1. THE FORMULA

By the time of Agnes Lyle, ballad language had lost its pristine purity, if indeed it ever possessed such. In particular, the language of this singer is sometimes “infected” with broadside diction, especially in “Geordie,” “Jamie Douglas,” and lines like the following from “The Eastmure King”:

But such use of broadside diction does not necessarily require that a broadside text be posited in the immediate ancestry of a particular ballad. Singers of the late classical period use broadside diction as one of the resources they draw upon to form a ballad and treat broadside-based couplets and stanzas in the same free way they treat other older formulas. Their ballads are, paradoxically, no less oral for an occasional echo of the broadside press. “The Eastmure King,” for instance, one of the most “infected” Lyle texts, is fully oral in its architectonics. Even the diction is full of the singer’s favorite phrases: “Milk-white steed,” “Hold your tongue,” “Out and spoke,” “When bells were rung and psalms were sung,” and so on. The entire text, in fact, is typical of the singer at her most controlled and most formulaic.

Whatever the feelings about broadside diction, however, no critic seriously quarrels with the judgment that ballad language in general is formulaic in some sense. The problem hitherto has been that theorists tried to devise and use a single univocal definition of the words formula and formulaic. This attempt has been especially troublesome for oral studies (see especially Andersen 1985, passim; Foley 1988, 63, 69-70, 100-102). Formula is an analogical term.¹ Because it is an analogical term, a definition such as Andersen’s, developed to describe phenomena on the level of commonplace, will fail when applied to phenomena on other levels. Andersen concludes from this failure that formulas are found only on the commonplace level. Such a conclusion and such a use of the word seems too restrictive. Corresponding phenomena on other levels are also formulas, each in an analogous though somewhat different sense, with distinctive morphology and distinctive modes of operation.

Many of these formulaic levels have been identified and discussed in the work of Buchan and Andersen, in preceding chapters of this study, and elsewhere. The work of Anders, in particular, provides a convenient catalog of many hundreds of formulaic expressions culled from Child and other significant collections. It is now possible to draw up a fairly complete typlogy of formulaic usage in the Anglo-Scottish ballad. The surest basis for such a classification of formulas is linguistic level, for characteristic formulaic elements can be identified on all linguistic levels from phoneme to total ballad text. But because the formula is usually a metrical as well as a linguistic unit it will sometimes be more convenient to denominate categories with metrical terms (e.g., half-line formulas).

Of course formulaic language is all interconnected. It is seldom possible to isolate formulas, except conceptually. And what is one kind of formula on one level may well function as part of a larger formula on another level, as in the case of the formulaic I alliterations in the letter commonplace, to be discussed below. Still, when independent free-floating formulas can be singled out in the following discussion, they will.

The simplest linguistic level is that of phoneme. Whether or not purely phonemic phenomena should be called formulas, they are certainly formulaic at times. Alliteration is one of these formulaic phenomena; it is often what makes a phrase spring immediately to the lips, a phrase such as “daughter dear,” “busk the bride,” “large and lang,” “cheer and charcoal clear,” “fair and full,” “gien me the geeks,” and hundreds more. In some cases the alliteration itself seems to be the formula, the matrix into which the expression must be fitted. A notable example occurs in the familiar commonplace describing the reading of a letter. The first line of this commonplace seems to have an I alliteration formula embedded in it. In “Johnie Scot” Agnes Lyle realizes this line thus:

The corresponding line in “Sir Patrick Spens” is similar. But in “Lord Derwentwater” she sings:

There is still an l alliteration, even without mention of the letter. The l alliteration is preserved, sometimes with and sometimes without mention of the letter, in the following examples from other Motherwell texts:

The alliteration frequently carries over into the second line, with mention of loud laughter. Of course there are realizations of the commonplace which do not include the alliteration. It is not a universal invariant in the commonplace but rather a formulaic element of the commonplace in the repertoires of some singers, and perhaps of some regions.

Rhymes, both internal and final, constitute another form of phonemic formula. The relatively limited selection of rhymes to serve a wide variety of purposes within a repertoire is an interesting example of formulaic economy. Singers seem to think in terms of their set of habitual rhymes when fitting their expression to the demands of rhymed strophes. Buchan (1972, 155-158) provides dramatic evidence of how a stanza can change in characters or action within the matrix of a common rhyme. In some ballads, such as the Yarrow ballads and “Barbara Allen,” a particular set of rhymes common to the tradition is a dominant formulaic element.

