no means always clear. Scholar after scholar has demonstrated an ‘oral-formulaic’ style in his chosen texts and taken this as a sign (whether tentative or definite) of ‘oral composition’. The model is often the Yugoslav singer’s oral composition described by Lord; in other cases it is left unclear, and the term even seems to take on a mystical aura of its own. But there are in fact different kinds of ‘oral composition’. Some is almost entirely oral, in the way that many Yugoslav singers composed/performed; but some may be based more or less directly on a written text (and this happens more often than is some times remembered with Yugoslav poetry) but is nevertheless recited orally. Other poems may be composed with oral performance as the aim, or composed without the initial use of writing for later written publication, or be specially dictated to a literate assistant—and so on. Which of these forms is involved is seldom explained in many analyses which deduce ‘oral composition’ from the style of the texts they have studied. What is needed in these cases is more discrimination of the exact meaning and application of the term ‘oral composition’—except that often it is because there is little or no direct evidence about the process of composition that the analyst argues from the ‘oral-formulaic’ style.
A second difficulty about the general approach is the exact definition of a ‘formula’. Here the doubts are not so much about the excessive claims of the approach, but about the method of analysis. Certain frequently repeated Homeric epithets like ‘grey-eyed Athene’ or ‘rosy-fingered dawn’ may seem obvious cases of’ formulaic units’, but it is a large extension when the term is used to cover any group of words ‘regularly employed under the same metrical conditions to represent a given essential idea’ (as Parry had it) or a phrase repeated only once or twice but intuitively regarded by the analyst as somehow ‘formulaic’. Even if ‘repetition’ is taken as basic in defining a ‘formula’, there are differences between analysts as to whether the repetition is, for instance, of metrical, syntactic or semantic elements, differences also about how long a ‘formula’ can or must be (see Watts, 1969, esp. chapter 3).
What do these various groupings have in common? Clearly there are in some sense recurrent patterns, a continuing structure within which the oral poet—or indeed any poet—composes. But to use the term ‘formula’ in an apparently exact sense, and to rely on it as the basis for a complex edifice of theory and supposedly empirical findings gives a misleading impression of precision. Does it really add to our understanding of the style or process of composition in a given piece to name certain repeated patterns of words, sounds or meanings as ‘formulae’? Or to suggest that the characteristic of oral style is that such formulae are ‘all-pervasive’ (as in Lord, 1968a, p. 47)? Or to propound tautologous definitions of the formula as, for instance,
‘a group of words, one half-line in length, which shows evidence of being the direct product of a formulaic system’ (Fry, 1967, p. 204)? Reading would-be scientific and rigorous analyses based on this concept of the formula, one is tempted to apply more widely the critique H. L. Rogers made of the oral-formulaic approach to Old English poetry: ‘The term “formula” becomes a portmanteau, enclosing within its ample capacity many different, and often undefined, sorts of lexical, morphological and syntactic similarities … One is forced to suspect that the growing dogmatism about the oral-formulaic character of Old English poetry owes more to faith and presumed psychological insight than to reason’
(Rogers, 1966, p. 102).
Without any exact or agreed definition of ‘the formula’ it is difficult to rely on the statistical analyses used either to demonstrate an ‘oral-formulaic’ style in a particular poem, or to compare percentages of ‘formularity’ across different kinds of poetry. Where the unit of measurement is unclear, such comparisons seem worth little. Further doubts have also been thrown on the kind of statistical conclusions that Parry and others derived from their analyses of the Homeric poems. Their data may at first look convincing (laying aside, that is, the fundamental point about the delimitation of ‘formula’ or ‘formulaic system’). But as a recent analysis (Russo, 1976) has demonstrated, their apparently solid conclusions were raised on a remarkably small statistical base. The opening lines of the Iliad and Odyssey have been analysed and measured several times, leading to a claim of ‘between 80 and 90 per cent formulaic content’; so have a very few other short passages. Otherwise no overall analysis has been completed, nor any systematic sampling undertaken. After all, the style of preludes is not necessarily representative of the style of the whole.
A more representative sampling might result in lower figures for ‘formulaic content’ than the ninety percent usually cited. Indeed the analysis by Russo (1976) suggests that more rigorous sampling might not support the higher claims for formulaic content in Homer and that the overall level of formularity, so far as it can be measured at all, may turn out to be more like that assumed as typical of literary texts generally. It appears, therefore, that the search for a scientific and precise measure of oral as against written composition may turn out to be a fruitless one.
