The English and Scottish Popular Ballads of Francis J. Child
by Thelma G. James
The Journal of American Folklore, Vol. 46, No. 179 (Jan. - Mar., 1933), pp. 51-68
[Footnotes moved to the end. Tables need to be added at the end before footnotes.]
THE ENGLISH AND SCOTTISH POPULAR BALLADS OF FRANCIS J. CHILD
BY THELMA G. JAMES
The English and Scottish Popular Ballads edited by Professor Francis James Child in 1882-1898 suggests two interesting problems: first, was Professor Child adhering to a fixed principle in selecting ballads for inclusion? Second, is it possible to derive a definition of the ballad from a study of this collection? Whether he followed a fixed notion is a matter of uncertainty among critics. Miss Louise Pound in her Poetic Origins and the Ballad declares that he fluctuated in his decisions: "The similarity in style of the pieces he included was the chief guide of Professor Francis James Child in his selection for his collection of English and Scottish Ballads... He would not have altered his decision concerning so many pieces had the test of style been so dependable as is usually assumed."[1] Dean Walter Morris Hart, on the contrary, feels that "the significant fact is that for at least forty years Professor Child retained without essential change his conception of the traditional ballad as a distinct literary type."[2] Professor Gummere thought: "It is clear that the notion of a traditional ballad existed in a very exact shape for Professor Child when one thinks of the hosts he rejected."[3]
A re-examination of the material included in the first and third editions[4] should resolve these differing opinions. In this study, the method was: first, to determine what was omitted from the first edition in compiling the third; second, to examine the new material added to the third edition, particularly those ballads available to Professor Child in 1858 which were not utilized until 1882; third, to examine, as far as possible by means of Professor Child's own comments, his reasons for these exclusions and inclusions; and finally, to attempt to reach some conclusion concerning the principles governing the selection. The task was complicated by the fact that in the third edition Professor Child completely ignored the two earlier editions. Further, the inconvenient numbering of the pieces in the first edition makes cross reference possible only through titles, which are more trouble than help. For example, one finds "The Enchanted Ring" under "Bonny Bee Home", and the "Birth of Robin Hood" under "Willie and Earl Richard's Daughter." Nor is it always easy to say whether a ballad has been included or excluded; for instance, "Waly, Waly, gin Love be Bonny" is given in an appendix to "Lord Jamie Douglas"; "The Gaberlunzie Man" is in an appendix to "The Jolly Beggar"; the "Playe of Robin Hood" in an appendix to "Robin Hood and the Potter". Does the assignment of such pieces to an appendix signify their inclusion, exclusion, or merely a dubious compromise bringing into question Professor Gummere's belief in the "very exact shape" of Professor Child's conception of a ballad? In the Preface to his first edition, Professor Child says of it: "It contains nearly all that is known to be left to us of the ancient ballads of England and Scotland, with a liberal selection of those which are of a later date. Of traditional ballads preserved in a variety of forms, all the important versions are given, and no genuine relic of olden minstrelsy, however mutilated or debased in its descent to our times, has been on that account excluded, if it was thought to be of value to the student of prose fiction. . . [The purpose was] to adhere to the originals as they stand in the printed collections."[5] In the Preface to the first volume of his third edition he wrote: "It was my wish not to begin to print The English and Scottish Ballads until this unrestricted title should be justified by my having at my command every valuable copy of every known ballad." In 188o, when the second edition was printed, some change in the general plan was already in Mr. Child's mind (although his definition of the ballad was not necessarily concerned), for he says in that Preface: ". . . the popular ballad deserves much more liberal treatment. Many of the older ones are mutilated, many more are miserably corrupted, but as long as any traces of their originals are left, they are worthy of attention and have received it."