One step up, on the level of morpheme, particular rhymes sometimes form a mold into which the expression of lines and even stanzas is fitted. One of the most common ways this happens is with word pairs analogous to the word pairs in Hebrew parallelism. The following example from “Earl Richard” will help clarify the concept. Stanza 13 runs:

Stanza 15 is an iterative repetition of this stanza:

In these correlated stanzas the words knee and hand function as a word pair. Substitution of one word for the other in a word pair permits iterative repetition in parallel ballad stanzas in a way analogous to the way word-pair substitution permits repetition in parallel verses of Hebrew poetry. A similar word pair, sleeve / gore, occurs in versions of “Johnie Scot,” including Mrs. Brown’s:

Seam and son in stanzas 15 and 16 of the Lyle “Sir Patrick Spens” constitute another such word pair. Other singers use other word pairs at this point in the narrative. The Child B and O texts, for instance, use the word pair hand / knee (ee, that is, eyes, in the B text) already seen in connection with “Earl Richard,” while other texts structure stanzas around rhymes on hand and hair.

The next highest level of linguistic organization on which formulas can occur is that of the phrase.² Formulas on this level include ballad clichés such as “cheer and charcoal,” “wee pen knife,” “beaten gold,” “milk-white steed,” “silver slippers,” “houkit a grave,” “wide and deep,” and so on. These phrases are often unvarying, but not always. The grave can be “dug” as well as “houkit,” and it may be “long and narrow” instead of “wide and deep.” Likewise, the lady may wear “silver shoon” instead of “slippers,” or offer “coal and candle light” instead of “cheer and charcoal.”

While many formulaic phrases are a half-line long, the true half-line formula seems to be a distinct type that includes many formula families. The most noticeable of these families is very simple: two one-syllable words are repeated once to make a half line, usually at the front rather than the back of the line: “Rise up, rise up,” “Come down, come down,” “Oh no, Oh no,” “A priest, a priest,” “A boon, a boon,” “Ye lie, ye lie,” and so on. A half-line formula of this family is usually followed by another half-line formula, either an epithet, a name, or an inquit: “My bonny boy,” “Little Mossgrey,” “Lord Barnabas said.”³ In expanded form the formula frequently spills over into the next line:

A related type of terracing occurs when a two-beat half-line formula is repeated to fill a whole line or even two lines:

As the foregoing examples make clear, the essential characteristic separating half-line formulas from formulaic phrases is utilization of the half-line as a distinct narrative unit.

Whole-line formulas may be four-beat formulas, three-beat formulas, or variable. Linguistically they are either phrases or clauses but not really full periods. Many free-floating four-beat formulas function as incipits or narrative connectors; e.g.,:

Free-floating formulas which exist only in a three-beat form are rare. The Lyle repertoire includes variations of “She sealed it with a ring” and “To her love in Dundee” (“To Will ayont the sea”). But the three-beat lines are also the place for formulas built around grammatical forms. Here occur longer prepositional phrases, such as “Beneath the garden tree,” “Since Saturday at morn,” “Upon the road so high,” “Instead o beaten gold,” and “For wantonness and play.” Here occur many participles: “The tears came trinkling down,” “But I am coming home,” “Rocking her eldest son,” “Thinking to be slain,” “Sits gabbing on the tree,” and “Coming marching in their sicht.” Here occur the relatively rare frequentatives, futures, and compound verbs, such as “About its neck did hing,” “Did him a traitor call,” “Shall ring for her the morn,” and “O busk and mak you braw.” And here abound infinitive phrases such as “To see men doing the same,” “To learn some unco lair,” “And for to kiss her sweet,” “To fight for King Jamie,” “My grave for to fill,” “The city for to see,” and “Torches for to burn.” Some “grammatical” formulas fall into easily identifiable patterns. In the last three infinitive examples just given, for example, there is an underlying structure which might be diagrammed thus:

(direct object) + for + (infinitive)

But in some cases there does not seem to be any underlying use of a widespread pattern: the formulaic idea seems to extend no further than filling out the short line by using an infinitive, participle, compound form, or preposition.