These are mostly detailed controversies within the oral-formulaic school. The basic insights remain stimulating and fruitful, and the demonstration that the oral bard composes with and within traditional patterns of various kinds will stand as a landmark in the study of oral literature. Provided that the more ambitious claims of some exponents are treated with caution, the Lord-Parry school provides a body of work which cannot be ignored by any student of comparative oral literature.
3.4 Prior composition, memorisation and performance
The demonstration that rote-memory need not be important in oral poetry has sometimes misled students of the subject (myself included) to assume that it is never important. Parry and Lord did not go so far as this, but their works can be read as implying it: ‘No graver mistake could be made’, writes Parry, ‘than to think the art of the singer calls only for memory … the oral poem even in the mouth of the same singer is ever in a state of change; and it is the same when his poetry is sung by others’ (Parry, 1932, pp. 14–15), and ‘Oral … does not mean merely oral presentation … what is important is not the oral performance but rather the composition during oral performance’ (Lord, 1968a, p. 5). Lord also goes out of his way to lay down a restrictive definition of ‘oral poetry’ which excludes the possibility of memorisation, when he asserts that texts which are preserved word-for-word ‘could not be oral in any except the most literal sense’ (ibid., p. 280), and defines ‘oral poetry’ as ‘poetry composed in oral performance’ (Lord, 1965, p. 591).
But memorisation and near-word-for-word reproduction sometimes are important in oral literature. We cannot ignore the possibility or define away its occurrence as not really involving oral poetry or as constituting an odd and perverse exception. It is true that one needs to be cautious of statements that a piece has been ‘repeated exactly’ on a different occasions or handed down word for word’ through a time: often, statements like this rest on no evidence beyond the dubious assumptions about ‘age-old tradition’ or ‘folk memories’ discussed in the last chapter, or the interesting but often untrustworthy or ambiguous claims of the bards themselves. But the caution necessary in assessing undemonstrated statements of this kind—even when allied to the kinds of expectations we now hold about oral literature as a result of the Parry-Lord research—must not prevent us from noting instances of exact reproduction when clear evidence is provided.
Some examples can illustrate this particular process. In these cases composition and performance are separated.
The first is the case of Somali poetry. In Somalia, in the Horn of Africa, oral poetry is a highly developed art. It has been extensively studied, both in its modern phase where radio and tapes as well as face-to-face delivery are employed, and in its earlier development. It includes many different genres (described in Andrzejewski and Lewis, 1964), from lengthy gabay poems, sometimes of several hundred lines, to the short one- or two-line compressed balwo lyrics.
Poetic composition is a prized and much-discussed art among the Somali, who have been described as ‘a nation of bards’, and an admired poet can become widely known. A Somali poem always arouses interest and discussion and attracts criticism if it is considered mediocre. Because they are aware of this, ‘Somali poets rarely perform their work until composition is completely finished in private’ (Johnson, 1971, p. 28) and ‘spend many hours, sometimes even days, composing their works’ before they perform them (Andrzejewski and Lewis, 1964, p. 45). A poet’s compositions, furthermore, become his own property, under his own name, and another poet reciting them has to acknowledge from whom he has learnt them. A good poet has an entourage of admirers some of whom learn his poems by heart and recite them. Others hear these recitations and memorise the poems they consider sufficiently beautiful and important. Andrzejewski and Lewis show that it is indeed memorisation rather than simultaneous composition/performance that is involved here.
While we may admire Somali poets for achieving worthwhile results in the very difficult medium of Somali prosody, we are no less impressed by feats of memory on the part of the poetry reciters, some of whom are poets themselves. Unaided by writing they learn long poems by heart and some have repertoires which are too great to be exhausted even by several evenings of continuous recitation. More over, some of them are endowed with such powers of memory that they can learn a poem by heart after hearing it only once, which is quite astonishing, even allowing for the fact that poems are chanted very slowly, and important lines are sometimes repeated. The reciters are not only capable of acquiring a wide repertoire but can store it in their memories for many years, sometimes for their lifetime. We have met poets who at a ripe age could still remember many poems which they learnt in their early youth.