In his article, Dean Hart says that 115 ballads were omitted from the first edition and my figures agree with his. On the basis of Professor Child's comments, these 115 ballads fall into the following divisions: 9 (numbers 1-9)[6] were rejected as having been unduly "edited" by modern editors; 16 (10-25) as romances; 41 (26-66) as of non-popular origin from broadsides, stalls, minstrels, re-rhymed literary plots; 10 (67-76) as translations of foreign parallels rather than native ballads; 4 (77-80) as types of ballads on non-popular subjects; 16 (81-96) as lyric types lacking a narrative element; 18 (97-115) as simple narrative poetry lacking ballad style.[7] Dean Hart summarizes: "It is not difficult to see why the 115 ballads are excluded from the later collection, and one gets the impression that, bad Professor Child chosen to enforce the conception of the ballad which he had already in mind, most of them would have been excluded from the earlier collection as well. This impression is deepened by an examination of the comments scattered through the ballads." [8]
The ninety ballads included for the first time in the third edition fall into two groups: first, those 37 ballads which were not available in 1858[9] ; and second, those 53 ballads which were available, but which were withheld until the third edition. Obviously no change in Professor Child's definition of the ballad can be deduced from his inclusion in the third edition of the 37 ballads to which he had not previously had access. If any change is to be noted, we shall find it in the 53 ballads which Professor Child knew in 1858 and rejected, although he later included them in 1882. It is highly significant that 35 of these 53 ballads were available in 1858 only through Buchan, an editor of whose methods Professor Child had grave suspicion. Only 9 Buchan texts were added to the third edition[10] without confirmation from other sources.[11] Even in the preparation of his first edition, Professor Child had serious distrust of the Buchan texts, for he says in that Preface: "Some resolution has been exercised and much disgust suppressed, in retaining certain pieces from Buchan's collection, so strong is the suspicion, that, after having been procured from very inferior sources, they were tampered with by the editor."[12] In the head notes to the third edition Professor Child seldom has more than scathing comment on the style and substance of the Buchan ballad texts: of the "Bonny Lass of Anglesey" (220) he says: "Buchan quite frightens one by what he says of this version... 'It is altogether a political piece and I do not wish to interfere much with it"; of "Auld Matrons" (249), "This piece was made by some one who had an acquaintance with the first fit of 'Adam Bell'... Stanzas 2-5 are hackneyed commonplaces"; of "Thomas of Yonderdale" (253), "This looks like a recent piece, fabricated with a certain amount of cheap mortar, from a recollection of 'Fair Annie', etc.,"; of "Willie's Fatal Visit" (255), "Stanzas 15-17, wherever they came from, are too good for the setting; nothing so spirited, word or deed, could have come from a ghost, wan, weary, and smiling"; of "The New Slain Knight" (263), "A large portion of this piece is imitated or taken outright from very well known ballads. . . This particular ballad, so far as it is original, is of very ordinary quality. The ninth stanza is pretty but not quite artless"; of "The White Fisher" (264), "But we need not trouble ourselves to make these counterfeits reasonable. Those who utter them rely confidently upon our taking jargon and folly as the marks of genuineness. The White Fisher is a frumpery fancy"; of "The Beggar Laddie" (280), ". . . it is inconceivable that any meddler should not have seen this [mistake]"; of "Child Owlet" (291), "The chain of gold in the first stanza and the penknife below the bed in the fourth have a false ring, and the story is of the tritest. The ballad seems at best to be a late one, and is perhaps a mere imitation, but, for an imitation the last two stanzas are unusually successful"; of "The Queen of Scotland" (301), "The insipid ballad may have been rhymed from some insipid tale"; of the "Holy Nunnery" (303), "The rest is wanting, and again we may doubt whether the balladist had not exhausted himself, whether a story so begun could be brought to any conclusion"; of "Brown Robin" (97), "The sequel to C is not at all beyond Buchan's blind beggar, and some other blind beggar may have contributed the cane, and the whale and the shooting and the hanging in B"; The general comment on "The Bent sae Brown" (71) summarizes Professor Child's feeling clearly: "The introduction and conclusion of some incidental descriptions.... are the outcome of the invention and piecing together of that humble but enterprising rhapsodist who has left his trail over so large a part of Buchan's volumes.... The silliness and fulsome vulgarity of Buchan's versions are often enough to make one wince and sicken, and many of them come through bad hands or mouths; we have even positive proof in one instance of imposture, though not of Buchan's being a conscious party to the imposture." [13] With all this in mind, it is not curious that Professor Child withheld so much Buchan material for later verification in manuscript and other sources.
