Ballad Classification
by D. K. Wilgus
Midwest Folklore, Vol. 5, No. 2 (Summer, 1955), pp. 95-100
BALLADC LASSIFICATION*
By D. K. WILGUS
The problem of ballad classification is difficult, yet crucial to the future of ballad scholarship. Leading scholars have wrestled with the problem or have left the answer blank. I do not propose in this paper to solve the problem. I wish merely to suggest a convenient method. When I propose to classify ballads by their narrative themes,
I do not deny the utility and necessity of other approaches to the
problem. Almost any index to the ballads will be productive of
fresh insights, and the more of them we have, the better. But I have
concluded that a thematic classification forms the most convenient
basis for ordering the riches of Anglo-American ballad collections.
I must begin with what is becoming accepted as a truism: the
great days of American folksong collecting are over. And another
truism: the results of collection are scattered, difficult of access,
and unevaluated. The crying need is for a syllabus which will organize
the materials for study, providing. classification, arrangement,
standard titles, etc. That such a syllabus cannot come immediately
has been demonstrated by recent attempts. My own suggestion is
that the end may be achieved by a progressive and continuingr
process, the construction of regional and partial syllabi productive
of tentative titles, classification, etc. From my attempts to lay the
groundwork for such a syllabus grow rny remarks, themselves tentative
and presented in the hope of correction.
I deal only with ballad texts, a restriction which calls for brief
explanation. Most important students of folk music recognize that,
for the present at least, texts and tunes can and must be studied
separately, however ideal it may be to deal with them in their basic
state of unstable marriage. But when I restrict myself to ballads, I take
on the obligation of answering at least one part of the question,
"What is a ballad?" That is, how to distinguish the folk ballad from
other types of folksong. I shall say simply that a ballad is a folksong
with explicit narrative content. Fortunately I shall not now find it
necessary to exercise nice distinctions in applying the definition, but
can move to the second half of the question, "What is a ballad?":
the classification of narrative folksongs.
Let me distinguish first between classification and arrangement.
By classification I mean the placing of ballad texts with their
relatives; by arrangement I mean the ordering of these groups of ballad
texts. There are two fundamental relationships between ballad texts:
narrative and textual. Both types must be taken into account, but
I wish to emphasize the classification by narrative relationship. In
this sense, a particular ballad is a narrative folksong which preserves
a particular story or tale-type. The approach is not new. Phillips
Barry suggested in 1913 [1] that we should classify ballads by their
themes. But, like so many of Barry's insights, the approach has
been neglected or inconsistently applied. A brief glance at the classification
and arrangement of published collections of American folksongs
may help explain the neglect and perhaps justify the application
of a narrative-type classification.
Because those editors who have attempted to organize their
offerings have made little distinction between classification and arrangement
(in my sense), discussion is difficult. Therefore I prefer
to speak of their efforts as organization. The organization of American
collections is largely of two types, the "Child and other" and the
"cute." The latter type, represented mainly by the work of Carl
Sandburg[2] and the Lomaxes,[3] strives for the attention of a popular
audience by the use of fanciful and perhaps fantastic categories, such
as "Critters and Chillun'," "Heroes and Hard Cases," and "Lonesome
Whistles." However, the divisions are not as "cute" as the
titles would indicate. The categories are chosen roughly by subject
matter and function, with origin thrown in, but cannot be logically
applied. "John Henry" may be a Negro song, a railroad song, or a
hero song. Because the categories are not mutually exclusive, placement
of an individual ballad is highly arbitrary, and the organization
does not even aid the reader. Who would expect to find "Little
Mohee" listed in "Songs from the Mountains"? But has scholarly
organization been much better?
The basic organization of the scholarly collection has been the
"Child and other" pattern. This presentation lists first, examples
of the English and Scottish popular ballads recognized by Francis
James Child, with perhaps the addition of one or two ballads judged
to be worthy of such company. Next come "other imported ballads
and songs," followed by native American products. The American
ballads are then divided among such categories as "Western Songs,"
"Criminal and Outlaw Songs," "Murder Ballads," "Humorous
Ballads," and-of course-"Miscellaneous Ballads and Songs." There
may be special sections devoted to religious songs, Negro (or pseudo-
Negro) songs, or nursery songs. Form, function, origin, and incidence
are hopelessly confused. While editors have made valiant attempts
to follow a logical organization, the old problems of overlapping
categories and arbitrary placement occur. Editors have unanimously
confessed their inability to apply logical categories to apparently
illogical materials, and the lack of criticism of ballad organization
may be attributed largely to a realization of the difficulty of the problem
and perhaps to a fear of criticism of the critic's own organization.