Three-beat formulas can be expanded to fill a longer line by the addition of a name, inquit, or epithet:

And four-beat formulas can be shrunk to fit the shorter line by dropping these optional elements, as well as in various other ways, as the following examples of repetition show:

Four-beat and three-beat formulaic phrases combine to form larger patterns. Thus, in the Lyle repertoire, the four-beat formula “Out and spak her ain bridegroom (. . . an auld rich knight,” etc.) is followed by the formula “And an angry man was he (. . . a sorry man,” etc). Even when the phraseology of the second line disappears, the ee rhyme remains, as in “An ill death may he die” (“Sir Patrick Spens,” 2), or “And a blithe blink from his ee” (“Lord William,” 18). Many other kinds of half-stanza formulaic expressions dot the repertoire. Some of these are fairly redundant, formed by various kinds of expansions and repetitions, as the discussion of smaller units has indicated. But some full-scale, two-line formulas convey fresh information in the second line. Compare, for instance, the following Lyle realizations of a single formula:

The first realization is a fairly straightforward description of William’s matriculation at a foreign university. The second is an ironic comment on Mary’s loss of innocence at what should have been a school of gentility. In each case the second line adds to the information contained in the first line.

The next level of linguistic organization, after phrase and clause, is full period or sentence. In the ballad a sentence usually fills out a stanza (cf. Lord 1985, 403). And as sentences are formed of phrases and clauses, so full-stanza formulas are invariably formed, at least in part, of smaller formulaic elements. One of Agnes Lyle’s favorite formulas is the “Come saddle to me” commonplace. Her four realizations, in “The Gypsy Laddie,” “Geordie,” “Lord William,” and “Lord Derwentwater,” are all multiforms of the same basic structure and idea, both over all and line for line:

Line 1: Call for a horse; e.g.,: “Come saddle to me my horse, he said.”

Line 2: Emphatic repetition; e.g.,: “Come saddle to me with speed.”

Line 3: Statement of length of journey; e.g.,: “For I must away to fair London town.”

Line 4: Statement of reason for journey; e.g.,: “For me was neer more need.”

As a unit this commonplace is found widely in Anglo-Scottish tradition and is one of the thirty or so commonplaces with distinct supra-narrative function to which Andersen assigns a privileged position in the tradition. But the first line, in three of the four Lyle realizations, also belongs to its own linear formula family. Mary Hamilton’s plea, “ ‘But bring to me a cup,’ she said,” belongs to the same family and like the realization in the “Come saddle to me” commonplace, is followed by a restatement of the line.

Some full-stanza formula families are quite diverse, including both commonplaces and nonce stanzas. The third stanza of “Fair Janet” is the commonplace:

This formula is related to the commonplace stanza 2 of “Mary Hamilton”:

Each of these commonplaces has multiform variants. Andersen considers them separate formulas, but on a deeper level they are united not only in structure and diction, but also in narrative idea. This essential narrative idea is that life has scarcely progressed a certain distance in time or space when (“Till”) something bad happens. The commonplace “She hadna pu’d a flower, a flower” is a close parallel on this deeper level. It may be assigned to the same larger family, as may the following context-bound formulations from the “Fair Janet” tradition:

The essential narrative idea in all these expressions of the formula is that life has scarcely progressed a certain distance in time or space when (“Till”) something bad happens.

Stanzaic formulas regularly form stanzaic pairs. The many ways these double formulas occur have already been indicated in the discussion of pairing techniques in Chapter 3 and in the discussion of word pairs, above. It is probably well, however, to distinguish pairs united only by some form of repetition from pairs united by the necessity of more than one stanza to convey the full narrative idea. The latter group includes some common ballad incipits and explicits. The “Rose and briar” commonplace, for instance, makes full narrative use of its two stanzas to describe the graves first and then the intertwining plants that grow from the graves. The “Hold your tongue” commonplace that ends a number of the ballads likewise takes two stanzas to express its complete idea:

Just as stanzaic formulas include lines and couplets which exist as formulas on their own level, so the two-stanza “Hold your tongue” commonplace includes a stanza with a life of its own as a one-stanza commonplace. Apart from the “Yarrow” rhymes—which come from the formulaic tradition of the “Yarrow” ballads, not of the “Hold your tongue” commonplace—the two stanzas are independent of each other in diction and sentence structure, and each conveys its own information. In addition to its occurrence in “The Braes o Yarrow,” this true two-stanza formula is found at the end of the Lyle “Jamie Douglas” and (arguably) “Mary Hamilton,” as well as (much modified) “Sheath and Knife,” and in many texts outside the repertoire.