In the nomadic interior whole villages move from place to place and there is constant traffic between villages, grazing camps, and towns. Poems spread very quickly over wide areas and in recent times motor transport and the radio have further accelerated the speed with which they are disseminated.
A poem passes from mouth to mouth. Between a young Somali who listens today to a poem composed fifty years ago, five hundred miles away, and its first audience there is a long chain of reciters who passed it one to another. It is only natural that in this process of transmission some distortion occurs, but comparison of different versions of the same poem usually shows a surprisingly high degree of fidelity to the original. This is due to a large extent to the formal rigidity of Somali poetry: if one word is substituted for another, for instance, it must still keep to the rules of alliteration, thus limiting very considerably the number of possible changes. The general trend of the poem, on the other hand, inhibits the omission or transposition of lines.
Another factor also plays an important role: the audience who listen to the poem would soon detect any gross departure from the style of the particular poet; moreover, among the audience there are often people who already know by heart the particular poem, having learnt it from another source. Heated disputes sometimes arise between a reciter and his audience concerning the purity of his version. It may even happen that the authorship of a poem is questioned by the audience, who carefully listen to the introductory phrases in which the reciter gives the name of the poet, and, if he is dead, says a prayer formula for his soul.
(Andrzejewski and Lewis, 1964, pp. 45–6)
In this case, then, memorisation is indeed involved, and the concept of a ‘correct’ version is locally recognised. This is not what one would expect if one relied mainly on the analogy of Yugoslav singers or the analyses of the oral-formulaic school. Somali poetic composition shows that there are other modes of composition in oral poetry.
Equally instructive is Gordon Innes’s research into Mandinka griots’ narrations (1973, 1974). He found a fascinating blend of stability and change, with both memorisation and fluidity involved.
These narrations are complicated, so far as style and genre are concerned, for a blend of ‘speech’, ‘recitation’ and ‘song modes’ are used in delivery (Innes, 1974, pp. 15ff). On the face of it, only the recitation and song modes seem to qualify as ‘poetry’, and the so-called ‘Sunjata epic’ would be largely prose since most of the narrative is in the ‘speech mode’. But it is a marginal case, for even the speech mode in these narrations has to fit the musical accompaniment which is usually built on a two-, four- or eight-bar phrase, characterised by a distinctive melodic, tonal or rhythmic structure (King in Innes, 1974, p. 18). In these respects, then, even the speech mode has constraints we associate with poetic composition. It is thus not irrelevant to include here some account of composition in the various versions of this ‘epic’ as providing a partial parallel to Yugoslav epic composition.
At first sight, the different versions of the ‘Sunjata epic’ throughout the Manding area of West Africa seem to offer an exact instance of the blend of composition and performance familiar in Lord’s writings. It concerns the exploits of the great hero Sunjata who established himself as king of Manding and Susu in the thirteenth century. Many versions of the story are extant even in the small area of the Gambia alone. As Innes writes, ‘From these and from published versions from elsewhere in the Manding area one almost has the impression that the Sunjata legend consists of a repertoire of various motifs, incidents, themes (call them what you will), and that each griot makes a selection which he strings together into a coherent narrative’ (Innes, 1973, p. 105). He goes on to show that this initial impression is wrong, for it suggests greater fluidity than is the case. Innes made a detailed comparison of a number of versions. One set was recorded by two brothers, Banna and Dembo Kanute, both regarded as outstanding performers. They had learnt their craft from their father, assisted (in the case of Banna) by Dembo, his elder brother. Innes notes the differences in their versions. Starting presumably from the same repertoire, their performances differed in the ground covered in two major incidents, and in a number of details. This is the sort of fluidity one would expect from the comparative Yugoslav material. Furthermore there are indications that each brother adapted his version to the situation in which he performed—when leading persons present in the audience, for instance, could trace their descent from figures in the Sunjata story.