In view of the recent attempts to provide syllabi of American folksong-
particularly that of Dr. G. Malcolm Laws, Jr.[4]-we must
move rapidly to some solution of the problem. Otherwise we shall
become saddled with Laws B 10 as we have been saddled with Child 7.
The primary problem is that of deference-deference to the
great scholar whose work has become an enemy because of its overinfluence.
As long as the group of ballads recognized by Child is
sealed off from other ballads, we can neither classify nor arrange
ballads rationally. Of course deference to Child has had a convenience,
and convenience is one of the primary tests of ballad
organization. Consider that even such an outspoken opponent of the
"Harvard School of Communalists" as Professor Louise Pound, herself
edited a "Child and" collection.[5] And anyone who has struggled
with such a book as Byron Arnold's Folksongs of Alabama (Birmingham,
Ala., 1950), which is organized by informant, turns with vast
relief to a "Child and" collection. But the time come to re-examine
the entire problem.
The Child collection contains ballads of different types and
styles, few of which are found exclusively within the covers of The
English and Scottish Popular Ballads. Child's tests of inclusion have
never been satisfactorily explained by his disciples and were in any
case formulated without the aid of material discovered in the twentieth
century and without attention to the needs of the contemporary student.
While the Child canon remains sacrosanct, we can attempt
no logical organization, for the canon cuts across all organizational
schemes.
The arrangement of the ballads which received Child's accolade
has had no appreciable effect on contemporary efforts and offers no
model. Child's first edition was arranged loosely according to subject
matter. Dissatisfied, he adopted for his final edition a modified
form of Grundtvig's proposal that the ballads be arranged by metrical
form, an arrangement assumed to be natural and historical.[6] This
organization is rather like indexing folk melodies according to their
opening phrases. Within the assumed historical arrangement, Child
attempted to group the ballads thematically, but the result is not
satisfactory.
Child's classification (in the strict sense of the word) has received
almost no comment from critics, precisely because most of them have
accepted Child's principle unconsciously. The principle is that of
origin and textual filiation. A ballad is considered as a group of
variants with a common origin and textual relationships. That
principle resulted in the separation of "Earl Brand" and "Erlinton,"
of "Kemp Owyne," "Allison Gross," and "The Laily Worm and
the Machrel of the Sea." Accordingly we now separate versions of
"The Gosport Tragedy" from versions of "Little Omie Wise"; Versions
of 'The Wittam Miller" from "On the Banks of the Ohio";
versions of "Pearl Bryan I" from versions of "Pearl Bryan II," "The
Jealous Lover," and "Lulu Viers." Yet they may be, in a sense, forms
of the same ballad.
I have displayed briefly factors which have led to the separation
of related ballads, or perhaps even forms of the same ballad. The
subject, the incidence, the origin, and the textual form have all had
this effect. A group of texts achieves a particular place in a collection
because its protagonist is a cowboy or a railroad engineer, because the
texts were sung by lumbermen or sailors, because the textual archetype
originated in England or America, or because members of the
ballad group have exclusive textual traits. The results are tragic
or comic, as you prefer. A ballad which is suddenly discovered to
have British antecedents must be immediately shifted to another
category. "Canaday-I-O" is a lumberman's ballad, but the related
"Buffalo Skinners" is placed in the "Cowboy and Frontier" section.
"The Cowboy's Lament," despite its important background is treated
as native American, while the "Pretty Polly" version of "The Gosport
Tragedy" is a foreigner. And the "St. James Infirmary" cousin of
"The Cowboy's Lament" does not even gain entrance to Laws'
Native American Ballads. Not only are "Earl Brand" and "Erlinton"
distinct because of textual differences and because Child separated
them, but "The Soldier's Wooing," telling the same basic story, is
a later broadside and must be separated from both; and the related
broadside, "New River Shore," is classified as separate, while one of
its American forms, "Red River Shore-" is a cowboy song. This will
never do.
But forces have been working toward a more convenient classification
of related ballads, or related forms of the same ballad. The very
concentration on the Child ballad has produced suggestive results.
Enthusiastic disciples once claimed as Child variants "Sally of London"
(related to "The Brown Girl," Child 295), "The Soldier's Wooing"
and "New River Shore" (previously mentioned), "The Prisoner
at the Bar" (possibly related to "Geordie," Child 209) and other
so-called derivatives. Thus arose the concept of the secondary ballad,
related to a Child item in theme, if not in text. More recently,
attention has been given to the appearance in ballads of such themes
as the returned lover and the murdered girl. Continuing notice of
thematic parallels and textural crossings, such as the studies of Anne
G. Gilchrist[7] and Ruth Harvey,[8] seems to point to a more useful
and perhaps logical organization of ballad texts-a classification by
narrative content and a convenient arrangement of the narrative
themes.