The remaining 18 ballads, which Professor Child knew in 1858 but reserved until 1882, show a period of waiting on his part for fuller, older, or more widely known forms. In short, the treatment he accorded them testifies to his scholarly approach to his problem. [14] His aims became more definite. In 1858 he intended to give a somewhat popular survey of balladry (a survey which I am strongly inclined to believe was modelled after Percy's Reliques); in 1882, he sought to print "every valuable copy of every known ballad". In 1858, he could find room for romances, translations, broadsides, and a variety of heterogeneous materials; in 1882, he excluded them. The rejections and inclusions in the first and third editions show no material change in Mr. Child's conception of the ballad; although, it may be granted, that they do show a marked change in his aims.
The demonstrated stability of Professor Child's ideas serves rather to confuse than to clarify the general problem of ballad definition, for his collection was made and printed without any pronouncement on this subject. The Advertisement to the first volume of the edition shows that he realized the need of some formal statement: "It was the Editor's wish and intention to insert in the concluding volume an essay on the History of Ballad Poetry." Two other opportunities slipped by, for the article in the Johnson Cyclopedia is merely a resume of ballad poetry in general; and the third edition of The English and Scottish Popular Ballads lacks the formal introduction promised by Professor Child. According to Mr. Gummere's note[15], such an essay had not been even remotely formulated at the time of Professor Child's death, although Professor Kittredge suggests in the Introduction to the first volume of the third edition that it had at least been planned.[16] Coming towards the completion of the work, such an essay would necessarily have been a defense, quite as much as an exposition, of the principles which dictated his selection. As it is, he nowhere seems to say anything more definite than this: "The word ballad in English signifies a narrative song, a short tale in lyric verse, which sense it has come to have, probably through the English, in some other languages ..... The popular ballad, for which our language has no unequivocal name, is a distinct and very important species of poetry... The fundamental characteristic of popular ballads is... the absence of subjectivity and self- consciousness.''[17]
In the absence, then, of a more specific definition from Professor Child's hand, later writers have sought to derive a definition from the ballads included in his collection. The endeavor has no doubt been stimulated by Gummere's description of the Child collection as "itself a definition of balladry".[18] Dean Hart does not attempt a formal definition, but rather seeks to give the positive and negative characteristics of the type as Professor Child presumably saw it: a ballad must tell a story, and that only partially; the transitions must be abrupt, although not incoherent; the introduction must be closely integrated with the story; there must be brevity; the action can seldom be carefully localized; description or exposition of the supernatural is omitted. The style must be artless, homely, without conceits or description of states of mind; it is marked by commonplaces which are retained in distorted form in several ballads on the same theme; it must be impressive, fine, spirited, pathetic, tender, and finally, lyrical. The subject matter must be of popular origin, and foreign parallels should exist; it may be pseudohistorical, must deal with heroic sentiment, occasionally it may be derived from other ballads. The ballad must not show extravagance or exaggeration; it must not be prosaic, over-refined, cynical, sophisticated, sentimental, moralizing; but a certain degree of probability is demanded of the plot, which must not be trite.[19]
Professor Kittredge, in the one volume edition of the Child ballads, says that the author must be of the folk, the material derived from popular sources, the structure moulded by inherited influences, the product early given to the folk, who subject it to the process of oral transmission. "That most of the 305 numbers in Mr. Child's collection satisfy these conditions is beyond question. In other words, most of these poems are genuine popular ballads within the limits of any reasonable definition of the term."[20] This, too, is a definition derived from the actual study of the texts. In a somewhat different way, Miss Pound generalizes the so-called "true" ballad of "the Child type", as differentiated from "other types of ballads and songs", by examining the 305 ballads.[21] The ballad might, she says, "at times be fitted to well-known dance tunes, or be utilized as a dance song", although this is an unessential characteristic.