After the stanzaic pair, the next highest level of formulaic organization is the cluster or run. A run always consists of a series of commonplaces. It can be lengthened or shortened by including or omitting commonplaces. But a given run does not really provide choice among commonplaces to be used or permit inclusion of other non-commonplace formulaic lines and stanzas. One common run includes the commonplaces “Where can I find a bonny boy,” “Here am I a bonny boy,” “When he came to the broken bridge,” and “What news, what news.” Other “Bonny boy” runs utilize other combinations of commonplaces. The “Long letter” commonplace, sometimes inaugurates a run that includes “The first line he read,” “Come saddle to me my horse,” and “Out and spoke.”

Runs sometimes function as “themes” in balladry, but the two can be distinguished. A run is a sequence of commonplaces, but a “theme” is a sequence of narrative elements; the formulas and commonplaces used to express that “thematic” sequence can vary from ballad to ballad and text to text. When Lord says, “The theme, even though it be verbal, is not a fixed set of words, but a grouping of ideas,” the student of ballads wants to add, “Not a fixed set of commonplaces either.” Another way to express the distinction is to say that a run is a surface structure, but a “theme” is a deep structure.⁴ Thus, while Agnes Lyle’s “King’s Letter Theme” invariably includes the “Long letter” commonplace, the development from there is different in each of the realizations of the “theme.” In “Johnie Scot” the king writes the letter in stanza 4, and the mother speaks out two stanzas later. In “Lord Derwentwater” the king writes the letter in stanza 1, and the lady gay does not speak out until four stanzas later. In “Lord Derwentwater” the hero responds to the letter by calling for his horse in the words of the appropriate commonplace, even though this commonplace does not occur in the other realizations of this “theme.” Nor do the nuncupative will and the “He has sent it not with a boy, a boy” commonplace. But let the last stanza of each “theme” stand for all the differences in realization among the Lyle variants, despite identity of basic narrative elements.

On the level of the whole ballad two types of formulas can be identified. The first is called by Andersen a formula ballad (1985, 78). His example is the song “Come Mother, Come Mother” (cf. MacColl and Seeger 1977, 112-115). Made up entirely of a stable series of well-known commonplaces, the song is like a run gone amok, a run so extended that it constitutes the whole song. The second type might be called the formulaic ballad, by analogy with the formulaic tale. Members of this type are one-formula pieces which do not vary greatly from performance to performance, though anyone who understands the basic formula can expand or compress the ballad at will. “The Maid Freed from the Gallows” is a prime example, remarkably stable across a wide linguistic and geographic range. The wit-combat songs and “Our Goodman” would also seem to belong to this category. Communalists were fond of citing ballads such as these in their expositions; it was easy to imagine a singing, dancing throng joining in, once the basic pattern set. But the single-formula character of these ballads is just the reason why they are not good fodder for discussions about ballad creation and transmission.

The above attempt at classification has demonstrated that many other schemes are possible. One might wish to distinguish grammatical formulas, incipits and inquits, formula families, free-floating formulas, or any of a dozen other categories. Formulas associated in one classification might well be separated in another. For instance, in Andersen’s category of “true ballad formulas” characterized by stable supra-narrative function, linear, half-stanza, stanza, and stanza-pair formulas associate quite happily.

One particularly necessary alternative to classification by linguistic level needs to be developed here. Formulas originate neither from the same impulse nor from the same mental process. This diversity has led to seemingly contradictory descriptions of the formulaic process. Lord, in The Singer of Tales, describes formulas as substitutionary. Opponents of the formulaic approach tend to see formulas as memorial. And Edwards (1983) and Andersen see them as generated on the deep level. But all three modes of formulating seem to occur in tradition, providing set or memorial formulas, substitutionary formulas, and generative formulas.