Such an organization will place under a single head those traditional
variants which converge in telling a single narrative, but
differentiate between forms of that narrative. Thus the theme of the
girl taken for a walk and murdered by her lover may be represented
by "The Wittam Miller," "The Gosport Tragedy," "Little Omie
Wise," "The Jealous Lover," "On the Banks of the Ohio," "Rose
Connoley," "Lulu Viers," and perhaps "Pearl Bryan" and "Poor
Ellen Smith." "Lord Lovell" may be a form of a ballad which
includes the sentimental "Cowboy Jack." The style and form in
which a theme is presented must not obscure the narrative itself.
One ballad age may prefer the confession style, another the impersonal.
The fact that the forms arose at different times, in different
places, and perhaps even independently is no barrier to such a
classification. It is barely possible that some of the numbers in the
Child canon have such a checkered background.
An arrangement of such classifications may be made on logical
or even arbitrary grounds. One has only to consult the Aarne-
Thompson folktale index to observe how a similar problem has
been met. "John Hardy" will appear with other ballads of criminals
instead of being separated because of its supposed Negro origin.
Almost any arrangement of classifications will be suitable if it be not
based on origin, style, textual form, incidence, or other non-essential
feature.
You may ask if there be not some philosophical position underlying this approach to ballad classification. Such certainly seems to have been true of Barry's oniginal suggestion. For when he referred to "a traditional ballad mythology, stereotyped ornamentations, and
details, suited to certain events,"[9] he apparently postulated a universal
unconscious factor in individuals which causes similar themes to
revert to common primitive form. But it is not necessary to accept
a psychological or even a Platonic theory to justify a convenient
arrangement.
You may then wonder if I am not forgetting that a ballad is
more than a tale. Certainly the various textual forms of a ballad
are exceedingly important and must be represented by the division
of a ballad into subtypes. But I cannot forget that a ballad is a
narrative and that its narrative substance differentiates it from other
varieties of folksong. The differing textual forms of a narrative
theme present one of the most interesting aspects of "the way of the
folk" in handling narrative substance, and a thematic organization
is one of the most convenient ways of approaching that aspect.
Instead of ignoring the textual form of the ballad, the proposed
classification emphasizes at least one aspect of the formal approach.
However, the virtues of the narrative-type approach must not
blind us to its limitations. It is not a ready-made last over which
we can cobble our ballads into shape. It is not the only valid approach
and will not answer all our problems. Although it is the most convenient
index, it represents only one trend in folk art. Texts split
as well as coalesce. Textual patterns are recomposed to fit widely
divergent narrative types. Singers make ballad mosaics of preexisting
stanzas, and many elements go into the making of one
ballad. The full tradition of "Casey Jones" must include "Rambling
John," "Jimmie Jones," "Jay Gould's Daughter," and we know not
what else. Each ballad is still a study in itself, and a useful syllabus
will require compromise and many cross-references.
Therefore no one index can provide us with all the necessary
insights. As Types of the Folk-Tale is supplemented by the Motif
Index, so must a narrative-type organization of ballads be supplemented
by many other indices. The various organizations we have
already observed have been of great aid in increasing our knowledge
of traditional song, for it is necessary to know what songs originate
or are sung among certain groups, what songs are now applied to
specific occupations, even what songs are known to a particular informant.
But the basic tool for classifying our ballads must be
thematic.
Western Kentucky State College Bowling Green, Kentucky
Footnotes:
* Read at the Folklore Section of the Seventh University of Kentucky Foreign Language Conference, April 24, 1954. Cf. my "A Syllabus of Kentucky Folksongs," Kentucky Folklore Record, I (1955), 31 ff.
1 MLN, XXVIII, 5.
2 The American Song Bag (New York, 1927).
3 John A. and Alan Lomax, American Ballads and Folk Songs (New
York, 1934); Folk Song: U. S. A. (New York, 1947).
4 Native American Balladry (Philadelphia, 1950).
5 American Ballads and Songs (New York, 1922).
6 Sigurd Bernhard Hustvedt, Ballad Books- and Ballad Men (Cambridge,
Mass., 1930) 219-220.
7"'Death and the Lady' in English Balladry," JEFDSS, IV (1941), 37
ff.; "The Song of Marvels (or Lies)," loc. cit., 113 ff.
8 "The Unquiet Grave," JEFDSS, IV, 49 ff.
9 ]AF, XXIII (1910), 454.