But when Innes came to a detailed comparison between two versions by the same griot—Bamba Suso—a different picture emerged. Bamba was one of the leading griots in the Gambia, with extensive historical knowledge. In his seventies, two recordings were made of his version of the Sunjata story, one for Radio Gambia, another at Brikama to an audience which included Innes himself. The most striking point to emerge from a comparison of the two is their close similarity, in places amounting to word-for-word repetition. Here are two passages quoted by Innes to illustrate the point:
In these extracts, there is certainly not word-for-word identity throughout. But there is much more verbal and line-for-line repetition than one might expect from the Yugoslav analogy. It is also clear that to some extent memorisation is involved here; at the very least it would be misleading to insist that ‘original composition’ played a large part in the performance. The model of simultaneous composition/performance must here be modified: much of the ‘composing’ must have preceded the performance—to a greater extent than in the composition/performance of the Yugoslav Avdo Mededović or the Kirghiz bards described by Radlov. One cannot, then, assume that the composition/performance process so well illustrated in the case of Eastern European singers by Lord and others is always characteristic of oral composition. Much (though not necessarily all) of the composing may take place before the moment of performance.
Hence the need for more careful and detailed research on the respective parts played by composition, memorisation and performance both in particular situations, and also by different individual singers—even by the same singers at different times. On the Mandinka evidence, Innes sums up his findings as follows:
At first sight the two pieces of evidence presented here seem to contradict each other. The evidence from the Kanute brothers shows that in the course of his professional career a griot’s version of the Sunjata legend may undergo considerable change. The evidence from Bamba, on the other hand, shows that a griot’s version may remain remarkably stable, both in content and language, over a period of time. Different interpretations of this evidence are no doubt possible, but, taken along with other evidence, it suggests to me a pattern of life in which a griot in his younger days travels extensively, listens to other griots and borrows selectively from them, repeatedly modifying his own version until eventually he arrives at a version which seems to him the most satisfying. With repetition, this version will become more or less fixed, and even the words will tend to become fixed to some extent. But even this version will of course vary from performance to performance, depending upon such factors as who happens to be present and in whose honour the performance is being given.
Many of the recorded cases where memorisation predominates rather than the composition-in-performance characteristic of the Yugoslav model, derive from lyric and shorter forms of poetry. Lengthy epic poetic narrations, lasting over several hours or nights of performance, are naturally likely to fit the model of composition-in-performance given by Parry and Lord. Yet Somali poems, it must be recalled, can extend to several hundreds of lines and involve at least some element of narration, while the lengthy Mandinka ‘epic’ is partly in ‘poetry’ in its recited and sung portions and arguably has poetic elements even in the more prosaic ‘spoken’ parts.
There are some relatively long quasi-narrative forms where memorisation and exact recollection are sometimes more important than creativity in performance. In western culture some ballad singers are in this category. They can be classified as ‘passive traditors’ of memorised words, aiming both in practice and in their own accounts at the exact reproduction of what they have heard from others. That this is not the only way in which ballads are transmitted and performed is well known (see also chapter 5 below), but it is nevertheless one accepted mode for the transmission of ballads and other oral poetry in the west. Some authorities go so far as to claim that, with few exceptions, the norm for the European-American folk performer is repetition (Glassie, 1970, p. 32) and suggest that this contrasts with the variability of the Afro-American tradition. But there are a number of other African examples where memorisation is important. The long panegyric poems of Ruanda and South Africa are often cited as outstanding examples of oral poetry. They commonly run to hundreds of lines and have an element of narrative, though the main emphasis is on praise. Yet in Ruanda there was often memorisation of received versions of the praise poems, with minimal variation in performance, and the original composers were remembered by name (Kagame, 1951). And for the Zulu a recent detailed study states categorically that the specialist praise singers attached to the courts were concerned more with ‘performance’ than ‘composition’: the singer ‘has to memorise [the praises of the chief and ancestors] so perfectly that on occasions of tribal importance they pour forth in a continuous stream or torrent. Although he may vary the order of the sections or stanzas of the praise-poem, he may not vary the praises themselves. He commits them to memory as he hears them, even if they are meaningless to him’ (Cope, 1968, pp. 27–8).
It is therefore clear that a single model of the relation of composition to performance will not cover all cases—perhaps not even all cases of narrative poetry. To accept uncritically Lord’s dictum that what is important in oral poetry is ‘the composition during oral performance’ would blind us to the differing ways in which the elements of composition, memorisation and performance may be in play in, or before, the delivery of a specific oral poem.
Once the possibility of prior composition followed by memorisation is conceded it becomes obvious that there are many cases. Many work songs are of this kind.
Fig. 3.1. Limba young men jointly singing a work song whilst rice threshing in northern Sierra Leone. This both keeps the threshers in time with each other (thus avoiding accidents) and adds musical impetus and enjoyment to the hard work. Photo by Ruth Finnegan, 1961.