There are, first of all, set formulas, which do not vary in wording, even over a wide geographical and temporal range. These are usually short epithets and phrases such as “wee pen knife,” “beaten gold,” “milk-white steed,” and so on. These ballad “clichés” have been noted and commented upon many times. There is no need for a complex model of formulation to deal with them. They are simply remembered.

On the next level of complexity, balladry is full of expressions such as “Rise up, rise up, ———,” “Out and spoke ———,” “It fell upon———,” “Fair ———,” “Sweet ———,” and “Merry ———,” One may fill in the blank with any appropriate and metrical expression. A simple substitutionary model seems sufficient to deal with these inquits, epithet-phrases, and other expressions, once they have acquired a relatively fixed surface structure.

More complex formulaic expressions call for a generative model with a unifying idea on the deep level and multiformity on the surface level.⁵ In some cases, in fact, it is possible to identify several levels of deep structure. In the “They had not” formula family discussed above, the commonplace idea is able to generate Agnes Lyle’s “They had not sailed a league, a league” stanza and many related realizations. But on a deeper level this formula family is able to generate this commonplace, the commonplace which Agnes Lyle renders “She had not been in the king’s court / A twelvemonth and a day,” and the commonplace “She hadna pu’d a flower, a flower,” as well as the several context-bound formulations in “Fair Janet,” all of which are related in narrative idea and structure. Once realized, such commonplaces can then set, on the surface level, at least in a particular region or in the practice of a particular performer, and become subject to substitutionary variation (“She had not . . . he had not”). Finally, they can be parodied on the surface level, without respect to the ideas and affects which are part of the deep (or the supra-narrative) structure. In “Lord Derwentwater” Agnes Lyle sings:

In “The Eastmure King” she sings:

Each of these stanzas seems to be modelled upon stanzas belonging to the formula family under question. And yet neither realizes the basic relationship of events characteristic of the formula family. One substitutes hierarchical progression for temporal or spatial progression. The other introduces an act that, though violent, is desirable. The stanzas, accordingly, are not realizations of the deepest structure of the formula but parodies on the surface level.

One may speak then of three modes of production for formulas. They may spring full blown from the memory. They may be created by substitution or parody on the surface level. Or they may be generated from a deep structure. Furthermore, a formulation which has been generated from a deep structure may subsequently be manipulated on the surface and even set in memory. The opposite process is also possible: A singer who intuits the deep structure latent in a remembered phrase may generate new expressions of the formulaic idea.

2. AGNES LYLE AND ANNA GORDON
BROWN

Prior to this study, the only classic oral ballad repertoire to receive extensive study had been that of Anna Gordon, Mrs. Brown of Falkland.⁶ Comparison of Agnes Lyle’s repertoire with that of Mrs. Brown can provide deeper understanding of oral ballad making.

Mrs. Brown tended to sing romantic and magical ballads, all but one of which, “Sir Hugh” (Child 155A), concerned the course of true and sometimes false love. Even the two “historical” ballads in her repertoire, “Bonny Baby Livingston” (Child 222A) and “The Baron of Brackley” (Child 203C), include distinctly romantic elements. Though the Lyle ballads include only one revenant piece, they likewise represent the more romantic branch of the Anglo-Scottish tradition. But the realizations of these ballads are completely different in tone and emphasis from Mrs. Brown’s realizations. Agnes Lyle, like some of her Kilbarchan neighbors, gives her ballads a nationalistic emphasis (Andersen 1985, 327). And she seasons this nationalism with a bitter cynicism unique to herself.

One of the most remarkable things about Mrs. Brown’s ballads is their uniformity. David Buchan’s careful and extended analysis shows that the most diverse stories, from the most diverse sources, have all been fitted into a common narrative mold, developed according to a single system of composition. This system, which seems to have been learned when Mrs. Brown was a child, produced ballads which may be described thus: stanzaic narrative poems involving three principal characters, a hero, a heroine, and a villain, in an amatory story—except in the case of “Sir Hugh”—told through stanzas of dialogue alternating with stanzas of narrative in a three-act structure, all elements arranged symmetrically in harmonic and contrapuntal annular, binary, and trinary patterns.