[22] She regards oral preservation as the test of inclusion, although its importance has been greatly exaggerated.[23] Much more important are "anonymity of authorship or traces of a medieval style".[24] Indeed, "the only dependable test elements in ballads are lyrical quality and a story element, and for traditional folk-ballads, anonymity of authorship".[25] "The lowly as against the aristocratic, hardly plays any part in the English and Scottish Popular Ballads."[26] The ballads are not so impersonal as is often assumed[27]; "there is only the uniformity of simplicity to be expected of popular songs of all types."[28] The ballads preserve in style and structure many archaic traits, use alliteration, some literary words, include occasionally a satirical legacy, shade off into literary verse or into some other type of folk verse or popular verse such as: the allegory, and epic chanson, romance, verse chronicle, dialogue, debate, aube, lyric, coronach, carol, theological discussion.[29] (All these shadings are represented in the Child collection, but notwithstanding, Miss Pound and other scholars persistently use the phrase "the Child type").[30] The ballad is further marked by the presence of strophes, meter, the use of refrain (in the later texts), by repetition, and parallelism. In summary, Miss Pound arrives at the following definition: "It is enough to say that in English we mean by a ballad a certain type of lyrical or narrative song or song-tale, which appears rather late in literary history; and we may discard as unessential for defining this type reference to the origin of such pieces in the dance."[31]
Another summary of the tests for inclusion is made by Professor Gordon Hall Gerould: "The tests by which it (the traditional ballad) must be judged, I take it, are three. The first is purely personal, the critical sense of the scholar who has learned by long continued and careful study to distinguish the false from the true, to separate the chaff from the wheat. The second is the external evidence with reference to the circumstances of discovery, whether the collector or collectors be trusted. The third is the source of the material, whether the narrative is the product of tradition or of some clever inventor."[32]
An examination of the 305 ballads as "constituting a definition of balladry" leaves one puzzled at their diversity. No single definition can include "The Carnal and the Crane" (55), "Dives and Lazarus" (156), "The Cherry Tree Carol" (54), "Sir Patrick Spens" (58), "Bonny Barbara Allen" (84), and "Thomas Rhymer" (37). There is scarcely the unity of style so frequently assumed, and certainly no unity of either spirit or subject. The individual pieces vary widely in worth.[33] If one were to use only those ballads upon which Professor Child commented favorably[34], there would still be a tremendous diversity of materials. As Professor Gerould points out, "That ballads of very various degrees of worth may be regarded as valuable to the study of the type is evidenced by comparing the contents of the last two volumes of Child with the earlier ones."[35]
The definitions which have been quoted above do not inspire one to formulate another. These definitions fail for various reasons: Dean Hart aims at being descriptive and loses himself in details which are not properly integrated nor subordinated to one another; Professor Kittredge is discussing the problems of ultimate origins; Miss Pound frequently uses the phrase "the Child type"[36] at the same time that she demonstrates that, for all practical purposes, there is no such single clearly defined classification possible. In view of such vague definitions of a mass of materials so nearly indefinable, it is not difficult to see why the warm discussion of communal origins, or of any other ballad theory for that matter, cannot be resolved. Obviously, the debate turns upon a different subject with each interpreter: it deals now with the origins of primitive popular poetry and now with the modern folk ballad, subjects between which there could be no possible confusion were the terms precise and clearly understood.[37]
In some mysterious way, these 305 ballads have come to be canonized as superior to all other folk-songs of the English people. This in the face of such evaluations as Professor Child himself put upon some of them sufficient to indicate their widely divergent values and types. Even the number 305 has come to possess a curiously magical connotation which has exerted sway for twenty-five years.[38] When Miss Pound says of ballad-making that "It is a closed account for ballads of the Child type",[39] there is much sound truth in the remark. It has been the purpose of this paper to show that a "Child ballad" means little more than one collected and approved by Professor Child.[40]
TABLE 1
Footnotes:
1 P. 146.
2 PMLA, XIV (19o6), 799.
3 Modern Philology, I (1903-04), 378.
4 The second edition is unimportant. Four new texts were added to it, "Brave Earl Brand and the King of England's Daughter" which merited only a footnote in the third edition. "St. Stephen and Herod" (22) was retained for the third edition; while "Gifts from over the Sea" and "The Hawthorne Tree" were discarded.
5 Italics mine.
6 Dean Hart has unfortunately failed to list the ballads completely in his otherwise excellent study. For my titles, see Tabulation I in the Appendix, Nos. 1-9, etc.
7 These classes are intended to be convenient and suggestive rather than rigid and mutually exclusive since a text might often fall into two ormore groups. A text might be lyric and at the same time be the victim of a modern editor. There might easily be debate about the assignment of any ballad to its special group, but the transfer of one or another text to a different group does not affect the general conclusion.
8 Loc. cit., 796.
9 Ballads not available in 1858 are marked (-) in Table II; see Appendix.
10 Ballads 71, 255, 261, 263, 268, 282, 296, 303. It may be important to note that all but one of these appear in the two final volumes.
11. Four other Buchan texts are retained only because of foreign parallels: Ballads 57, 115, 171, 268.
12 p. IX.
13 This would seem to answer adequately the recent attempt to place Professor Child in the position of a Buchan defender. See: Greig, Gavin, Last Leaves of the Traditional Ballads and Ballad Airs, (1925), XXX.
14 This point is elaborated by Professor Taylor in Mod. Phil., XXV (1928), 481-491.
15 Mod. Phil., I (1903-04), 378 n.
16 I, XXXI.
17 Johnson's Cyclopedia, I (1895), 464.
18 Mod. Phil., I (1903-04), 378.
19 Loc. Cit.
20 p. XXVII.
21 Poetic Origins and the Ballad, 1921, p. IOI.
22 Ibid., pp. 67, 100, I17.
23 Ibid., p. 88.
24 Ibid., p. 95.
25 Ibid., p. 124.
26 Ibid., p. 10I.
27 Ibid., p. IoI.
28 Ibid., p. 153. Miss Pound discusses this subject also in Mod. Lang. Notes, XXV (1920), 65ff.
29 Ibid., p. 108ff.
30. Hustvedt is drawn into this inconsistency: ".. .there is no such thing as a canon of popular ballads, no nearly infallible means of setting a popular ballad apart from one that is not, just as there is no accepted definition of the popular ballad." A few lines later, on the same page, he says: "The traditional canon has been established for practical purposes to lie within the covers of such repositories as those of Child, Grundtvig. ... Working definitions, implicit or explicit, will be found there too. For the purpose of this history, popular ballad may be taken to mean the sort of verse so named by Child, in whatever dialect it may happen to be recorded." Hustvedt, Sigurd. Ballad Books and Ballad Men (Harvard, 1930), 4.
31 Ibid., p. 117. Miss Pound reaffirms her theories of ballad origins and structure in a recent article: "A Recent Theory of the Making of Ballads", PMLA, XLVI (1929), 622ff.
32 PMLA, XXIII (I9O8), I4If.
33. See above, Professor Child's comments on the Buchan texts. Concerning "Young Ronald" (304) he remarks: "In this and not a few other cases I have suppressed disgust, and admitted an actually worthless and a manifestly... at least in part.,. spurious ballad, because of a remote possibility that it might contain relies or be a debased representative of something better. Such was the advice of my lamented friend, Grundtvig, in more instances than those in which I have brought myself to defer to his judgment." Of "Outlaw Murray" (305): "That it was not originally intended to insert "The Outlaw Murray' in this collection will be apparent from the position it occupies.