To be sure, they often provide an inventive and skilled leader with the opportunity for improvisation and elaboration that gives play to his originality. But often the work involved is so demanding or the choral element so strong—for instance in some of the Texas prison work songs discussed in Jackson (1972)—that for the most part the words are merely repeated from memory: and yet they are surely still ‘oral poetry’. The same goes for many Irish popular songs, Christian hymns, liturgical poetry, or genealogical verse—all of which have some element of ‘oralness’ about them. In performance they are all subject at times to variation and adaptation—and perhaps it is this opportunity rather than actual variation that is ‘typical’ of oral compositions. But it must be accepted that in many performances songs are not in practice much changed: the performance is from memory. The naive model of a child learning by heart in order to recite is after all not always misleading.
To admit the possibility of memorisation in oral literature is not, however, to go back to the idea of passive reception from memorised ‘tradition’. The prior composition involved is not infrequently known to be by named and individual poets, consciously labouring over the difficult task of constructing the words (and sometimes music) of their poetry.
There are many recorded cases of oral poetry where its creation does not fit either of the two extreme cases—improvisation at the moment of performance or blind acceptance of ‘tradition’—but is due to long deliberation by the individual poet before the performance. This is particularly the case with shorter, more lyrical poetry, rather than the narrative poems mainly stressed by the oral-formulaic school. Though Lord asserts at one point that he is only concerned with narrative poets, it is easy to come away with the impression that all oral poetry is subsumed under the same general rubric, and that ‘oral composition’ is only oral if it takes place at the moment of performance. This impression is reinforced by Lord’s generalised definition of ‘oral poetry’ in his encyclopaedia article on the topic—‘poetry composed in oral performance’ (1965, p. 591)—and by the many analyses in ‘oral-formulaic’ terms of lyrical and ballad poetry as well as epic. Since this impression that all oral poetry is composed on the Yugoslav model is not correct it is worth illustrating the process of deliberate composition prior to performance with some further examples.
Inuit poetry provides an excellent instance. Long and careful consideration is given to the composition of the words of many Inuit poems before their performance, and the Inuit are extremely articulate about the problems and delights of composition; indeed it is a constant preoccupation of many of the poems. Here the poet Piuvkaq compares the difficulties of fishing with those of poetic composition:
... Why, I wonder
My song-to-be that I wish to use
My song-to-be that I wish to put together
I wonder why it will not come to me?
At Sioraq it was, at a fishing hole in the ice,
A little trout I could feel on the line
And then it was gone, I stood jigging
But why is that so difficult, I wonder? ...
Fig. 3.2. Photo of an Iñupiat Inuit mother, father, and son, photographed in Noatak, Alaska by Edward Sheriff Curtis, circa 1929. Wikimedia, https://en.wikipedia.org/wiki/File:Inupiat_Family_from_Noatak,_Alaska,_1929,_Edward_S._Curtis_(restored).jpg
.jpg){kind=link}
The Inuit combine their awareness of the deliberate and conscious struggle involved in ‘putting together words into a song’ with the concept of poetic inspiration. One of the outstanding Inuit poets was Orpingalik, who used to call his songs his ‘comrades in solitude’ and ‘his breath’. He described some of the processes of composition to Rasmussen in these terms
Songs are thoughts, sung out with the breath when people are moved by great forces and ordinary speech no longer suffices.
Man is moved just like the ice floe sailing here and there out in the current. His thoughts are driven by a flowing force when he feels joy, when he feels fear, when he feels sorrow. Thoughts can wash over him like a flood, making his breath come in gasps and his heart throb. Something, like an abatement in the weather, will keep him thawed up. And then it will happen that we, who always think we are small, will feel still smaller. And we will fear to use his words. When the words we want to use shoot up of themselves—we get a new song.