The most remarkable evidence of the consistency of this narrative technique is that these thirty-three or more ballads admit of classification into three types solely on the basis of the role played by the villain! In the largest class, ballads of family opposition, the villain is a member of the family. In the second class, ballads of other love, the villain is a rival. In the third class, murder ballads, the villain is the murderer. In this third class, as well, the villain may overshadow one of the other characters in importance. In “Sir Hugh,” for instance, the Jew’s daughter is more important than Sir Hugh’s mother, the “good” female character. This procrustean uniformity seems to have been imposed upon the ballads by Mrs. Brown. Variants of these ballads by other singers often have different deep character structures or a different number of acts, to name but two of the more obvious differences. Mrs. Brown was asked for songs of a particular type. She looked on the songs of her childhood as belonging to that type. Apparently she reconstructed those all-but-forgotten songs not in all their diversity but according to a single system, a system likewise learned in her childhood. In all probability she was the first and last singer ever to sing “The Twa Sisters,” for instance, in that elaborate form produced by her system.⁷

The circumstances under which Agnes Lyle sang for William Motherwell were rather different from the circumstances under which Mrs. Brown dictated her ballads. And there is nothing like the same uniformity of technique in the pieces she sang. Presumably Agnes Lyle’s texts were still a part of her daily life. For the most part she did not have to reach into her past to revive a long-dormant skill. Nor did she have to reconstruct her favorite songs “the way they must have been.” Nor, again, did she distinguish among the songs she sang in the same way Mrs. Brown obviously distinguished ballads of her childhood from other songs. Mrs. Brown was consciously singing one particular type of song. While it would be a mistake to exaggerate her level of education, her concept of ballad was rather sophisticated, and she knew her Percy. Consequently, even without stepping outside the traditional technique or system that she had mastered in childhood, she could impose that sophisticated concept upon the ballads she sang. Agnes Lyle, on the other hand, simply knew songs. She knew all kinds of songs. The evidence suggests that she did not consciously reflect on the differences between the various types of songs she knew. On 24 August 1825, for instance, she produced “The Cruel Mother” and “Hind Horn,” rather different sorts of couplet ballads; “Johnie Scot,” an orally recreated standard-quatrain ballad with political overtones; “The Wee Wee Man,” a song; “The Baffled Knight,” a slightly bawdy broadside piece; and the ill-remembered “The Sweet Trinity.” The ballads of Mrs. Brown represent a conscious selection from songs of childhood by a sophisticated informant, whereas the ballads and other songs of Agnes Lyle represent a cross section of the current repertoire, active and passive, of a singer for whom ballads are a vital part of the singing experience of daily life and contemporary politics.

Agnes Lyle knew the technique of the orally recreative ballad singer, but she did not draw on that technique every time she sang. She was equally capable of remembering a fixed text such as “The Wee Wee Man,” or loosely recreating a half-remembered piece such as “The Twa Sisters.” When she did produce a full oral recreation her technique was in many ways like that of Mrs. Brown. It involved the same binary, trinary, and annular patterns, the same stanzaic, character, narrative, and sometimes even tonal structures. It was an expressive system into which an individual ballad might be assimilated and in terms of which it might be uttered. It was, however, a more limited and less monolithic system than that of Mrs. Brown, whose every ballad was totally assimilated into the system and expressed totally in terms of that system.

Mrs. Brown’s technique somehow transcended metrical considerations. Of her technique Buchan says (1972, p. 142), “As the one corpus contains both couplet and quatrain texts demonstrably structured by the same methods, the traditional maker must have been able to move with relative ease between the two forms.” Agnes Lyle’s technique depended on the standard ballad quatrain, and in that meter she was more than competent. But she did not experience “relative ease” when she moved into singing couplet ballads. The nearest to success she came in this regard was with “Hind Horn.” For “Babylon” and “The Cruel Mother” she went outside the system to discover a completely separate principle of structure. The four ballads in nonstandard quatrains provided other difficulties in addition to the uncongenial meter. In one case, “The Braes o Yarrow,” she achieved an outstanding success, difficulties notwithstanding; in the other three cases her success was only moderate. In general, the ballads in couplets or non-standard quatrains suffered most stress in the stanzaic structure, while retaining most control in narrative structure, confirming that the difficulty was not with the story but with the stanza in which the singer had to tell the story.