34 For example: "Judas" (23) "Brown Robins' Confession"' (57), "Edward" (I3), "Earl Brand" (7), "Young Beichan" (53), "Walter Lesley" (296), "The Gray Cock" (248).
35 Loc. cit.
36 Poetic Origins and the Ballad, pp. 41, 27, 88, 107, 171. The term persists even in so recent an article as PMLA, June 1929, p. 622ff.
37 The fatal attraction of the discussion of origins as a substitute for critical analysis of the problems of definition is illustrated by Mr. Arthur Kyle Davis' Introduction to the Traditional Ballads of Virginia. (1929). The editor devotes a scant two pages, to the matter of definition and is thereby launched upon a five page discussion of the theory of origins. A still more marked disproportion appears in Hustvedt's Ballad Books and Ballad Men (Harvard, 1930) which devotes two pages to definition and sixteen to a review of the theories of origins.
I am convinced that it did not begin its existence as a popular ballad, and I am not convinced that (as Scott asserts) 'it had been for ages a popular song in Selkirkshire.[38] But the 'song' gained a place in oral tradition, as we see from B, C, and I prefer to err by including rather than excluding." He also says of it, "I cannot assent to the praise bestowed by Scott on 'The Outlaw Murray'. The story lacks point, and the style is affected-not that of the unconscious poet of the real traditional ballad. Of the "Crafty Farmer" (283): "This very ordinary ballad has enjoyed great popularity, and is given for that reason and as a specimen of its class."
38 Mr. Kittredge in 1904 says in the Introduction to the one volume edition (p. XIII): "Professor Child's great collection, The English and Scottish Popular Ballads, in five volumes (Boston, 1882-1898), comprises the whole extant mass of this material. It includes three hundred and five pieces, most of them in a number of different versions, with full collations and other pertinent apparatus. A few variants of this or that ballad have come to light since the publication of this admirable work, but no additional ballads have been discovered." Professor Ker in his On the History of the Ballads (1910), says (p. 3): "The English [ballads] are all together in Child's five volumes, The English and Scottish Popular Ballads, 1882-1898." In 1928, Professor Reed Smith says (South Carolina Ballads, Cambridge, (1928), p. 65); "The surviving ballads are 305 in number and are known to enthusiasts almost as familiarly by their numbers in Child as by their titles. . . [Child's collection] is the authoritative resting-place of what has survived of the splendid body of English and Scottish ballads." One or two other ballads have made a plea for entrance to the closed ranks, notably "The Bitter Withy", whose cause has been supported by Professor Gerould (PMLA, XXIII (1908), 141-167. The same advocate has urged the claims for "A Ballad of the Twelfth Day" (MLR, VIII (1913), 65ff.; and IX, (1914), 235-6). Miss Pound points out the value of the Hill Ms., (E. E. T. S. 101, 1907) as a valuable source for "ballad-like" material (Poetic Origins, p. 126 n.). But even Professor Gerould prefaces his arguments with: "From the completion of Professor Child's magnificent work up to the present no ballad has been discovered which would merit insertion under a new title in that corpus. Variants of ballads already known continue to be unearthed with gratifying frequency, but so well did the great collector glean the field that it can seldom fall to the lot of any follower to bring to light a new specimen." Greig insists (Last Leaves, p. XXXI): "Child had all of the English and Scottish Traditional Ballads (305) except 'Young Betrice', No. 5 in William Tytler's lost Brown Ms., (v. 397), which "may possibly be a version of 'Hugh Spencer's Feats in France'."
39. PMLA, XLVI (1929), 629.
40. That there may be some contemporary chafing under the restraint imposed in the canon of the Child ballads is attested by the Foreword to Barry, Eckstrom and Smyth's British Ballads from Maine (1929) wherein the editors point out (p. XVIII) that "in some respects it has been impossible to be bound by Professor Child." Professor Davis also says (p. 11), "The editor is not inclined to draw too hard and fast a distinction between the Child ballads and all other ballads."