The same emphasis on inspiration and waiting for the right words to be born is given in an Alaskan Inuit’s description of how one must wait in silence and ‘stillness’ for the poems to come. Then ‘they take shape in the minds of men and rise up like bubbles from the depths of the sea, bubbles that seek the air to burst in the light’ (quoted in Freuchen, 1962, pp. 280–1). This process is quite unlike composition in the heat of performance such as we associate with Yugoslav epic singers; for this waiting for inspiration is expected to take place, not at the moment of public performance, but in the dark, in deep silence, as an act of artistic concentration. Walking about outside in solitude is another common occasion for Inuit poetic composition. Rasmussen describes how great pains are taken to put the words together skilfully so that ‘there is melody in them, while at the same time they are pertinent in expression’, and how ‘a man who wants to compose a song may long walk to and fro in some solitary place, arranging his words while humming a melody which he also has to make up himself’ (Rasmussen, 1931, p. 320).
Perhaps the most vivid expression of the combination of hard work and of heightened emotive perceptiveness inherent in Inuit poetic composition is the Inuit poet Sadlaqé’s account of trying to compose a song
Once when I was quite young, I wished to sing a song about my village, and one winter evening when the moon was shining, I was walking back and forth to put words together that could fit into a tune I was humming. Beautiful words I found, words that should tell my friends about the greatness of the mountains and everything else that I enjoyed every time I came outside and opened my eyes. I walked, and I continued walking over the frozen snow, and I was so busy with my thoughts that I forgot where I was. Suddenly, I stood still and lifted my head up, and looked: In front of me was the huge mountain of my settlement, greater and steeper than I had ever seen it. It was almost as if it grew slowly out of the earth and began to lean out over me, deadly dangerous and menacing. And I heard a voice from the air that cried out: ‘Little human! The echo of your words has reached me! Do you really think that I can be comprehended in your song?’
(Freuchen, 1962, pp. 279–80)
Improvisation or adaptation in the moment of performance is not completely unknown in Inuit poetry—witness for instance the special song an old woman sang to welcome the visiting Rasmussen. But the emphasis seems to be on deliberate and studied composition, with recognised personal ownership of particular songs.
Another clear instance of self-conscious and painstaking composition separated from the act of performance is documented for oral poets in the Gilbert Islands of the Southern Pacific. This Gilbertese love song gives a first impression of apparent spontaneity:
How deep are my thoughts as I sit on the point of land
Thinking of her tonight,
Her feet are luminous over dark ways,
Even as the moon stepping between clouds,
Her shoulders shine like Kaama in the South1
Her hands, in the sitting dance,
Trouble my eyes as the flicker of stars;
And at the lifting of her eyes to mine I am abashed,
I, who have looked undaunted into the sun.
But such a poem rests on a long process of deliberate composition. Grimble has described how when a Gilbertese poet ‘feels the divine spark of inspiration once more stirring within him’, he leaves the village and goes off to some lonely place where he can do the initial work on his composition alone: ‘This is his “house of song”, wherein he will sit in travail with the poem that is yet unborn. All the next night he squats there, bolt upright, facing east, while the song quickens within him’. Next morning he returns to the village to collect a group of friends to help him. It is their job to criticise and assess the poem—‘to interrupt, criticize, interject suggestions, applaud, or howl down, according to their taste. Very often they do howl him down, too, for they are themselves poets. On the other hand, if the poem, in their opinion, shows beauty they are indefatigable in abetting its perfection’. They spend the whole day with the poet, working with him on his ‘rough draft’—‘searching for the right word, the balance, the music that will convert it into a finished work of art’ (Grimble, 1957, pp. 204–5). After a day spent in this joint process, the friends leave and the poet is left on his own once more. ‘He remains alone again—probably for several days—to reflect upon their advice, accept, reject, accommodate, improve, as his genius dictates. The responsibility for the completed poem will be entirely his’ (Grimble, 1957, p. 205).
The result of this long-drawn out process of oral composition is that the poem as finally produced has been worked and re-worked over many days. This process results, in Arthur Grimble’s words, in ‘clear-cut gems of diction, polished and repolished with loving care, according to the canons of a technique as exacting as it is beautiful’ (p. 200). The Gilbertese, he holds, are ‘consummate poets’ who, ‘sincerely convinced of beauty, enlisted every artifice of balance, form and rhythm to express it worthily. The island poet thrills as subtly as our own to the exquisite values of words, labouring as patiently after the perfect epithet’ (Grimble, 1957, p. 200).