With regard to diction, both singers had favorite phrases which they used again and again. Agnes Lyle, however, was more free in her realization of formulas. She seldom used the same wording from stanza to stanza, except in the case of close emphatic or incremental repetition. In particular, her realizations of commonplaces were amazingly multiform, as a comparison of the four versions of the “Come saddle to me” stanza will demonstrate. Andersen’s research confirms that such freedom in the realization of commonplaces is characteristic of the oral poet (passim, but especially chs. 3 and 5).

The rhyming practice which Agnes Lyle shares with Mrs. Brown provides further evidence of the orality of the repertoire (see Buchan 1972, 151-155). The rhyme ee occurs exactly the same percentage of times in the standard quatrain Lyle ballads as it does in the Brown ballads, twenty-eight percent. As in the Brown repertoire, the largest group of consonantal rhymes involves n sounds (and, an, in, etc.). D and t rhymes are also common in both repertoires, though the r rhymes, so common in the Brown repertoire, are rare in the Lyle. Like Mrs. Brown, Agnes Lyle draws on more than one dialect for rhymes, rhyming away with day, for instance, and awa with braw. The principal difference in rhyming practice between the two women is that Agnes Lyle is considerably more liberal in what she admits as rhyme, rhyming head with sleep, for instance, green with down and gone, and slain with within, dream, and on. Mrs. Brown’s more conservative rhyming practice probably reflects her literate education and her experience reading and composing literary verse, which would lead her to avoid slant rhymes.⁸

In Agnes Lyle’s ballads, as in those of Mrs. Brown, alliteration is wholly traditional. Though a principle decorative element in the verse, it is never used in purely decorative fashion. It regularly functions to relate words and ideas, as in “grass-green growing,” “lay the saddle saft,” and “come to Mary’s Kirk.” In those ancient phrases, such as “busk the bride,” “lang letter,” and “hunting horn,” that have become part of the standard language of the ballad singer, the alliteration also functions to make the phrase memorable and easily recalled in moments of need.

Consideration of the similarities and differences in the oral styles of these two singers can lead to a deeper understanding of the process of oral creativity itself. The principles at work in the practice of each singer are the same. The techniques for developing these principles and the degree to which songs are assimilated to technique vary considerably. It is dangerous, therefore, to speak of the way oral performers recreate ballads.

3. THE WAY ORAL PERFORMERS
RECREATE BALLADS

One of the great triumphs of the Parry-Lord school has been the demonstration that for epic the moment of creation, the moment of performance, and the moment of transmission coincide; all three of these processes go on at the same time, in fact, are in some sense a single epic process. Aspects of a single process or not, however, the three can be distinguished conceptually and, at least in the case of some genres, separated as well. To discuss how oral performers recreate ballads, the processes involved must be distinguished. Generally speaking the processes are the acquisition of oral recreative technique, the learning of a traditional ballad story, the development of that story within the limits and according to the structures of the technique, and the performance of that ballad, often for an audience which includes a future singer of the ballad. During the time when the singer is becoming competent the first three processes often coincide. During this apprenticeship period the singer, like Lord’s apprentice epic singer (1960, 20-29), probably practices repeatedly prior to performance. Even competent singers will practice or sing for themselves more than for an audience. In a genre as short as the ballad, such private practice will tend to stabilize (not “fix”) the text.⁹

About such stable texts several observations are in order. First, they will betray their origins in their overall oral architectonic and in their pervasive formulicity. Second, not being “fixed,” they will always be open to improvisation and innovation. The inspiration of the moment, a better idea, or the influence of another version can at any performance induce changes, minor or major. Third, a person exposed to frequent repetitions of such a stable text, such as the child of a ballad singer, might learn the particular version more or less word for word, in a more or less memorized way, and so pass on a now-fixed text. Fourth, even if this child or other person is habitually an orally recreative singer, nothing is to stop the singer from learning that particular piece in a memorial way, the same way the hymns or psalms are learned in church, for instance;¹⁰ nor, on the contrary, is there anything to stop the singer from learning it freely and orally, the way other ballad stories are learned, and fashioning therefrom a new rendition.

Once singers become proficient it is quite possible for a singing session involving two or more such singers to include a performance that is simultaneously composition and transmission as well. Even during the learning period, processes can not always be separated, for the singer learns the technique by listening, as well as by singing, and there can be no singing without at least one or two ballads to sing.

Like the story, the tune is learned as an idea, rather than a fixed series of sounds. The Bronson anthology demonstrates the tremendous variation that tune ideas can undergo, variation in range, in scale, in mode, in tempo, and in harmonic structure. Learning the basic tune idea, each singer realizes it according to personal ability and taste, one singer producing something rich and subtle, another producing something simple and effective, a third producing something crude and monotonous. Moreover, though text and tune are integrated, not all great ballad storytellers are great ballad musicians. The two surviving tunes from Agnes Lyle suggest that she belongs to the “simple and effective” school. Mrs. Brown’s tunes are more troubling, though it is hard to know if the trouble originates with the singer or with the transcriber.

The fewer mentors chosen or attended to, the more singer will sing like mentor. For a time. But great singers will eventually develop their own style and approach, as Agnes Lyle did. This style may be conservative or innovative, monolithic or adaptive, stable or improvisatory. These variations are not the significant thing. What is important is that the ballad has been forged in the oral crucible, and embodies an oral aesthetic. It tells the traditional story in formulaic language on all levels from phoneme up, structures its elements according to the binary, trinary, and annular principles of an oral architectonic, and expresses its unique cultural and thematic insights in terms of the binary oppositions of the traditional worldview. As such, it is never fixed for the orally recreative singer, but at any point, even after the hundredth performance, it can change to accommodate a new insight or a new view of the material.

The difference between singing fixed and singing stable songs is in some ways like the difference between going somewhere in a strange city by following memorized or written directions and going somewhere routine in a familiar city. The driver in the strange city must look for just the landmarks and make just the turns prescribed in the directions or risk never getting to the destination. The driver taking a routine trip in a familiar city probably goes the same way regularly, out of habit. But this driver, at any time, can go some other way because of traffic, desire for change, curiosity to see if it’s faster, or mere whim. The first driver drives by rote. The second driver knows the way, and can vary that way at will. And, just as the driver, once that driver knows where something is in a familiar city, can go there without needing directions, so the ballad singer, once that singer knows where a song goes, can go there too. The singer may go the same way every time, or may vary. What Lord says of the epic singer seems to apply here as well: “In a truly oral tradition of song there is no guarantee that even the apparently most stable ‘runs’ will always be word-for-word the same in performance” (1960, 125). The oral tradition is always open to change.

It is usually easy to distinguish ballads forged in an oral crucible from broadside texts, ballad imitations, and other types of folk or popular songs. But the immediate singer is not always the original creator of the version. Sometimes, as has been noted, a stable text in one generation becomes a fixed text in the next. Such fixed texts then take on an existence of their own. They may even become the dominant tradition, as may have happened in North America. In such a tradition stabilized ballads texts will live, grow, and die according to a law of their own. That law deserves a book of its own, many books, but the law of transmission of fixed texts must not be confused with the law of oral recreation that has been the subject of this study.

I began this study by proposing to survey the field of contention in the new “ballad wars.” What this survey has shown is that the debate between the proponents and opponents of the oral approach becomes almost a debate about vocabulary, not about actuality. When the oral model is applied carefully, with due regard to what happens in the case of short sequences of formulas, the description of balladry that emerges is remarkably like the description of balladry that emerges from moderate and sensible contemporary applications of a text memory model, such, for instance, as that of Eleanor Long. The oral model has the advantage, however, of accounting for the generation of the text and the oral architechtonics and aesthetics therein embodied.

In balladry, and probably in other short oral genres as well, the issues of generation and transmission seem separable in ways that they are not in longer genres, such as epic, wayang kulit, or Navajo sing. In the case of a fairly stable ballad the performance or performances at which a second singer learns the ballad probably do not coincide with the series of practices and performances during which the first singer stabilizes her version. And the second singer, once she learns the ballad, may sing it much as she has heard it, or she may generate something quite different in structure, in diction, in emotional content, and even in melody. So, on the issue of transmission the memorial and the oral approaches, unlike the high road and the low road, arrive at Loch Lomond together. But on the issue of generation of the text the oral formulaic approach goes beyond the text memory approach to suggest how the text is generated in the first place, and what are the uniquely oral qualities of that text, both structurally and aesthetically.