There are many other documented examples of deliberate and protracted composition, divorced from the act of performance. Ila and Tonga women in Zambia make personal songs where the owner, working out the words and tune of the song, ‘sings it in her heart’ until it is time to stand up and sing it in public (Jones, 1943, pp. 11–12). There is a Ruanda custom of memorising praise poems by other named poets, whose prior composition is recognised (Kagame, 1951). The Dinka use an expert to make up a song to their requirements, for later performance by the ‘owner’ rather than the composer (Deng, 1973, p. 85). The Pueblo poet puts much preparatory work into composing a new song for an approaching festival:
Yellow butterflies,
Over the blossoming virgin corn,
With pollen-painted faces
Chase one another in brilliant throng...
(Curtis, 1907, p. 484)
Mediaeval Gaelic court poets composed their poems orally in a darkened room. When the poem was complete it was recited or chanted to the chief not by the poet but by a bard who, according to a near-contemporary source, ‘got it well by heart, and now pronounc’d it orderly’ (Knott and Murphy, 1967, p. 64).
Where, as in the Gaelic case, there is a distinction between poets (responsible for composition) and reciters (responsible for performance) the situation is self-evidently unlike the Yugoslav model where ‘singing, performing, composing are facets of the same act’ (Lord, 1968a, p. 13). The distinction is known in other poetic traditions; compare the one sometimes made between the mediaeval European trobador (composer) and joglar (performer), the Ruanda and Somali poets, as opposed to reciters (Johnson, 1971, p. 29), or the Dinka expert composer who is sometimes distinct from the performers of his poem (Deng, 1973, p. 85).
Where writing enters in—and (see chapter 1) it is hard to draw a strict line excluding it—possible variations in modes of composition become even more complex. There are cases of prior composition with some reliance on writing which is then used, in varying degrees, as an aid to memory for later oral performance. Both written narratives and brief notes were sometimes used by mediaeval Chinese ballad singers (Doleželová-Velingerová and Crump, 1971, pp. 2, 8), and texts of Irish street songs or handwritten ‘ballets’ of English, Scottish and American ballads have often formed one basis for later oral performance. All these can play some part in the process of composing ‘oral poetry’.
The situation is made even more complex—and the possibility of a single generalisation about the nature of ‘oral composition’ the more remote—because of differences between ‘cultures’ or poetic traditions in the wide sense, and also between individual composers and differing genres within one poetic culture. American collectors have grown accustomed to working with the distinction between ‘active’ and ‘passive’ singer, and a number of studies have been devoted to the differing styles and creativity of individual composers. As early as 1908 Percy Grainger was discussing the ‘Impress of personality on traditional singing’, and more recently Wolf has related the amount and nature of creativity in ballads to the differing personalities of singers (Wolf, 1967). Innes’s research into the stability and change in Mandinka griots’ narrations is another case in point.
It is also well known that in many poetic traditions some genres are recognised as less innovative and creative—with more emphasis on memorisation, less on composition—whereas in others there can be a high degree of creativity by the individual poet. Among the Zulu, for instance, the praise poems of long-dead chiefs were fixed in form, and their recitation depended on memorisation and repetition; but some more recent praise poems are ‘uttered on the spur of the moment by an imbongi inspired by the presence of his chief or eager to incite his audience to loyalty for the chief bear the individual stamp of the singer’ (Opland, 1971, p. 172): here composition-in-performance is the expected form. Another variation is found among the Dinka. There some types of song have to be composed in a short time—for instance, initiation and ‘cathartic’ songs—and it is common for an expert to compose rapidly while his listeners help by memorising the song for him, the composer himself often being unable to remember it when he has finished. With songs where less haste is necessary, as with ox songs, that help is not needed and the ‘composers create at leisure’ (Deng, 1973, p. 85).
More than one person may be involved in composition. I refer here to something over and above the effect that audiences can have on the performing poet—as described by Radlov for Kirghiz minstrels—or even beyond the help that friends and colleagues of a poet can provide, as in the Gilbertese instance. There are cases where two or more people seem to be clearly credited with, and hold the responsibility for, the composition of a particular piece. Emeneau describes how in Toda poetry there are different ways in which a song can be composed and performed.
Solo composition is only one of the manners of delivery, and perhaps not the commonest. It must be obvious that with all details of composition closely dictated by the technique, duet and choral delivery is always possible. All the performers will have a good knowledge of the technique and will know what is being sung about. The first unit, even the first syllable of the first unit that is uttered by the chief performer almost always gives a certain clue to the limited possibilities of the two-dimensional structures that he intends to use; a quick intelligence on the part of his accompanists does the rest. In the dances there is usually a chief composer assisted by one companion; they shout in unison. If a song is being sung, the composer whistles the tune first, and those who sing with him can then accompany him in unison, or he may sing the first half of each two-dimensional structure and a single accompanist may sing the other half antiphonally, or a large group may likewise split up to perform antiphonally. My impression is that group performances are preferred to solo work.
Similar co-operation and interaction between two authors can be seen in a number of dialogue songs,2 or the kind of poetic duels which can be interpreted as resulting in a single poetic composition, produced jointly by two composers. Larger groups too can be involved. If we leave on one side general theories about the ‘public’ or the ‘folk’ gradually affecting a composition over time, there is still evidence that a number of people sometimes take an active part in the composition of a piece. Ben-Amos mentions that song composers from Benin in Midwestern Nigeria explained that ‘they often composed a song alone, but that the group of singers to whom they belonged reworked it afterwards until everybody was pleased’ (Ben-Amos, 1972, p. 7). A more rigorous system is reported for Hawaiian oral poetry.
A single poet working alone might produce the panegyric, but for the longer and more important songs of occasion a group got together, the theme was proposed and either submitted to a single composer or required line by line from each member of the group. In this way each line as it was composed was offered for criticism lest any ominous allusion creep in to mar the whole by bringing disaster upon the person celebrated, and as it was perfected it was committed to memory by the entire group, thus insuring it against loss. Protective criticism, therefore, and exact transmission were secured by group composition.
(Beckwith, 1919, p. 28)
Evidence of this kind of joint authorship is not very clear-cut. But we do have to consider dual or multiple authorship as one possible variable in a study of the processes of oral composition.
It becomes clear that if we scrutinise the concept of ‘oral composition’ in the light of the comparative empirical evidence it turns out not to be one single and unique process as is often implied by scholars of the oral-formulaic school. The Yugoslav evidence about the process of composition-in-performance can be a useful analogy in cases where (as with Kirghiz epic) the processes of composition and performance are closely fused—but there are cases in which the analogy is not relevant, and we cannot assume in advance that it will necessarily apply in a given case.
The processes of composition, memorisation and performance in oral poetry turn out to be more complex than was once supposed. We can no longer accept Lord’s definitive generalisation about composition in oral poetry—that ‘with oral poetry we are dealing with a particular and distinctive process in which oral learning, oral composition, and oral transmission almost merge; they seem to be different facets of the same process’ (Lord, 1968a, p. 5). The reality is more interesting than any monolithic theory. There turn out to be different combinations of the processes of composition, memorisation and performance, with differing relationships between them according to cultural traditions, genres and individual poets. There are several ways—and not just one determined way suitable for ‘the oral mind’—in which human beings can engage in the complex processes of poetic composition.
3.5 Conclusion
So the relationship between composition, memorisation and performance in oral poetry is more open than the definitive-sounding term ‘oral composition’ seems to imply. The ‘oral-formulaic’ style of composition (as depicted in The Singer of Tales and similar work) is not a sufficient indication for concluding that a given work is ‘oral’, nor a necessary condition for the creation of ‘oral poetry’.
Yet Lord’s insistence on moving away from the concept of a written model has illuminated and stimulated all studies of oral poetry; for showing this and for indirectly conveying to western readers the approaches of earlier Russian scholars on the poetic creativity of oral performers there is much to thank the oral-formulaic school. Even if the definition of ‘formula’ has been both too vague and too inflexible and limited to cover other formative constraints on, and opportunities of, an oral composer—some are discussed in the next chapter—it is in large measure due to the proponents of the term that we can now grasp so much better how an oral poet is both constrained and free in his composition, representing at once ‘old’ and ‘new’, ‘tradition’ and ‘originality’. We can understand more fully how the modern improvising composer in the west can stress that he both engages in free improvisation and at the same time works within an accepted background, or appreciate the ending of the mediaeval Chinese singer’s rendering of ‘The ballad of the hidden dragon’:
I was asked to make a new tale from the old.
For this worthy, intelligent assembly
I was happy to unfold
The story of Liu Chih-yüan
From the beginning to the end
And with absolutely nothing left untold.
(Doleželová-Velingerová and Crump, 1971, p. 113)